calendars0000644000000000000000000001123415105703306007664 0ustar00----- Calendrical issues ----- As mentioned in Theory.html, although calendrical issues are out of scope for tzdb, they indicate the sort of problems that we would run into if we extended tzdb further into the past. The following information and sources go beyond Theory.html's brief discussion. They sometimes disagree. France Gregorian calendar adopted 1582-12-20. French Revolutionary calendar used 1793-11-24 through 1805-12-31, and (in Paris only) 1871-05-06 through 1871-05-23. Russia Soviet Russia adopted the Gregorian calendar on 1918-02-14. It also used 5- and 6-day work weeks at times, in parallel with the Gregorian calendar; see . Sweden (and Finland) From: Mark Brader Subject: Re: Gregorian reform - a part of locale? Date: 1996-07-06 In 1700, Denmark made the transition from Julian to Gregorian. Sweden decided to *start* a transition in 1700 as well, but rather than have one of those unsightly calendar gaps :-), they simply decreed that the next leap year after 1696 would be in 1744 - putting the whole country on a calendar different from both Julian and Gregorian for a period of 40 years. However, in 1704 something went wrong and the plan was not carried through; they did, after all, have a leap year that year. And one in 1708. In 1712 they gave it up and went back to Julian, putting 30 days in February that year!... Then in 1753, Sweden made the transition to Gregorian in the usual manner, getting there only 13 years behind the original schedule. (A previous posting of this story was challenged, and Swedish readers produced the following references to support it: "Tideräkning och historia" by Natanael Beckman (1924) and "Tid, en bok om tideräkning och kalenderväsen" by Lars-Olof Lodén (1968). Grotefend's data From: "Michael Palmer" [with two obvious typos fixed] Subject: Re: Gregorian Calendar (was Re: Another FHC related question Newsgroups: soc.genealogy.german Date: Tue, 9 Feb 1999 02:32:48 -800 ... The following is a(n incomplete) listing, arranged chronologically, of European states, with the date they converted from the Julian to the Gregorian calendar: 04/15 Oct 1582 - Italy (with exceptions), Spain, Portugal, Poland (Roman Catholics and Danzig only) 09/20 Dec 1582 - France, Lorraine 21 Dec 1582/ 01 Jan 1583 - Holland, Brabant, Flanders, Hennegau 10/21 Feb 1583 - bishopric of Liege (Lüttich) 13/24 Feb 1583 - bishopric of Augsburg 04/15 Oct 1583 - electorate of Trier 05/16 Oct 1583 - Bavaria, bishoprics of Freising, Eichstedt, Regensburg, Salzburg, Brixen 13/24 Oct 1583 - Austrian Oberelsaß and Breisgau 20/31 Oct 1583 - bishopric of Basel 02/13 Nov 1583 - duchy of Jülich-Berg 02/13 Nov 1583 - electorate and city of Köln 04/15 Nov 1583 - bishopric of Würzburg 11/22 Nov 1583 - electorate of Mainz 16/27 Nov 1583 - bishopric of Strassburg and the margraviate of Baden 17/28 Nov 1583 - bishopric of Münster and duchy of Cleve 14/25 Dec 1583 - Steiermark 06/17 Jan 1584 - Austria and Bohemia 11/22 Jan 1584 - Lucerne, Uri, Schwyz, Zug, Freiburg, Solothurn 12/23 Jan 1584 - Silesia and the Lausitz 22 Jan/ 02 Feb 1584 - Hungary (legally on 21 Oct 1587) Jun 1584 - Unterwalden 01/12 Jul 1584 - duchy of Westfalen 16/27 Jun 1585 - bishopric of Paderborn 14/25 Dec 1590 - Transylvania 22 Aug/ 02 Sep 1612 - duchy of Prussia 13/24 Dec 1614 - Pfalz-Neuburg 1617 - duchy of Kurland (reverted to the Julian calendar in 1796) 1624 - bishopric of Osnabrück 1630 - bishopric of Minden 15/26 Mar 1631 - bishopric of Hildesheim 1655 - Kanton Wallis 05/16 Feb 1682 - city of Strassburg 18 Feb/ 01 Mar 1700 - Protestant Germany (including Swedish possessions in Germany), Denmark, Norway 30 Jun/ 12 Jul 1700 - Gelderland, Zutphen 10 Nov/ 12 Dec 1700 - Utrecht, Overijssel 31 Dec 1700/ 12 Jan 1701 - Friesland, Groningen, Zürich, Bern, Basel, Geneva, Thurgau, and Schaffhausen 1724 - Glarus, Appenzell, and the city of St. Gallen 01 Jan 1750 - Pisa and Florence 02/14 Sep 1752 - Great Britain 17 Feb/ 01 Mar 1753 - Sweden 1760-1812 - Graubünden The Russian empire (including Finland and the Baltic states) did not convert to the Gregorian calendar until the Soviet revolution of 1917. Source: H. Grotefend, _Taschenbuch der Zeitrechnung des deutschen Mittelalters und der Neuzeit_, herausgegeben von Dr. O. Grotefend (Hannover: Hahnsche Buchhandlung, 1941), pp. 26-28. ----- This file is in the public domain, so clarified as of 2009-05-17 by Arthur David Olson. ----- Local Variables: coding: utf-8 End: CONTRIBUTING0000644000000000000000000000727615242710656007662 0ustar00# Contributing to the tz code and data Please do not create issues or pull requests on GitHub, as the proper procedure for proposing and distributing patches is via email as described below. The time zone database is by no means authoritative: governments change timekeeping rules erratically and sometimes with little warning, the data entries do not cover all of civil time before 1970, and undoubtedly errors remain in the code and data. Feel free to fill gaps or fix mistakes, and please email improvements to for use in the future. In your email, please give reliable sources that reviewers can check. The mailing list and its archives are public, so please do not send confidential information. ## Contributing technical changes To email small changes, please run a POSIX shell command like ‘diff -u old/europe new/europe >myfix.patch’, and attach ‘myfix.patch’ to the email. For more-elaborate or possibly controversial changes, such as renaming, adding or removing zones, please read “Theory and pragmatics of the tz code and data” It is also good to browse the mailing list archives for examples of patches that tend to work well. Changes should contain commentary citing reliable sources. Citations should use ‘https:’ URLs if available. For changes that fix sensitive security-related bugs, please see the distribution’s SECURITY file. Please submit changes against either the latest release or the main branch of the development repository. The latter is preferred. ## Sample Git workflow for developing contributions If you use Git the following workflow may be helpful: * Copy the development repository. git clone https://github.com/eggert/tz.git cd tz * Get current with the main branch. git checkout main git pull * Switch to a new branch for the changes. Choose a different branch name for each change set. git checkout -b mybranch * Sleuth by using ‘git blame’. For example, when fixing data for Africa/Sao_Tome, if the command ‘git blame africa’ outputs a line ‘2951fa3b (Paul Eggert 2018-01-08 09:03:13 -0800 1068) Zone Africa/Sao_Tome 0:26:56 - LMT 1884’, commit 2951fa3b should provide some justification for the ‘Zone Africa/Sao_Tome’ line. * Edit source files. Include commentary that justifies the changes by citing reliable sources. * Debug the changes locally, e.g.: make TOPDIR=$PWD/tz clean check install ./zdump -v America/Los_Angeles Although builds assume only basic POSIX, they use extra features if available. ‘make check’ accesses validator.w3.org unless you lack ‘curl’ or use ‘make CURL=:’. If you have the latest GCC, ‘make CFLAGS='$(GCC_DEBUG_FLAGS)'’ does extra checking. * For each separable change, commit it in the new branch, e.g.: git add northamerica git commit See recent ‘git log’ output for the commit-message style. * Create patch files 0001-..., 0002-..., ... git format-patch main * Check that the patch files and your email setup contain only information that you want to make public. * After reviewing the patch files, send the patches to for others to review. git send-email main For an archived example of such an email, see “[PROPOSED] Fix off-by-1 error for Jamaica and T&C before 1913” . * Start anew by getting current with the main branch again (the second step above). ----- This file is in the public domain. LICENSE0000644000000000000000000000037413111575741007022 0ustar00Unless specified below, all files in the tz code and data (including this LICENSE file) are in the public domain. If the files date.c, newstrftime.3, and strftime.c are present, they contain material derived from BSD and use the BSD 3-clause license. Makefile0000644000000000000000000015626515244743474007501 0ustar00# Make and install tzdb code and data. # This file is in the public domain, so clarified as of # 2009-05-17 by Arthur David Olson. # Request POSIX conformance; this must be the first non-comment line. .POSIX: # By default, builds of code and data assume POSIX.1-2001 or later; # this assumption can be relaxed by tailoring the build as described below. # On older platforms you may need to scrounge for POSIX conformance. # For example, on Solaris 10 (2005) with Sun Studio 12 aka Sun C 5.9 (2007), # use 'PATH=/usr/xpg4/bin:$PATH make CC=c99'. # Reproducible builds of distribution tarballs also need a copy of the # Git repository, and assume the behavior of the following programs # (or later versions): # Git 2.7.0 (2016) # GNU Coreutils 6.3 (2006) # GNU Tar 1.14 (2004) # GnuPG 1.4 (2004) # Although tzdb does not come with a software bill of materials, # you should be able to construct one based on the above information, # your platform, and the way you use this Makefile. # To affect how this Makefile works, you can run a shell script like this: # # #!/bin/sh # make CFLAGS='-O2 -DHAVE_GETTEXT=0' "$@" # # This example script is appropriate for a GNU/Linux system # which needs more optimization than default, and which does not want # gettext's internationalization of diagnostics. # # Alternatively, you can simply edit this Makefile to tailor the following # macro definitions. ############################################################################### # Start of macros that one plausibly might want to tailor. # Package name for the code distribution. PACKAGE= tzcode # Version number for the distribution, overridden in the 'tarballs' rule below. VERSION= unknown # Email address for bug reports. BUGEMAIL= tz@iana.org # DATAFORM selects the data format. # Available formats represent essentially the same data, albeit # possibly with minor discrepancies that users are not likely to notice. # To get new features and the best data right away, use: # DATAFORM= vanguard # To wait a while before using new features, to give downstream users # time to upgrade zic (the default), use: # DATAFORM= main # To wait even longer for new features, use: # DATAFORM= rearguard # Rearguard users might also want "ZFLAGS = -b fat"; see below. DATAFORM= main # Change the line below for your timezone (after finding the one you want in # one of the $(TDATA) source files, or adding it to a source file). # Alternatively, if you discover you've got the wrong timezone, you can just # 'zic -l -' to remove it, or 'zic -l rightzone' to change it. # Use the command # make zonenames # to get a list of the values you can use for LOCALTIME. LOCALTIME= Factory # Installation locations. # # The defaults are suitable for Debian, except that if REDO is # posix_right or right_posix then files that Debian puts under # /usr/share/zoneinfo/posix and /usr/share/zoneinfo/right are instead # put under /usr/share/zoneinfo-posix and /usr/share/zoneinfo-leaps, # respectively. Problems with the Debian approach are discussed in # the commentary for the right_posix rule (below). # Destination directory, which can be used for staging. # 'make DESTDIR=/stage install' installs under /stage (e.g., to # /stage/etc/localtime instead of to /etc/localtime). Files under # /stage are not intended to work as-is, but can be copied by hand to # the root directory later. If DESTDIR is empty, 'make install' does # not stage, but installs directly into production locations. DESTDIR = # Everything is installed into subdirectories of TOPDIR, and used there. # TOPDIR should be empty (meaning the root directory), # or a directory name that does not end in "/". # TOPDIR should be empty or an absolute name unless you're just testing. TOPDIR = # The default local timezone is taken from the file TZDEFAULT. TZDEFAULT = $(TOPDIR)/etc/localtime # The subdirectory containing installed program and data files, and # likewise for installed files that can be shared among architectures. # These should be relative file names. USRDIR = usr USRSHAREDIR = $(USRDIR)/share # "Compiled" timezone information is placed in the "TZDIR" directory # (and subdirectories). # TZDIR_BASENAME should not contain "/" and should not be ".", ".." or empty. TZDIR_BASENAME= zoneinfo TZDIR = $(TOPDIR)/$(USRSHAREDIR)/$(TZDIR_BASENAME) # The "tzselect" and (if you do "make INSTALL") "date" commands go in: BINDIR = $(TOPDIR)/$(USRDIR)/bin # The "zdump" command goes in: ZDUMPDIR = $(BINDIR) # The "zic" command goes in: ZICDIR = $(TOPDIR)/$(USRDIR)/sbin # Manual pages go in subdirectories of: MANDIR = $(TOPDIR)/$(USRSHAREDIR)/man # Library functions are put in an archive in LIBDIR. LIBDIR = $(TOPDIR)/$(USRDIR)/lib # Types to try, as an alternative to time_t. TIME_T_ALTERNATIVES = $(TIME_T_ALTERNATIVES_HEAD) $(TIME_T_ALTERNATIVES_TAIL) TIME_T_ALTERNATIVES_HEAD = int_least64_t.ck TIME_T_ALTERNATIVES_TAIL = int_least32_t.ck uint_least32_t.ck \ uint_least64_t.ck # What kind of TZif data files to generate. (TZif is the binary time # zone data format that zic generates; see Internet RFC 9636.) # If you want only POSIX time, with time values interpreted as # seconds since the epoch (not counting leap seconds), use # REDO= posix_only # below. If you want only "right" time, with values interpreted # as seconds since the epoch (counting leap seconds), use # REDO= right_only # below. If you want both sets of data available, with leap seconds not # counted normally, use # REDO= posix_right # below. If you want both sets of data available, with leap seconds counted # normally, use # REDO= right_posix # below. POSIX mandates that leap seconds not be counted, and a # nonnegative TZ_CHANGE_INTERVAL also assumes this, so to be compatible with # these, use "posix_only" or "posix_right". Use POSIX time on systems with # leap smearing; this can work better than unsmeared "right" time with # applications that are not aware of leap seconds, and is closer to unsmeared # "right" time than unsmeared POSIX time is (e.g., 0.5 vs 1.0 s max error). REDO= posix_only # Whether to put an "Expires" line in the leapseconds file. # Use EXPIRES_LINE=1 to put the line in, 0 to omit it. # The EXPIRES_LINE value matters only if REDO's value contains "right". # If you change EXPIRES_LINE, remove the leapseconds file before running "make". # zic's support for the Expires line was introduced in tzdb 2020a, # and was modified in tzdb 2021b to generate version 4 TZif files. # EXPIRES_LINE defaults to 0 for now so that the leapseconds file # can be given to pre-2020a zic implementations and so that TZif files # built by newer zic implementations can be read by pre-2021b libraries. EXPIRES_LINE= 0 # To install data in text form that has all the information of the TZif data, # (optionally incorporating leap second information), use # TZDATA_TEXT= tzdata.zi leapseconds # To install text data without leap second information (e.g., because # REDO='posix_only'), use # TZDATA_TEXT= tzdata.zi # To avoid installing text data, use # TZDATA_TEXT= TZDATA_TEXT= leapseconds tzdata.zi # For backward-compatibility links and zones for old names, use # BACKWARD= backward # To omit these links and zones, use # BACKWARD= BACKWARD= backward # If you want out-of-scope and often-wrong data from the file 'backzone', # but only for entries listed in the backward-compatibility file zone.tab, use # PACKRATDATA= backzone # PACKRATLIST= zone.tab # If you want all the 'backzone' data, use # PACKRATDATA= backzone # PACKRATLIST= # To omit this data, use # PACKRATDATA= # PACKRATLIST= PACKRATDATA= PACKRATLIST= # The name of a locale using the UTF-8 encoding, used during self-tests. # The tests are skipped if the name does not appear to work on this system. UTF8_LOCALE= en_US.utf8 # Extra flags for producing man page files like tzfile.5.txt. # These flags are used only if groff (or mandoc) is present. # Each option should begin with "-" and should lack shell metacharacters. # Plausible options include -Tascii and -Tutf8. MANFLAGS= -Tutf8 # Non-default libraries needed to link. # On some hosts, this should have -lintl unless CFLAGS has -DHAVE_GETTEXT=0. LDLIBS= # Add the following to an uncommented "CFLAGS=" line as needed # to override defaults specified in the source code or by the system. # "-DFOO" is equivalent to "-DFOO=1". # -DDEPRECATE_TWO_DIGIT_YEARS for optional runtime warnings about strftime # formats that generate only the last two digits of year numbers # -DEPOCH_LOCAL if the 'time' function returns local time not UT # -DEPOCH_OFFSET=N if the 'time' function returns a value N greater # than what POSIX specifies, assuming local time is UT. # For example, N is 252460800 on AmigaOS. # -DFREE_PRESERVES_ERRNO=[01] if the 'free' function munges or preserves errno # (default is guessed) # -DHAVE_DECL_ASCTIME_R=0 if does not declare asctime_r # on POSIX platforms predating POSIX.1-2024 # -DHAVE_DECL_ENVIRON if declares 'environ' # -DHAVE_DECL_TIMEGM=0 if does not declare timegm # -DHAVE_DIRECT_H if mkdir needs (MS-Windows) # -DHAVE_FCHMOD=0 if your system lacks the fchmod function # -DHAVE__GENERIC=0 if _Generic does not work* # -DHAVE_GETEUID=0 if gete?[ug]id do not work # -DHAVE_GETRANDOM if getrandom works (e.g., GNU/Linux), # -DHAVE_GETRANDOM=0 to avoid using getrandom # -DHAVE_GETRESUID=0 if getres[ug]id do not work # -DHAVE_GETTEXT if gettext works (e.g., GNU/Linux, FreeBSD, Solaris), # where LDLIBS also needs to contain -lintl on some hosts; # -DHAVE_GETTEXT=0 to avoid using gettext # -DHAVE_INCOMPATIBLE_CTIME_R if your system's time.h declares # ctime_r and asctime_r incompatibly with POSIX.1-2017 and earlier # (Solaris when _POSIX_PTHREAD_SEMANTICS is not defined). # -DHAVE_INTTYPES_H=0 if does not work*+ # -DHAVE_ISSETUGID=1 if issetugid works, 0 otherwise (default is guessed) # If 0, you may also use -DHAVE_SYS_AUXV_H=1 if works, # 0 otherwise (default is guessed). # -DHAVE_LINK=0 if your system lacks a link function # -DHAVE_LOCALTIME_R=0 if your system lacks a localtime_r function # -DHAVE_LOCALTIME_RZ=0 if you do not want zdump to use localtime_rz # localtime_rz can make zdump significantly faster, but is nonstandard. # -DHAVE_MALLOC_ERRNO=0 if malloc etc. do not set errno on failure. # -DHAVE_MEMPCPY=1 if your system has mempcpy, 0 if not (default is guessed) # -DHAVE_POSIX_DECLS=0 if your system's include files do not declare # variables like 'tzname' required by POSIX # -DHAVE_PWD_H=0 if your system lacks pwd.h, grp.h and corresponding functions # If 0, you may also need -Dgid_t=G -Duid_t=U # to define gid_t and uid_t to be types G and U. # -DHAVE_SETENV=0 if your system lacks the setenv function # -DHAVE_SETMODE=[01] if your system lacks or has the setmode and getmode # functions (default is guessed) # -DHAVE_SNPRINTF=0 if your system lacks the snprintf function+ # -DHAVE_STDCKDINT_H=0 if neither nor substitutes like # __builtin_add_overflow work* # -DHAVE_STDINT_H=0 if does not work*+ # -DHAVE_STRFTIME_L if declares locale_t and strftime_l # -DHAVE_STRDUP=0 if your system lacks the strdup function # -DHAVE_STRNLEN=0 if your system lacks the strnlen function+ # -DHAVE_STRTOLL=0 if your system lacks the strtoll function+ # -DHAVE_STRUCT_STAT_ST_CTIM=0 if struct stat lacks a status-change member # of type struct timespec, so code should use st_ctime instead; # but if the status-change member name is st_ctimespec, # use -Dst_ctim=st_ctimespec instead (default is guessed)+ # -DHAVE_STRUCT_TIMESPEC=0 if your system lacks struct timespec+ # -DHAVE_SYMLINK=0 if your system lacks the symlink function # -DHAVE_SYS_STAT_H=0 if does not work* # If 0, you may also need -Dmode_t=M to define mode_t to be type M. # -DHAVE_TZSET=0 if your system lacks a tzset function # -DHAVE_UNISTD_H=0 if does not work* # -DHAVE_UTMPX_H=0 if does not work* # -Dlocale_t=XXX if your system uses XXX instead of locale_t # -DMKTIME_MIGHT_OVERFLOW if mktime might fail due to time_t overflow # -DOPENAT_TZDIR if tzset should use openat on TZDIR then a relative open. # See localtime.c for details. # -DPORT_TO_C89 if tzcode should also run on mostly-C89 platforms+ # Typically it is better to use a later standard. For example, # with GCC 4.9.4 (2016), prefer '-std=gnu11' to '-DPORT_TO_C89'. # Even with -DPORT_TO_C89, the code needs at least one C99 # feature (integers at least 64 bits wide) and maybe more. # -DRESERVE_STD_EXT_IDS if your platform reserves standard identifiers # with external linkage, e.g., applications cannot define 'localtime'. # -Dssize_t=int on hosts like MS-Windows that lack ssize_t # -DSUPPORT_C89=0 if the tzcode library should not support C89 callers # Although -DSUPPORT_C89=0 might work around latent bugs in callers, # it does not conform to POSIX. # -DSUPPORT_POSIX2008 if the library should support older POSIX callers+ # However, this might cause problems in POSIX.1-2024-or-later callers. # -DSUPPRESS_TZDIR to not prepend TZDIR to file names; this has # security implications and is not recommended for general use # -DTHREAD_SAFE to make localtime.c thread-safe, as POSIX requires; # not needed by the main-program tz code, which is single-threaded. # Append other compiler flags as needed, e.g., -pthread on GNU/Linux. # With -DTHREAD_SAFE the following options can also be used: # -DTHREAD_PREFER_SINGLE to prefer speed in single-threaded apps, # at some cost in CPU time and energy in multi-threaded apps. # With -DTHREAD_PREFER_SINGLE the following options can also be used: # -DHAVE___ISTHREADED=1 if there is an extern int __isthreaded # variable, 0 otherwise (default is guessed) # -DHAVE_SYS_SINGLE_THREADED_H=0 if works, # 0 otherwise (default is guessed) # -DTHREAD_RWLOCK to use read-write locks instead of mutexes. # This can improve parallelism and thus save real time # if many threads call tzcode functions simultaneously. # It also costs CPU time and thus energy. # -DTHREAD_TM_MULTI to have gmtime, localtime, and offtime # return different struct tm * addresses in different threads. # This supports nonportable programs that call # gmtime/localtime/offtime when they should call # gmtime_r/localtime_r/offtime_r to avoid races. # Because the corresponding storage is freed on thread exit, # this option is incompatible with POSIX.1-2024 and earlier. # It also costs CPU time and memory. # -Dtime_tz=\"T\" to use T as the time_t type, rather than the system time_t # This is intended for internal use only; it mangles external names. # -DTZ_CHANGE_INTERVAL=N if functions depending on TZ should check # no more often than every N seconds for TZif file changes. # If N is negative (the default), no such checking is done. # This option is intended for platforms that want localtime etc. # to respond to changes to a file selected by TZ, including to # TZDEFAULT (normally /etc/localtime) if TZ is unset. # On these platforms, REDO should be "posix_only" or "posix_right". # This option does not affect tzalloc-allocated objects. # -DTZ_DOMAIN=\"foo\" to use "foo" for gettext domain name; default is "tz" # -DTZ_DOMAINDIR=\"/path\" to use "/path" for gettext directory; # the default is system-supplied, typically "/usr/lib/locale" # -DTZ_RUNTIME_LEAPS=0 to disable runtime support for leap seconds. # This conforms to POSIX, shrinks tzcode's attack surface, # and is more efficient. However, it fails to support Internet # RFC 9636's leap seconds. # -DTZDEFRULESTRING=\",date/time,date/time\" to default to the specified # DST transitions for proleptic format TZ strings lacking them. # If not specified, it defaults to US rules for future DST transitions. # This mishandles some past timestamps, as US DST rules have changed. # It also mishandles settings like TZ='EET-2EEST' for eastern Europe, # as Europe and US DST rules differ. # -DTZNAME_MAXIMUM=N to limit time zone abbreviations to N bytes (default 254) # -DUNINIT_TRAP if reading uninitialized storage can cause problems # other than simply getting garbage data # -DUSE_LTZ=0 to build zdump with the system time zone library # Also set TZDOBJS=zdump.o and CHECK_TIME_T_ALTERNATIVES= below. # -DZIC_BLOAT_DEFAULT=\"fat\" to default zic's -b option to "fat", and # similarly for "slim". Fat TZif files work around incompatibilities # and bugs in some TZif readers, notably older ones that # ignore or otherwise mishandle 64-bit data in TZif files; # however, fat TZif files may trigger bugs in newer TZif readers. # Slim TZif files are more efficient, and are the default. # -DZIC_MAX_ABBR_LEN_WO_WARN=3 # (or some other number) to set the maximum time zone abbreviation length # that zic will accept without a warning (the default is 6) # -g to generate symbolic debugging info # -Idir to include from directory 'dir' # -O0 to disable optimization; other -O options to enable more optimization # -Uname to remove any definition of the macro 'name' # $(GCC_DEBUG_FLAGS) if you are using recent GCC and want lots of checking # # * Options marked "*" can be omitted if your compiler is C23 compatible. # * Options marked "+" are obsolescent and are planned to be removed # once the code assumes C99 or later (say in the year 2029) # and POSIX.1-2024 or later (say in the year 2034). # # Select instrumentation via "make GCC_INSTRUMENT='whatever'". GCC_INSTRUMENT = \ -fsanitize=undefined -fsanitize-address-use-after-scope \ -fsanitize-trap=all -fstack-protector # Omit -fanalyzer from GCC_DEBUG_FLAGS, as it makes GCC too slow. GCC_DEBUG_FLAGS = -DGCC_LINT -g3 -O3 \ $(GCC_INSTRUMENT) \ -Wall -Wextra \ -Walloc-size-larger-than=100000 -Warray-bounds=2 \ -Wbad-function-cast -Wbidi-chars=any,ucn -Wcast-align=strict -Wcast-qual \ -Wdate-time \ -Wdeclaration-after-statement -Wdouble-promotion \ -Wduplicated-branches -Wduplicated-cond -Wflex-array-member-not-at-end \ -Wformat=2 -Wformat-overflow=2 -Wformat-signedness -Wformat-truncation \ -Wfree-labels -Wimplicit-fallthrough=5 -Winit-self \ -Wkeyword-macro -Wlogical-op \ -Wmissing-declarations -Wmissing-prototypes \ -Wmissing-variable-declarations -Wnested-externs \ -Wnull-dereference \ -Wold-style-definition -Woverlength-strings -Wpointer-arith \ -Wshadow -Wshift-overflow=2 \ -Wstrict-prototypes -Wstringop-overflow=4 \ -Wsuggest-attribute=cold \ -Wsuggest-attribute=const -Wsuggest-attribute=format \ -Wsuggest-attribute=malloc \ -Wsuggest-attribute=noreturn -Wsuggest-attribute=pure \ -Wtrampolines -Wundef -Wunused-macros -Wuse-after-free=3 \ -Wvariadic-macros -Wvla -Wwrite-strings \ -Wzero-as-null-pointer-constant \ -Wno-format-nonliteral -Wno-sign-compare -Wno-type-limits # # If your system has a "GMT offset" field in its "struct tm"s # (or if you decide to add such a field in your system's "time.h" file), # add the name to a define such as # -DTM_GMTOFF=tm_gmtoff # to the end of the "CFLAGS=" line. If not defined, the code attempts to # guess TM_GMTOFF from other macros; define NO_TM_GMTOFF to suppress this. # Similarly, if your system has a "zone abbreviation" field, define # -DTM_ZONE=tm_zone # and define NO_TM_ZONE to suppress any guessing. # Although POSIX.1-2024 requires these fields and they are widely available # on GNU/Linux and BSD systems, some older systems lack them. # # The next batch of options control support for external variables # exported by tzcode. In practice these variables are less useful # than TM_GMTOFF and TM_ZONE. However, most of them are standardized. # # # # To omit or support the external variable "tzname", add one of: # # -DHAVE_TZNAME=0 # do not support "tzname" # # -DHAVE_TZNAME=1 # support "tzname", which is defined by system library # # -DHAVE_TZNAME=2 # support and define "tzname" # # to the "CFLAGS=" line. Although "tzname" is required by POSIX.1-1988 # # and later, its contents are unspecified if you use a geographical TZ # # and the variable is planned to be removed in a future POSIX edition. # # If not defined, the code attempts to guess HAVE_TZNAME from other macros. # # Warning: unless time_tz is also defined, HAVE_TZNAME=1 can cause # # crashes when combined with some platforms' standard libraries, # # presumably due to memory allocation issues. # # # # To omit or support the external variables "timezone" and "daylight", add # # -DUSG_COMPAT=0 # do not support # # -DUSG_COMPAT=1 # support, and variables are defined by system library # # -DUSG_COMPAT=2 # support and define variables # # to the "CFLAGS=" line; "timezone" and "daylight" are inspired by Unix # # Systems Group code and are required by POSIX.1-2008 and later (with XSI), # # although their contents are unspecified if you use a geographical TZ # # and the variables are planned to be removed in a future edition of POSIX. # # If not defined, the code attempts to guess USG_COMPAT from other macros. # # # # To support the external variable "altzone", add # # -DALTZONE=0 # do not support # # -DALTZONE=1 # support "altzone", which is defined by system library # # -DALTZONE=2 # support and define "altzone" # # to the end of the "CFLAGS=" line; although "altzone" appeared in # # System V Release 3.1 it has not been standardized. # # If not defined, the code attempts to guess ALTZONE from other macros. # # If you want functions that were inspired by early versions of X3J11's work, # add # -DSTD_INSPIRED # to the end of the "CFLAGS=" line. This arranges for the following # functions to be added to the time conversion library. # "offtime" is like "gmtime" except that it accepts a second (long) argument # that gives an offset to add to the time_t when converting it. # "offtime_r" is to "offtime" what "gmtime_r" is to "gmtime". # I.e., "offtime" and "offtime_r" are like calling "localtime_rz" # with a fixed-offset zone. # "timelocal" is nearly equivalent to "mktime". # "timeoff" is like "timegm" except that it accepts a second (long) argument # that gives an offset to use when converting to a time_t. # I.e., "timeoff" is like calling "mktime_z" with a fixed-offset zone. # "posix2time" and "time2posix" are described in an included manual page. # X3J11's work does not describe any of these functions. # These functions may well disappear in future releases of the time # conversion package. # # If you don't want functions that were inspired by NetBSD, add # -DNETBSD_INSPIRED=0 # to the end of the "CFLAGS=" line. Otherwise, the functions # "localtime_rz", "mktime_z", "tzalloc", and "tzfree" are added to the # time library, and if STD_INSPIRED is also defined to nonzero the functions # "posix2time_z" and "time2posix_z" are added as well. # The functions ending in "_z" (or "_rz") are like their unsuffixed # (or suffixed-by-"_r") counterparts, except with an extra first # argument of opaque type timezone_t that specifies the timezone. # "tzalloc" allocates a timezone_t value, and "tzfree" frees it. # # If you want to allocate state structures in localtime, add # -DALL_STATE # to the end of the "CFLAGS=" line. Storage is obtained by calling malloc. # # NIST-PCTS:151-2, Version 1.4, (1993-12-03) is a test suite put # out by the National Institute of Standards and Technology # which claims to test C and POSIX conformance. If you want to pass PCTS, add # -DPCTS # to the end of the "CFLAGS=" line. # # If you want strict compliance with XPG4 as of 1994-04-09, add # -DXPG4_1994_04_09 # to the end of the "CFLAGS=" line. This causes "strftime" to always return # 53 as a week number (rather than 52 or 53) for January days before # January's first Monday when a "%V" format is used and January 1 # falls on a Friday, Saturday, or Sunday. # # POSIX says CFLAGS defaults to "-O 1". # Uncomment the following line and edit its contents as needed. #CFLAGS= -O 1 # The name of a POSIX-like library archiver, its flags, C compiler, # linker flags, and 'make' utility. Ordinarily the defaults suffice. # The commented-out values are the defaults specified by POSIX.1-2024. #AR = ar #ARFLAGS = -rv #CC = c17 #LDFLAGS = #MAKE = make # Where to fetch leap-seconds.list from. leaplist_URI = \ https://hpiers.obspm.fr/iers/bul/bulc/ntp/leap-seconds.list # The file is generated by the IERS Earth Orientation Centre, in Paris. leaplist_TZ = Europe/Paris # # To fetch leap-seconds.list from NIST via a less-secure protocol # and with less-volatile metadata, use these settings: #leaplist_URI = ftp://ftp.boulder.nist.gov/pub/time/leap-seconds.list #leaplist_TZ = America/Denver # The zic command and its arguments. zic= ./zic ZIC= $(zic) $(ZFLAGS) # To shrink the size of installed TZif files, # append "-r @N" to omit data before N-seconds-after-the-Epoch. # To grow the files and work around bugs in older applications, # possibly at the expense of introducing bugs in newer ones, # append "-b fat"; see ZIC_BLOAT_DEFAULT above. # See the zic man page for more about -b and -r. ZFLAGS= # How to use zic to install TZif files. ZIC_INSTALL= $(ZIC) -d '$(DESTDIR)$(TZDIR)' # The name of a POSIX-compliant 'awk' on your system. # mawk 1.3.3 and Solaris 10 /usr/bin/awk do not work. # Also, it is better (though not essential) if 'awk' supports UTF-8, # and unfortunately mawk and busybox awk do not support UTF-8. # Try AWK=gawk or AWK=nawk if your awk has the problems mentioned above. AWK= awk # The full path name of a POSIX-compliant shell, preferably one that supports # the Korn shell's 'select' statement as an extension. # These days, Bash is the most popular. # It should be OK to set this to /bin/sh, on platforms where /bin/sh # lacks 'select' or doesn't completely conform to POSIX, but /bin/bash # is typically nicer if it works. KSHELL= /bin/bash # Name of curl , used for HTML validation # and to fetch leap-seconds.list from upstream. # Set CURL=: to disable use of the Internet. CURL= curl # Name of GNU Privacy Guard , used to sign distributions. GPG= gpg # This expensive test requires USE_LTZ. # To suppress it, define this macro to be empty. CHECK_TIME_T_ALTERNATIVES = check_time_t_alternatives # SAFE_CHAR is a regular expression that matches a safe character. # Some parts of this distribution are limited to safe characters; # others can use any UTF-8 character. # For now, the safe characters are a safe subset of ASCII. # The caller must set the shell variable 'sharp' to the character '#', # since Makefile macros cannot contain '#'. # TAB_CHAR is a single tab character, in single quotes. TAB_CHAR= ' ' SAFE_CHARSET1= $(TAB_CHAR)' !\"'$$sharp'$$%&'\''()*+,./0123456789:;<=>?@' SAFE_CHARSET2= 'ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\^_`' SAFE_CHARSET3= 'abcdefghijklmnopqrstuvwxyz{|}~' SAFE_CHARSET= $(SAFE_CHARSET1)$(SAFE_CHARSET2)$(SAFE_CHARSET3) SAFE_CHAR= '[]'$(SAFE_CHARSET)'-]' # These non-alphabetic, non-ASCII printable characters are # used in commentary or in generated *.txt files # and are not likely to cause confusion. UNUSUAL_OK_CHARSET= §«°±»½¾¿×–‘’“”•→−≤★⟨⟩⯪ # Put this in a bracket expression to match spaces. s = [:space:] # OK_CHAR matches any character allowed in the distributed files. # This is the same as SAFE_CHAR, except that UNUSUAL_OK_CHARSET and # multibyte letters are also allowed so that commentary can contain a # few safe symbols and people's names and can quote non-English sources. OK_CHAR= '[][:alpha:]$(UNUSUAL_OK_CHARSET)'$(SAFE_CHARSET)'-]' # SAFE_LINE matches a line of safe characters. # SAFE_SHARP_LINE is similar, except any OK character can follow '#'; # this is so that comments can contain non-ASCII characters. # OK_LINE matches a line of OK characters. SAFE_LINE= '^'$(SAFE_CHAR)'*$$' SAFE_SHARP_LINE='^'$(SAFE_CHAR)'*('$$sharp$(OK_CHAR)'*)?$$' OK_LINE= '^'$(OK_CHAR)'*$$' # Flags to give 'tar' when making a distribution. # Try to use flags appropriate for GNU tar. GNUTARFLAGS= --format=pax --pax-option=delete=atime,delete=ctime \ --numeric-owner --owner=0 --group=0 \ --mode=go+u,go-w --sort=name SETUP_TAR= \ export LC_ALL=C && \ if tar $(GNUTARFLAGS) --version >/dev/null 2>&1; then \ TAR='tar $(GNUTARFLAGS)'; \ else \ TAR=tar; \ fi # Flags to give 'gzip' when making a distribution. GZIPFLAGS= -9n # When comparing .tzs files, use GNU diff's -F'^TZ=' option if supported. # This makes it easier to see which Zone has been affected. SETUP_DIFF_TZS = \ if diff -u -F'^TZ=' - - <>/dev/null >&0 2>&1; then \ DIFF_TZS='diff -u -F^TZ='; \ else \ DIFF_TZS='diff -u'; \ fi # ':' on typical hosts; 'ranlib' on the ancient hosts that still need ranlib. RANLIB= : # POSIX prohibits defining or using SHELL. However, csh users on systems # that use the user shell for Makefile commands may need to define SHELL. #SHELL= /bin/sh # End of macros that one plausibly might want to tailor. ############################################################################### TZCOBJS= zic.o TZDOBJS= zdump.o localtime.o strftime.o DATEOBJS= date.o localtime.o strftime.o LIBSRCS= localtime.c asctime.c difftime.c strftime.c LIBOBJS= localtime.o asctime.o difftime.o strftime.o HEADERS= tzfile.h private.h NONLIBSRCS= zic.c zdump.c NEWUCBSRCS= date.c SOURCES= $(HEADERS) $(LIBSRCS) $(NONLIBSRCS) $(NEWUCBSRCS) \ tzselect.ksh workman.sh MANS= newctime.3 newstrftime.3 newtzset.3 time2posix.3 \ tzfile.5 tzselect.8 zic.8 zdump.8 MANTXTS= newctime.3.txt newstrftime.3.txt newtzset.3.txt \ time2posix.3.txt \ tzfile.5.txt tzselect.8.txt zic.8.txt zdump.8.txt \ date.1.txt COMMON= calendars CONTRIBUTING LICENSE Makefile \ NEWS README SECURITY theory.html version WEB_PAGES= tz-art.html tz-how-to.html tz-link.html CHECK_WEB_PAGES=theory.ck tz-art.ck tz-how-to.ck tz-link.ck DOCS= $(MANS) date.1 $(MANTXTS) $(WEB_PAGES) PRIMARY_YDATA= africa antarctica asia australasia \ europe northamerica southamerica YDATA= $(PRIMARY_YDATA) etcetera NDATA= factory TDATA_TO_CHECK= $(YDATA) $(NDATA) backward TDATA= $(YDATA) $(NDATA) $(BACKWARD) ZONETABLES= zone.tab zone1970.tab zonenow.tab TABDATA= iso3166.tab $(TZDATA_TEXT) $(ZONETABLES) LEAP_DEPS= leapseconds.awk leap-seconds.list TZDATA_ZI_DEPS= ziguard.awk zishrink.awk version $(TDATA) \ $(PACKRATDATA) $(PACKRATLIST) DSTDATA_ZI_DEPS= ziguard.awk $(TDATA) $(PACKRATDATA) $(PACKRATLIST) DATA= $(TDATA_TO_CHECK) backzone iso3166.tab leap-seconds.list \ leapseconds $(ZONETABLES) AWK_SCRIPTS= checklinks.awk checknow.awk checktab.awk leapseconds.awk \ ziguard.awk zishrink.awk MISC= $(AWK_SCRIPTS) TZS_YEAR= 2050 TZS_CUTOFF_FLAG= -c $(TZS_YEAR) TZS= to$(TZS_YEAR).tzs TZS_NEW= to$(TZS_YEAR)new.tzs TZS_DEPS= $(YDATA) localtime.c private.h \ strftime.c tzfile.h zdump.c zic.c TZDATA_DIST = $(COMMON) $(DATA) $(MISC) # EIGHT_YARDS is just a yard short of the whole ENCHILADA. EIGHT_YARDS = $(TZDATA_DIST) $(DOCS) $(SOURCES) tzdata.zi ENCHILADA = $(EIGHT_YARDS) $(TZS) # Consult these files when deciding whether to rebuild the 'version' file. # This list is not the same as the output of 'git ls-files', since # .gitignore is not distributed. VERSION_DEPS= \ calendars CONTRIBUTING LICENSE Makefile NEWS README SECURITY \ africa antarctica asctime.c asia australasia \ backward backzone \ checklinks.awk checknow.awk checktab.awk \ date.1 date.c difftime.c \ etcetera europe factory iso3166.tab \ leap-seconds.list leapseconds.awk localtime.c \ newctime.3 newstrftime.3 newtzset.3 northamerica \ private.h southamerica strftime.c theory.html \ time2posix.3 tz-art.html tz-how-to.html tz-link.html \ tzfile.5 tzfile.h tzselect.8 tzselect.ksh \ workman.sh zdump.8 zdump.c zic.8 zic.c \ ziguard.awk zishrink.awk \ zone.tab zone1970.tab zonenow.tab all: tzselect zic zdump libtz.a $(TABDATA) \ vanguard.zi main.zi rearguard.zi ALL: all date $(ENCHILADA) install: all $(DATA) $(REDO) $(MANS) mkdir -p '$(DESTDIR)$(BINDIR)' \ '$(DESTDIR)$(ZDUMPDIR)' '$(DESTDIR)$(ZICDIR)' \ '$(DESTDIR)$(LIBDIR)' \ '$(DESTDIR)$(MANDIR)/man3' '$(DESTDIR)$(MANDIR)/man5' \ '$(DESTDIR)$(MANDIR)/man8' $(ZIC_INSTALL) -l $(LOCALTIME) \ -t '$(DESTDIR)$(TZDEFAULT)' cp -f $(TABDATA) '$(DESTDIR)$(TZDIR)/.' cp tzselect '$(DESTDIR)$(BINDIR)/.' cp zdump '$(DESTDIR)$(ZDUMPDIR)/.' cp zic '$(DESTDIR)$(ZICDIR)/.' cp libtz.a '$(DESTDIR)$(LIBDIR)/.' $(RANLIB) '$(DESTDIR)$(LIBDIR)/libtz.a' cp -f newctime.3 newtzset.3 '$(DESTDIR)$(MANDIR)/man3/.' cp -f tzfile.5 '$(DESTDIR)$(MANDIR)/man5/.' cp -f tzselect.8 zdump.8 zic.8 '$(DESTDIR)$(MANDIR)/man8/.' INSTALL: ALL install date.1 mkdir -p '$(DESTDIR)$(BINDIR)' '$(DESTDIR)$(MANDIR)/man1' cp date '$(DESTDIR)$(BINDIR)/.' cp -f date.1 '$(DESTDIR)$(MANDIR)/man1/.' # Calculate version number from git, if available. # Otherwise, use $(VERSION) unless it is "unknown" and there is already # a 'version' file, in which case reuse the existing 'version' contents # and append "-dirty" if the contents do not already end in "-dirty". version: $(VERSION_DEPS) { (type git) >/dev/null 2>&1 && \ V=$$(git describe --match '[0-9][0-9][0-9][0-9][a-z]*' \ --abbrev=7 --dirty) || \ if test '$(VERSION)' = unknown && read -r V <$@; then \ V=$${V%-dirty}-dirty; \ else \ V='$(VERSION)'; \ fi; } && \ printf '%s\n' "$$V" >$@.out mv $@.out $@ # These files can be tailored by setting BACKWARD, PACKRATDATA, PACKRATLIST. vanguard.zi main.zi rearguard.zi: $(DSTDATA_ZI_DEPS) $(AWK) \ -v DATAFORM=$(@:.zi=) \ -v PACKRATDATA='$(PACKRATDATA)' \ -v PACKRATLIST='$(PACKRATLIST)' \ -f ziguard.awk \ $(TDATA) $(PACKRATDATA) >$@.out mv $@.out $@ # This file has a version comment that attempts to capture any tailoring # via BACKWARD, DATAFORM, PACKRATDATA, PACKRATLIST, and REDO. tzdata.zi: $(DATAFORM).zi version zishrink.awk read -r version $@.out mv $@.out $@ tzdir.h: printf '%s\n' >$@.out \ '#ifndef TZDEFAULT' \ '# define TZDEFAULT "$(TZDEFAULT)" /* default zone */' \ '#endif' \ '#ifndef TZDIR' \ '# define TZDIR "$(TZDIR)" /* TZif directory */' \ '#endif' mv $@.out $@ version.h: version read -r VERSION $@.out mv $@.out $@ zdump: $(TZDOBJS) $(CC) -o $@ $(CFLAGS) $(LDFLAGS) $(TZDOBJS) $(LDLIBS) zic: $(TZCOBJS) $(CC) -o $@ $(CFLAGS) $(LDFLAGS) $(TZCOBJS) $(LDLIBS) leapseconds: $(LEAP_DEPS) $(AWK) -v EXPIRES_LINE=$(EXPIRES_LINE) \ -f leapseconds.awk leap-seconds.list >$@.out mv $@.out $@ # Awk script to extract a Git-style author from leap-seconds.list comments. EXTRACT_AUTHOR = \ author_line { sub(/^.[[:space:]]*/, ""); \ sub(/:[[:space:]]*/, " <"); \ printf "%s>\n", $$0; \ success = 1; \ exit \ } \ /Questions or comments to:/ { author_line = 1 } \ END { exit !success } # Fetch leap-seconds.list from upstream. fetch-leap-seconds.list: $(CURL) -OR $(leaplist_URI) # Fetch leap-seconds.list from upstream and commit it to the local repository. commit-leap-seconds.list: fetch-leap-seconds.list author=$$($(AWK) '$(EXTRACT_AUTHOR)' leap-seconds.list) && \ date=$$(TZ=$(leaplist_TZ) stat -c%y leap-seconds.list) && \ git commit --author="$$author" --date="$$date" -m'make $@' \ leap-seconds.list # Arguments to pass to submakes. # They can be overridden by later submake arguments. INSTALLARGS = \ BACKWARD='$(BACKWARD)' \ DESTDIR='$(DESTDIR)' \ PACKRATDATA='$(PACKRATDATA)' \ PACKRATLIST='$(PACKRATLIST)' \ TZDEFAULT='$(TZDEFAULT)' \ TZDIR='$(TZDIR)' \ ZIC='$(ZIC)' INSTALL_DATA_DEPS = zic leapseconds tzdata.zi posix_only: $(INSTALL_DATA_DEPS) $(ZIC_INSTALL) tzdata.zi right_only: $(INSTALL_DATA_DEPS) $(ZIC_INSTALL) -L leapseconds tzdata.zi # In earlier versions of this makefile, the other two directories were # subdirectories of $(TZDIR). However, this led to configuration errors. # For example, with posix_right under the earlier scheme, # TZ='right/Australia/Adelaide' got you localtime with leap seconds, # but gmtime without leap seconds, which led to problems with applications # like sendmail that subtract gmtime from localtime. # Therefore, the other two directories are now siblings of $(TZDIR). # You must replace all of $(TZDIR) to switch from not using leap seconds # to using them, or vice versa. right_posix: right_only rm -fr '$(DESTDIR)$(TZDIR)-leaps' ln -s '$(TZDIR_BASENAME)' '$(DESTDIR)$(TZDIR)-leaps' || \ $(MAKE) $(INSTALLARGS) TZDIR='$(TZDIR)-leaps' right_only $(MAKE) $(INSTALLARGS) TZDIR='$(TZDIR)-posix' posix_only posix_right: posix_only rm -fr '$(DESTDIR)$(TZDIR)-posix' ln -s '$(TZDIR_BASENAME)' '$(DESTDIR)$(TZDIR)-posix' || \ $(MAKE) $(INSTALLARGS) TZDIR='$(TZDIR)-posix' posix_only $(MAKE) $(INSTALLARGS) TZDIR='$(TZDIR)-leaps' right_only zones: $(REDO) # dummy.zd is not a real file; it is mentioned here only so that the # top-level 'make' does not have a syntax error. ZDS = dummy.zd # Rule used only by submakes invoked by the $(TZS_NEW) rule. # It is separate so that GNU 'make -j' can run instances in parallel. $(ZDS): zdump ./zdump -i $(TZS_CUTOFF_FLAG) "$$PWD/$(@:.zd=)" >$@ TZS_NEW_DEPS = tzdata.zi zdump zic $(TZS_NEW): $(TZS_NEW_DEPS) rm -fr tzs$(TZS_YEAR).dir mkdir tzs$(TZS_YEAR).dir $(zic) -d tzs$(TZS_YEAR).dir tzdata.zi $(AWK) '/^L/{print "Link\t" $$2 "\t" $$3}' \ tzdata.zi | LC_ALL=C sort >$@.out x=$$($(AWK) '/^Z/{print "tzs$(TZS_YEAR).dir/" $$2 ".zd"}' \ tzdata.zi \ | LC_ALL=C sort -t . -k 2,2) && \ set x $$x && \ shift && \ ZDS=$$* && \ $(MAKE) TZS_CUTOFF_FLAG="$(TZS_CUTOFF_FLAG)" \ ZDS="$$ZDS" $$ZDS && \ sed 's,^TZ=".*\.dir/,TZ=",' $$ZDS >>$@.out rm -fr tzs$(TZS_YEAR).dir mv $@.out $@ # If $(TZS) exists but 'make tzs.ck' fails, a maintainer should inspect the # failed output and fix the inconsistency, perhaps by running 'make force_tzs'. $(TZS): touch $@ force_tzs: $(TZS_NEW) cp $(TZS_NEW) $(TZS) libtz.a: $(LIBOBJS) rm -f $@ $(AR) $(ARFLAGS) $@ $(LIBOBJS) $(RANLIB) $@ date: $(DATEOBJS) $(CC) -o $@ $(CFLAGS) $(LDFLAGS) $(DATEOBJS) $(LDLIBS) tzselect: tzselect.ksh version read -r VERSION $@.out chmod +x $@.out mv $@.out $@ check: check_mild back.ck now.ck check_mild: check_web check_zishrink \ character-set.ck white-space.ck links.ck mainguard.ck \ name-lengths.ck news.ck slashed-abbrs.ck sorted.ck \ tables.ck ziguard.ck tzs.ck # True if UTF8_LOCALE does not work; # otherwise, false but with LC_ALL set to $(UTF8_LOCALE). UTF8_LOCALE_MISSING = \ { test ! '$(UTF8_LOCALE)' \ || ! printf 'A\304\200B\n' \ | LC_ALL='$(UTF8_LOCALE)' grep -q '^A[[:alpha:]]B$$' >/dev/null 2>&1 \ || { export LC_ALL='$(UTF8_LOCALE)'; false; }; } character-set.ck: $(ENCHILADA) $(UTF8_LOCALE_MISSING) || { \ sharp='#' && \ ! grep -Env $(SAFE_LINE) $(MANS) date.1 \ $(MISC) $(SOURCES) \ LICENSE \ version tzdata.zi && \ ! grep -Env $(SAFE_LINE)'|^UNUSUAL_OK_'$(OK_CHAR)'*$$' \ Makefile && \ ! grep -Env $(SAFE_SHARP_LINE) $(TDATA_TO_CHECK) backzone \ leapseconds zone.tab && \ ! grep -Env $(OK_LINE) $(ENCHILADA); \ } touch $@ white-space.ck: $(ENCHILADA) $(UTF8_LOCALE_MISSING) || { \ patfmt=' \t|[\f\r\v]' && pat=$$(printf "$$patfmt\\n") && \ ! grep -En "$$pat|[$s]\$$" \ $(ENCHILADA:leap-seconds.list=); \ } touch $@ PRECEDES_FILE_NAME = ^(Zone|Link[$s]+[^$s]+)[$s]+ FILE_NAME_COMPONENT_TOO_LONG = $(PRECEDES_FILE_NAME)[^$s]*[^/$s]{15} name-lengths.ck: $(TDATA_TO_CHECK) backzone :;! grep -En '$(FILE_NAME_COMPONENT_TOO_LONG)' \ $(TDATA_TO_CHECK) backzone touch $@ mainguard.ck: main.zi test '$(PACKRATLIST)' || \ cat $(TDATA) $(PACKRATDATA) | diff -u - main.zi touch $@ PRECEDES_STDOFF = ^(Zone[$s]+[^$s]+)?[$s]+ STDOFF = [-+]?[0-9:.]+ RULELESS_SAVE = (-|$(STDOFF)[sd]?) RULELESS_SLASHED_ABBRS = \ $(PRECEDES_STDOFF)$(STDOFF)[$s]+$(RULELESS_SAVE)[$s]+[^$s]*/ slashed-abbrs.ck: $(TDATA_TO_CHECK) :;! grep -En '$(RULELESS_SLASHED_ABBRS)' $(TDATA_TO_CHECK) touch $@ CHECK_CC_LIST = { n = split($$1,a,/,/); for (i=2; i<=n; i++) print a[1], a[i]; } sorted.ck: backward backzone $(AWK) '/^Link/ {printf "%.5d %s\n", g, $$3} !/./ {g++}' \ backward | LC_ALL=C sort -cu $(AWK) '/^Zone.*\// {print $$2}' backzone | LC_ALL=C sort -cu touch $@ back.ck: checklinks.awk $(TDATA_TO_CHECK) $(AWK) \ -v DATAFORM=$(DATAFORM) \ -v backcheck=backward \ -f checklinks.awk $(TDATA_TO_CHECK) touch $@ links.ck: checklinks.awk tzdata.zi $(AWK) \ -v DATAFORM=$(DATAFORM) \ -f checklinks.awk tzdata.zi touch $@ # Check timestamps from now through 28 years from now, to make sure # that zonenow.tab contains all sequences of planned timestamps, # without any duplicate sequences. In theory this might require # 2800+ years but that would take a long time to check. CHECK_NOW_TIMESTAMP = $$(./date +%s) CHECK_NOW_FUTURE_YEARS = 28 CHECK_NOW_FUTURE_SECS = $(CHECK_NOW_FUTURE_YEARS) * 366 * 24 * 60 * 60 now.ck: checknow.awk date tzdata.zi zdump zic zone1970.tab zonenow.tab rm -fr $@d mkdir $@d ./zic -d $@d tzdata.zi now=$(CHECK_NOW_TIMESTAMP) && \ future=$$(($(CHECK_NOW_FUTURE_SECS) + $$now)) && \ ./zdump -i -t $$now,$$future \ $$(find "$$PWD/$@d"/????*/ -type f -o -type l) \ >$@d/zdump-now.tab && \ ./zdump -i -t 0,$$future \ $$(find "$$PWD/$@d" -name Etc -prune \ -o \( -type f -o -type l \) ! -name '*.tab' -print) \ >$@d/zdump-1970.tab && \ $(AWK) \ -v now=$$now \ -v now_out=$@.out \ -v zdump_table=$@d/zdump-now.tab \ -f checknow.awk zonenow.tab $(AWK) \ 'BEGIN { \ SysVzone["EST5EDT"] = 1; SysVzone["CST6CDT"] = 1; \ SysVzone["MST7MDT"] = 1; SysVzone["PST8PDT"] = 1; \ print "-\t-\tUTC" \ } \ $$1 == "Zone" && !SysVzone[$$2] {print "-\t-\t" $$2}' \ $(PRIMARY_YDATA) backward factory | \ $(AWK) \ -v zdump_table=$@d/zdump-1970.tab \ -f checknow.awk rm -fr $@d touch $@.out mv $@.out $@ tables.ck: checktab.awk $(YDATA) backward zone.tab zone1970.tab for tab in $(ZONETABLES); do \ test "$$tab" = zone.tab && links='$(BACKWARD)' || links=''; \ $(AWK) -f checktab.awk -v zone_table=$$tab $(YDATA) $$links \ || exit; \ done touch $@ tzs.ck: $(TZS) $(TZS_NEW) if test -s $(TZS); then \ $(SETUP_DIFF_TZS) && $$DIFF_TZS $(TZS) $(TZS_NEW); \ else \ cp $(TZS_NEW) $(TZS); \ fi touch $@ check_web: $(CHECK_WEB_PAGES) .SUFFIXES: .ck .html .html.ck: { ! ($(CURL) --version) >/dev/null 2>&1 || \ $(CURL) -sS --url https://validator.w3.org/nu/ -F out=gnu \ -F file=@$<; } >$@.out && \ test ! -s $@.out || { cat $@.out; exit 1; } mv $@.out $@ ziguard.ck: rearguard.zi vanguard.zi ziguard.awk $(AWK) -v DATAFORM=rearguard -f ziguard.awk vanguard.zi | \ diff -u rearguard.zi - $(AWK) -v DATAFORM=vanguard -f ziguard.awk rearguard.zi | \ diff -u vanguard.zi - touch $@ # Check that zishrink.awk does not alter the data, and that ziguard.awk # preserves main-format data. check_zishrink: zishrink-posix.ck zishrink-right.ck zishrink-posix.ck zishrink-right.ck: \ zic leapseconds $(PACKRATDATA) $(PACKRATLIST) \ $(TDATA) $(DATAFORM).zi tzdata.zi rm -fr $@d t-$@d shrunk-$@d mkdir $@d t-$@d shrunk-$@d case $@ in \ *right*) leap='-L leapseconds';; \ *) leap=;; \ esac && \ $(ZIC) $$leap -d $@d $(DATAFORM).zi && \ $(ZIC) $$leap -d shrunk-$@d tzdata.zi && \ case $(DATAFORM),$(PACKRATLIST) in \ main,) \ $(ZIC) $$leap -d t-$@d $(TDATA) && \ $(AWK) '/^Rule/' $(TDATA) | \ $(ZIC) $$leap -d t-$@d - $(PACKRATDATA) && \ diff -r $@d t-$@d;; \ esac diff -r $@d shrunk-$@d rm -fr $@d t-$@d shrunk-$@d touch $@ # Check that NEWS has data release versions and dates in reverse order. news.ck: NEWS grep '^Release [0-9][0-9][0-9][0-9]' NEWS | LC_ALL=C sort -cru sed -n '/ -0000$$/!s/^Release [^ ]*//p' NEWS|LC_ALL=C sort -cru touch $@ clean_misc: rm -fr *.ckd *.dir rm -f *.ck *.core *.o *.out *.t core core.* \ date tzdir.h tzselect version.h zdump zic libtz.a clean: clean_misc rm -fr tzdb-*/ rm -f *.zi $(TZS_NEW) maintainer-clean: clean @echo 'This command is intended for maintainers to use; it' @echo 'deletes files that may need special tools to rebuild.' rm -f leapseconds version $(MANTXTS) $(TZS) *.asc *.tar.* names: @echo $(ENCHILADA) public: check public.ck $(CHECK_TIME_T_ALTERNATIVES) \ tarballs signatures date.1.txt: date.1 newctime.3.txt: newctime.3 newstrftime.3.txt: newstrftime.3 newtzset.3.txt: newtzset.3 time2posix.3.txt: time2posix.3 tzfile.5.txt: tzfile.5 tzselect.8.txt: tzselect.8 zdump.8.txt: zdump.8 zic.8.txt: zic.8 $(MANTXTS): workman.sh LC_ALL=C sh workman.sh $(MANFLAGS) $(@:.txt=) >$@.out mv $@.out $@ # Set file timestamps deterministically if possible, # so that tarballs containing the timestamps are reproducible. # # '$(SET_TIMESTAMP_N) N DEST A B C ...' sets the timestamp of the # file DEST to the maximum of the timestamps of the files A B C ..., # plus N if GNU ls and touch are available. SET_TIMESTAMP_N = sh -c '\ n=$$0 dest=$$1; shift; \ <"$$dest" && \ if test $$n != 0 && \ lsout=$$(ls -nt --time-style="+%s" "$$@" 2>/dev/null); then \ set x $$lsout && \ timestamp=$$(($$7 + $$n)) && \ echo "+ touch -md @$$timestamp $$dest" && \ touch -md @$$timestamp "$$dest"; \ else \ newest=$$(ls -t "$$@" | sed 1q) && \ echo "+ touch -mr $$newest $$dest" && \ touch -mr "$$newest" "$$dest"; \ fi' # If DEST depends on A B C ... in this Makefile, callers should use # $(SET_TIMESTAMP_DEP) DEST A B C ..., for the benefit of any # downstream 'make' that considers equal timestamps to be out of date. # POSIX allows this 'make' behavior, although only HP-UX 'make' # (which is no longer supported) did things that way. # If all that matters is that the timestamp be reproducible # and plausible, use $(SET_TIMESTAMP). SET_TIMESTAMP = $(SET_TIMESTAMP_N) 0 SET_TIMESTAMP_DEP = $(SET_TIMESTAMP_N) 1 # Set the timestamps to those of the git repository, if available, # and if the files have not changed since then. # This uses GNU 'ls --time-style=+%s', which outputs the seconds count, # and GNU 'touch -d@N FILE', where N is the number of seconds since 1970. # If git or GNU is absent, don't bother to sync with git timestamps. # Also, set the timestamp of each prebuilt file like 'leapseconds' # to be the maximum of the files it depends on. set-timestamps.out: $(EIGHT_YARDS) rm -f $@ if (type git) >/dev/null 2>&1 && \ files=$$(git ls-files $(EIGHT_YARDS)) && \ touch -md @1 test.out; then \ rm -f test.out && \ for file in $$files; do \ if git diff --quiet HEAD $$file; then \ time=$$(TZ=UTC0 git log -1 \ --format='tformat:%cd' \ --date='format-local:%Y-%m-%dT%H:%M:%SZ' \ $$file) && \ echo "+ touch -md $$time $$file" && \ touch -md $$time $$file; \ else \ echo >&2 "$$file: warning: does not match repository"; \ fi || exit; \ done; \ fi $(SET_TIMESTAMP_DEP) leapseconds $(LEAP_DEPS) for file in $(MANTXTS); do \ $(SET_TIMESTAMP_DEP) $$file $${file%.txt} workman.sh || \ exit; \ done $(SET_TIMESTAMP_DEP) version $(VERSION_DEPS) $(SET_TIMESTAMP_DEP) tzdata.zi $(TZDATA_ZI_DEPS) touch $@ set-tzs-timestamp.out: $(TZS) $(SET_TIMESTAMP_DEP) $(TZS) $(TZS_DEPS) touch $@ # The zics below ensure that each data file can stand on its own. # We also do an all-files run to catch links to links. public.ck: $(VERSION_DEPS) rm -fr $@d mkdir $@d ln $(VERSION_DEPS) $@d cd $@d \ && $(MAKE) CFLAGS='$(GCC_DEBUG_FLAGS)' TZDIR='$(TZDIR)' ALL for i in $(TDATA_TO_CHECK) \ tzdata.zi vanguard.zi main.zi rearguard.zi; \ do \ $@d/zic -v -d $@d/zoneinfo $@d/$$i || exit; \ done $@d/zic -v -d $@d/zoneinfo-all $(TDATA_TO_CHECK) : : Also check 'backzone' syntax. rm $@d/main.zi cd $@d && $(MAKE) PACKRATDATA=backzone main.zi $@d/zic -d $@d/zoneinfo main.zi rm $@d/main.zi cd $@d && \ $(MAKE) PACKRATDATA=backzone PACKRATLIST=zone.tab main.zi $@d/zic -d $@d/zoneinfo main.zi : rm -fr $@d touch $@ # Check that the code works under various alternative # implementations of time_t. check_time_t_alternatives: $(TIME_T_ALTERNATIVES) $(TIME_T_ALTERNATIVES_TAIL): $(TIME_T_ALTERNATIVES_HEAD) $(TIME_T_ALTERNATIVES): $(VERSION_DEPS) rm -fr $@d mkdir $@d ln $(VERSION_DEPS) $@d case $@ in \ *32_t*) range=-2147483648,2147483648;; \ u*) range=0,4294967296;; \ *) range=-4294967296,4294967296;; \ esac && \ wd=$$PWD && \ zones=$$($(AWK) '/^[^#]/ { print $$3 }' /dev/null; then \ quiet_option='-q'; \ else \ quiet_option=''; \ fi && \ diff $$quiet_option -r $(TIME_T_ALTERNATIVES_HEAD)d/etc \ $@d/etc && \ diff $$quiet_option -r \ $(TIME_T_ALTERNATIVES_HEAD)d/usr/share \ $@d/usr/share; \ } touch $@ TRADITIONAL_ASC = \ tzcode$(VERSION).tar.gz.asc.t \ tzdata$(VERSION).tar.gz.asc.t REARGUARD_ASC = \ tzdata$(VERSION)-rearguard.tar.gz.asc.t ALL_ASC = $(TRADITIONAL_ASC) $(REARGUARD_ASC) \ tzdb-$(VERSION).tar.lz.asc.t tarballs rearguard_tarballs tailored_tarballs traditional_tarballs \ signatures rearguard_signatures traditional_signatures: \ version set-timestamps.out rearguard.zi vanguard.zi read -r VERSION $(@:.t=) $(SET_TIMESTAMP) $(@:.t=) $(COMMON) $(DOCS) $(SOURCES) touch $@ tzdata$(VERSION).tar.gz.t: set-timestamps.out $(SETUP_TAR) && \ $$TAR -cf - $(TZDATA_DIST) | \ gzip $(GZIPFLAGS) >$(@:.t=) $(SET_TIMESTAMP) $(@:.t=) $(TZDATA_DIST) touch $@ # Create empty files with a reproducible timestamp. CREATE_EMPTY = TZ=UTC0 touch -mt 202010122253.00 # The obsolescent *rearguard* targets and related macros are present # for backwards compatibility with tz releases 2018e through 2022a. # They should go away eventually. To build rearguard tarballs you # can instead use 'make DATAFORM=rearguard tailored_tarballs'. tzdata$(VERSION)-rearguard.tar.gz.t: rearguard.zi set-timestamps.out rm -fr $@.dir mkdir $@.dir ln $(TZDATA_DIST) $@.dir cd $@.dir && rm -f $(TDATA) $(PACKRATDATA) version for f in $(TDATA) $(PACKRATDATA); do \ rearf=$@.dir/$$f; \ $(AWK) -v DATAFORM=rearguard -f ziguard.awk $$f >$$rearf && \ $(SET_TIMESTAMP_DEP) $$rearf ziguard.awk $$f || exit; \ done sed '1s/$$/-rearguard/' $@.dir/version : The dummy pacificnew pacifies TZUpdater 2.3.1 and earlier. $(CREATE_EMPTY) $@.dir/pacificnew touch -mr version $@.dir/version $(SETUP_TAR) && \ (cd $@.dir && \ $$TAR -cf - \ $(TZDATA_DIST) pacificnew | \ gzip $(GZIPFLAGS)) >$(@:.t=) $(SET_TIMESTAMP) $(@:.t=) \ $$(cd $@.dir && \ ls $(TZDATA_DIST) pacificnew | sed 's,^,$@.dir/,') touch $@ # Create a tailored tarball suitable for TZUpdater and compatible tools. # For example, 'make DATAFORM=vanguard tailored_tarballs' makes a tarball # useful for testing whether TZUpdater supports vanguard form. # The generated tarball is not byte-for-byte equivalent to a hand-tailored # traditional tarball, as data entries are put into 'etcetera' even if they # came from some other source file. However, the effect should be the same # for ordinary use, which reads all the source files. tzdata$(VERSION)-tailored.tar.gz.t: set-timestamps.out rm -fr $@.dir mkdir $@.dir : The dummy pacificnew pacifies TZUpdater 2.3.1 and earlier. if test $(DATAFORM) = vanguard; then \ pacificnew=; \ else \ pacificnew=pacificnew; \ fi && \ cd $@.dir && \ $(CREATE_EMPTY) $(PRIMARY_YDATA) $(NDATA) backward \ $$pacificnew (sed '/^#/!d' tzdata.zi && echo && cat $(DATAFORM).zi) \ >$@.dir/etcetera touch -mr tzdata.zi $@.dir/etcetera sed -n \ -e '/^# *version *\(.*\)/h' \ -e '/^# *ddeps */H' \ -e '$$!d' \ -e 'g' \ -e 's/^# *version *//' \ -e 's/\n# *ddeps */-/' \ -e 's/ /-/g' \ -e 'p' \ $@.dir/version touch -mr version $@.dir/version links= && \ for file in $(TZDATA_DIST); do \ test -f $@.dir/$$file || links="$$links $$file"; \ done && \ ln $$links $@.dir $(SETUP_TAR) && \ (cd $@.dir && \ $$TAR -cf - *) | gzip $(GZIPFLAGS) >$(@:.t=) $(SET_TIMESTAMP) $(@:.t=) \ $$(cd $@.dir && ls * | sed 's,^,$@.dir/,') touch $@ tzdb-$(VERSION).tar.lz.t: set-timestamps.out set-tzs-timestamp.out rm -fr tzdb-$(VERSION) mkdir tzdb-$(VERSION) ln $(ENCHILADA) tzdb-$(VERSION) $(SET_TIMESTAMP) tzdb-$(VERSION) tzdb-$(VERSION)/* $(SETUP_TAR) && \ $$TAR -cf - tzdb-$(VERSION) | lzip -9 >$(@:.t=) $(SET_TIMESTAMP) $(@:.t=) tzdb-$(VERSION) touch $@ tzcode$(VERSION).tar.gz.asc.t: tzcode$(VERSION).tar.gz.t tzdata$(VERSION).tar.gz.asc.t: tzdata$(VERSION).tar.gz.t tzdata$(VERSION)-rearguard.tar.gz.asc.t: tzdata$(VERSION)-rearguard.tar.gz.t tzdb-$(VERSION).tar.lz.asc.t: tzdb-$(VERSION).tar.lz.t $(ALL_ASC): $(GPG) --armor --detach-sign $(?:.t=) $(SET_TIMESTAMP) $(@:.t=) $(?:.t=) touch $@ zonenames: tzdata.zi @$(AWK) '/^Z/ { print $$2 } /^L/ { print $$3 }' tzdata.zi asctime.o: private.h date.o: private.h difftime.o: private.h localtime.o: private.h tzdir.h tzfile.h strftime.o: localtime.c private.h tzdir.h tzfile.h zdump.o: private.h version.h zic.o: private.h tzdir.h tzfile.h version.h .PHONY: ALL INSTALL all .PHONY: check check_mild check_time_t_alternatives .PHONY: check_web check_zishrink .PHONY: clean clean_misc commit-leap-seconds.list dummy.zd .PHONY: fetch-leap-seconds.list force_tzs .PHONY: install maintainer-clean names .PHONY: posix_only posix_right public .PHONY: rearguard_signatures rearguard_signatures_version .PHONY: rearguard_tarballs rearguard_tarballs_version .PHONY: right_only right_posix signatures signatures_version .PHONY: tarballs tarballs_version .PHONY: traditional_signatures traditional_signatures_version .PHONY: traditional_tarballs traditional_tarballs_version .PHONY: tailored_tarballs tailored_tarballs_version .PHONY: zonenames zones .PHONY: $(ZDS) NEWS0000644000000000000000000077572515257052270006537 0ustar00News for the tz database Release 2026e - 2026-09-29 17:14:38 -0700 Briefly: Manitoba moves to permanent -05 on 2026-10-31. Changes to future timestamps Manitoba’s 2026-03-08 spring forward was its last foreseeable clock change, as it moved to permanent -05 thereafter. Model this with its traditional abbreviation EST. Although the change to permanent -05 legally takes place on 2026-10-31, temporarily model the change to occur on 2026-11-01 at 02:00 for the same reason as other recent temporary hacks. (Caution: see “NOTE FOR 2026b TEMPORARY HACK FOR CLDR AND CANADA” below.) As the change affects both America/Winnipeg and its backward compatibility link, the obsolescent setting TZ="Canada/Central" will now use the abbreviation EST for affected timestamps, akin to TZ="Canada/Pacific" behavior introduced in 2026b. Another TZDB release may be needed soon if any of the portions of northwestern Ontario historically aligned with America/Winnipeg do not follow. For now, given the short leadtime with Manitoba’s announcement, assume that they will, to avoid creating new zones that may not be needed if no divergence occurs. Changes to past timestamps In 1925 Ireland fell back on 09-20 not 10-04 (thanks to Stan Ulbrych). Changes to commentary Discussion on timekeeping practices in northwestern Ontario has been expanded in light of Manitoba’s recent announcement. It is not yet known whether or to what extent these areas may adapt their own timekeeping practices in response. Release 2026d - 2026-09-11 15:21:07 -0700 Briefly: Canada’s Northwest Territories moved to permanent -06 on 2026-08-21. Obsolescent settings like TZ="EST5EDT" now conform better to POSIX. Fix security, performance and porting bugs in zic and localtime. Changes to future timestamps Canada’s Northwest Territories will not fall back on 2026-11-01 and will stay on -06 year-round, matching Alberta’s recent change. Model this with its traditional abbreviation CST. Although the change to permanent -06 legally took place on 2026-08-21, temporarily model the change to occur on 2026-11-01 at 02:00 for the same reason as other recent temporary hacks. (Caution: see “NOTE FOR 2026b TEMPORARY HACK FOR CLDR AND CANADA” below.) This affects only America/Inuvik as the rest of the territory is covered by America/Edmonton, for which the equivalent change was released in 2026c. Changes to past timestamps Colombia’s 1992-05-02 spring forward was at 00:00, not 24:00. Iran’s 1979-05-26 spring forward was at 00:00, not 24:00. (Thanks to N.F. Hase.) The backward-compatibility names EST5EDT, CST6CDT, MST7MDT, and PST8PDT now conform better to POSIX. For example, EST5EDT now always uses the abbreviation "EST" for standard time (now always 5 hours behind UT) and "EDT" for daylight saving time, whereas it formerly had different UT offsets before standard time was introduced and sometimes used abbreviations like "LMT", "EWT" and "EPT", all contrary to POSIX. Also, though not required by POSIX these names now use US federal rules rather than rules of places like New York, reverting to 2024a behavior. This change affects only timestamps before 1966-10-30 at 01:00 standard time. Other data changes The temporary hacks used for North American timekeeping changes now work around a libstdc++ std::chrono bug in GCC 14.1-14.4, 15.1-15.2, and 16.1; see GCC bug 124851. This data change does not affect TZif files or timestamps. The change does not work around the related but less serious GCC bugs 116110 and 124513. These GCC bugs are all fixed in GCC 16.2. Changes to code zic now rejects Link targets that would have invalid names, and more efficiently processes Expires, Leap and Rule lines with years far in the past or future. (Thanks to Darren Carreras.) zic now ports to systems that report lack of link support via EINVAL, ENOSYS or EPERM errno values. (Thanks to Tom Lane.) When tzset and related functions encounter a TZif file that is too large for them, they now consistently fail instead of sometimes silently ignoring excess parts of the file. localtime-related functions no longer mishandle extreme timestamps when given TZif files holding some unlikely timezone histories. (Problem reported by David Sarkisyan.) localtime-related functions no longer check the values of TZif files’ standard/wall and UT/local indicators, which these functions have not used since 2026a’s removal of the old posixrules feature. tzcode has been ported to Haiku. localtime.c now works again by default on AIX and DragonFly BSD. zic now rejects ‘:’ and ‘\’ in Zone and Link names when running on Microsoft Windows. (Problem reported by David Diaz.) Changes to documentation URLs for release tarballs in tz-link.html have been updated to reflect their new canonical URLs on data.iana.org. Release 2026c - 2026-07-08 10:23:58 -0700 Briefly: Alberta moved to permanent -06 on 2026-06-18. Morocco moves to permanent +00 on 2026-09-20. More integer overflow bugs have been fixed in zic. Changes to future timestamps Alberta’s 2026-03-08 spring forward was its last foreseeable clock change, as it moved to permanent -06 thereafter. (Thanks to Roozbeh Pournader and others.) Model this with its traditional abbreviation CST. Although the change to permanent -06 legally took place on 2026-06-18, temporarily model the change to occur on 2026-11-01 at 02:00 instead, for the same reason we introduced a similarly temporary hack for British Columbia in 2026b. (Caution: see “NOTE FOR 2026b TEMPORARY HACK FOR CLDR AND CANADA” below.) As the change affects both America/Edmonton and its backward compatibility link, the obsolescent setting TZ="Canada/Mountain" will now use the abbreviation CST for affected timestamps, akin to TZ="Canada/Pacific" behavior introduced in 2026b. Although another TZDB release will likely be needed soon because Northwest Territories will likely follow Alberta, the legal formalities have not yet taken place. Morocco plans to move back to permanent UTC, without daylight saving time transitions, on 2026-09-20 at 02:00. This also affects Western Sahara. Changes to code zic no longer overflows integers when processing outlandish input like ‘Zone Ouch 0 - LMT 9223372036854775807’, ‘Zone Ouch 0 2562047788015215 LMT’, ‘Zone Ouch -2562047788015215:30:08 - LMT’, and ‘Zone Ouch -2562047788015215:30:08 - %%z’. This avoids undefined behavior in C. (Problems reported by Naveed Khan.) On platforms that have EFTYPE, tzalloc now fails with errno set to EFTYPE, not EINVAL, if it detects that the TZif file has an invalid format or is not a regular file. Formerly it did this only on NetBSD, and only when the file was not a regular file. Unprivileged programs no longer require TZif files to be regular files or reject relative names containing ".." components. This reverts to the more-permissive 2025b behavior, as the stricter behavior did not catch on in FreeBSD. zic now reports any failure to remove a temporary file when cleaning up after a previous failure. (Problem reported by Tom Lane.) Changes to commentary Northwest Territories is expected to move to permanent -06 prior to 2026-11-01 02:00, when clocks would otherwise fall back. (Thanks to Tim Parenti and James Bellaire.) Model this with its traditional abbreviation CST. Unfortunately the change is not yet official, so it is currently present only as comments that can be uncommented as needed. Changes to build procedure The undocumented ‘typecheck’ Makefile check rule has been removed. It stopped working in 2025a and evidently nobody noticed. The rule was superseded by ‘check_time_t_alternatives’ in 2013d. Release 2026b - 2026-04-22 23:06:43 -0700 Briefly: British Columbia moved to permanent -07 on 2026-03-09. zic no longer mishandles a last transition to a new time type. Some more overflow bugs have been fixed in zic. Changes to future timestamps British Columbia’s 2026-03-08 spring forward was its last foreseeable clock change, as it moved to permanent -07 thereafter. (Thanks to Arthur David Olson.) Although the change to permanent -07 legally took place on 2026-03-09, temporarily model the change to occur on 2026-11-01 at 02:00 instead. This works around a limitation in CLDR 48.1 (2026-01-08). This temporary hack is planned to be removed after CLDR is fixed. To match longstanding practice for -07 standard time in North America, such as for America/Dawson, America/Dawson_Creek, America/Fort_Nelson and America/Whitehorse, affected BC timestamps use the abbreviation MST. As the change affects both America/Vancouver and its backward compatibility link, the obsolescent setting TZ="Canada/Pacific" will now use the abbreviation MST. Future TZDB releases may change the abbreviation if something else becomes popular across areas observing -07, and may change it to "-07" if no alphabetic consensus emerges. Changes to code zic no longer mishandles a last transition to a new time type. NOTE FOR 2026b TEMPORARY HACK FOR CLDR AND CANADA: This zic fix is needed for the temporary hack (mentioned above) that works around the Canadian timekeeping bug in Unicode CLDR. Without the fix, the temporary hack causes zic versions 2023d through 2026a, in their default mode that generates slim output, to generate a TZif file that violates Internet RFC 9636 §3.3. The buggy file in turn causes some TZif readers, including tzcode itself, to ignore America/Vancouver’s 2026-11-01 02:00 transition from PDT (tm_isdst=1) to MST (tm_isdst=0). Although the buggy file does not cause any known TZif reader to mishandle UT offsets, caution is advised when using zic 2023d through 2026a to compile data from more-recent tz releases. To work around this problem when using these older zic versions, use ‘zic -b fat’. zic no longer overflows a buffer when generating a TZ string like "PST-167:59:58PDT-167:59:59,M11.5.6/-167:59:59,M12.5.6/-167:59:59", which can occur with adversarial input. (Thanks to Naveed Khan.) zic no longer generates a longer TZif file than necessary when an earlier time zone abbreviation is a suffix of a later one. As a nice side effect, zic no longer overflows a buffer when given a long series of abbreviations, each a suffix of the next. (Buffer overflow reported by Arthur Chan.) zic no longer overflows an int when processing input like ‘Zone Ouch 2147483648:00:00 - LMT’. The int overflow can lead to buffer overflow in adversarial cases. (Thanks to Naveed Khan.) zic now checks for signals more often. Release 2026a - 2026-03-01 22:59:49 -0800 Briefly: Moldova has used EU transition times since 2022. The “right” TZif files are no longer installed by default. -DTZ_RUNTIME_LEAPS=0 disables runtime support for leap seconds. TZif files are no longer limited to 50 bytes of abbreviations. zic is no longer limited to 50 leap seconds. Several integer overflow bugs have been fixed. Changes to past and future timestamps Since 2022 Moldova has observed EU transition times, that is, it has sprung forward at 03:00, not 02:00, and has fallen back at 04:00, not 03:00. (Thanks to Heitor David Pinto.) Changes to data Remove Europe/Chisinau from zonenow.tab, as it now agrees with Europe/Athens for future timestamps. Changes to build procedure The Makefile no longer by default installs an alternate set of TZif files for system clocks that count leap seconds. Install with ‘make REDO=posix_right’ to get the old default, which is rarely used in major downstream distributions. If your system clock counts leap seconds (contrary to POSIX), it is better to install with ‘make REDO=right_only’. This change does not affect the leapseconds file, which is still installed as before. The Makefile’s POSIXRULES option, which was declared obsolete in release 2019b, has been removed. The Makefile’s build procedure thus no longer optionally installs the obsolete posixrules file. Changes to code Compiling with the new option -DTZ_RUNTIME_LEAPS=0 disables runtime support for leap seconds. Although this conforms to POSIX, shrinks tzcode’s attack surface, and is more efficient, it fails to support Internet RFC 9636’s leap seconds. zic now can generate, and localtime.c can now use, TZif files that hold up to 256 bytes of abbreviations, counting trailing NULs. The previous limit was 50 bytes, and some tzdata TZif files were already consuming 40 bytes. zic -v warns if it generates a file that exceeds the old 50-byte limit. zic -L can now generate TZif files with more than 50 leap seconds. This helps test TZif readers not limited to 50 leap seconds, as tzcode’s localtime.c is; it has little immediate need for practical timekeeping as there have been only 27 leap seconds and possibly there will be no more, due to planned changes to UTC. zic -v warns if its output exceeds the old 50-second limit. localtime.c no longer accesses the posixrules file generated by zic -p. Hence for obsolete and nonconforming settings like TZ="AST4ADT" it now typically falls back on US DST rules, rather than attempting to override this fallback with the contents of the posixrules file. This removes library support that was declared obsolete in release 2019b, and fixes some undefined behavior. (Undefined behavior reported by Naveed Khan.) The posix2time, posix2time_z, time2posix, and time2posix_z functions now set errno=EOVERFLOW and return ((time_t) -1) if the result is not representable. Formerly they had undefined behavior that could in practice result in crashing, looping indefinitely, or returning an incorrect result. As before, these functions are defined only when localtime.c is compiled with the -DSTD_INSPIRED option. Some other undefined behavior, triggered by TZif files containing outlandish but conforming UT offsets or leap second corrections, has also been fixed. (Some of these bugs reported by Naveed Khan.) localtime.c no longer rejects TZif files that exactly fit in its internal structures, fixing off-by-one typos introduced in 2014g. zic no longer generates a no-op transition when simultaneous Rule and Zone changes cancel each other out. This occurs in tzdata only in Asia/Tbilisi on 1997-03-30. (Thanks to Renchunhui for a test case showing the bug.) zic no longer assumes you can fflush a read-only stream. (Problem reported by Christos Zoulas.) zic no longer generates UT offsets equal to -2**31 and localtime.c no longer accepts them, as they can cause trouble in both localtime.c and its callers. RFC 9636 prohibits such offsets. zic -p now warns that the -p option is obsolete and likely ineffective. Release 2025c - 2025-12-10 14:42:37 -0800 Briefly: Several code changes for compatibility with FreeBSD. Changes to past timestamps Baja California agreed with California’s DST rules in 1953 and in 1961 through 1975, instead of observing standard time all year. (Thanks to Alois Treindl.) Changes to build procedure Files in distributed tarballs now have correct commit times. Formerly, the committer’s time zone was incorrectly ignored. Distribution products (*.asc, *.gz, and *.lz) now have reproducible timestamps. Formerly, only the contents of the compressed tarballs had reproducible timestamps. By default, distributed formatted man pages (*.txt) now use UTF-8 and are left-adjusted more consistently. A new Makefile macro MANFLAGS can override these defaults. (Thanks to G. Branden Robinson for inspiring these changes.) Changes to code An unset TZ is no longer invalid when /etc/localtime is missing, and is abbreviated "UTC" not "-00". This reverts to 2024b behavior. (Problem and patch reported by Dag-Erling Smørgrav.) New function offtime_r, short for fixed-offset localtime_rz. It is defined if STD_INSPIRED is defined. (Patch from Dag-Erling Smørgrav.) tzset etc. are now more cautious about questionable TZ settings. Privileged programs now reject TZ settings that start with '/', unless they are TZDEFAULT (default "/etc/localtime") or start with TZDIR then '/' (default "/usr/share/zoneinfo/"). Unprivileged programs now require files to be regular files and reject relative names containing ".." directory components; formerly, only privileged programs did those two things. These changes were inspired by similar behavior in FreeBSD. On NetBSD, unprivileged programs now use O_REGULAR to check whether a TZ setting starting with '/' names a regular file, avoiding a minor security race still present elsewhere. TZ strings taken from tzalloc arguments are now treated with no less caution than TZ strings taken from the environment, as the old undocumented behavior would have been hard to explain. tzset etc. no longer use the ‘access’ system call to check access; instead they now use the system calls issetugid, getauxval, getresuid/getresgid, and geteuid/getegid/getuid/getgid (whichever first works) to test whether a program is privileged. Compile with -DHAVE_SYS_AUXV_H=[01] to enable or disable which (if it defines AT_SECURE) enables getauxval, and compile with -DHAVE_ISSETUGID=[01], -DHAVE_GETRESUID=[01], and -DHAVE_GETEUID=[01] to enable or disable the other calls’ use. The new CFLAGS option -DTZ_CHANGE_INTERVAL=N makes tzset etc. check for TZif file changes if the in-memory data are N seconds old or more, and are derived from the TZ environment variable. This is intended for platforms that want tzset etc. to reflect changes to whatever file TZ selects (including changes to /etc/localtime if TZ is unset). If N is negative (the default) these checks are omitted; this is the traditional behavior. The new CFLAGS options -DHAVE_STRUCT_STAT_ST_CTIM=0 and -DHAVE_STRUCT_TIMESPEC=0 port to non-POSIX.1-2008 platforms that lack st_ctim and struct timespec, respectively. On these platforms, the code falls back on st_ctime to implement -DTZ_CHANGE_INTERVAL=N. tzset etc. now treat ' ' like '_' in time zone abbreviations, just as they treat other invalid bytes. This continues the transition begun in release 96k, which removed spaces in tzdata because the spaces break time string parsers. The new CFLAGS option -DTHREAD_PREFER_SINGLE causes tzcode in single-threaded processes to avoid locks, as FreeBSD does. This can save time in single-threaded apps. The threadedness testing costs CPU time and energy in multi-threaded apps. New options -DHAVE___ISTHREADED and -DHAVE_SYS_SINGLE_THREADED_H can help configure how to test for single-threadedness. The new CFLAGS option -DTHREAD_RWLOCK uses read-write locks, as macOS does, instead of mutexes. This saves real time when TZ is rarely changing and many threads call tzcode simultaneously. It costs more CPU time and energy. The new CFLAGS option -DTHREAD_TM_MULTI causes localtime to return a pointer to thread-specific memory, as FreeBSD does, instead of to the same memory in all threads. This supports nonportable programs that incorrectly use localtime instead of localtime_r. This option affects gmtime and offtime similarly to localtime. Because the corresponding storage is freed on thread exit, this option is incompatible with POSIX.1-2024 and earlier. It also costs CPU time and memory. tzfree now preserves errno, consistently with POSIX.1-2024 ‘free’. tzcode now uses mempcpy if available, guessing its availability. Compile with -DHAVE_MEMPCPY=1 or 0 to override the guess. tzcode now uses strnlen to improve asymptotic performance a bit. Compile with -DHAVE_STRNLEN=0 if your platform lacks it. tzcode now hand-declares unistd.h-provided symbols like getopt if HAVE_UNISTD_H=0, not if HAVE_POSIX_DECLS=0. tzset etc. now have an experimental OPENAT_TZDIR option; see Makefile and localtime.c for details. On platforms like GNU/Hurd that do not define PATH_MAX, exceedingly long TZ strings no longer fail merely because they exceed an arbitrary file name length limit imposed by tzcode. zic has new options inspired by FreeBSD. ‘-D’ skips creation of output ancestor directories, ‘-m MODE’ sets output files’ mode, and ‘-u OWNER[:GROUP]’ sets output files’ owner and group. zic now uses the fdopen function, which was standardized by POSIX.1-1988 and is now safe to use in portable code. This replaces its use of the older umask function, which complicated maintenance. Changes to commentary The leapseconds file contains commentary about the IERS and NIST last-modified and expiration timestamps for leap second data. (Thanks to Judah Levine.) Commentary now also uses characters from the set –‘’“”•≤ as this can be useful and should work with current applications. This also affects data in iso3166.tab and zone1970.tab, which now contain strings like “Côte d’Ivoire” instead of “Côte d'Ivoire”. Release 2025b - 2025-03-22 13:40:46 -0700 Briefly: New zone for Aysén Region in Chile which moves from -04/-03 to -03. Changes to future timestamps Chile's Aysén Region moves from -04/-03 to -03 year-round, joining Magallanes Region. The region will not change its clocks on 2025-04-05 at 24:00, diverging from America/Santiago and creating a new zone America/Coyhaique. (Thanks to Yonathan Dossow.) Model this as a change to standard offset effective 2025-03-20. Changes to past timestamps Iran switched from +04 to +0330 on 1978-11-10 at 24:00, not at year end. (Thanks to Roozbeh Pournader.) Changes to code 'zic -l TIMEZONE -d . -l /some/other/file/system' no longer attempts to create an incorrect symlink, and no longer has a read buffer underflow. (Problem reported by Evgeniy Gorbanev.) Release 2025a - 2025-01-15 10:47:24 -0800 Briefly: Paraguay adopted permanent -03 starting spring 2024. Improve pre-1991 data for the Philippines. Etc/Unknown is now reserved. Changes to future timestamps Paraguay stopped changing its clocks after the spring-forward transition on 2024-10-06, so it is now permanently at -03. (Thanks to Heitor David Pinto and Even Scharning.) This affects timestamps starting 2025-03-22, as well as the obsolescent tm_isdst flags starting 2024-10-15. Changes to past timestamps Correct timestamps for the Philippines before 1900, and from 1937 through 1990. (Thanks to P Chan for the heads-up and citations.) This includes adjusting local mean time before 1899; fixing transitions in September 1899, January 1937, and June 1954; adding transitions in December 1941, November 1945, March and September 1977, and May and July 1990; and removing incorrect transitions in March and September 1978. Changes to data Add zone1970.tab lines for the Concordia and Eyre Bird Observatory research stations. (Thanks to Derick Rethans and Jule Dabars.) Changes to code strftime %s now generates the correct numeric string even when the represented number does not fit into time_t. This is better than generating the numeric equivalent of (time_t) -1, as strftime did in TZDB releases 96a (when %s was introduced) through 2020a and in releases 2022b through 2024b. It is also better than failing and returning 0, as strftime did in releases 2020b through 2022a. strftime now outputs an invalid conversion specifier as-is, instead of eliding the leading '%', which confused debugging. An invalid TZ now generates the time zone abbreviation "-00", not "UTC", to help the user see that an error has occurred. (Thanks to Arthur David Olson for suggesting a "wrong result".) mktime and timeoff no longer incorrectly fail merely because a struct tm component near INT_MIN or INT_MAX overflows when a lower-order component carries into it. TZNAME_MAXIMUM, the maximum number of bytes in a proleptic TZ string's time zone abbreviation, now defaults to 254 not 255. This helps reduce the size of internal state from 25480 to 21384 on common platforms. This change should not be a problem, as nobody uses such long "abbreviations" and the longstanding tzcode maximum was 16 until release 2023a. For those who prefer no arbitrary limits, you can now specify TZNAME_MAXIMUM values up to PTRDIFF_MAX, a limit forced by C anyway; formerly tzcode silently misbehaved unless TZNAME_MAXIMUM was less than INT_MAX. tzset and related functions no longer leak a file descriptor if another thread forks or execs at about the same time and if the platform has O_CLOFORK and O_CLOEXEC respectively. Also, the functions no longer let a TZif file become a controlling terminal. 'zdump -' now reads TZif data from /dev/stdin. (From a question by Arthur David Olson.) Changes to documentation The name Etc/Unknown is now reserved: it will not be used by TZDB. This is for compatibility with CLDR, which uses the string "Etc/Unknown" for an unknown or invalid timezone. (Thanks to Justin Grant, Mark Davis, and Guy Harris.) Cite Internet RFC 9636, which obsoletes RFC 8536 for TZif format. Release 2024b - 2024-09-04 12:27:47 -0700 Briefly: Improve historical data for Mexico, Mongolia, and Portugal. System V names are now obsolescent. The main data form now uses %z. The code now conforms to RFC 8536 for early timestamps. Support POSIX.1-2024, which removes asctime_r and ctime_r. Assume POSIX.2-1992 or later for shell scripts. SUPPORT_C89 now defaults to 1. Changes to past timestamps Asia/Choibalsan is now an alias for Asia/Ulaanbaatar rather than being a separate Zone with differing behavior before April 2008. This seems better given our wildly conflicting information about Mongolia's time zone history. (Thanks to Heitor David Pinto.) Historical transitions for Mexico have been updated based on official Mexican decrees. The affected timestamps occur during the years 1921-1927, 1931, 1945, 1949-1970, and 1981-1997. The affected zones are America/Bahia_Banderas, America/Cancun, America/Chihuahua, America/Ciudad_Juarez, America/Hermosillo, America/Mazatlan, America/Merida, America/Mexico_City, America/Monterrey, America/Ojinaga, and America/Tijuana. (Thanks to Heitor David Pinto.) Historical transitions for Portugal, represented by Europe/Lisbon, Atlantic/Azores, and Atlantic/Madeira, have been updated based on a close reading of old Portuguese legislation, replacing previous data mainly originating from Whitman and Shanks & Pottenger. These changes affect a few transitions in 1917-1921, 1924, and 1940 throughout these regions by a few hours or days, and various timestamps between 1977 and 1993 depending on the region. In particular, the Azores and Madeira did not observe DST from 1977 to 1981. Additionally, the adoption of standard zonal time in former Portuguese colonies have been adjusted: Africa/Maputo in 1909, and Asia/Dili by 22 minutes at the start of 1912. (Thanks to Tim Parenti.) Changes to past tm_isdst flags The period from 1966-04-03 through 1966-10-02 in Portugal is now modeled as DST, to more closely reflect how contemporaneous changes in law entered into force. Changes to data Names present only for compatibility with UNIX System V (last released in the 1990s) have been moved to 'backward'. These names, which for post-1970 timestamps mostly just duplicate data of geographical names, were confusing downstream uses. Names moved to 'backward' are now links to geographical names. This affects behavior for TZ='EET' for some pre-1981 timestamps, for TZ='CET' for some pre-1947 timestamps, and for TZ='WET' for some pre-1996 timestamps. Also, TZ='MET' now behaves like TZ='CET' and so uses the abbreviation "CET" rather than "MET". Those needing the previous TZDB behavior, which does not match any real-world clocks, can find the old entries in 'backzone'. (Problem reported by Justin Grant.) The main source files' time zone abbreviations now use %z, supported by zic since release 2015f and used in vanguard form since release 2022b. For example, America/Sao_Paulo now contains the zone continuation line "-3:00 Brazil %z", which is less error prone than the old "-3:00 Brazil -03/-02". This does not change the represented data: the generated TZif files are unchanged. Rearguard form still avoids %z, to support obsolescent parsers. Asia/Almaty has been removed from zonenow.tab as it now agrees with Asia/Tashkent for future timestamps, due to Kazakhstan's 2024-02-29 time zone change. Similarly, America/Scoresbysund has been removed, as it now agrees with America/Nuuk due to its 2024-03-31 time zone change. Changes to code localtime.c now always uses a TZif file's time type 0 to handle timestamps before the file's first transition. Formerly, localtime.c sometimes inferred a different time type, in order to handle problematic data generated by zic 2018e or earlier. As it is now safe to assume more recent versions of zic, there is no longer a pressing need to fail to conform RFC 8536 section 3.2, which requires using time type 0 in this situation. This change does not affect behavior when reading TZif files generated by zic 2018f and later. POSIX.1-2024 removes asctime_r and ctime_r and does not let libraries define them, so remove them except when needed to conform to earlier POSIX. These functions are dangerous as they can overrun user buffers. If you still need them, add -DSUPPORT_POSIX2008 to CFLAGS. The SUPPORT_C89 option now defaults to 1 instead of 0, fixing a POSIX-conformance bug introduced in 2023a. tzselect now supports POSIX.1-2024 proleptic TZ strings. Also, it assumes POSIX.2-1992 or later, as practical porting targets now all support that, and it uses some features from POSIX.1-2024 if available. Changes to build procedure 'make check' no longer requires curl and Internet access. The build procedure now assumes POSIX.2-1992 or later, to simplify maintenance. To build on Solaris 10, the only extant system still defaulting to pre-POSIX, prepend /usr/xpg4/bin to PATH. Changes to documentation The documentation now reflects POSIX.1-2024. Changes to commentary Commentary about historical transitions in Portugal and her former colonies has been expanded with links to relevant legislation. (Thanks to Tim Parenti.) Release 2024a - 2024-02-01 09:28:56 -0800 Briefly: Kazakhstan unifies on UTC+5 beginning 2024-03-01. Palestine springs forward a week later after Ramadan. zic no longer pretends to support indefinite-past DST. localtime no longer mishandles Ciudad Juárez in 2422. Changes to future timestamps Kazakhstan unifies on UTC+5. This affects Asia/Almaty and Asia/Qostanay which together represent the eastern portion of the country that will transition from UTC+6 on 2024-03-01 at 00:00 to join the western portion. (Thanks to Zhanbolat Raimbekov.) Palestine springs forward a week later than previously predicted in 2024 and 2025. (Thanks to Heba Hamad.) Change spring-forward predictions to the second Saturday after Ramadan, not the first; this also affects other predictions starting in 2039. Changes to past timestamps Asia/Ho_Chi_Minh's 1955-07-01 transition occurred at 01:00 not 00:00. (Thanks to Đoàn Trần Công Danh.) From 1947 through 1949, Toronto's transitions occurred at 02:00 not 00:00. (Thanks to Chris Walton.) In 1911 Miquelon adopted standard time on June 15, not May 15. Changes to code The FROM and TO columns of Rule lines can no longer be "minimum" or an abbreviation of "minimum", because TZif files do not support DST rules that extend into the indefinite past - although these rules were supported when TZif files had only 32-bit data, this stopped working when 64-bit TZif files were introduced in 1995. This should not be a problem for realistic data, since DST was first used in the 20th century. As a transition aid, FROM columns like "minimum" are now diagnosed and then treated as if they were the year 1900; this should suffice for TZif files on old systems with only 32-bit time_t, and it is more compatible with bugs in 2023c-and-earlier localtime.c. (Problem reported by Yoshito Umaoka.) localtime and related functions no longer mishandle some timestamps that occur about 400 years after a switch to a time zone with a DST schedule. In 2023d data this problem was visible for some timestamps in November 2422, November 2822, etc. in America/Ciudad_Juarez. (Problem reported by Gilmore Davidson.) strftime %s now uses tm_gmtoff if available. (Problem and draft patch reported by Dag-Erling Smørgrav.) Changes to build procedure The leap-seconds.list file is now copied from the IERS instead of from its downstream counterpart at NIST, as the IERS version is now in the public domain too and tends to be more up-to-date. (Thanks to Martin Burnicki for liaising with the IERS.) Changes to documentation The strftime man page documents which struct tm members affect which conversion specs, and that tzset is called. (Problems reported by Robert Elz and Steve Summit.) Release 2023d - 2023-12-21 20:02:24 -0800 Briefly: Ittoqqortoormiit, Greenland changes time zones on 2024-03-31. Vostok, Antarctica changed time zones on 2023-12-18. Casey, Antarctica changed time zones five times since 2020. Code and data fixes for Palestine timestamps starting in 2072. A new data file zonenow.tab for timestamps starting now. Changes to future timestamps Ittoqqortoormiit, Greenland (America/Scoresbysund) joins most of the rest of Greenland's timekeeping practice on 2024-03-31, by changing its time zone from -01/+00 to -02/-01 at the same moment as the spring-forward transition. Its clocks will therefore not spring forward as previously scheduled. The time zone change reverts to its common practice before 1981. (Thanks to Jule Dabars.) Fix predictions for DST transitions in Palestine in 2072-2075, correcting a typo introduced in 2023a. (Thanks to Jule Dabars.) Changes to past and future timestamps Vostok, Antarctica changed to +05 on 2023-12-18. It had been at +07 (not +06) for years. (Thanks to Zakhary V. Akulov.) Change data for Casey, Antarctica to agree with timeanddate.com, by adding five time zone changes since 2020. Casey is now at +08 instead of +11. Changes to past tm_isdst flags Much of Greenland, represented by America/Nuuk, changed its standard time from -03 to -02 on 2023-03-25, not on 2023-10-28. This does not affect UTC offsets, only the tm_isdst flag. (Thanks to Thomas M. Steenholdt.) New data file A new data file zonenow.tab helps configure applications that use timestamps dated from now on. This simplifies configuration, since users choose from a smaller Zone set. The file's format is experimental and subject to change. Changes to code localtime.c no longer mishandles TZif files that contain a single transition into a DST regime. Previously, it incorrectly assumed DST was in effect before the transition too. (Thanks to Alois Treindl for debugging help.) localtime.c's timeoff no longer collides with OpenBSD 7.4. The C code now uses _Generic only if __STDC_VERSION__ says the compiler is C11 or later. tzselect now optionally reads zonenow.tab, to simplify when configuring only for timestamps dated from now on. tzselect no longer creates temporary files. tzselect no longer mishandles the following: Spaces and most other special characters in BUGEMAIL, PACKAGE, TZDIR, and VERSION. TZ strings when using mawk 1.4.3, which mishandles regular expressions of the form /X{2,}/. ISO 6709 coordinates when using an awk that lacks the GNU extension of newlines in -v option-arguments. Non UTF-8 locales when using an iconv command that lacks the GNU //TRANSLIT extension. zic no longer mishandles data for Palestine after the year 2075. Previously, it incorrectly omitted post-2075 transitions that are predicted for just before and just after Ramadan. (Thanks to Ken Murchison for debugging help.) zic now works again on Linux 2.6.16 and 2.6.17 (2006). (Problem reported by Rune Torgersen.) Changes to build procedure The Makefile is now more compatible with POSIX: * It no longer defines AR, CC, CFLAGS, LDFLAGS, and SHELL. * It no longer uses its own 'cc' in place of CC. * It now uses ARFLAGS, with default specified by POSIX. * It does not use LFLAGS incompatibly with POSIX. * It uses the special .POSIX target. * It quotes special characters more carefully. * It no longer mishandles builds in an ISO 8859 locale. Due to the CC changes, TZDIR is now #defined in a file tzdir.h built by 'make', not in a $(CC) -D option. Also, TZDEFAULT is now treated like TZDIR as they have similar roles. Changes to commentary Limitations and hazards of the optional support for obsolescent C89 platforms are documented better, along with a tentative schedule for removing this support. Release 2023c - 2023-03-28 12:42:14 -0700 Changes to past and future timestamps Model Lebanon's DST chaos by reverting data to tzdb 2023a. (Thanks to Rany Hany for the heads-up.) Release 2023b - 2023-03-23 19:50:38 -0700 Changes to future timestamps This year Lebanon springs forward April 20/21 not March 25/26. (Thanks to Saadallah Itani.) [This was reverted in 2023c.] Release 2023a - 2023-03-22 12:39:33 -0700 Briefly: Egypt now uses DST again, from April through October. This year Morocco springs forward April 23, not April 30. Palestine delays the start of DST this year. Much of Greenland still uses DST from 2024 on. America/Yellowknife now links to America/Edmonton. tzselect can now use current time to help infer timezone. The code now defaults to C99 or later. Fix use of C23 attributes. Changes to future timestamps Starting in 2023, Egypt will observe DST from April's last Friday through October's last Thursday. (Thanks to Ahmad ElDardiry.) Assume the transition times are 00:00 and 24:00, respectively. In 2023 Morocco's spring-forward transition after Ramadan will occur April 23, not April 30. (Thanks to Milamber.) Adjust predictions for future years accordingly. This affects predictions for 2023, 2031, 2038, and later years. This year Palestine will delay its spring forward from March 25 to April 29 due to Ramadan. (Thanks to Heba Hamad.) Make guesses for future Ramadans too. Much of Greenland, represented by America/Nuuk, will continue to observe DST using European Union rules. When combined with Greenland's decision not to change the clocks in fall 2023, America/Nuuk therefore changes from -03/-02 to -02/-01 effective 2023-10-29 at 01:00 UTC. (Thanks to Thomas M. Steenholdt.) This change from 2022g doesn't affect timestamps until 2024-03-30, and doesn't affect tm_isdst until 2023-03-25. Changes to past timestamps America/Yellowknife has changed from a Zone to a backward compatibility Link, as it no longer differs from America/Edmonton since 1970. (Thanks to Almaz Mingaleev.) This affects some pre-1948 timestamps. The old data are now in 'backzone'. Changes to past time zone abbreviations When observing Moscow time, Europe/Kirov and Europe/Volgograd now use the abbreviations MSK/MSD instead of numeric abbreviations, for consistency with other timezones observing Moscow time. Changes to code You can now tell tzselect local time, to simplify later choices. Select the 'time' option in its first prompt. You can now compile with -DTZNAME_MAXIMUM=N to limit time zone abbreviations to N bytes (default 255). The reference runtime library now rejects POSIX-style TZ strings that contain longer abbreviations, treating them as UTC. Previously the limit was platform dependent and abbreviations were silently truncated to 16 bytes even when the limit was greater than 16. The code by default is now designed for C99 or later. To build on a mostly-C89 platform, compile with -DPORT_TO_C89; this should work on C89 platforms that also support C99 'long long' and perhaps a few other extensions to C89. To support C89 callers of tzcode's library, compile with -DSUPPORT_C89; however, this could trigger latent bugs in C99-or-later callers. The two new macros are transitional aids planned to be removed in a future version (say, in 2029), when C99 or later will be required. The code now builds again on pre-C99 platforms, if you compile with -DPORT_TO_C89. This fixes a bug introduced in 2022f. On C23-compatible platforms tzcode no longer uses syntax like 'static [[noreturn]] void usage(void);'. Instead, it uses '[[noreturn]] static void usage(void);' as strict C23 requires. (Problem reported by Houge Langley.) The code's functions now constrain their arguments with the C 'restrict' keyword consistently with their documentation. This may allow future optimizations. zdump again builds standalone with ckdadd and without setenv, fixing a bug introduced in 2022g. (Problem reported by panic.) leapseconds.awk can now process a leap seconds file that never expires; this might be useful if leap seconds are discontinued. Changes to commentary tz-link.html has a new section "Coordinating with governments and distributors". (Thanks to Neil Fuller for some of the text.) To improve tzselect diagnostics, zone1970.tab's comments column is now limited to countries that have multiple timezones. Note that there are plans to discontinue leap seconds by 2035. Release 2022g - 2022-11-29 08:58:31 -0800 Briefly: The northern edge of Chihuahua changes to US timekeeping. Much of Greenland stops changing clocks after March 2023. Fix some pre-1996 timestamps in northern Canada. C89 is now deprecated; please use C99 or later. Portability fixes for AIX, libintl, MS-Windows, musl, z/OS In C code, use more C23 features if available. C23 timegm now supported by default Fixes for unlikely integer overflows Changes to future timestamps In the Mexican state of Chihuahua, the border strip near the US will change to agree with nearby US locations on 2022-11-30. The strip's western part, represented by Ciudad Juárez, switches from -06 all year to -07/-06 with US DST rules, like El Paso, TX. The eastern part, represented by Ojinaga, will observe US DST next year, like Presidio, TX. (Thanks to Heitor David Pinto.) A new Zone America/Ciudad_Juarez splits from America/Ojinaga. Much of Greenland, represented by America/Nuuk, stops observing winter time after March 2023, so its daylight saving time becomes standard time. (Thanks to Jonas Nyrup and Jürgen Appel.) Changes to past timestamps Changes for pre-1996 northern Canada (thanks to Chris Walton): Merge America/Iqaluit and America/Pangnirtung into the former, with a backward compatibility link for the latter name. There is no good evidence the two locations differ since 1970. This change affects pre-1996 America/Pangnirtung timestamps. Cambridge Bay, Inuvik, Iqaluit, Rankin Inlet, Resolute and Yellowknife did not observe DST in 1965, and did observe DST from 1972 through 1979. Whitehorse moved from -09 to -08 on 1966-02-27, not 1967-05-28. Colombia's 1993 fallback was 02-06 24:00, not 04-04 00:00. (Thanks to Alois Treindl.) Singapore's 1981-12-31 change was at 16:00 UTC (23:30 local time), not 24:00 local time. (Thanks to Geoff Clare via Robert Elz.) Changes to code Although tzcode still works with C89, bugs found in recent routine maintenance indicate that bitrot has set in and that in practice C89 is no longer used to build tzcode. As it is a maintenance burden, support for C89 is planned to be removed soon. Instead, please use compilers compatible with C99, C11, C17, or C23. timegm, which tzcode implemented in 1989, will finally be standardized 34 years later as part of C23, so timegm is now supported even if STD_INSPIRED is not defined. Fix bug in zdump's tzalloc emulation on hosts that lack tm_zone. (Problem reported by Đoàn Trần Công Danh.) Fix bug in zic on hosts where malloc(0) yields NULL on success. (Problem reported by Tim McBrayer for AIX 6.1.) Fix zic configuration to avoid linkage failures on some platforms. (Problems reported by Gilmore Davidson and Igor Ivanov.) Work around MS-Windows nmake incompatibility with POSIX. (Problem reported by Manuela Friedrich.) Port mktime and strftime to debugging platforms where accessing uninitialized data has undefined behavior (strftime problem reported by Robert Elz). Check more carefully for unlikely integer overflows, preferring C23 to overflow checking by hand, as the latter has had obscure bugs. Changes to build procedure New Makefile rule check_mild that skips checking whether Link lines are in the file 'backward'. (Inspired by a suggestion from Stephen Colebourne.) Release 2022f - 2022-10-28 18:04:57 -0700 Briefly: Mexico will no longer observe DST except near the US border. Chihuahua moves to year-round -06 on 2022-10-30. Fiji no longer observes DST. Move links to 'backward'. In vanguard form, GMT is now a Zone and Etc/GMT a link. zic now supports links to links, and vanguard form uses this. Simplify four Ontario zones. Fix a Y2438 bug when reading TZif data. Enable 64-bit time_t on 32-bit glibc platforms. Omit large-file support when no longer needed. In C code, use some C23 features if available. Remove no-longer-needed workaround for Qt bug 53071. Changes to future timestamps Mexico will no longer observe DST after 2022, except for areas near the US border that continue to observe US DST rules. On 2022-10-30 at 02:00 the Mexican state of Chihuahua moves from -07 (-06 with DST) to year-round -06, thus not changing its clocks that day. The new law states that Chihuahua near the US border no longer observes US DST. (Thanks to gera for the heads-up about Chihuahua.) Fiji will not observe DST in 2022/3. (Thanks to Shalvin Narayan.) For now, assume DST is suspended indefinitely. Changes to data Move links to 'backward' to ease and simplify link maintenance. This affects generated data only if you use 'make BACKWARD='. GMT is now a Zone and Etc/GMT a link instead of vice versa, as GMT is needed for leap second support whereas Etc/GMT is not. However, this change exposes a bug in TZUpdater 2.3.2 so it is present only in vanguard form for now. Vanguard form now uses links to links, as zic now supports this. Changes to past timestamps Simplify four Ontario zones, as most of the post-1970 differences seem to have been imaginary. (Problem reported by Chris Walton.) Move America/Nipigon, America/Rainy_River, and America/Thunder_Bay to 'backzone'; backward-compatibility links still work, albeit with some different timestamps before November 2005. Changes to code zic now supports links to links regardless of input line order. For example, if Australia/Sydney is a Zone, the lines Link Australia/Canberra Australia/ACT Link Australia/Sydney Australia/Canberra now work correctly, even though the shell commands ln Australia/Canberra Australia/ACT ln Australia/Sydney Australia/Canberra would fail because the first command attempts to use a link Australia/Canberra that does not exist until after the second command is executed. Previously, zic had unspecified behavior if a Link line's target was another link, and zic often misbehaved if a Link line's target was a later Link line. Fix line number in zic's diagnostic for a link to a link. Fix a bug that caused localtime to mishandle timestamps starting in the year 2438 when reading data generated by 'zic -b fat' when distant-future DST transitions occur at times given in standard time or in UT, not the usual case of local time. This occurs when the corresponding .zi Rule lines specify DST transitions with TO columns of 'max' and AT columns that end in 's' or 'u'. The number 2438 comes from the 32-bit limit in the year 2038, plus the 400-year Gregorian cycle. (Problem reported by Bradley White.) On glibc 2.34 and later, which optionally supports 64-bit time_t on platforms like x86 where time_t was traditionally 32 bits, default time_t to 64 instead of 32 bits. This lets functions like localtime support timestamps after the year 2038, and fixes year-2038 problems in zic when accessing files dated after 2038. To continue to limit time_t to 32 bits on these platforms, use "make CFLAGS='-D_TIME_BITS=32'". In C code, do not enable large-file support on platforms like AIX and macOS that no longer need it now that tzcode does not use off_t or related functions like 'stat'. Large-file support is still enabled by default on GNU/Linux, as it is needed for 64-bit time_t support. In C code, prefer C23 keywords to pre-C23 macros for alignof, bool, false, and true. Also, use the following C23 features if available: __has_include, unreachable. zic no longer works around Qt bug 53071, as the relevant Qt releases have been out of support since 2019. This change affects only fat TZif files, as thin files never had the workaround. zdump no longer modifies the environ vector when compiled on platforms lacking tm_zone or when compiled with -DUSE_LTZ=0. This avoid undefined behavior on POSIX platforms. Release 2022e - 2022-10-11 11:13:02 -0700 Briefly: Jordan and Syria switch from +02/+03 with DST to year-round +03. Changes to future timestamps Jordan and Syria are abandoning the DST regime and are changing to permanent +03, so they will not fall back from +03 to +02 on 2022-10-28. (Thanks to Steffen Thorsen and Issam Al-Zuwairi.) Changes to past timestamps On 1922-01-01 Tijuana adopted standard time at 00:00, not 01:00. Changes to past time zone abbreviations and DST flags The temporary advancement of clocks in central Mexico in summer 1931 is now treated as daylight saving time, instead of as two changes to standard time. Release 2022d - 2022-09-23 12:02:57 -0700 Briefly: Palestine transitions are now Saturdays at 02:00. Simplify three Ukraine zones into one. Changes to future timestamps Palestine now springs forward and falls back at 02:00 on the first Saturday on or after March 24 and October 24, respectively. This means 2022 falls back 10-29 at 02:00, not 10-28 at 01:00. (Thanks to Heba Hamad.) Changes to past timestamps Simplify three Ukraine zones to one, since the post-1970 differences seem to have been imaginary. Move Europe/Uzhgorod and Europe/Zaporozhye to 'backzone'; backward-compatibility links still work, albeit with different timestamps before October 1991. Release 2022c - 2022-08-15 17:47:18 -0700 Briefly: Work around awk bug in FreeBSD, macOS, etc. Improve tzselect on intercontinental Zones. Changes to code Work around a bug in onetrueawk that broke commands like 'make traditional_tarballs' on FreeBSD, macOS, etc. (Problem reported by Deborah Goldsmith.) Add code to tzselect that uses experimental structured comments in zone1970.tab to clarify whether Zones like Africa/Abidjan and Europe/Istanbul cross continent or ocean boundaries. (Inspired by a problem reported by Peter Krefting.) Fix bug with 'zic -d /a/b/c' when /a is unwritable but the directory /a/b already exists. Remove zoneinfo2tdf.pl, as it was unused and triggered false malware alarms on some email servers. Release 2022b - 2022-08-10 15:38:32 -0700 Briefly: Chile's DST is delayed by a week in September 2022. Iran no longer observes DST after 2022. Rename Europe/Kiev to Europe/Kyiv. New zic -R option Vanguard form now uses %z. Finish moving duplicate-since-1970 zones to 'backzone'. New build option PACKRATLIST New tailored_tarballs target, replacing rearguard_tarballs Changes to future timestamps Chile's 2022 DST start is delayed from September 4 to September 11. (Thanks to Juan Correa.) Iran plans to stop observing DST permanently, after it falls back on 2022-09-21. (Thanks to Ali Mirjamali.) Changes to past timestamps Finish moving to 'backzone' the location-based zones whose timestamps since 1970 are duplicates; adjust links accordingly. This change ordinarily affects only pre-1970 timestamps, and with the new PACKRATLIST option it does not affect any timestamps. In this round the affected zones are Antarctica/Vostok, Asia/Brunei, Asia/Kuala_Lumpur, Atlantic/Reykjavik, Europe/Amsterdam, Europe/Copenhagen, Europe/Luxembourg, Europe/Monaco, Europe/Oslo, Europe/Stockholm, Indian/Christmas, Indian/Cocos, Indian/Kerguelen, Indian/Mahe, Indian/Reunion, Pacific/Chuuk, Pacific/Funafuti, Pacific/Majuro, Pacific/Pohnpei, Pacific/Wake and Pacific/Wallis, and the affected links are Arctic/Longyearbyen, Atlantic/Jan_Mayen, Iceland, Pacific/Ponape, Pacific/Truk, and Pacific/Yap. From fall 1994 through fall 1995, Shanks wrote that Crimea's DST transitions were at 02:00 standard time, not at 00:00. (Thanks to Michael Deckers.) Iran adopted standard time in 1935, not 1946. In 1977 it observed DST from 03-21 23:00 to 10-20 24:00; its 1978 transitions were on 03-24 and 08-05, not 03-20 and 10-20; and its spring 1979 transition was on 05-27, not 03-21. (Thanks to Roozbeh Pournader and Francis Santoni.) Chile's observance of -04 from 1946-08-29 through 1947-03-31 was considered DST, not standard time. Santiago and environs had moved their clocks back to rejoin the rest of mainland Chile; put this change at the end of 1946-08-28. (Thanks to Michael Deckers.) Some old, small clock transitions have been removed, as people at the time did not change their clocks. This affects Asia/Hong_Kong in 1904, Asia/Ho_Chi_Minh in 1906, and Europe/Dublin in 1880. Changes to zone name Rename Europe/Kiev to Europe/Kyiv, as "Kyiv" is more common in English now. Spelling of other names in Ukraine has not yet demonstrably changed in common English practice so for now these names retain old spellings, as in other countries (e.g., Europe/Prague not "Praha", and Europe/Sofia not "Sofiya"). Changes to code zic has a new option '-R @N' to output explicit transitions < N. (Need suggested by Almaz Mingaleev.) 'zic -r @N' no longer outputs bad data when N < first transition. (Problem introduced in 2021d and reported by Peter Krefting.) zic now checks its input for NUL bytes and unterminated lines, and now supports input line lengths up to 2048 (not 512) bytes. gmtime and related code now use the abbreviation "UTC" not "GMT". POSIX is being revised to require this. When tzset and related functions set vestigial static variables like tzname, they now prefer specified timestamps to unspecified ones. (Problem reported by Almaz Mingaleev.) zic no longer complains "can't determine time zone abbreviation to use just after until time" when a transition to a new standard time occurs simultaneously with the first DST fallback transition. Changes to build procedure Source data in vanguard form now uses the %z notation, introduced in release 2015f. For example, for America/Sao_Paulo vanguard form contains the zone continuation line "-3:00 Brazil %z", which is simpler and more reliable than the line "-3:00 Brazil -03/-02" used in main and rearguard forms. The plan is for the main form to use %z eventually; in the meantime maintainers of zi parsers are encouraged to test the parsers on vanguard.zi. The Makefile has a new PACKRATLIST option to select a subset of 'backzone'. For example, 'make PACKRATDATA=backzone PACKRATLIST=zone.tab' now generates TZif files identical to those of the global-tz project. The Makefile has a new tailored_tarballs target for generating special-purpose tarballs. It generalizes and replaces the rearguard_tarballs target and related targets and macros, which are now obsolescent. 'make install' now defaults LOCALTIME to Factory not GMT, which means the default abbreviation is now "-00" not "GMT". Remove the posix_packrat target, marked obsolescent in 2016a. Release 2022a - 2022-03-15 23:02:01 -0700 Briefly: Palestine will spring forward on 2022-03-27, not -03-26. zdump -v now outputs better failure indications. Bug fixes for code that reads corrupted TZif data. Changes to future timestamps Palestine will spring forward on 2022-03-27, not 2022-03-26. (Thanks to Heba Hamad.) Predict future transitions for first Sunday >= March 25. Additionally, predict fallbacks to be the first Friday on or after October 23, not October's last Friday, to be more consistent with recent practice. The first differing fallback prediction is on 2025-10-24, not 2025-10-31. Changes to past timestamps From 1992 through spring 1996, Ukraine's DST transitions were at 02:00 standard time, not at 01:00 UTC. (Thanks to Alois Treindl.) Chile's Santiago Mean Time and its LMT precursor have been adjusted eastward by 1 second to align with past and present law. Changes to commentary Add several references for Chile's 1946/1947 transitions, some of which only affected portions of the country. Changes to code Fix bug when mktime gets confused by truncated TZif files with unspecified local time. (Problem reported by Almaz Mingaleev.) Fix bug when 32-bit time_t code reads malformed 64-bit TZif data. (Problem reported by Christos Zoulas.) When reading a version 2 or later TZif file, the TZif reader now validates the version 1 header and data block only enough to skip over them, as recommended by RFC 8536 section 4. Also, the TZif reader no longer mistakenly attempts to parse a version 1 TZIf file header as a TZ string. zdump -v now outputs "(localtime failed)" and "(gmtime failed)" when local time and UT cannot be determined for a timestamp. Changes to build procedure Distribution tarballs now use standard POSIX.1-1988 ustar format instead of GNU format. Although the formats are almost identical for these tarballs, ustar headers' magic fields contain "ustar" instead of "ustar ", and their version fields contain "00" instead of " ". The two formats are planned to diverge more significantly for tzdb releases after 2242-03-16 12:56:31 UTC, when the ustar format becomes obsolete and the tarballs switch to pax format, an extension of ustar. For details about these formats, please see "pax - portable archive interchange", IEEE Std 1003.1-2017, . Release 2021e - 2021-10-21 18:41:00 -0700 Changes to future timestamps Palestine will fall back 10-29 (not 10-30) at 01:00. (Thanks to P Chan and Heba Hemad.) Release 2021d - 2021-10-15 13:48:18 -0700 Briefly: Fiji suspends DST for the 2021/2022 season. 'zic -r' marks unspecified timestamps with "-00". Changes to future timestamps Fiji will suspend observance of DST for the 2021/2022 season. Assume for now that it will return next year. (Thanks to Jashneel Kumar and P Chan.) Changes to code 'zic -r' now uses "-00" time zone abbreviations for intervals with UT offsets that are unspecified due to -r truncation. This implements a change in draft Internet RFC 8536bis. Release 2021c - 2021-10-01 14:21:49 -0700 Briefly: Revert most 2021b changes to 'backward'. Fix 'zic -b fat' bug in pre-1970 32-bit data. Fix two Link line typos. Distribute SECURITY file. This release is intended as a bugfix release, to fix compatibility problems and typos reported since 2021b was released. Changes to Link directives Revert almost all of 2021b's changes to the 'backward' file, by moving Link directives back to where they were in 2021a. Although 'zic' doesn't care which source file contains a Link directive, some downstream uses ran into trouble with the move. (Problem reported by Stephen Colebourne for Joda-Time.) Fix typo that linked Atlantic/Jan_Mayen to the wrong location (problem reported by Chris Walton). Fix 'backzone' typo that linked America/Virgin to the wrong location (problem reported by Michael Deckers). Changes to code Fix a bug in 'zic -b fat' that caused old timestamps to be mishandled in 32-bit-only readers (problem reported by Daniel Fischer). Changes to documentation Distribute the SECURITY file (problem reported by Andreas Radke). Release 2021b - 2021-09-24 16:23:00 -0700 Briefly: Jordan now starts DST on February's last Thursday. Samoa no longer observes DST. Merge more location-based Zones whose timestamps agree since 1970. Move some backward-compatibility links to 'backward'. Rename Pacific/Enderbury to Pacific/Kanton. Correct many pre-1993 transitions in Malawi, Portugal, etc. zic now creates each output file or link atomically. zic -L no longer omits the POSIX TZ string in its output. zic fixes for truncation and leap second table expiration. zic now follows POSIX for TZ strings using all-year DST. Fix some localtime crashes and bugs in obscure cases. zdump -v now outputs more-useful boundary cases. tzfile.5 better matches a draft successor to RFC 8536. A new file SECURITY. This release is prompted by recent announcements by Jordan and Samoa. It incorporates many other changes that had accumulated since 2021a. However, it omits most proposed changes that merged all Zones agreeing since 1970, as concerns were raised about doing too many of these changes at once. It does keeps some of these changes in the interest of making tzdb more equitable one step at a time; see "Merge more location-based Zones" below. Changes to future timestamps Jordan now starts DST on February's last Thursday. (Thanks to Steffen Thorsen.) Samoa no longer observes DST. (Thanks to Geoffrey D. Bennett.) Changes to zone name Rename Pacific/Enderbury to Pacific/Kanton. When we added Enderbury in 1993, we did not know that it is uninhabited and that Kanton (population two dozen) is the only inhabited location in that timezone. The old name is now a backward-compatibility link. Changes to past timestamps Correct many pre-1993 transitions, fixing entries originally derived from Shanks, Whitman, and Mundell. The fixes include: - Barbados: standard time was introduced in 1911, not 1932; and DST was observed in 1942-1944 - Cook Islands: In 1899 they switched from east to west of GMT, celebrating Christmas for two days. They (and Niue) switched to standard time in 1952, not 1901. - Guyana: corrected LMT for Georgetown; the introduction of standard time in 1911, not 1915; and corrections to 1975 and 1992 transitions - Kanton: uninhabited before 1937-08-31 - Niue: only observed -11:20 from 1952 through 1964, then went to -11 instead of -11:30 - Portugal: DST was observed in 1950 - Tonga: corrected LMT; the introduction of standard time in 1945, not 1901; and corrections to the transition from +12:20 to +13 in 1961, not 1941 Additional fixes to entries in the 'backzone' file include: - Enderbury: inhabited only 1860/1885 and 1938-03-06/1942-02-09 - The Gambia: 1933 and 1942 transitions - Malawi: several 1911 through 1925 transitions - Sierra Leone: several 1913 through 1941 transitions, and DST was NOT observed in 1957 through 1962 (Thanks to P Chan, Michael Deckers, Alexander Krivenyshev and Alois Treindl.) Merge more location-based Zones whose timestamps agree since 1970, as pre-1970 timestamps are out of scope. This is part of a process that has been ongoing since 2013. This does not affect post-1970 timestamps, and timezone historians who build with 'make PACKRATDATA=backzone' should see no changes to pre-1970 timestamps. When merging, keep the most-populous location's data, and move data for other locations to 'backzone' with a backward link in 'backward'. For example, move America/Creston data to 'backzone' with a link in 'backward' from America/Phoenix because the two timezones' timestamps agree since 1970; this change affects some pre-1968 timestamps in America/Creston because Creston and Phoenix disagreed before 1968. The affected Zones are Africa/Accra, America/Atikokan, America/Blanc-Sablon, America/Creston, America/Curacao, America/Nassau, America/Port_of_Spain, Antarctica/DumontDUrville, and Antarctica/Syowa. Changes to maintenance procedure The new file SECURITY covers how to report security-related bugs. Several backward-compatibility links have been moved to the 'backward' file. These links, which range from Africa/Addis_Ababa to Pacific/Saipan, are only for compatibility with now-obsolete guidelines suggesting an entry for every ISO 3166 code. The intercontinental convenience links Asia/Istanbul and Europe/Nicosia have also been moved to 'backward'. Changes to code zic now creates each output file or link atomically, possibly by creating a temporary file and then renaming it. This avoids races where a TZ setting would temporarily stop working while zic was installing a replacement file or link. zic -L no longer omits the POSIX TZ string in its output. Starting with 2020a, zic -L truncated its output according to the "Expires" directive or "#expires" comment in the leapseconds file. The resulting TZif files omitted daylight saving transitions after the leap second table expired, which led to far less accurate predictions of times after the expiry. Although future timestamps cannot be converted accurately in the presence of leap seconds, it is more accurate to convert near-future timestamps with a few seconds error than with an hour error, so zic -L no longer truncates output in this way. Instead, when zic -L is given the "Expires" directive, it now outputs the expiration by appending a no-change entry to the leap second table. Although this should work well with most TZif readers, it does not conform to Internet RFC 8536 and some pickier clients (including tzdb 2017c through 2021a) reject it, so "Expires" directives are currently disabled by default. To enable them, set the EXPIRES_LINE Makefile variable. If a TZif file uses this new feature it is marked with a new TZif version number 4, a format intended to be documented in a successor to RFC 8536. The old-format "#expires" comments are now treated solely as comments and have no effect on the TZif files. zic -L LEAPFILE -r @LO no longer generates an invalid TZif file that omits leap second information for the range LO..B when LO falls between two leap seconds A and B. Instead, it generates a TZif version 4 file that represents the previously missing information. The TZif reader now allows the leap second table to begin with a correction other than -1 or +1, and to contain adjacent transitions with equal corrections. This supports TZif version 4. The TZif reader now lets leap seconds occur less than 28 days apart. This supports possible future TZif extensions. Fix bug that caused 'localtime' etc. to crash when TZ was set to a all-year DST string like "EST5EDT4,0/0,J365/25" that does not conform to POSIX but does conform to Internet RFC 8536. Fix another bug that caused 'localtime' etc. to crash when TZ was set to a POSIX-conforming but unusual TZ string like "EST5EDT4,0/0,J365/0", where almost all the year is DST. Fix yet another bug that caused 'localtime' etc. to mishandle slim TZif files containing leap seconds after the last explicit transition in the table, or when handling far-future timestamps in slim TZif files lacking leap seconds. Fix localtime misbehavior involving positive leap seconds. This change affects only behavior for "right" system time, which contains leap seconds, and only if the UT offset is not a multiple of 60 seconds when a positive leap second occurs. (No such timezone exists in tzdb, luckily.) Without the fix, the timestamp was ambiguous during a positive leap second. With the fix, any seconds occurring after a positive leap second and within the same localtime minute are counted through 60, not through 59; their UT offset (tm_gmtoff) is the same as before. Here is how the fix affects timestamps in a timezone with UT offset +01:23:45 (5025 seconds) and with a positive leap second at 1972-06-30 23:59:60 UTC (78796800): time_t without the fix with the fix 78796800 1972-07-01 01:23:45 1972-07-01 01:23:45 (leap second) 78796801 1972-07-01 01:23:45 1972-07-01 01:23:46 ... 78796815 1972-07-01 01:23:59 1972-07-01 01:23:60 78796816 1972-07-01 01:24:00 1972-07-01 01:24:00 Fix an unlikely bug that caused 'localtime' etc. to misbehave if civil time changes a few seconds before time_t wraps around, when leap seconds are enabled. Fix bug in zic -r; in some cases, the dummy time type after the last time transition disagreed with the TZ string, contrary to Internet RFC 8563 section 3.3. Fix a bug with 'zic -r @X' when X is a negative leap second that has a nonnegative correction. Without the fix, the output file was truncated so that X appeared to be a positive leap second. Fix a similar, even less likely bug when truncating at a positive leap second that has a nonpositive correction. zic -r now reports an error if given rolling leap seconds, as this usage has never generally worked and is evidently unused. zic now generates a POSIX-conforming TZ string for TZif files where all-year DST is predicted for the indefinite future. For example, for all-year Eastern Daylight Time, zic now generates "XXX3EDT4,0/0,J365/23" where it previously generated "EST5EDT,0/0,J365/25" or "". (Thanks to Michael Deckers for noting the possibility of POSIX conformance.) zic.c no longer requires sys/wait.h (thanks to spazmodius for noting it wasn't needed). When reading slim TZif files, zdump no longer mishandles leap seconds on the rare platforms where time_t counts leap seconds, fixing a bug introduced in 2014g. zdump -v now outputs timestamps at boundaries of what localtime and gmtime can represent, instead of the less useful timestamps one day after the minimum and one day before the maximum. (Thanks to Arthur David Olson for prototype code, and to Manuela Friedrich for debugging help.) zdump's -c and -t options are now consistently inclusive for the lower time bound and exclusive for the upper. Formerly they were inconsistent. (Confusion noted by Martin Burnicki.) Changes to build procedure You can now compile with -DHAVE_MALLOC_ERRNO=0 to port to non-POSIX hosts where malloc doesn't set errno. (Problem reported by Jan Engelhardt.) Changes to documentation tzfile.5 better matches a draft successor to RFC 8536 . Release 2021a - 2021-01-24 10:54:57 -0800 Changes to future timestamps South Sudan changes from +03 to +02 on 2021-02-01 at 00:00. (Thanks to Steffen Thorsen.) Release 2020f - 2020-12-29 00:17:46 -0800 Change to build procedure 'make rearguard_tarballs' no longer generates a bad rearguard.zi, fixing a 2020e bug. (Problem reported by Deborah Goldsmith.) Release 2020e - 2020-12-22 15:14:34 -0800 Briefly: Volgograd switches to Moscow time on 2020-12-27 at 02:00. Changes to future timestamps Volgograd changes time zone from +04 to +03 on 2020-12-27 at 02:00. (Thanks to Alexander Krivenyshev and Stepan Golosunov.) Changes to past timestamps Correct many pre-1986 transitions, fixing entries originally derived from Shanks. The fixes include: - Australia: several 1917 through 1971 transitions - The Bahamas: several 1941 through 1945 transitions - Bermuda: several 1917 through 1956 transitions - Belize: several 1942 through 1968 transitions - Ghana: several 1915 through 1956 transitions - Israel and Palestine: several 1940 through 1985 transitions - Kenya and adjacent: several 1908 through 1960 transitions - Nigeria and adjacent: correcting LMT in Lagos, and several 1905 through 1919 transitions - Seychelles: the introduction of standard time in 1907, not 1906 - Vanuatu: DST in 1973-1974, and a corrected 1984 transition (Thanks to P Chan.) Because of the Australia change, Australia/Currie (King Island) is no longer needed, as it is identical to Australia/Hobart for all timestamps since 1970 and was therefore created by mistake. Australia/Currie has been moved to the 'backward' file and its corrected data moved to the 'backzone' file. Changes to past time zone abbreviations and DST flags To better match legislation in Turks and Caicos, the 2015 shift to year-round observance of -04 is now modeled as AST throughout before returning to Eastern Time with US DST in 2018, rather than as maintaining EDT until 2015-11-01. (Thanks to P Chan.) Changes to documentation The zic man page now documents zic's coalescing of transitions when a zone falls back just before DST springs forward. Release 2020d - 2020-10-21 11:24:13 -0700 Briefly: Palestine ends DST earlier than predicted, on 2020-10-24. Changes to past and future timestamps Palestine ends DST on 2020-10-24 at 01:00, instead of 2020-10-31 as previously predicted (thanks to Sharef Mustafa.) Its 2019-10-26 fall-back was at 00:00, not 01:00 (thanks to Steffen Thorsen.) Its 2015-10-23 transition was at 01:00 not 00:00, and its spring 2020 transition was on March 28 at 00:00, not March 27 (thanks to Pierre Cashon.) This affects Asia/Gaza and Asia/Hebron. Assume future spring and fall transitions will be on the Saturday preceding the last Sunday of March and October, respectively. Release 2020c - 2020-10-16 11:15:53 -0700 Briefly: Fiji starts DST later than usual, on 2020-12-20. Changes to future timestamps Fiji will start DST on 2020-12-20, instead of 2020-11-08 as previously predicted. DST will still end on 2021-01-17. (Thanks to Raymond Kumar and Alan Mintz.) Assume for now that the later-than-usual start date is a one-time departure from the recent pattern. Changes to build procedure Rearguard tarballs now contain an empty file pacificnew. Some older downstream software expects this file to exist. (Problem reported by Mike Cullinan.) Release 2020b - 2020-10-06 18:35:04 -0700 Briefly: Revised predictions for Morocco's changes starting in 2023. Canada's Yukon changes to -07 on 2020-11-01, not 2020-03-08. Macquarie Island has stayed in sync with Tasmania since 2011. Casey, Antarctica is at +08 in winter and +11 in summer. zic no longer supports -y, nor the TYPE field of Rules. Changes to future timestamps Morocco's spring-forward after Ramadan is now predicted to occur no sooner than two days after Ramadan, instead of one day. (Thanks to Milamber.) The first altered prediction is for 2023, now predicted to spring-forward on April 30 instead of April 23. Changes to past and future timestamps Casey Station, Antarctica has been using +08 in winter and +11 in summer since 2018. The most recent transition from +08 to +11 was 2020-10-04 00:01. Also, Macquarie Island has been staying in sync with Tasmania since 2011. (Thanks to Steffen Thorsen.) Changes to past and future time zone abbreviations and DST flags Canada's Yukon, represented by America/Whitehorse and America/Dawson, changes its time zone rules from -08/-07 to permanent -07 on 2020-11-01, not on 2020-03-08 as 2020a had it. This change affects only the time zone abbreviation (MST vs PDT) and daylight saving flag for the period between the two dates. (Thanks to Andrew G. Smith.) Changes to past timestamps Correct several transitions for Hungary for 1918/1983. For example, the 1983-09-25 fall-back was at 01:00, not 03:00. (Thanks to Géza Nyáry.) Also, the 1890 transition to standard time was on 11-01, not 10-01 (thanks to Michael Deckers). The 1891 French transition was on March 16, not March 15. The 1911-03-11 French transition was at midnight, not a minute later. Monaco's transitions were on 1892-06-01 and 1911-03-29, not 1891-03-15 and 1911-03-11. (Thanks to Michael Deckers.) Changes to code Support for zic's long-obsolete '-y YEARISTYPE' option has been removed and, with it, so has support for the TYPE field in Rule lines, which is now reserved for compatibility with earlier zic. These features were previously deprecated in release 2015f. (Thanks to Tim Parenti.) zic now defaults to '-b slim' instead of to '-b fat'. zic's new '-l -' and '-p -' options uninstall any existing localtime and posixrules files, respectively. The undocumented and ineffective tzsetwall function has been removed. Changes to build procedure The Makefile now defaults POSIXRULES to '-', so the posixrules feature (obsolete as of 2019b) is no longer installed by default. Changes to documentation and commentary The long-obsolete files pacificnew, systemv, and yearistype.sh have been removed from the distribution. (Thanks to Tim Parenti.) Release 2020a - 2020-04-23 16:03:47 -0700 Briefly: Morocco springs forward on 2020-05-31, not 2020-05-24. Canada's Yukon advanced to -07 year-round on 2020-03-08. America/Nuuk renamed from America/Godthab. zic now supports expiration dates for leap second lists. Changes to future timestamps Morocco's second spring-forward transition in 2020 will be May 31, not May 24 as predicted earlier. (Thanks to Semlali Naoufal.) Adjust future-year predictions to use the first Sunday after the day after Ramadan, not the first Sunday after Ramadan. Canada's Yukon, represented by America/Whitehorse and America/Dawson, advanced to -07 year-round, beginning with its spring-forward transition on 2020-03-08, and will not fall back on 2020-11-01. Although a government press release calls this "permanent Pacific Daylight Saving Time", we prefer MST for consistency with nearby Dawson Creek, Creston, and Fort Nelson. (Thanks to Tim Parenti.) Changes to past timestamps Shanghai observed DST in 1919. (Thanks to Phake Nick.) Changes to timezone identifiers To reflect current usage in English better, America/Godthab has been renamed to America/Nuuk. A backwards-compatibility link remains for the old name. Changes to code localtime.c no longer mishandles timestamps after the last transition in a TZif file with leap seconds and with daylight saving time transitions projected into the indefinite future. For example, with TZ='America/Los_Angeles' with leap seconds, zdump formerly reported a DST transition on 2038-03-14 from 01:59:32.999... to 02:59:33 instead of the correct transition from 01:59:59.999... to 03:00:00. zic -L now supports an Expires line in the leapseconds file, and truncates the TZif output accordingly. This propagates leap second expiration information into the TZif file, and avoids the localtime.c bug mentioned above, as well as similar bugs present in many client implementations. If no Expires line is present, zic -L instead truncates the TZif output based on the #expires comment present in leapseconds files distributed by tzdb 2018f and later; however, this usage is obsolescent. For now, the distributed leapseconds file has an Expires line that is commented out, so that the file can be fed to older versions of zic which ignore the commented-out line. Future tzdb distributions are planned to contain a leapseconds file with an Expires line. The configuration macros HAVE_TZNAME and USG_COMPAT should now be set to 1 if the system library supports the feature, and 2 if not. As before, these macros are nonzero if tzcode should support the feature, zero otherwise. The configuration macro ALTZONE now has the same values with the same meaning as HAVE_TZNAME and USG_COMPAT. The code's defense against CRLF in leap-seconds.list is now portable to POSIX awk. (Problem reported by Deborah Goldsmith.) Although the undocumented tzsetwall function is not changed in this release, it is now deprecated in preparation for removal in future releases. Due to POSIX requirements, tzsetwall has not worked for some time. Any code that uses it should instead use tzalloc(NULL) or, if portability trumps thread-safety, should unset the TZ environment variable. Changes to commentary The Îles-de-la-Madeleine and the Listuguj reserve are noted as following America/Halifax, and comments about Yukon's "south" and "north" have been corrected to say "east" and "west". (Thanks to Jeffery Nichols.) Release 2019c - 2019-09-11 08:59:48 -0700 Briefly: Fiji observes DST from 2019-11-10 to 2020-01-12. Norfolk Island starts observing Australian-style DST. Changes to future timestamps Fiji's next DST transitions will be 2019-11-10 and 2020-01-12 instead of 2019-11-03 and 2020-01-19. (Thanks to Raymond Kumar.) Adjust future guesses accordingly. Norfolk Island will observe Australian-style DST starting in spring 2019. The first transition is on 2019-10-06. (Thanks to Kyle Czech and Michael Deckers.) Changes to past timestamps Many corrections to time in Turkey from 1940 through 1985. (Thanks to Oya Vulaş via Alois Treindl, and to Kıvanç Yazan.) The Norfolk Island 1975-03-02 transition was at 02:00 standard time, not 02:00 DST. (Thanks to Michael Deckers.) South Korea observed DST from 1948 through 1951. Although this info was supposed to appear in release 2014j, a typo inadvertently suppressed the change. (Thanks to Alois Treindl.) Detroit observed DST in 1967 and 1968 following the US DST rules, except that its 1967 DST began on June 14 at 00:01. (Thanks to Alois Treindl for pointing out that the old data entries were probably wrong.) Fix several errors in pre-1970 transitions in Perry County, IN. (Thanks to Alois Treindl for pointing out the 1967/9 errors.) Edmonton did not observe DST in 1967 or 1969. In 1946 Vancouver ended DST on 09-29 not 10-13, and Vienna ended DST on 10-07 not 10-06. In 1945 Königsberg (now Kaliningrad) switched from +01/+02 to +02/+03 on 04-10 not 01-01, and its +02/+03 is abbreviated EET/EEST, not CET/CEST. (Thanks to Alois Treindl.) In 1946 Königsberg switched to +03 on 04-07 not 01-01. In 1946 Louisville switched from CST to CDT on 04-28 at 00:01, not 01-01 at 00:00. (Thanks to Alois Treindl and Michael Deckers.) Also, it switched from CST to CDT on 1950-04-30, not 1947-04-27. The 1892-05-01 transition in Brussels was at 00:17:30, not at noon. (Thanks to Michael Deckers.) Changes to past time zone abbreviations and DST flags Hong Kong Winter Time, observed from 1941-10-01 to 1941-12-25, is now flagged as DST and is abbreviated HKWT not HKT. Changes to code leapseconds.awk now relies only on its input data, rather than also relying on its comments. (Inspired by code from Dennis Ferguson and Chris Woodbury.) The code now defends against CRLFs in leap-seconds.list. (Thanks to Brian Inglis and Chris Woodbury.) Changes to documentation and commentary theory.html discusses leap seconds. (Thanks to Steve Summit.) Nashville's newspapers dueled about the time of day in the 1950s. (Thanks to John Seigenthaler.) Liechtenstein observed Swiss DST in 1941/2. (Thanks to Alois Treindl.) Release 2019b - 2019-07-01 00:09:53 -0700 Briefly: Brazil no longer observes DST. 'zic -b slim' outputs smaller TZif files; please try it out. Palestine's 2019 spring-forward transition was on 03-29, not 03-30. Changes to future timestamps Brazil has canceled DST and will stay on standard time indefinitely. (Thanks to Steffen Thorsen, Marcus Diniz, and Daniel Soares de Oliveira.) Predictions for Morocco now go through 2087 instead of 2037, to work around a problem on newlib when using TZif files output by zic 2019a or earlier. (Problem reported by David Gauchard.) Changes to past and future timestamps Palestine's 2019 spring transition was 03-29 at 00:00, not 03-30 at 01:00. (Thanks to Sharef Mustafa and Even Scharning.) Guess future transitions to be March's last Friday at 00:00. Changes to past timestamps Hong Kong's 1941-06-15 spring-forward transition was at 03:00, not 03:30. Its 1945 transition from JST to HKT was on 11-18 at 02:00, not 09-15 at 00:00. In 1946 its spring-forward transition was on 04-21 at 00:00, not the previous day at 03:30. From 1946 through 1952 its fall-back transitions occurred at 04:30, not at 03:30. In 1947 its fall-back transition was on 11-30, not 12-30. (Thanks to P Chan.) Changes to past time zone abbreviations Italy's 1866 transition to Rome Mean Time was on December 12, not September 22. This affects only the time zone abbreviation for Europe/Rome between those dates. (Thanks to Stephen Trainor and Luigi Rosa.) Changes affecting metadata only Add info about the Crimea situation in zone1970.tab and zone.tab. (Problem reported by Serhii Demediuk.) Changes to code zic's new -b option supports a way to control data bloat and to test for year-2038 bugs in software that reads TZif files. 'zic -b fat' and 'zic -b slim' generate larger and smaller output; for example, changing from fat to slim shrinks the Europe/London file from 3648 to 1599 bytes, saving about 56%. Fat and slim files represent the same set of timestamps and use the same TZif format as documented in tzfile(5) and in Internet RFC 8536. Fat format attempts to work around bugs or incompatibilities in older software, notably software that mishandles 64-bit TZif data or uses obsolete TZ strings like "EET-2EEST" that lack DST rules. Slim format is more efficient and does not work around 64-bit bugs or obsolete TZ strings. Currently zic defaults to fat format unless you compile with -DZIC_BLOAT_DEFAULT=\"slim\"; this out-of-the-box default is intended to change in future releases as the buggy software often mishandles timestamps anyway. zic no longer treats a set of rules ending in 2037 specially. Previously, zic assumed that such a ruleset meant that future timestamps could not be predicted, and therefore omitted a POSIX-like TZ string in the TZif output. The old behavior is no longer needed for current tzdata, and caused problems with newlib when used with older tzdata (reported by David Gauchard). zic no longer generates some artifact transitions. For example, Europe/London no longer has a no-op transition in January 1996. Changes to build procedure tzdata.zi now assumes zic 2017c or later. This shrinks tzdata.zi by a percent or so. Changes to documentation and commentary The Makefile now documents the POSIXRULES macro as being obsolete, and similarly, zic's -p POSIXRULES option is now documented as being obsolete. Although the POSIXRULES feature still exists and works as before, in practice it is rarely used for its intended purpose, and it does not work either in the default reference implementation (for timestamps after 2037) or in common implementations such as GNU/Linux (for contemporary timestamps). Since POSIXRULES was designed primarily as a temporary transition facility for System V platforms that died off decades ago, it is being decommissioned rather than institutionalized. New info on Bonin Islands and Marcus (thanks to Wakaba and Phake Nick). Release 2019a - 2019-03-25 22:01:33 -0700 Briefly: Palestine "springs forward" on 2019-03-30 instead of 2019-03-23. Metlakatla "fell back" to rejoin Alaska Time on 2019-01-20 at 02:00. Changes to past and future timestamps Palestine will not start DST until 2019-03-30, instead of 2019-03-23 as previously predicted. Adjust our prediction by guessing that spring transitions will be between 24 and 30 March, which matches recent practice since 2016. (Thanks to Even Scharning and Tim Parenti.) Metlakatla ended its observance of Pacific standard time, rejoining Alaska Time, on 2019-01-20 at 02:00. (Thanks to Ryan Stanley and Tim Parenti.) Changes to past timestamps Israel observed DST in 1980 (08-02/09-13) and 1984 (05-05/08-25). (Thanks to Alois Treindl and Isaac Starkman.) Changes to time zone abbreviations Etc/UCT is now a backward-compatibility link to Etc/UTC, instead of being a separate zone that generates the abbreviation "UCT", which nowadays is typically a typo. (Problem reported by Isiah Meadows.) Changes to code zic now has an -r option to limit the time range of output data. For example, 'zic -r @1000000000' limits the output data to timestamps starting 1000000000 seconds after the Epoch. This helps shrink output size and can be useful for applications not needing the full timestamp history, such as TZDIST truncation; see Internet RFC 8536 section 5.1. (Inspired by a feature request from Christopher Wong, helped along by bug reports from Wong and from Tim Parenti.) Changes to documentation Mention Internet RFC 8536 (February 2019), which documents TZif. tz-link.html now cites tzdata-meta . Release 2018i - 2018-12-30 11:05:43 -0800 Briefly: São Tomé and Príncipe switches from +01 to +00 on 2019-01-01. Changes to future timestamps Due to a change in government, São Tomé and Príncipe switches back from +01 to +00 on 2019-01-01 at 02:00. (Thanks to Vadim Nasardinov and Michael Deckers.) Release 2018h - 2018-12-23 17:59:32 -0800 Briefly: Qyzylorda, Kazakhstan moved from +06 to +05 on 2018-12-21. New zone Asia/Qostanay because Qostanay, Kazakhstan didn't move. Metlakatla, Alaska observes PST this winter only. Guess Morocco will continue to adjust clocks around Ramadan. Add predictions for Iran from 2038 through 2090. Changes to future timestamps Guess that Morocco will continue to fall back just before and spring forward just after Ramadan, the practice since 2012. (Thanks to Maamar Abdelkader.) This means Morocco will observe negative DST during Ramadan in main and vanguard formats, and in rearguard format it stays in the +00 timezone and observes ordinary DST in all months other than Ramadan. As before, extend this guesswork to the year 2037. As a consequence, Morocco is scheduled to observe three DST transitions in some Gregorian years (e.g., 2033) due to the mismatch between the Gregorian and Islamic calendars. The table of exact transitions for Iranian DST has been extended. It formerly cut off before the year 2038 in a nod to 32-bit time_t. It now cuts off before 2091 as there is doubt about how the Persian calendar will treat 2091. This change predicts DST transitions in 2038-9, 2042-3, and 2046-7 to occur one day later than previously predicted. As before, post-cutoff transitions are approximated. Changes to past and future timestamps Qyzylorda (aka Kyzylorda) oblast in Kazakhstan moved from +06 to +05 on 2018-12-21. This is a zone split as Qostanay (aka Kostanay) did not switch, so create a zone Asia/Qostanay. Metlakatla moved from Alaska to Pacific standard time on 2018-11-04. It did not change clocks that day and remains on -08 this winter. (Thanks to Ryan Stanley.) It will revert to the usual Alaska rules next spring, so this change affects only timestamps from 2018-11-04 through 2019-03-10. Change to past timestamps Kwajalein's 1993-08-20 transition from -12 to +12 was at 24:00, not 00:00. I transcribed the time incorrectly from Shanks. (Thanks to Phake Nick.) Nauru's 1979 transition was on 02-10 at 02:00, not 05-01 at 00:00. (Thanks to Phake Nick.) Guam observed DST irregularly from 1959 through 1977. (Thanks to Phake Nick.) Hong Kong observed DST in 1941 starting 06-15 (not 04-01), then on 10-01 changed standard time to +08:30 (not +08). Its transition back to +08 after WWII was on 1945-09-15, not the previous day. Its 1904-10-30 change took effect at 01:00 +08 (not 00:00 LMT). (Thanks to Phake Nick, Steve Allen, and Joseph Myers.) Also, its 1952 fallback was on 11-02 (not 10-25). This release contains many changes to timestamps before 1946 due to Japanese possession or occupation of Pacific/Chuuk, Pacific/Guam, Pacific/Kosrae, Pacific/Kwajalein, Pacific/Majuro, Pacific/Nauru, Pacific/Palau, and Pacific/Pohnpei. (Thanks to Phake Nick.) Assume that the Spanish East Indies was like the Philippines and observed American time until the end of 1844. This affects Pacific/Chuuk, Pacific/Kosrae, Pacific/Palau, and Pacific/Pohnpei. Changes to past tm_isdst flags For the recent Morocco change, the tm_isdst flag should be 1 from 2018-10-27 00:00 to 2018-10-28 03:00. (Thanks to Michael Deckers.) Give a URL to the official decree. (Thanks to Matt Johnson.) Release 2018g - 2018-10-26 22:22:45 -0700 Briefly: Morocco switches to permanent +01 on 2018-10-28. Changes to future timestamps Morocco switches from +00/+01 to permanent +01 effective 2018-10-28, so its clocks will not fall back as previously scheduled. (Thanks to Mohamed Essedik Najd and Brian Inglis.) Changes to code When generating TZif files with leap seconds, zic no longer uses a format that trips up older 32-bit clients, fixing a bug introduced in 2018f. (Reported by Daniel Fischer.) Also, the zic workaround for QTBUG-53071 now also works for TZif files with leap seconds. The translator to rearguard format now rewrites the line "Rule Japan 1948 1951 - Sep Sat>=8 25:00 0 S" to "Rule Japan 1948 1951 - Sep Sun>=9 1:00 0 S". This caters to zic before 2007 and to Oracle TZUpdater 2.2.0 and earlier. (Reported by Christos Zoulas.) Changes to past time zone abbreviations Change HDT to HWT/HPT for WWII-era abbreviations in Hawaii. This reverts to 2011h, as the abbreviation change in 2011i was likely inadvertent. Changes to documentation tzfile.5 has new sections on interoperability issues. Release 2018f - 2018-10-18 00:14:18 -0700 Briefly: Volgograd moves from +03 to +04 on 2018-10-28. Fiji ends DST 2019-01-13, not 2019-01-20. Most of Chile changes DST dates, effective 2019-04-06. Changes to future timestamps Volgograd moves from +03 to +04 on 2018-10-28 at 02:00. (Thanks to Alexander Fetisov and Stepan Golosunov.) Fiji ends DST 2019-01-13 instead of the 2019-01-20 previously predicted. (Thanks to Raymond Kumar.) Adjust future predictions accordingly. Most of Chile will end DST on the first Saturday in April at 24:00 mainland time, and resume DST on the first Saturday in September at 24:00 mainland time. The changes are effective from 2019-04-06, and do not affect the Magallanes region modeled by America/Punta_Arenas. (Thanks to Juan Correa and Tim Parenti.) Adjust future predictions accordingly. Changes to past timestamps The 2018-05-05 North Korea 30-minute time zone change took place at 23:30 the previous day, not at 00:00 that day. China's 1988 spring-forward transition was on April 17, not April 10. Its DST transitions in 1986/91 were at 02:00, not 00:00. (Thanks to P Chan.) Fix several issues for Macau before 1992. Macau's pre-1904 LMT was off by 10 s. Macau switched to +08 in 1904 not 1912, and temporarily switched to +09/+10 during World War II. Macau observed DST in 1942/79, not 1961/80, and there were several errors for transition times and dates. (Thanks to P Chan.) The 1948-1951 fallback transitions in Japan were at 25:00 on September's second Saturday, not at 24:00. (Thanks to Phake Nick.) zic turns this into 01:00 on the day after September's second Saturday, which is the best that POSIX or C platforms can do. Incorporate 1940-1949 Asia/Shanghai DST transitions from a 2014 paper by Li Yu, replacing more-questionable data from Shanks. Changes to time zone abbreviations Use "PST" and "PDT" for Philippine time. (Thanks to Paul Goyette.) Changes to code zic now always generates TZif files where time type 0 is used for timestamps before the first transition. This simplifies the reading of TZif files and should not affect behavior of existing TZif readers because the same set of time types is used; only their internal indexes may have changed. This affects only the legacy zones EST5EDT, CST6CDT, MST7MDT, PST8PDT, CET, MET, and EET, which previously used nonzero types for these timestamps. Because of the type 0 change, zic no longer outputs a dummy transition at time -2**59 (before the Big Bang), as clients should no longer need this to handle historical timestamps correctly. This reverts a change introduced in 2013d and shrinks most TZif files by a few bytes. zic now supports negative time-of-day in Rule and Leap lines, e.g., "Rule X min max - Apr lastSun -6:00 1:00 -" means the transition occurs at 18:00 on the Saturday before the last Sunday in April. This behavior was documented in 2018a but the code did not entirely match the documentation. localtime.c no longer requires at least one time type in TZif files that lack transitions or have a POSIX-style TZ string. This future-proofs the code against possible future extensions to the format that would allow TZif files with POSIX-style TZ strings and without transitions or time types. A read-access subscript error in localtime.c has been fixed. It could occur only in TZif files with timecnt == 0, something that does not happen in practice now but could happen in future versions. localtime.c no longer ignores TZif POSIX-style TZ strings that specify only standard time. Instead, these TZ strings now override the default time type for timestamps after the last transition (or for all timestamps if there are no transitions), just as DST strings specifying DST have always done. leapseconds.awk now outputs "#updated" and "#expires" comments, and supports leap seconds at the ends of months other than June and December. (Inspired by suggestions from Chris Woodbury.) Changes to documentation New restrictions: A Rule name must start with a character that is neither an ASCII digit nor "-" nor "+", and an unquoted name should not use characters in the set "!$%&'()*,/:;<=>?@[\]^`{|}~". The latter restriction makes room for future extensions (a possibility noted by Tom Lane). tzfile.5 now documents what time types apply before the first and after the last transition, if any. Documentation now uses the spelling "timezone" for a TZ setting that determines timestamp history, and "time zone" for a geographic region currently sharing the same standard time. The name "TZif" is now used for the tz binary data format. tz-link.htm now mentions the A0 TimeZone Migration utilities. (Thanks to Aldrin Martoq for the link.) Changes to build procedure New 'make' target 'rearguard_tarballs' to build the rearguard tarball only. This is a convenience on platforms that lack lzip if you want to build the rearguard tarball. (Problem reported by Deborah Goldsmith.) tzdata.zi is now more stable from release to release. (Problem noted by Tom Lane.) It is also a bit shorter. tzdata.zi now can contain comment lines documenting configuration information, such as which data format was selected, which input files were used, and how leap seconds are treated. (Problems noted by Lester Caine and Brian Inglis.) If the Makefile defaults are used these comment lines are absent, for backward compatibility. A redistributor intending to alter its copy of the files should also append "-LABEL" to the 'version' file's first line, where "LABEL" identifies the redistributor's change. Release 2018e - 2018-05-01 23:42:51 -0700 Briefly: North Korea switches back to +09 on 2018-05-05. The main format uses negative DST again, for Ireland etc. 'make tarballs' now also builds a rearguard tarball. New 's' and 'd' suffixes in SAVE columns of Rule and Zone lines. Changes to past and future timestamps North Korea switches back from +0830 to +09 on 2018-05-05. (Thanks to Kang Seonghoon, Arthur David Olson, Seo Sanghyeon, and Tim Parenti.) Bring back the negative-DST changes of 2018a, except be more compatible with data parsers that do not support negative DST. Also, this now affects historical timestamps in Namibia and the former Czechoslovakia, not just Ireland. The main format now uses negative DST to model timestamps in Europe/Dublin (from 1971 on), Europe/Prague (1946/7), and Africa/Windhoek (1994/2017). This does not affect UT offsets, only time zone abbreviations and the tm_isdst flag. Also, this does not affect rearguard or vanguard formats; effectively the main format now uses vanguard instead of rearguard format. Data parsers that do not support negative DST can still use data from the rearguard tarball described below. Changes to build procedure The command 'make tarballs' now also builds the tarball tzdataVERSION-rearguard.tar.gz, which is like tzdataVERSION.tar.gz except that it uses rearguard format intended for trailing-edge data parsers. Changes to data format and to code The SAVE column of Rule and Zone lines can now have an 's' or 'd' suffix, which specifies whether the adjusted time is standard time or daylight saving time. If no suffix is given, daylight saving time is used if and only if the SAVE column is nonzero; this is the longstanding behavior. Although this new feature is not used in tzdata, it could be used to specify the legal time in Namibia 1994-2017, as opposed to the popular time (see below). Changes to past timestamps From 1994 through 2017 Namibia observed DST in winter, not summer. That is, it used negative DST, as Ireland still does. This change does not affect UTC offsets; it affects only the tm_isdst flag and the abbreviation used during summer, which is now CAT, not WAST. Although (as noted by Michael Deckers) summer and winter time were both simply called "standard time" in Namibian law, in common practice winter time was considered to be DST (as noted by Stephen Colebourne). The full effect of this change is only in vanguard and main format; in rearguard format, the tm_isdst flag is still zero in winter and nonzero in summer. In 1946/7 Czechoslovakia also observed negative DST in winter. The full effect of this change is only in vanguard and main formats; in rearguard format, it is modeled as plain GMT without daylight saving. Also, the dates of some 1944/5 DST transitions in Czechoslovakia have been changed. Release 2018d - 2018-03-22 07:05:46 -0700 Briefly: Palestine starts DST a week earlier in 2018. Add support for vanguard and rearguard data consumers. Add subsecond precision to source data format, though not to data. Changes to future timestamps In 2018, Palestine starts DST on March 24, not March 31. Adjust future predictions accordingly. (Thanks to Sharef Mustafa.) Changes to past and future timestamps Casey Station in Antarctica changed from +11 to +08 on 2018-03-11 at 04:00. (Thanks to Steffen Thorsen.) Changes to past timestamps Historical transitions for Uruguay, represented by America/Montevideo, have been updated per official legal documents, replacing previous data mainly originating from the inventions of Shanks & Pottenger. This has resulted in adjustments ranging from 30 to 90 minutes in either direction over at least two dozen distinct periods ranging from one day to several years in length. A mere handful of pre-1991 transitions are unaffected; data since then has come from more reliable contemporaneous reporting. These changes affect various timestamps in 1920-1923, 1936, 1939, 1942-1943, 1959, 1966-1970, 1972, 1974-1980, and 1988-1990. Additionally, Uruguay's pre-standard-time UT offset has been adjusted westward by 7 seconds, from UT-03:44:44 to UT-03:44:51, to match the location of the Observatory of the National Meteorological Institute in Montevideo. (Thanks to Jeremie Bonjour, Tim Parenti, and Michael Deckers.) East Kiribati skipped New Year's Eve 1994, not New Year's Day 1995. (Thanks to Kerry Shetline.) Fix the 1912-01-01 transition for Portugal and its colonies. This transition was at 00:00 according to the new UT offset, not according to the old one. Also assume that Cape Verde switched on the same date as the rest, not in 1907. This affects Africa/Bissau, Africa/Sao_Tome, Asia/Macau, Atlantic/Azores, Atlantic/Cape_Verde, Atlantic/Madeira, and Europe/Lisbon. (Thanks to Michael Deckers.) Fix an off-by-1 error for pre-1913 timestamps in Jamaica and in Turks & Caicos. Changes to past time zone abbreviations MMT took effect in Uruguay from 1908-06-10, not 1898-06-28. There is no clock change associated with the transition. Changes to build procedure The new DATAFORM macro in the Makefile lets the installer choose among three source data formats. The idea is to lessen downstream disruption when data formats are improved. * DATAFORM=vanguard installs from the latest, bleeding-edge format. DATAFORM=main (the default) installs from the format used in the 'africa' etc. files. DATAFORM=rearguard installs from a trailing-edge format. Eventually, elements of today's vanguard format should move to the main format, and similarly the main format's features should eventually move to the rearguard format. * In the current version, the main and rearguard formats are identical and match that of 2018c, so this change does not affect default behavior. The vanguard format currently contains one feature not in the main format: negative SAVE values. This improves support for Ireland, which uses Irish Standard Time (IST, UTC+01) in summer and GMT (UTC) in winter. tzcode has supported negative SAVE values for decades, and this feature should move to the main format soon. However, it will not move to the rearguard format for quite some time because some downstream parsers do not support it. * The build procedure constructs three files vanguard.zi, main.zi, and rearguard.zi, one for each format. Although the files represent essentially the same data, they may have minor discrepancies that users are not likely to notice. The files are intended for downstream data consumers and are not installed. Zoneinfo parsers that do not support negative SAVE values should start using rearguard.zi, so that they will be unaffected when the negative-DST feature moves from vanguard to main. Bleeding-edge Zoneinfo parsers that support the new features already can use vanguard.zi; in this respect, current tzcode is bleeding-edge. The Makefile should now be safe for parallelized builds, and 'make -j to2050new.tzs' is now much faster on a multiprocessor host with GNU Make. When built with -DSUPPRESS_TZDIR, the tzcode library no longer prepends TZDIR/ to file names that do not begin with '/'. This is not recommended for general use, due to its security implications. (From a suggestion by Manuela Friedrich.) Changes to code zic now accepts subsecond precision in expressions like 00:19:32.13, which is approximately the legal time of the Netherlands from 1835 to 1937. However, because it is questionable whether the few recorded uses of non-integer offsets had subsecond precision in practice, there are no plans for tzdata to use this feature. (Thanks to Steve Allen for pointing out the limitations of historical data in this area.) The code is a bit more portable to MS-Windows. Installers can compile with -DRESERVE_STD_EXT_IDS on MS-Windows platforms that reserve identifiers like 'localtime'. (Thanks to Manuela Friedrich.) Changes to documentation and commentary theory.html now outlines tzdb's extensions to POSIX's model for civil time, and has a section "POSIX features no longer needed" that lists POSIX API components that are now vestigial. (From suggestions by Steve Summit.) It also better distinguishes time zones from tz regions. (From a suggestion by Guy Harris.) Commentary is now more consistent about using the phrase "daylight saving time", to match the C name tm_isdst. Daylight saving time need not occur in summer, and need not have a positive offset from standard time. Commentary about historical transitions in Uruguay has been expanded with links to many relevant legal documents. (Thanks to Tim Parenti.) Commentary now uses some non-ASCII characters with Unicode value less than U+0100, as they can be useful and should work even with older editors such as XEmacs. Release 2018c - 2018-01-22 23:00:44 -0800 Briefly: Revert Irish changes that relied on negative SAVE values. Changes to tm_isdst Revert the 2018a change to Europe/Dublin. As before, this change does not affect UT offsets or abbreviations; it affects only whether timestamps are considered to be standard time or daylight-saving time, as expressed in the tm_isdst flag of C's struct tm type. This reversion is intended to be a temporary workaround for problems discovered with downstream uses of releases 2018a and 2018b, which implemented Irish time by using negative SAVE values in the Eire rules of the 'europe' file. Although negative SAVE values have been part of tzcode for many years and are supported by many platforms, they were not documented before 2018a and ICU and OpenJDK do not currently support them. A mechanism to export data to platforms lacking support for negative DST is planned to be developed before the change is reapplied. (Problems reported by Deborah Goldsmith and Stephen Colebourne.) Changes to past timestamps Japanese DST transitions (1948-1951) were Sundays at 00:00, not Saturdays or Sundays at 02:00. (Thanks to Takayuki Nikai.) Changes to build procedure The build procedure now works around mawk 1.3.3's lack of support for character class expressions. (Problem reported by Ohyama.) Release 2018b - 2018-01-17 23:24:48 -0800 Briefly: Fix a packaging problem in tz2018a, which was missing 'pacificnew'. Changes to build procedure The distribution now contains the file 'pacificnew' again. This file was inadvertently omitted in the 2018a distribution. (Problem reported by Matias Fonzo.) Release 2018a - 2018-01-12 22:29:21 -0800 Briefly: São Tomé and Príncipe switched from +00 to +01. Brazil's DST will now start on November's first Sunday. Ireland's standard time is now in the summer, not the winter. Use Debian-style installation locations, instead of 4.3BSD-style. New zic option -t. Changes to past and future timestamps São Tomé and Príncipe switched from +00 to +01 on 2018-01-01 at 01:00. (Thanks to Steffen Thorsen and Michael Deckers.) Changes to future timestamps Starting in 2018 southern Brazil will begin DST on November's first Sunday instead of October's third Sunday. (Thanks to Steffen Thorsen.) Changes to past timestamps A discrepancy of 4 s in timestamps before 1931 in South Sudan has been corrected. The 'backzone' and 'zone.tab' files did not agree with the 'africa' and 'zone1970.tab' files. (Problem reported by Michael Deckers.) The abbreviation invented for Bolivia Summer Time (1931-2) is now BST instead of BOST, to be more consistent with the convention used for Latvian Summer Time (1918-9) and for British Summer Time. Changes to tm_isdst Change Europe/Dublin so that it observes Irish Standard Time (UT +01) in summer and GMT (as negative daylight-saving) in winter, instead of observing standard time (GMT) in winter and Irish Summer Time (UT +01) in summer. This change does not affect UT offsets or abbreviations; it affects only whether timestamps are considered to be standard time or daylight-saving time, as expressed in the tm_isdst flag of C's struct tm type. (Discrepancy noted by Derick Rethans.) Changes to build procedure The default installation locations have been changed to mostly match Debian circa 2017, instead of being designed as an add-on to 4.3BSD circa 1986. This affects the Makefile macros TOPDIR, TZDIR, MANDIR, and LIBDIR. New Makefile macros TZDEFAULT, USRDIR, USRSHAREDIR, BINDIR, ZDUMPDIR, and ZICDIR let installers tailor locations more precisely. (This responds to suggestions from Brian Inglis and from Steve Summit.) The default installation procedure no longer creates the backward-compatibility link US/Pacific-New, which causes confusion during user setup (e.g., see Debian bug 815200). Use 'make BACKWARD="backward pacificnew"' to create the link anyway, for now. Eventually we plan to remove the link entirely. tzdata.zi now contains a version-number comment. (Suggested by Tom Lane.) The Makefile now quotes values like BACKWARD more carefully when passing them to the shell. (Problem reported by Zefram.) Builders no longer need to specify -DHAVE_SNPRINTF on platforms that have snprintf and use pre-C99 compilers. (Problem reported by Jon Skeet.) Changes to code zic has a new option -t FILE that specifies the location of the file that determines local time when TZ is unset. The default for this location can be configured via the new TZDEFAULT makefile macro, which defaults to /etc/localtime. Diagnostics and commentary now distinguish UT from UTC more carefully; see theory.html for more information about UT vs UTC. zic has been ported to GCC 8's -Wstringop-truncation option. (Problem reported by Martin Sebor.) Changes to documentation and commentary The zic man page now documents the longstanding behavior that times and years can be out of the usual range, with negative times counting backwards from midnight and with year 0 preceding year 1. (Problem reported by Michael Deckers.) The theory.html file now mentions the POSIX limit of six chars per abbreviation, and lists alphabetic abbreviations used. The files tz-art.htm and tz-link.htm have been renamed to tz-art.html and tz-link.html, respectively, for consistency with other file names and to simplify web server configuration. Release 2017c - 2017-10-20 14:49:34 -0700 Briefly: Northern Cyprus switches from +03 to +02/+03 on 2017-10-29. Fiji ends DST 2018-01-14, not 2018-01-21. Namibia switches from +01/+02 to +02 on 2018-04-01. Sudan switches from +03 to +02 on 2017-11-01. Tonga likely switches from +13/+14 to +13 on 2017-11-05. Turks & Caicos switches from -04 to -05/-04 on 2018-11-04. A new file tzdata.zi now holds a small text copy of all data. The zic input format has been regularized slightly. Changes to future timestamps Northern Cyprus has decided to resume EU rules starting 2017-10-29, thus reinstituting winter time. Fiji ends DST 2018-01-14 instead of the 2018-01-21 previously predicted. (Thanks to Dominic Fok.) Adjust future predictions accordingly. Namibia will switch from +01 with DST to +02 all year on 2017-09-03 at 02:00. This affects UT offsets starting 2018-04-01 at 02:00. (Thanks to Steffen Thorsen.) Sudan will switch from +03 to +02 on 2017-11-01. (Thanks to Ahmed Atyya and Yahia Abdalla.) South Sudan is not switching, so Africa/Juba is no longer a link to Africa/Khartoum. Tonga has likely ended its experiment with DST, and will not adjust its clocks on 2017-11-05. Although Tonga has not announced whether it will continue to observe DST, the IATA is assuming that it will not. (Thanks to David Wade.) Turks & Caicos will switch from -04 all year to -05 with US DST on 2018-03-11 at 03:00. This affects UT offsets starting 2018-11-04 at 02:00. (Thanks to Steffen Thorsen.) Changes to past timestamps Namibia switched from +02 to +01 on 1994-03-21, not 1994-04-03. (Thanks to Arthur David Olson.) Detroit did not observe DST in 1967. Use railway time for Asia/Kolkata before 1941, by switching to Madras local time (UT +052110) in 1870, then to IST (UT +0530) in 1906. Also, treat 1941-2's +0630 as DST, like 1942-5. Europe/Dublin's 1946 and 1947 fallback transitions occurred at 02:00 standard time, not 02:00 DST. (Thanks to Michael Deckers.) Pacific/Apia and Pacific/Pago_Pago switched from Antipodean to American time in 1892, not 1879. (Thanks to Michael Deckers.) Adjust the 1867 transition in Alaska to better reflect the historical record, by changing it to occur on 1867-10-18 at 15:30 Sitka time rather than at the start of 1867-10-17 local time. Although strictly speaking this is accurate only for Sitka, the rest of Alaska's blanks need to be filled in somehow. Fix off-by-one errors in UT offsets for Adak and Nome before 1867. (Thanks to Michael Deckers.) Add 7 s to the UT offset in Asia/Yangon before 1920. Changes to zone names Remove Canada/East-Saskatchewan from the 'backward' file, as it exceeded the 14-character limit and was an unused misnomer anyway. Changes to build procedure To support applications that prefer to read time zone data in text form, two zic input files tzdata.zi and leapseconds are now installed by default. The commands 'zic tzdata.zi' and 'zic -L leapseconds tzdata.zi' can reproduce the tzdata binary files without and with leap seconds, respectively. To prevent these two new files from being installed, use 'make TZDATA_TEXT=', and to suppress leap seconds from the tzdata text installation, use 'make TZDATA_TEXT=tzdata.zi'. 'make BACKWARD=' now suppresses backward-compatibility names like 'US/Pacific' that are defined in the 'backward' and 'pacificnew' files. 'make check' now works on systems that lack a UTF-8 locale, or that lack the nsgmls program. Set UTF8_LOCALE to configure the name of a UTF-8 locale, if you have one. Y2K runtime checks are no longer enabled by default. Add -DDEPRECATE_TWO_DIGIT_YEARS to CFLAGS to enable them, instead of adding -DNO_RUN_TIME_WARNINGS_ABOUT_YEAR_2000_PROBLEMS_THANK_YOU to disable them. (New name suggested by Brian Inglis.) The build procedure for zdump now works on AIX 7.1. (Problem reported by Kees Dekker.) Changes to code zic and the reference runtime now reject multiple leap seconds within 28 days of each other, or leap seconds before the Epoch. As a result, support for double leap seconds, which was obsolescent and undocumented, has been removed. Double leap seconds were an error in the C89 standard; they have never existed in civil timekeeping. (Thanks to Robert Elz and Bradley White for noticing glitches in the code that uncovered this problem.) zic now warns about use of the obsolescent and undocumented -y option, and about use of the obsolescent TYPE field of Rule lines. zic now allows unambiguous abbreviations like "Sa" and "Su" for weekdays; formerly it rejected them due to a bug. Conversely, zic no longer considers non-prefixes to be abbreviations; for example, it no longer accepts "lF" as an abbreviation for "lastFriday". Also, zic warns about the undocumented usage with a "last-" prefix, e.g., "last-Fri". Similarly, zic now accepts the unambiguous abbreviation "L" for "Link" in ordinary context and for "Leap" in leap-second context. Conversely, zic no longer accepts non-prefixes such as "La" as abbreviations for words like "Leap". zic no longer accepts leap second lines in ordinary input, or ordinary lines in leap second input. Formerly, zic sometimes warned about this undocumented usage and handled it incorrectly. The new macro HAVE_TZNAME governs whether the tzname external variable is exported, instead of USG_COMPAT. USG_COMPAT now governs only the external variables "timezone" and "daylight". This change is needed because the three variables are not in the same category: although POSIX requires tzname, it specifies the other two variables as optional. Also, USG_COMPAT is now 1 or 0: if not defined, the code attempts to guess it from other macros. localtime.c and difftime.c no longer require stdio.h, and .c files other than zic.c no longer require sys/wait.h. zdump.c no longer assumes snprintf. (Reported by Jonathan Leffler.) Calculation of time_t extrema works around a bug in GCC 4.8.4 (Reported by Stan Shebs and Joseph Myers.) zic.c no longer mistranslates formats of line numbers in non-English locales. (Problem reported by Benno Schulenberg.) Several minor changes have been made to the code to make it a bit easier to port to MS-Windows and Solaris. (Thanks to Kees Dekker for reporting the problems.) Changes to documentation and commentary The two new files 'theory.html' and 'calendars' contain the contents of the removed file 'Theory'. The goal is to document tzdb theory more accessibly. The zic man page now documents abbreviation rules. tz-link.htm now covers how to apply tzdata changes to clients. (Thanks to Jorge Fábregas for the AIX link.) It also mentions MySQL. The leap-seconds.list URL has been updated to something that is more reliable for tzdb. (Thanks to Tim Parenti and Brian Inglis.) Release 2017b - 2017-03-17 07:30:38 -0700 Briefly: Haiti has resumed DST. Changes to past and future timestamps Haiti resumed observance of DST in 2017. (Thanks to Steffen Thorsen.) Changes to past timestamps Liberia changed from -004430 to +00 on 1972-01-07, not 1972-05-01. Use "MMT" to abbreviate Liberia's time zone before 1972, as "-004430" is one byte over the POSIX limit. (Problem reported by Derick Rethans.) Changes to code The reference localtime implementation now falls back on the current US daylight-saving transition rules rather than the 1987-2006 rules. This fallback occurs only when (1) the TZ environment variable has a value like "AST4ADT" that asks for daylight saving time but does not specify the rules, (2) there is no file by that name, and (3) the TZDEFRULES file cannot be loaded. (Thanks to Tom Lane.) Release 2017a - 2017-02-28 00:05:36 -0800 Briefly: Southern Chile moves from -04/-03 to -03, and Mongolia discontinues DST. Changes to future timestamps Mongolia no longer observes DST. (Thanks to Ganbold Tsagaankhuu.) Chile's Region of Magallanes moves from -04/-03 to -03 year-round. Its clocks diverge from America/Santiago starting 2017-05-13 at 23:00, hiving off a new zone America/Punta_Arenas. Although the Chilean government says this change expires in May 2019, for now assume it's permanent. (Thanks to Juan Correa and Deborah Goldsmith.) This also affects Antarctica/Palmer. Changes to past timestamps Fix many entries for historical timestamps for Europe/Madrid before 1979, to agree with tables compiled by Pere Planesas of the National Astronomical Observatory of Spain. As a side effect, this changes some timestamps for Africa/Ceuta before 1929, which are probably guesswork anyway. (Thanks to Steve Allen and Pierpaolo Bernardi for the heads-ups, and to Michael Deckers for correcting the 1901 transition.) Ecuador observed DST from 1992-11-28 to 1993-02-05. (Thanks to Alois Treindl.) Asia/Atyrau and Asia/Oral were at +03 (not +04) before 1930-06-21. (Thanks to Stepan Golosunov.) Changes to past and future time zone abbreviations Switch to numeric time zone abbreviations for South America, as part of the ongoing project of removing invented abbreviations. This avoids the need to invent an abbreviation for the new Chilean new zone. Similarly, switch from invented to numeric time zone abbreviations for Afghanistan, American Samoa, the Azores, Bangladesh, Bhutan, the British Indian Ocean Territory, Brunei, Cape Verde, Chatham Is, Christmas I, Cocos (Keeling) Is, Cook Is, Dubai, East Timor, Eucla, Fiji, French Polynesia, Greenland, Indochina, Iran, Iraq, Kiribati, Lord Howe, Macquarie, Malaysia, the Maldives, Marshall Is, Mauritius, Micronesia, Mongolia, Myanmar, Nauru, Nepal, New Caledonia, Niue, Norfolk I, Palau, Papua New Guinea, the Philippines, Pitcairn, Qatar, Réunion, St Pierre & Miquelon, Samoa, Saudi Arabia, Seychelles, Singapore, Solomon Is, Tokelau, Tuvalu, Wake, Vanuatu, Wallis & Futuna, and Xinjiang; for 20-minute daylight saving time in Ghana before 1943; for half-hour daylight saving time in Belize before 1944 and in the Dominican Republic before 1975; and for Canary Islands before 1946, for Guinea-Bissau before 1975, for Iceland before 1969, for Indian Summer Time before 1942, for Indonesia before around 1964, for Kenya before 1960, for Liberia before 1973, for Madeira before 1967, for Namibia before 1943, for the Netherlands in 1937-9, for Pakistan before 1971, for Western Sahara before 1977, and for Zaporozhye in 1880-1924. For Alaska time from 1900 through 1967, instead of "CAT" use the abbreviation "AST", the abbreviation commonly used at the time (Atlantic Standard Time had not been standardized yet). Use "AWT" and "APT" instead of the invented abbreviations "CAWT" and "CAPT". Use "CST" and "CDT" instead of invented abbreviations for Macau before 1999 and Taiwan before 1938, and use "JST" instead of the invented abbreviation "JCST" for Japan and Korea before 1938. Change to database entry category Move the Pacific/Johnston link from 'australasia' to 'backward', since Johnston is now uninhabited. Changes to code zic no longer mishandles some transitions in January 2038 when it attempts to work around Qt bug 53071. This fixes a bug affecting Pacific/Tongatapu that was introduced in zic 2016e. localtime.c now contains a workaround, useful when loading a file generated by a buggy zic. (Problem and localtime.c fix reported by Bradley White.) zdump -i now outputs non-hour numeric time zone abbreviations without a colon, e.g., "+0530" rather than "+05:30". This agrees with zic %z and with common practice, and simplifies auditing of zdump output. zdump is now buildable again with -DUSE_LTZ=0. (Problem reported by Joseph Myers.) zdump.c now always includes private.h, to avoid code duplication with private.h. (Problem reported by Kees Dekker.) localtime.c no longer mishandles early or late timestamps when TZ is set to a POSIX-style string that specifies DST. (Problem reported by Kees Dekker.) date and strftime now cause %z to generate "-0000" instead of "+0000" when the UT offset is zero and the time zone abbreviation begins with "-". Changes to documentation and commentary The 'Theory' file now better documents choice of historical time zone abbreviations. (Problems reported by Michael Deckers.) tz-link.htm now covers leap smearing, which is popular in clouds. Release 2016j - 2016-11-22 23:17:13 -0800 Briefly: Saratov, Russia moves from +03 to +04 on 2016-12-04. Changes to future timestamps Saratov, Russia switches from +03 to +04 on 2016-12-04 at 02:00. This hives off a new zone Europe/Saratov from Europe/Volgograd. (Thanks to Yuri Konotopov and Stepan Golosunov.) Changes to past timestamps The new zone Asia/Atyrau for Atyraū Region, Kazakhstan, is like Asia/Aqtau except it switched from +05/+06 to +04/+05 in spring 1999, not fall 1994. (Thanks to Stepan Golosunov.) Changes to past time zone abbreviations Asia/Gaza and Asia/Hebron now use "EEST", not "EET", to denote summer time before 1948. The old use of "EET" was a typo. Changes to code zic no longer mishandles file systems that lack hard links, fixing bugs introduced in 2016g. (Problems reported by Tom Lane.) Also, when the destination already contains symbolic links, zic should now work better on systems where the 'link' system call does not follow symbolic links. Changes to documentation and commentary tz-link.htm now documents the relationship between release version numbers and development-repository commit tags. (Suggested by Paul Koning.) The 'Theory' file now documents UT. iso3166.tab now accents "Curaçao", and commentary now mentions the names "Cabo Verde" and "Czechia". (Thanks to Jiří Boháč.) Release 2016i - 2016-11-01 23:19:52 -0700 Briefly: Cyprus split into two time zones on 2016-10-30, and Tonga reintroduces DST on 2016-11-06. Changes to future timestamps Pacific/Tongatapu begins DST on 2016-11-06 at 02:00, ending on 2017-01-15 at 03:00. Assume future observances in Tonga will be from the first Sunday in November through the third Sunday in January, like Fiji. (Thanks to Pulu ʻAnau.) Switch to numeric time zone abbreviations for this zone. Changes to past and future timestamps Northern Cyprus is now +03 year round, causing a split in Cyprus time zones starting 2016-10-30 at 04:00. This creates a zone Asia/Famagusta. (Thanks to Even Scharning and Matt Johnson.) Antarctica/Casey switched from +08 to +11 on 2016-10-22. (Thanks to Steffen Thorsen.) Changes to past timestamps Several corrections were made for pre-1975 timestamps in Italy. These affect Europe/Malta, Europe/Rome, Europe/San_Marino, and Europe/Vatican. First, the 1893-11-01 00:00 transition in Italy used the new UT offset (+01), not the old (+00:49:56). (Thanks to Michael Deckers.) Second, rules for daylight saving in Italy were changed to agree with Italy's National Institute of Metrological Research (INRiM) except for 1944, as follows (thanks to Pierpaolo Bernardi, Brian Inglis, and Michael Deckers): The 1916-06-03 transition was at 24:00, not 00:00. The 1916-10-01, 1919-10-05, and 1920-09-19 transitions were at 00:00, not 01:00. The 1917-09-30 and 1918-10-06 transitions were at 24:00, not 01:00. The 1944-09-17 transition was at 03:00, not 01:00. This particular change is taken from Italian law as INRiM's table, (which says 02:00) appears to have a typo here. Also, keep the 1944-04-03 transition for Europe/Rome, as Rome was controlled by Germany then. The 1967-1970 and 1972-1974 fallback transitions were at 01:00, not 00:00. Changes to code The code should now be buildable on AmigaOS merely by setting the appropriate Makefile variables. (From a patch by Carsten Larsen.) Release 2016h - 2016-10-19 23:17:57 -0700 Changes to future timestamps Asia/Gaza and Asia/Hebron end DST on 2016-10-29 at 01:00, not 2016-10-21 at 00:00. (Thanks to Sharef Mustafa.) Predict that future fall transitions will be on the last Saturday of October at 01:00, which is consistent with predicted spring transitions on the last Saturday of March. (Thanks to Tim Parenti.) Changes to past timestamps In Turkey, transitions in 1986-1990 were at 01:00 standard time not at 02:00, and the spring 1994 transition was on March 20, not March 27. (Thanks to Kıvanç Yazan.) Changes to past and future time zone abbreviations Asia/Colombo now uses numeric time zone abbreviations like "+0530" instead of alphabetic ones like "IST" and "LKT". Various English-language sources use "IST", "LKT" and "SLST", with no working consensus. (Usage of "SLST" mentioned by Sadika Sumanapala.) Changes to code zic no longer mishandles relativizing file names when creating symbolic links like /etc/localtime, when these symbolic links are outside the usual directory hierarchy. This fixes a bug introduced in 2016g. (Problem reported by Andreas Stieger.) Changes to build procedure New rules 'traditional_tarballs' and 'traditional_signatures' for building just the traditional-format distribution. (Requested by Deborah Goldsmith.) The file 'version' is now put into the tzdata tarball too. (Requested by Howard Hinnant.) Changes to documentation and commentary The 'Theory' file now has a section on interface stability. (Requested by Paul Koning.) It also mentions features like tm_zone and localtime_rz that have long been supported by the reference code. tz-link.htm has improved coverage of time zone boundaries suitable for geolocation. (Thanks to heads-ups from Evan Siroky and Matt Johnson.) The US commentary now mentions Allen and the "day of two noons". The Fiji commentary mentions the government's 2016-10-03 press release. (Thanks to Raymond Kumar.) Release 2016g - 2016-09-13 08:56:38 -0700 Changes to future timestamps Turkey switched from EET/EEST (+02/+03) to permanent +03, effective 2016-09-07. (Thanks to Burak AYDIN.) Use "+03" rather than an invented abbreviation for the new time. New leap second 2016-12-31 23:59:60 UTC as per IERS Bulletin C 52. (Thanks to Tim Parenti.) Changes to past timestamps For America/Los_Angeles, spring-forward transition times have been corrected from 02:00 to 02:01 in 1948, and from 02:00 to 01:00 in 1950-1966. For zones using Soviet time on 1919-07-01, transitions to UT-based time were at 00:00 UT, not at 02:00 local time. The affected zones are Europe/Kirov, Europe/Moscow, Europe/Samara, and Europe/Ulyanovsk. (Thanks to Alexander Belopolsky.) Changes to past and future time zone abbreviations The Factory zone now uses the time zone abbreviation -00 instead of a long English-language string, as -00 is now the normal way to represent an undefined time zone. Several zones in Antarctica and the former Soviet Union, along with zones intended for ships at sea that cannot use POSIX TZ strings, now use numeric time zone abbreviations instead of invented or obsolete alphanumeric abbreviations. The affected zones are Antarctica/Casey, Antarctica/Davis, Antarctica/DumontDUrville, Antarctica/Mawson, Antarctica/Rothera, Antarctica/Syowa, Antarctica/Troll, Antarctica/Vostok, Asia/Anadyr, Asia/Ashgabat, Asia/Baku, Asia/Bishkek, Asia/Chita, Asia/Dushanbe, Asia/Irkutsk, Asia/Kamchatka, Asia/Khandyga, Asia/Krasnoyarsk, Asia/Magadan, Asia/Omsk, Asia/Sakhalin, Asia/Samarkand, Asia/Srednekolymsk, Asia/Tashkent, Asia/Tbilisi, Asia/Ust-Nera, Asia/Vladivostok, Asia/Yakutsk, Asia/Yekaterinburg, Asia/Yerevan, Etc/GMT-14, Etc/GMT-13, Etc/GMT-12, Etc/GMT-11, Etc/GMT-10, Etc/GMT-9, Etc/GMT-8, Etc/GMT-7, Etc/GMT-6, Etc/GMT-5, Etc/GMT-4, Etc/GMT-3, Etc/GMT-2, Etc/GMT-1, Etc/GMT+1, Etc/GMT+2, Etc/GMT+3, Etc/GMT+4, Etc/GMT+5, Etc/GMT+6, Etc/GMT+7, Etc/GMT+8, Etc/GMT+9, Etc/GMT+10, Etc/GMT+11, Etc/GMT+12, Europe/Kaliningrad, Europe/Minsk, Europe/Samara, Europe/Volgograd, and Indian/Kerguelen. For Europe/Moscow the invented abbreviation MSM was replaced by +05, whereas MSK and MSD were kept as they are not our invention and are widely used. Changes to zone names Rename Asia/Rangoon to Asia/Yangon, with a backward compatibility link. (Thanks to David Massoud.) Changes to code zic no longer generates binary files containing POSIX TZ-like strings that disagree with the local time type after the last explicit transition in the data. This fixes a bug with Africa/Casablanca and Africa/El_Aaiun in some year-2037 timestamps on the reference platform. (Thanks to Alexander Belopolsky for reporting the bug and suggesting a way forward.) If the installed localtime and/or posixrules files are symbolic links, zic now keeps them symbolic links when updating them, for compatibility with platforms like OpenSUSE where other programs configure these files as symlinks. zic now avoids hard linking to symbolic links, avoids some unnecessary mkdir and stat system calls, and uses shorter file names internally. zdump has a new -i option to generate transitions in a smaller but still human-readable format. This option is experimental, and the output format may change in future versions. (Thanks to Jon Skeet for suggesting that an option was needed, and thanks to Tim Parenti and Chris Rovick for further comments.) Changes to build procedure An experimental distribution format is available, in addition to the traditional format which will continue to be distributed. The new format is a tarball tzdb-VERSION.tar.lz with signature file tzdb-VERSION.tar.lz.asc. It unpacks to a top-level directory tzdb-VERSION containing the code and data of the traditional two-tarball format, along with extra data that may be useful. (Thanks to Antonio Diaz Diaz, Oscar van Vlijmen, and many others for comments about the experimental format.) The release version number is now more accurate in the usual case where releases are built from a Git repository. For example, if 23 commits and some working-file changes have been made since release 2016g, the version number is now something like '2016g-23-g50556e3-dirty' instead of the misleading '2016g'. Tagged releases use the same version number format as before, e.g., '2016g'. To support the more accurate version number, its specification has moved from a line in the Makefile to a new source file 'version'. The experimental distribution contains a file to2050.tzs that contains what should be the output of 'zdump -i -c 2050' on primary zones. If this file is available, 'make check' now checks that zdump generates this output. 'make check_web' now works on Fedora-like distributions. Changes to documentation and commentary tzfile.5 now documents the new restriction on POSIX TZ-like strings that is now implemented by zic. Comments now cite URLs for some 1917-1921 Russian DST decrees. (Thanks to Alexander Belopolsky.) tz-link.htm mentions JuliaTime (thanks to Curtis Vogt) and Time4J (thanks to Meno Hochschild) and ThreeTen-Extra, and its description of Java 8 has been brought up to date (thanks to Stephen Colebourne). Its description of local time on Mars has been updated to match current practice, and URLs have been updated and some obsolete ones removed. Release 2016f - 2016-07-05 16:26:51 +0200 Changes affecting future timestamps The Egyptian government changed its mind on short notice, and Africa/Cairo will not introduce DST starting 2016-07-07 after all. (Thanks to Mina Samuel.) Asia/Novosibirsk switches from +06 to +07 on 2016-07-24 at 02:00. (Thanks to Stepan Golosunov.) Changes to past and future timestamps Asia/Novokuznetsk and Asia/Novosibirsk now use numeric time zone abbreviations instead of invented ones. Changes affecting past timestamps Europe/Minsk's 1992-03-29 spring-forward transition was at 02:00 not 00:00. (Thanks to Stepan Golosunov.) Release 2016e - 2016-06-14 08:46:16 -0700 Changes affecting future timestamps Africa/Cairo observes DST in 2016 from July 7 to the end of October. Guess October 27 and 24:00 transitions. (Thanks to Steffen Thorsen.) For future years, guess April's last Thursday to October's last Thursday except for Ramadan. Changes affecting past timestamps Locations while uninhabited now use '-00', not 'zzz', as a placeholder time zone abbreviation. This is inspired by Internet RFC 3339 and is more consistent with numeric time zone abbreviations already used elsewhere. The change affects several arctic and antarctic locations, e.g., America/Cambridge_Bay before 1920 and Antarctica/Troll before 2005. Asia/Baku's 1992-09-27 transition from +04 (DST) to +04 (non-DST) was at 03:00, not 23:00 the previous day. (Thanks to Michael Deckers.) Changes to code zic now outputs a dummy transition at time 2**31 - 1 in zones whose POSIX-style TZ strings contain a '<'. This mostly works around Qt bug 53071 . (Thanks to Zhanibek Adilbekov for reporting the Qt bug.) Changes affecting documentation and commentary tz-link.htm says why governments should give plenty of notice for time zone or DST changes, and refers to Matt Johnson's blog post. tz-link.htm mentions Tzdata for Elixir. (Thanks to Matt Johnson.) Release 2016d - 2016-04-17 22:50:29 -0700 Changes affecting future timestamps America/Caracas switches from -0430 to -04 on 2016-05-01 at 02:30. (Thanks to Alexander Krivenyshev for the heads-up.) Asia/Magadan switches from +10 to +11 on 2016-04-24 at 02:00. (Thanks to Alexander Krivenyshev and Matt Johnson.) New zone Asia/Tomsk, split off from Asia/Novosibirsk. It covers Tomsk Oblast, Russia, which switches from +06 to +07 on 2016-05-29 at 02:00. (Thanks to Stepan Golosunov.) Changes affecting past timestamps New zone Europe/Kirov, split off from Europe/Volgograd. It covers Kirov Oblast, Russia, which switched from +04/+05 to +03/+04 on 1989-03-26 at 02:00, roughly a year after Europe/Volgograd made the same change. (Thanks to Stepan Golosunov.) Russia and nearby locations had daylight-saving transitions on 1992-03-29 at 02:00 and 1992-09-27 at 03:00, instead of on 1992-03-28 at 23:00 and 1992-09-26 at 23:00. (Thanks to Stepan Golosunov.) Many corrections to historical time in Kazakhstan from 1991 through 2005. (Thanks to Stepan Golosunov.) Replace Kazakhstan's invented time zone abbreviations with numeric abbreviations. Changes to commentary Mention Internet RFCs 7808 (TZDIST) and 7809 (CalDAV time zone references). Release 2016c - 2016-03-23 00:51:27 -0700 Changes affecting future timestamps Azerbaijan no longer observes DST. (Thanks to Steffen Thorsen.) Chile reverts from permanent to seasonal DST. (Thanks to Juan Correa for the heads-up, and to Tim Parenti for corrections.) Guess that future transitions are August's and May's second Saturdays at 24:00 mainland time. Also, call the period from 2014-09-07 through 2016-05-14 daylight saving time instead of standard time, as that seems more appropriate now. Changes affecting past timestamps Europe/Kaliningrad and Europe/Vilnius changed from +03/+04 to +02/+03 on 1989-03-26, not 1991-03-31. Europe/Volgograd changed from +04/+05 to +03/+04 on 1988-03-27, not 1989-03-26. (Thanks to Stepan Golosunov.) Changes to commentary Several updates and URLs for historical and proposed Russian changes. (Thanks to Stepan Golosunov, Matt Johnson, and Alexander Krivenyshev.) Release 2016b - 2016-03-12 17:30:14 -0800 Compatibility note Starting with release 2016b, some data entries cause zic implementations derived from tz releases 2005j through 2015e to issue warnings like "time zone abbreviation differs from POSIX standard (+03)". These warnings should not otherwise affect zic's output and can safely be ignored on today's platforms, as the warnings refer to a restriction in POSIX.1-1988 that was removed in POSIX.1-2001. One way to suppress the warnings is to upgrade to zic derived from tz releases 2015f and later. Changes affecting future timestamps New zones Europe/Astrakhan and Europe/Ulyanovsk for Astrakhan and Ulyanovsk Oblasts, Russia, both of which will switch from +03 to +04 on 2016-03-27 at 02:00 local time. They need distinct zones since their post-1970 histories disagree. New zone Asia/Barnaul for Altai Krai and Altai Republic, Russia, which will switch from +06 to +07 on the same date and local time. The Astrakhan change is already official; the others have passed the first reading in the State Duma and are extremely likely. Also, Asia/Sakhalin moves from +10 to +11 on 2016-03-27 at 02:00. (Thanks to Alexander Krivenyshev for the heads-up, and to Matt Johnson and Stepan Golosunov for followup.) As a trial of a new system that needs less information to be made up, the new zones use numeric time zone abbreviations like "+04" instead of invented abbreviations like "ASTT". Haiti will not observe DST in 2016. (Thanks to Jean Antoine via Steffen Thorsen.) Palestine's spring-forward transition on 2016-03-26 is at 01:00, not 00:00. (Thanks to Hannah Kreitem.) Guess future transitions will be March's last Saturday at 01:00, not March's last Friday at 24:00. Changes affecting past timestamps Europe/Chisinau observed DST during 1990, and switched from +04 to +03 at 1990-05-06 02:00, instead of switching from +03 to +02. (Thanks to Stepan Golosunov.) 1991 abbreviations in Europe/Samara should be SAMT/SAMST, not KUYT/KUYST. (Thanks to Stepan Golosunov.) Changes to code tzselect's diagnostics and checking, and checktab.awk's checking, have been improved. (Thanks to J William Piggott.) tzcode now builds under MinGW. (Thanks to Ian Abbott and Esben Haabendal.) tzselect now tests Julian-date TZ settings more accurately. (Thanks to J William Piggott.) Changes to commentary Comments in zone tables have been improved. (Thanks to J William Piggott.) tzselect again limits its menu comments so that menus fit on a 24×80 alphanumeric display. A new web page tz-how-to.html. (Thanks to Bill Seymour.) In the Theory file, the description of possible time zone abbreviations in tzdata has been cleaned up, as the old description was unclear and inconsistent. (Thanks to Alain Mouette for reporting the problem.) Release 2016a - 2016-01-26 23:28:02 -0800 Changes affecting future timestamps America/Cayman will not observe daylight saving this year after all. Revert our guess that it would. (Thanks to Matt Johnson.) Asia/Chita switches from +0800 to +0900 on 2016-03-27 at 02:00. (Thanks to Alexander Krivenyshev.) Asia/Tehran now has DST predictions for the year 2038 and later, to be March 21 00:00 to September 21 00:00. This is likely better than predicting no DST, albeit off by a day every now and then. Changes affecting past and future timestamps America/Metlakatla switched from PST all year to AKST/AKDT on 2015-11-01 at 02:00. (Thanks to Steffen Thorsen.) America/Santa_Isabel has been removed, and replaced with a backward compatibility link to America/Tijuana. Its contents were apparently based on a misreading of Mexican legislation. Changes affecting past timestamps Asia/Karachi's two transition times in 2002 were off by a minute. (Thanks to Matt Johnson.) Changes affecting build procedure An installer can now combine leap seconds with use of the backzone file, e.g., with 'make PACKRATDATA=backzone REDO=posix_right zones'. The old 'make posix_packrat' rule is now marked as obsolescent. (Thanks to Ian Abbott for an initial implementation.) Changes affecting documentation and commentary A new file LICENSE makes it easier to see that the code and data are mostly public-domain. (Thanks to James Knight.) The three non-public-domain files now use the current (3-clause) BSD license instead of older versions of that license. tz-link.htm mentions the BDE library (thanks to Andrew Paprocki), CCTZ (thanks to Tim Parenti), TimeJones.com, and has a new section on editing tz source files (with a mention of Sublime zoneinfo, thanks to Gilmore Davidson). The Theory and asia files now mention the 2015 book "The Global Transformation of Time, 1870-1950", and cite a couple of reviews. The America/Chicago entry now documents the informal use of US central time in Fort Pierre, South Dakota. (Thanks to Rick McDermid, Matt Johnson, and Steve Jones.) Release 2015g - 2015-10-01 00:39:51 -0700 Changes affecting future timestamps Turkey's 2015 fall-back transition is scheduled for Nov. 8, not Oct. 25. (Thanks to Fatih.) Norfolk moves from +1130 to +1100 on 2015-10-04 at 02:00 local time. (Thanks to Alexander Krivenyshev.) Fiji's 2016 fall-back transition is scheduled for January 17, not 24. (Thanks to Ken Rylander.) Fort Nelson, British Columbia will not fall back on 2015-11-01. It has effectively been on MST (-0700) since it advanced its clocks on 2015-03-08. New zone America/Fort_Nelson. (Thanks to Matt Johnson.) Changes affecting past timestamps Norfolk observed DST from 1974-10-27 02:00 to 1975-03-02 02:00. Changes affecting code localtime no longer mishandles America/Anchorage after 2037. (Thanks to Bradley White for reporting the bug.) On hosts with signed 32-bit time_t, localtime no longer mishandles Pacific/Fiji after 2038-01-16 14:00 UTC. The localtime module allows the variables 'timezone', 'daylight', and 'altzone' to be in common storage shared with other modules, and declares them in case the system does not. (Problems reported by Kees Dekker.) On platforms with tm_zone, strftime.c now assumes it is not NULL. This simplifies the code and is consistent with zdump.c. (Problem reported by Christos Zoulas.) Changes affecting documentation The tzfile man page now documents that transition times denote the starts (not the ends) of the corresponding time periods. (Ambiguity reported by Bill Seymour.) Release 2015f - 2015-08-10 18:06:56 -0700 Changes affecting future timestamps North Korea switches to +0830 on 2015-08-15. (Thanks to Steffen Thorsen.) The abbreviation remains "KST". (Thanks to Robert Elz.) Uruguay no longer observes DST. (Thanks to Steffen Thorsen and Pablo Camargo.) Changes affecting past and future timestamps Moldova starts and ends DST at 00:00 UTC, not at 01:00 UTC. (Thanks to Roman Tudos.) Changes affecting data format and code zic's '-y YEARISTYPE' option is no longer documented. The TYPE field of a Rule line should now be '-'; the old values 'even', 'odd', 'uspres', 'nonpres', 'nonuspres' were already undocumented. Although the implementation has not changed, these features do not work in the default installation, they are not used in the data, and they are now considered obsolescent. zic now checks that two rules don't take effect at the same time. (Thanks to Jon Skeet and Arthur David Olson.) Constraints on simultaneity are now documented. The two characters '%z' in a zone format now stand for the UT offset, e.g., '-07' for seven hours behind UT and '+0530' for five hours and thirty minutes ahead. This better supports time zone abbreviations conforming to POSIX.1-2001 and later. Changes affecting installed data files Comments for America/Halifax and America/Glace_Bay have been improved. (Thanks to Brian Inglis.) Data entries have been simplified for Atlantic/Canary, Europe/Simferopol, Europe/Sofia, and Europe/Tallinn. This yields slightly smaller installed data files for Europe/Simferopol and Europe/Tallinn. It does not affect timestamps. (Thanks to Howard Hinnant.) Changes affecting code zdump and zic no longer warn about valid time zone abbreviations like '-05'. Some Visual Studio 2013 warnings have been suppressed. (Thanks to Kees Dekker.) 'date' no longer sets the time of day and its -a, -d, -n and -t options have been removed. Long obsolescent, the implementation of these features had porting problems. Builders no longer need to configure HAVE_ADJTIME, HAVE_SETTIMEOFDAY, or HAVE_UTMPX_H. (Thanks to Kees Dekker for pointing out the problem.) Changes affecting documentation The Theory file mentions naming issues earlier, as these seem to be poorly publicized (thanks to Gilmore Davidson for reporting the problem). tz-link.htm mentions Time Zone Database Parser (thanks to Howard Hinnant). Mention that Herbert Samuel introduced the term "Summer Time". Release 2015e - 2015-06-13 10:56:02 -0700 Changes affecting future timestamps Morocco will suspend DST from 2015-06-14 03:00 through 2015-07-19 02:00, not 06-13 and 07-18 as we had guessed. (Thanks to Milamber.) Assume Cayman Islands will observe DST starting next year, using US rules. Although it isn't guaranteed, it is the most likely. Changes affecting data format The file 'iso3166.tab' now uses UTF-8, so that its entries can better spell the names of Åland Islands, Côte d'Ivoire, and Réunion. Changes affecting code When displaying data, tzselect converts it to the current locale's encoding if the iconv command works. (Problem reported by random832.) tzselect no longer mishandles Dominica, fixing a bug introduced in Release 2014f. (Problem reported by Owen Leibman.) zic -l no longer fails when compiled with -DTZDEFAULT=\"/etc/localtime\". This fixes a bug introduced in Release 2014f. (Problem reported by Leonardo Chiquitto.) Release 2015d - 2015-04-24 08:09:46 -0700 Changes affecting future timestamps Egypt will not observe DST in 2015 and will consider canceling it permanently. For now, assume no DST indefinitely. (Thanks to Ahmed Nazmy and Tim Parenti.) Changes affecting past timestamps America/Whitehorse switched from UT -09 to -08 on 1967-05-28, not 1966-07-01. Also, Yukon's time zone history is documented better. (Thanks to Brian Inglis and Dennis Ferguson.) Change affecting past and future time zone abbreviations The abbreviations for Hawaii-Aleutian standard and daylight times have been changed from HAST/HADT to HST/HDT, as per US Government Printing Office style. This affects only America/Adak since 1983, as America/Honolulu was already using the new style. Changes affecting code zic has some minor performance improvements. Release 2015c - 2015-04-11 08:55:55 -0700 Changes affecting future timestamps Egypt's spring-forward transition is at 24:00 on April's last Thursday, not 00:00 on April's last Friday. 2015's transition will therefore be on Thursday, April 30 at 24:00, not Friday, April 24 at 00:00. Similar fixes apply to 2026, 2037, 2043, etc. (Thanks to Steffen Thorsen.) Changes affecting past timestamps The following changes affect some pre-1991 Chile-related timestamps in America/Santiago, Antarctica/Palmer, and Pacific/Easter. The 1910 transition was January 10, not January 1. The 1918 transition was September 10, not September 1. The UT -04 time observed from 1932 to 1942 is now considered to be standard time, not year-round DST. Santiago observed DST (UT -03) from 1946-07-15 through 1946-08-31, then reverted to standard time, then switched to -05 on 1947-04-01. Assume transitions before 1968 were at 00:00, since we have no data saying otherwise. The spring 1988 transition was 1988-10-09, not 1988-10-02. The fall 1990 transition was 1990-03-11, not 1990-03-18. Assume no UT offset change for Pacific/Easter on 1890-01-01, and omit all transitions on Pacific/Easter from 1942 through 1946 since we have no data suggesting that they existed. One more zone has been turned into a link, as it differed from an existing zone only for older timestamps. As usual, this change affects UT offsets in pre-1970 timestamps only. The zone's old contents have been moved to the 'backzone' file. The affected zone is America/Montreal. Changes affecting commentary Mention the TZUpdater tool. Mention "The Time Now". (Thanks to Brandon Ramsey.) Release 2015b - 2015-03-19 23:28:11 -0700 Changes affecting future timestamps Mongolia will start observing DST again this year, from the last Saturday in March at 02:00 to the last Saturday in September at 00:00. (Thanks to Ganbold Tsagaankhuu.) Palestine will start DST on March 28, not March 27. Also, correct the fall 2014 transition from September 26 to October 24. Adjust future predictions accordingly. (Thanks to Steffen Thorsen.) Changes affecting past timestamps The 1982 zone shift in Pacific/Easter has been corrected, fixing a 2015a regression. (Thanks to Stuart Bishop for reporting the problem.) Some more zones have been turned into links, when they differed from existing zones only for older timestamps. As usual, these changes affect UT offsets in pre-1970 timestamps only. Their old contents have been moved to the 'backzone' file. The affected zones are: America/Antigua, America/Cayman, Pacific/Midway, and Pacific/Saipan. Changes affecting time zone abbreviations Correct the 1992-2010 DST abbreviation in Volgograd from "MSK" to "MSD". (Thanks to Hank W.) Changes affecting code Fix integer overflow bug in reference 'mktime' implementation. (Problem reported by Jörg Richter.) Allow -Dtime_tz=time_t compilations, and allow -Dtime_tz=... libraries to be used in the same executable as standard-library time_t functions. (Problems reported by Bradley White.) Changes affecting commentary Cite the recent Mexican decree changing Quintana Roo's time zone. (Thanks to Carlos Raúl Perasso.) Likewise for the recent Chilean decree. (Thanks to Eduardo Romero Urra.) Update info about Mars time. Release 2015a - 2015-01-29 22:35:20 -0800 Changes affecting future timestamps The Mexican state of Quintana Roo, represented by America/Cancun, will shift from Central Time with DST to Eastern Time without DST on 2015-02-01 at 02:00. (Thanks to Steffen Thorsen and Gwillim Law.) Chile will not change clocks in April or thereafter; its new standard time will be its old daylight saving time. This affects America/Santiago, Pacific/Easter, and Antarctica/Palmer. (Thanks to Juan Correa.) New leap second 2015-06-30 23:59:60 UTC as per IERS Bulletin C 49. (Thanks to Tim Parenti.) Changes affecting past timestamps Iceland observed DST in 1919 and 1921, and its 1939 fallback transition was Oct. 29, not Nov. 29. Remove incorrect data from Shanks about time in Iceland between 1837 and 1908. Some more zones have been turned into links, when they differed from existing zones only for older timestamps. As usual, these changes affect UT offsets in pre-1970 timestamps only. Their old contents have been moved to the 'backzone' file. The affected zones are: Asia/Aden, Asia/Bahrain, Asia/Kuwait, and Asia/Muscat. Changes affecting code tzalloc now scrubs time zone abbreviations compatibly with the way that tzset always has, by replacing invalid bytes with '_' and by shortening too-long abbreviations. tzselect ports to POSIX awk implementations, no longer mishandles POSIX TZ settings when GNU awk is used, and reports POSIX TZ settings to the user. (Thanks to Stefan Kuhn.) Changes affecting build procedure 'make check' now checks for links to links in the data. One such link (for Africa/Asmera) has been fixed. (Thanks to Stephen Colebourne for pointing out the problem.) Changes affecting commentary The leapseconds file commentary now mentions the expiration date. (Problem reported by Martin Burnicki.) Update Mexican Library of Congress URL. Release 2014j - 2014-11-10 17:37:11 -0800 Changes affecting current and future timestamps Turks & Caicos' switch from US eastern time to UT -04 year-round did not occur on 2014-11-02 at 02:00. It's currently scheduled for 2015-11-01 at 02:00. (Thanks to Chris Walton.) Changes affecting past timestamps Many pre-1989 timestamps have been corrected for Asia/Seoul and Asia/Pyongyang, based on sources for the Korean-language Wikipedia entry for time in Korea. (Thanks to Sanghyuk Jung.) Also, no longer guess that Pyongyang mimicked Seoul time after World War II, as this is politically implausible. Some more zones have been turned into links, when they differed from existing zones only for older timestamps. As usual, these changes affect UT offsets in pre-1970 timestamps only. Their old contents have been moved to the 'backzone' file. The affected zones are: Africa/Addis_Ababa, Africa/Asmara, Africa/Dar_es_Salaam, Africa/Djibouti, Africa/Kampala, Africa/Mogadishu, Indian/Antananarivo, Indian/Comoro, and Indian/Mayotte. Changes affecting commentary The commentary is less enthusiastic about Shanks as a source, and is more careful to distinguish UT from UTC. Release 2014i - 2014-10-21 22:04:57 -0700 Changes affecting future timestamps Pacific/Fiji will observe DST from 2014-11-02 02:00 to 2015-01-18 03:00. (Thanks to Ken Rylander for the heads-up.) Guess that future years will use a similar pattern. A new Zone Pacific/Bougainville, for the part of Papua New Guinea that plans to switch from UT +10 to +11 on 2014-12-28 at 02:00. (Thanks to Kiley Walbom for the heads-up.) Changes affecting time zone abbreviations Since Belarus is not changing its clocks even though Moscow is, the time zone abbreviation in Europe/Minsk is changing from FET to its more traditional value MSK on 2014-10-26 at 01:00. (Thanks to Alexander Bokovoy for the heads-up about Belarus.) The new abbreviation IDT stands for the pre-1976 use of UT +08 in Indochina, to distinguish it better from ICT (+07). Changes affecting past timestamps Many timestamps have been corrected for Asia/Ho_Chi_Minh before 1976 (thanks to Trần Ngọc Quân for an indirect pointer to Trần Tiến Bình's authoritative book). Asia/Ho_Chi_Minh has been added to zone1970.tab, to give tzselect users in Vietnam two choices, since north and south Vietnam disagreed after our 1970 cutoff. Asia/Phnom_Penh and Asia/Vientiane have been turned into links, as they differed from existing zones only for older timestamps. As usual, these changes affect pre-1970 timestamps only. Their old contents have been moved to the 'backzone' file. Changes affecting code The time-related library functions now set errno on failure, and some crashes in the new tzalloc-related library functions have been fixed. (Thanks to Christos Zoulas for reporting most of these problems and for suggesting fixes.) If USG_COMPAT is defined and the requested timestamp is standard time, the tz library's localtime and mktime functions now set the extern variable timezone to a value appropriate for that timestamp; and similarly for ALTZONE, daylight saving time, and the altzone variable. This change is a companion to the tzname change in 2014h, and is designed to make timezone and altzone more compatible with tzname. The tz library's functions now set errno to EOVERFLOW if they fail because the result cannot be represented. ctime and ctime_r now return NULL and set errno when a timestamp is out of range, rather than having undefined behavior. Some bugs associated with the new 2014g functions have been fixed. This includes a bug that largely incapacitated the new functions time2posix_z and posix2time_z. (Thanks to Christos Zoulas.) It also includes some uses of uninitialized variables after tzalloc. The new code uses the standard type 'ssize_t', which the Makefile now gives porting advice about. Changes affecting commentary Updated URLs for NRC Canada (thanks to Matt Johnson and Brian Inglis). Release 2014h - 2014-09-25 18:59:03 -0700 Changes affecting past timestamps America/Jamaica's 1974 spring-forward transition was Jan. 6, not Apr. 28. Shanks says Asia/Novokuznetsk switched from LMT (not "NMT") on 1924-05-01, not 1920-01-06. The old entry was based on a misinterpretation of Shanks. Some more zones have been turned into links, when they differed from existing zones only for older timestamps. As usual, these changes affect UT offsets in pre-1970 timestamps only. Their old contents have been moved to the 'backzone' file. The affected zones are: Africa/Blantyre, Africa/Bujumbura, Africa/Gaborone, Africa/Harare, Africa/Kigali, Africa/Lubumbashi, Africa/Lusaka, Africa/Maseru, and Africa/Mbabane. Changes affecting code zdump -V and -v now output gmtoff= values on all platforms, not merely on platforms defining TM_GMTOFF. The tz library's localtime and mktime functions now set tzname to a value appropriate for the requested timestamp, and zdump now uses this on platforms not defining TM_ZONE, fixing a 2014g regression. (Thanks to Tim Parenti for reporting the problem.) The tz library no longer sets tzname if localtime or mktime fails. zdump -c no longer mishandles transitions near year boundaries. (Thanks to Tim Parenti for reporting the problem.) An access to uninitialized data has been fixed. (Thanks to Jörg Richter for reporting the problem.) When THREAD_SAFE is defined, the code ports to the C11 memory model. A memory leak has been fixed if ALL_STATE and THREAD_SAFE are defined and two threads race to initialize data used by gmtime-like functions. (Thanks to Andy Heninger for reporting the problems.) Changes affecting build procedure 'make check' now checks better for properly sorted data. Changes affecting documentation and commentary zdump's gmtoff=N output is now documented, and its isdst=D output is now documented to possibly output D values other than 0 or 1. zdump -c's treatment of years is now documented to use the Gregorian calendar and Universal Time without leap seconds, and its behavior at cutoff boundaries is now documented better. (Thanks to Arthur David Olson and Tim Parenti for reporting the problems.) Programs are now documented to use the proleptic Gregorian calendar. (Thanks to Alan Barrett for the suggestion.) Fractional-second GMT offsets have been documented for civil time in 19th-century Chennai, Jakarta, and New York. Release 2014g - 2014-08-28 12:31:23 -0700 Changes affecting future timestamps Turks & Caicos is switching from US eastern time to UT -04 year-round, modeled as a switch on 2014-11-02 at 02:00. [As noted in 2014j, this switch was later delayed.] Changes affecting past timestamps Time in Russia or the USSR before 1926 or so has been corrected by a few seconds in the following zones: Asia/Irkutsk, Asia/Krasnoyarsk, Asia/Omsk, Asia/Samarkand, Asia/Tbilisi, Asia/Vladivostok, Asia/Yakutsk, Europe/Riga, Europe/Samara. For Asia/Yekaterinburg the correction is a few minutes. (Thanks to Vladimir Karpinsky.) The Portuguese decree of 1911-05-26 took effect on 1912-01-01. This affects 1911 timestamps in Africa/Bissau, Africa/Luanda, Atlantic/Azores, and Atlantic/Madeira. Also, Lisbon's pre-1912 GMT offset was -0:36:45 (rounded from -0:36:44.68), not -0:36:32. (Thanks to Stephen Colebourne for pointing to the decree.) Asia/Dhaka ended DST on 2009-12-31 at 24:00, not 23:59. A new file 'backzone' contains data which may appeal to connoisseurs of old timestamps, although it is out of scope for the tz database, is often poorly sourced, and contains some data that is known to be incorrect. The new file is not recommended for ordinary use and its entries are not installed by default. (Thanks to Lester Caine for the high-quality Jersey, Guernsey, and Isle of Man entries.) Some more zones have been turned into links, when they differed from existing zones only for older timestamps. As usual, these changes affect UT offsets in pre-1970 timestamps only. Their old contents have been moved to the 'backzone' file. The affected zones are: Africa/Bangui, Africa/Brazzaville, Africa/Douala, Africa/Kinshasa, Africa/Libreville, Africa/Luanda, Africa/Malabo, Africa/Niamey, and Africa/Porto-Novo. Changes affecting code Unless NETBSD_INSPIRED is defined to 0, the tz library now supplies functions for creating and using objects that represent timezones. The new functions are tzalloc, tzfree, localtime_rz, mktime_z, and (if STD_INSPIRED is also defined) posix2time_z and time2posix_z. They are intended for performance: for example, localtime_rz (unlike localtime_r) is trivially thread-safe without locking. (Thanks to Christos Zoulas for proposing NetBSD-inspired functions, and to Alan Barrett and Jonathan Lennox for helping to debug the change.) zdump now builds with the tz library unless USE_LTZ is defined to 0, This lets zdump use tz features even if the system library lacks them. To build zdump with the system library, use 'make CFLAGS=-DUSE_LTZ=0 TZDOBJS=zdump.o CHECK_TIME_T_ALTERNATIVES='. zdump now uses localtime_rz if available, as it's significantly faster, and it can help zdump better diagnose invalid timezone names. Define HAVE_LOCALTIME_RZ to 0 to suppress this. HAVE_LOCALTIME_RZ defaults to 1 if NETBSD_INSPIRED && USE_LTZ. When localtime_rz is not available, zdump now uses localtime_r and tzset if available, as this is a bit cleaner and faster than plain localtime. Compile with -DHAVE_LOCALTIME_R=0 and/or -DHAVE_TZSET=0 if your system lacks these two functions. If THREAD_SAFE is defined to 1, the tz library is now thread-safe. Although not needed for tz's own applications, which are single-threaded, this supports POSIX better if the tz library is used in multithreaded apps. Some crashes have been fixed when zdump or the tz library is given invalid or outlandish input. The tz library no longer mishandles leap seconds on platforms with unsigned time_t in timezones that lack ordinary transitions after 1970. The tz code now attempts to infer TM_GMTOFF and TM_ZONE if not already defined, to make it easier to configure on common platforms. Define NO_TM_GMTOFF and NO_TM_ZONE to suppress this. Unless the new macro UNINIT_TRAP is defined to 1, the tz code now assumes that reading uninitialized memory yields garbage values but does not cause other problems such as traps. If TM_GMTOFF is defined and UNINIT_TRAP is 0, mktime is now more likely to guess right for ambiguous timestamps near transitions where tm_isdst does not change. If HAVE_STRFTIME_L is defined to 1, the tz library now defines strftime_l for compatibility with recent versions of POSIX. Only the C locale is supported, though. HAVE_STRFTIME_L defaults to 1 on recent POSIX versions, and to 0 otherwise. tzselect -c now uses a hybrid distance measure that works better in Africa. (Thanks to Alan Barrett for noting the problem.) The C source code now ports to NetBSD when GCC_DEBUG_FLAGS is used, or when time_tz is defined. When HAVE_UTMPX_H is set the 'date' command now builds on systems whose file does not define WTMPX_FILE, and when setting the date it updates the wtmpx file if _PATH_WTMPX is defined. This affects GNU/Linux and similar systems. For easier maintenance later, some C code has been simplified, some lint has been removed, and the code has been tweaked so that plain 'make' is more likely to work. The C type 'bool' is now used for boolean values, instead of 'int'. The long-obsolete LOCALE_HOME code has been removed. The long-obsolete 'gtime' function has been removed. Changes affecting build procedure 'zdump' no longer links in ialloc.o, as it's not needed. 'make check_time_t_alternatives' no longer assumes GNU diff. Changes affecting distribution tarballs The files checktab.awk and zoneinfo2tdf.pl are now distributed in the tzdata tarball instead of the tzcode tarball, since they help maintain the data. The NEWS and Theory files are now also distributed in the tzdata tarball, as they're relevant for data. (Thanks to Alan Barrett for pointing this out.) Also, the leapseconds.awk file is no longer distributed in the tzcode tarball, since it belongs in the tzdata tarball (where 2014f inadvertently also distributed it). Changes affecting documentation and commentary A new file CONTRIBUTING is distributed. (Thanks to Tim Parenti for suggesting a CONTRIBUTING file, and to Tony Finch and Walter Harms for debugging it.) The man pages have been updated to use function prototypes, to document thread-safe variants like localtime_r, and to document the NetBSD-inspired functions tzalloc, tzfree, localtime_rz, and mktime_z. The fields in Link lines have been renamed to be more descriptive and more like the parameters of 'ln'. LINK-FROM has become TARGET, and LINK-TO has become LINK-NAME. tz-link.htm mentions the IETF's tzdist working group; Windows Runtime etc. (thanks to Matt Johnson); and HP-UX's tztab. Some broken URLs have been fixed in the commentary. (Thanks to Lester Caine.) Commentary about Philippines DST has been updated, and commentary on pre-1970 time in India has been added. Release 2014f - 2014-08-05 17:42:36 -0700 Changes affecting future timestamps Russia will subtract an hour from most of its time zones on 2014-10-26 at 02:00 local time. (Thanks to Alexander Krivenyshev.) There are a few exceptions: Magadan Oblast (Asia/Magadan) and Zabaykalsky Krai are subtracting two hours; conversely, Chukotka Autonomous Okrug (Asia/Anadyr), Kamchatka Krai (Asia/Kamchatka), Kemerovo Oblast (Asia/Novokuznetsk), and the Samara Oblast and the Udmurt Republic (Europe/Samara) are not changing their clocks. The changed zones are Europe/Kaliningrad, Europe/Moscow, Europe/Simferopol, Europe/Volgograd, Asia/Yekaterinburg, Asia/Omsk, Asia/Novosibirsk, Asia/Krasnoyarsk, Asia/Irkutsk, Asia/Yakutsk, Asia/Vladivostok, Asia/Khandyga, Asia/Sakhalin, and Asia/Ust-Nera; Asia/Magadan will have two hours subtracted; and Asia/Novokuznetsk's time zone abbreviation is affected, but not its UTC offset. Two zones are added: Asia/Chita (split from Asia/Yakutsk, and also with two hours subtracted) and Asia/Srednekolymsk (split from Asia/Magadan, but with only one hour subtracted). (Thanks to Tim Parenti for much of the above.) Changes affecting time zone abbreviations Australian eastern time zone abbreviations are now AEST/AEDT not EST, and similarly for the other Australian zones. That is, for eastern standard and daylight saving time the abbreviations are AEST and AEDT instead of the former EST for both; similarly, ACST/ACDT, ACWST/ACWDT, and AWST/AWDT are now used instead of the former CST, CWST, and WST. This change does not affect UT offsets, only time zone abbreviations. (Thanks to Rich Tibbett and many others.) Asia/Novokuznetsk shifts from NOVT to KRAT (remaining on UT +07) effective 2014-10-26 at 02:00 local time. The time zone abbreviation for Xinjiang Time (observed in Ürümqi) has been changed from URUT to XJT. (Thanks to Luther Ma.) Prefer MSK/MSD for Moscow time in Russia, even in other cities. Similarly, prefer EET/EEST for eastern European time in Russia. Change time zone abbreviations in (western) Samoa to use "ST" and "DT" suffixes, as this is more likely to match common practice. Prefix "W" to (western) Samoa time when its standard-time offset disagrees with that of American Samoa. America/Metlakatla now uses PST, not MeST, to abbreviate its time zone. Time zone abbreviations have been updated for Japan's two time zones used 1896-1937. JWST now stands for Western Standard Time, and JCST for Central Standard Time (formerly this was CJT). These abbreviations are now used for time in Korea, Taiwan, and Sakhalin while controlled by Japan. Changes affecting past timestamps China's five zones have been simplified to two, since the post-1970 differences in the other three seem to have been imaginary. The zones Asia/Harbin, Asia/Chongqing, and Asia/Kashgar have been removed; backwards-compatibility links still work, albeit with different behaviors for timestamps before May 1980. Asia/Urumqi's 1980 transition to UT +08 has been removed, so that it is now at +06 and not +08. (Thanks to Luther Ma and to Alois Treindl; Treindl sent helpful translations of two papers by Guo Qingsheng.) Some zones have been turned into links, when they differed from existing zones only for older UT offsets where data entries were likely invented. These changes affect UT offsets in pre-1970 timestamps only. This is similar to the change in release 2013e, except this time for western Africa. The affected zones are: Africa/Bamako, Africa/Banjul, Africa/Conakry, Africa/Dakar, Africa/Freetown, Africa/Lome, Africa/Nouakchott, Africa/Ouagadougou, Africa/Sao_Tome, and Atlantic/St_Helena. This also affects the backwards-compatibility link Africa/Timbuktu. (Thanks to Alan Barrett, Stephen Colebourne, Tim Parenti, and David Patte for reporting problems in earlier versions of this change.) Asia/Shanghai's pre-standard-time UT offset has been changed from 8:05:57 to 8:05:43, the location of Xujiahui Observatory. Its transition to standard time has been changed from 1928 to 1901. Asia/Taipei switched to JWST on 1896-01-01, then to JST on 1937-10-01, then to CST on 1945-09-21 at 01:00, and did not observe DST in 1945. In 1946 it observed DST from 05-15 through 09-30; in 1947 from 04-15 through 10-31; and in 1979 from 07-01 through 09-30. (Thanks to Yu-Cheng Chuang.) Asia/Riyadh's transition to standard time is now 1947-03-14, not 1950. Europe/Helsinki's 1942 fall-back transition was 10-04 at 01:00, not 10-03 at 00:00. (Thanks to Konstantin Hyppönen.) Pacific/Pago_Pago has been changed from UT -11:30 to -11 for the period from 1911 to 1950. Pacific/Chatham has been changed to New Zealand standard time plus 45 minutes for the period before 1957, reflecting a 1956 remark in the New Zealand parliament. Europe/Budapest has several pre-1946 corrections: in 1918 the transition out of DST was on 09-16, not 09-29; in 1919 it was on 11-24, not 09-15; in 1945 it was on 11-01, not 11-03; in 1941 the transition to DST was 04-08 not 04-06 at 02:00; and there was no DST in 1920. Africa/Accra is now assumed to have observed DST from 1920 through 1935. Time in Russia before 1927 or so has been corrected by a few seconds in the following zones: Europe/Moscow, Asia/Irkutsk, Asia/Tbilisi, Asia/Tashkent, Asia/Vladivostok, Asia/Yekaterinburg, Europe/Helsinki, and Europe/Riga. Also, Moscow's location has been changed to its Kilometer 0 point. (Thanks to Vladimir Karpinsky for the Moscow changes.) Changes affecting data format A new file 'zone1970.tab' supersedes 'zone.tab' in the installed data. The new file's extended format allows multiple country codes per zone. The older file is still installed but is deprecated; its format is not changing and it will still be distributed for a while, but new applications should use the new file. The new file format simplifies maintenance of obscure locations. To test this, it adds coverage for the Crozet Islands and the Scattered Islands. (Thanks to Tobias Conradi and Antoine Leca.) The file 'iso3166.tab' is planned to switch from ASCII to UTF-8. It is still ASCII now, but commentary about the switch has been added. The new file 'zone1970.tab' already uses UTF-8. Changes affecting code 'localtime', 'mktime', etc. now use much less stack space if ALL_STATE is defined. (Thanks to Elliott Hughes for reporting the problem.) 'zic' no longer mishandles input when ignoring case in locales that are not compatible with English, e.g., unibyte Turkish locales when compiled with HAVE_GETTEXT. Error diagnostics of 'zic' and 'yearistype' have been reworded so that they no longer use ASCII '-' as if it were a dash. 'zic' now rejects output file names that contain '.' or '..' components. (Thanks to Tim Parenti for reporting the problem.) 'zic -v' now warns about output file names that do not follow POSIX rules, or that contain a digit or '.'. (Thanks to Arthur David Olson for starting the ball rolling on this.) Some lint has been removed when using GCC_DEBUG_FLAGS with GCC 4.9.0. Changes affecting build procedure 'zic' no longer links in localtime.o and asctime.o, as they're not needed. (Thanks to John Cochran.) Changes affecting documentation and commentary The 'Theory' file documents legacy names, the longstanding exceptions to the POSIX-inspired file name rules. The 'zic' documentation clarifies the role of time types when interpreting dates. (Thanks to Arthur David Olson.) Documentation and commentary now prefer UTF-8 to US-ASCII, allowing the use of proper accents in foreign words and names. Code and data have not changed because of this. (Thanks to Garrett Wollman, Ian Abbott, and Guy Harris for helping to debug this.) Non-HTML documentation and commentary now use plain-text URLs instead of HTML insertions, and are more consistent about bracketing URLs when they are not already surrounded by white space. (Thanks to suggestions by Steffen Nurpmeso.) There is new commentary about Xujiahui Observatory, the five time-zone project in China from 1918 to 1949, timekeeping in Japanese-occupied Shanghai, and Tibet Time in the 1950s. The sharp-eyed can spot the warlord Jin Shuren in the data. Commentary about the coverage of each Russian zone has been standardized. (Thanks to Tim Parenti.) There is new commentary about contemporary timekeeping in Ethiopia. Obsolete comments about a 2007 proposal for DST in Kuwait has been removed. There is new commentary about time in Poland in 1919. Proper credit has been given to DST inventor George Vernon Hudson. Commentary about time in Metlakatla, AK and Resolute, NU has been improved, with a new source for the former. In zone.tab, Pacific/Easter no longer mentions Salas y Gómez, as it is uninhabited. Commentary about permanent Antarctic bases has been updated. Several typos have been corrected. (Thanks to Tim Parenti for contributing some of these fixes.) tz-link.htm now mentions the JavaScript libraries Moment Timezone, TimezoneJS.Date, Walltime-js, and Timezone. (Thanks to a heads-up from Matt Johnson.) Also, it mentions the Go 'latlong' package. (Thanks to a heads-up from Dirkjan Ochtman.) The files usno1988, usno1989, usno1989a, usno1995, usno1997, and usno1998 have been removed. These obsolescent US Naval Observatory entries were no longer helpful for maintenance. (Thanks to Tim Parenti for the suggestion.) Release 2014e - 2014-06-12 21:53:52 -0700 Changes affecting near-future timestamps Egypt's 2014 Ramadan-based transitions are June 26 and July 31 at 24:00. (Thanks to Imed Chihi.) Guess that from 2015 on Egypt will temporarily switch to standard time at 24:00 the last Thursday before Ramadan, and back to DST at 00:00 the first Friday after Ramadan. Similarly, Morocco's are June 28 at 03:00 and August 2 at 02:00. (Thanks to Milamber Space Network.) Guess that from 2015 on Morocco will temporarily switch to standard time at 03:00 the last Saturday before Ramadan, and back to DST at 02:00 the first Saturday after Ramadan. Changes affecting past timestamps The abbreviation "MSM" (Moscow Midsummer Time) is now used instead of "MSD" for Moscow's double daylight time in summer 1921. Also, a typo "VLASST" has been repaired to be "VLAST" for Vladivostok summer time in 1991. (Thanks to Hank W. for reporting the problems.) Changes affecting commentary tz-link.htm now cites RFC 7265 for jCal, mentions PTP and the draft CalDAV extension, updates URLs for TSP, TZInfo, IATA, and removes stale pointers to World Time Explorer and WORLDTIME. Release 2014d - 2014-05-27 21:34:40 -0700 Changes affecting code zic no longer generates files containing timestamps before the Big Bang. This works around GNOME glib bug 878 (Thanks to Leonardo Chiquitto for reporting the bug, and to Arthur David Olson and James Cloos for suggesting improvements to the fix.) Changes affecting documentation tz-link.htm now mentions GNOME. Release 2014c - 2014-05-13 07:44:13 -0700 Changes affecting near-future timestamps Egypt observes DST starting 2014-05-15 at 24:00. (Thanks to Ahmad El-Dardiry and Gunther Vermier.) Details have not been announced, except that DST will not be observed during Ramadan. Guess that DST will stop during the same Ramadan dates as Morocco, and that Egypt's future spring and fall transitions will be the same as 2010 when it last observed DST, namely April's last Friday at 00:00 to September's last Thursday at 23:00 standard time. Also, guess that Ramadan transitions will be at 00:00 standard time. Changes affecting code zic now generates transitions for minimum time values, eliminating guesswork when handling low-valued timestamps. (Thanks to Arthur David Olson.) Port to Cygwin sans glibc. (Thanks to Arthur David Olson.) Changes affecting commentary and documentation Remove now-confusing comment about Jordan. (Thanks to Oleksii Nochovnyi.) Release 2014b - 2014-03-24 21:28:50 -0700 Changes affecting near-future timestamps Crimea switches to Moscow time on 2014-03-30 at 02:00 local time. (Thanks to Alexander Krivenyshev.) Move its zone.tab entry from UA to RU. New entry for Troll station, Antarctica. (Thanks to Paul-Inge Flakstad and Bengt-Inge Larsson.) This is currently an approximation; a better version will require the zic and localtime fixes mentioned below, and the plan is to wait for a while until at least the zic fixes propagate. Changes affecting code 'zic' and 'localtime' no longer reject locations needing four transitions per year for the foreseeable future. (Thanks to Andrew Main (Zefram).) Also, 'zic' avoids some unlikely failures due to integer overflow. Changes affecting build procedure 'make check' now detects Rule lines defined but never used. The NZAQ rules, an instance of this problem, have been removed. Changes affecting commentary and documentation Fix Tuesday/Thursday typo in description of time in Israel. (Thanks to Bert Katz via Pavel Kharitonov and Mike Frysinger.) Microsoft Windows 8.1 doesn't support tz database names. (Thanks to Donald MacQueen.) Instead, the Microsoft Windows Store app library supports them. Add comments about Johnston Island time in the 1960s. (Thanks to Lyle McElhaney.) Morocco's 2014 DST start will be as predicted. (Thanks to Sebastien Willemijns.) Release 2014a - 2014-03-07 23:30:29 -0800 Changes affecting near-future timestamps Turkey begins DST on 2014-03-31, not 03-30. (Thanks to Faruk Pasin for the heads-up, and to Tim Parenti for simplifying the update.) Changes affecting past timestamps Fiji ended DST on 2014-01-19 at 02:00, not the previously scheduled 03:00. (Thanks to Steffen Thorsen.) Ukraine switched from Moscow to Eastern European time on 1990-07-01 (not 1992-01-01), and observed DST during the entire next winter. (Thanks to Vladimir in Moscow via Alois Treindl.) In 1988 Israel observed DST from 04-10 to 09-04, not 04-09 to 09-03. (Thanks to Avigdor Finkelstein.) Changes affecting code A uninitialized-storage bug in 'localtime' has been fixed. (Thanks to Logan Chien.) Changes affecting the build procedure The settings for 'make check_web' now default to Ubuntu 13.10. Changes affecting commentary and documentation The boundary of the US Pacific time zone is given more accurately. (Thanks to Alan Mintz.) Chile's 2014 DST will be as predicted. (Thanks to José Miguel Garrido.) Paraguay's 2014 DST will be as predicted. (Thanks to Carlos Raúl Perasso.) Better descriptions of countries with same time zone history as Trinidad and Tobago since 1970. (Thanks to Alan Barrett for suggestion.) Several changes affect tz-link.htm, the main web page. Mention Time.is (thanks to Even Scharning) and WX-now (thanks to David Braverman). Mention xCal (Internet RFC 6321) and jCal. Microsoft has some support for tz database names. CLDR data formats include both XML and JSON. Mention Maggiolo's map of solar vs standard time. (Thanks to Arthur David Olson.) Mention TZ4Net. (Thanks to Matt Johnson.) Mention the timezone-olson Haskell package. Mention zeitverschiebung.net. (Thanks to Martin Jäger.) Remove moribund links to daylight-savings-time.info and to Simple Timer + Clocks. Update two links. (Thanks to Oscar van Vlijmen.) Fix some formatting glitches, e.g., remove random newlines from abbr elements' title attributes. Release 2013i - 2013-12-17 07:25:23 -0800 Changes affecting near-future timestamps: Jordan switches back to standard time at 00:00 on December 20, 2013. The 2006-2011 transition schedule is planned to resume in 2014. (Thanks to Steffen Thorsen.) Changes affecting past timestamps: In 2004, Cuba began DST on March 28, not April 4. (Thanks to Steffen Thorsen.) Changes affecting code The compile-time flag NOSOLAR has been removed, as nowadays the benefit of slightly shrinking runtime table size is outweighed by the cost of disallowing potential future updates that exceed old limits. Changes affecting documentation and commentary The files solar87, solar88, and solar89 are no longer distributed. They were a negative experiment - that is, a demonstration that tz data can represent solar time only with some difficulty and error. Their presence in the distribution caused confusion, as Riyadh civil time was generally not solar time in those years. tz-link.htm now mentions Noda Time. (Thanks to Matt Johnson.) Release 2013h - 2013-10-25 15:32:32 -0700 Changes affecting current and future timestamps: Libya has switched its UT offset back to +02 without DST, instead of +01 with DST. (Thanks to Even Scharning.) Western Sahara (Africa/El_Aaiun) uses Morocco's DST rules. (Thanks to Gwillim Law.) Changes affecting future timestamps: Acre and (we guess) western Amazonas will switch from UT -04 to -05 on 2013-11-10. This affects America/Rio_Branco and America/Eirunepe. (Thanks to Steffen Thorsen.) Add entries for DST transitions in Morocco in the year 2038. This avoids some year-2038 glitches introduced in 2013g. (Thanks to Yoshito Umaoka for reporting the problem.) Changes affecting API The 'tzselect' command no longer requires the 'select' command, and should now work with /bin/sh on more platforms. It also works around a bug in BusyBox awk before version 1.21.0. (Thanks to Patrick 'P. J.' McDermott and Alan Barrett.) Changes affecting code Fix localtime overflow bugs with 32-bit unsigned time_t. zdump no longer assumes sscanf returns maximal values on overflow. Changes affecting the build procedure The builder can specify which programs to use, if any, instead of 'ar' and 'ranlib', and libtz.a is now built locally before being installed. (Thanks to Michael Forney.) A dependency typo in the 'zdump' rule has been fixed. (Thanks to Andrew Paprocki.) The Makefile has been simplified by assuming that 'mkdir -p' and 'cp -f' work as specified by POSIX.2-1992 or later; this is portable nowadays. 'make clean' no longer removes 'leapseconds', since it's host-independent and is part of the distribution. The unused makefile macros TZCSRCS, TZDSRCS, DATESRCS have been removed. Changes affecting documentation and commentary tz-link.htm now mentions TC TIMEZONE's draft time zone service protocol (thanks to Mike Douglass) and TimezoneJS.Date (thanks to Jim Fehrle). Update URLs in tz-link page. Add URLs for Microsoft Windows, since 8.1 introduces tz support. Remove URLs for Tru64 and UnixWare (no longer maintained) and for old advisories. SOFA now does C. Release 2013g - 2013-09-30 21:08:26 -0700 Changes affecting current and near-future timestamps Morocco now observes DST from the last Sunday in March to the last Sunday in October, not April to September respectively. (Thanks to Steffen Thorsen.) Changes affecting 'zic' 'zic' now runs on platforms that lack both hard links and symlinks. (Thanks to Theo Veenker for reporting the problem, for MinGW.) Also, fix some bugs on platforms that lack hard links but have symlinks. 'zic -v' again warns that Asia/Tehran has no POSIX environment variable to predict the far future, fixing a bug introduced in 2013e. Changes affecting the build procedure The 'leapseconds' file is again put into the tzdata tarball. Also, 'leapseconds.awk', so tzdata is self-contained. (Thanks to Matt Burgess and Ian Abbott.) The timestamps of these and other dependent files in tarballs are adjusted more consistently. Changes affecting documentation and commentary The README file is now part of the data tarball as well as the code. It now states that files are public domain unless otherwise specified. (Thanks to Andrew Main (Zefram) for asking for clarifications.) Its details about the 1989 release moved to a place of honor near the end of NEWS. Release 2013f - 2013-09-24 23:37:36 -0700 Changes affecting near-future timestamps Tocantins will very likely not observe DST starting this spring. (Thanks to Steffen Thorsen.) Jordan will likely stay at UT +03 indefinitely, and will not fall back this fall. Palestine will fall back at 00:00, not 01:00. (Thanks to Steffen Thorsen.) Changes affecting API The types of the global variables 'timezone' and 'altzone' (if present) have been changed back to 'long'. This is required for 'timezone' by POSIX, and for 'altzone' by common practice, e.g., Solaris 11. These variables were originally 'long' in the tz code, but were mistakenly changed to 'time_t' in 1987; nobody reported the incompatibility until now. The difference matters on x32, where 'long' is 32 bits and 'time_t' is 64. (Thanks to Elliott Hughes.) Changes affecting the build procedure Avoid long strings in leapseconds.awk to work around a mawk bug. (Thanks to Cyril Baurand.) Changes affecting documentation and commentary New file 'NEWS' that contains release notes like this one. Paraguay's law does not specify DST transition time; 00:00 is customary. (Thanks to Waldemar Villamayor-Venialbo.) Minor capitalization fixes. Changes affecting version-control only The experimental GitHub repository now contains annotated and signed tags for recent releases, e.g., '2013e' for Release 2013e. Releases are tagged starting with 2012e; earlier releases were done differently, and tags would either not have a simple name or not exactly match what was released. 'make set-timestamps' is now simpler and a bit more portable. Release 2013e - 2013-09-19 23:50:04 -0700 Changes affecting near-future timestamps This year Fiji will start DST on October 27, not October 20. (Thanks to David Wheeler for the heads-up.) For now, guess that Fiji will continue to spring forward the Sunday before the fourth Monday in October. Changes affecting current and future time zone abbreviations Use WIB/WITA/WIT rather than WIT/CIT/EIT for alphabetic Indonesian time zone abbreviations since 1932. (Thanks to George Ziegler, Priyadi Iman Nurcahyo, Zakaria, Jason Grimes, Martin Pitt, and Benny Lin.) This affects Asia/Dili, Asia/Jakarta, Asia/Jayapura, Asia/Makassar, and Asia/Pontianak. Use ART (UT -03, standard time), rather than WARST (also -03, but daylight saving time) for San Luis, Argentina since 2009. Changes affecting Godthåb timestamps after 2037 if version mismatch Allow POSIX-like TZ strings where the transition time's hour can range from -167 through 167, instead of the POSIX-required 0 through 24. E.g., TZ='FJT-12FJST,M10.3.1/146,M1.3.4/75' for the new Fiji rules. This is a more compact way to represent far-future timestamps for America/Godthab, America/Santiago, Antarctica/Palmer, Asia/Gaza, Asia/Hebron, Asia/Jerusalem, Pacific/Easter, and Pacific/Fiji. Other zones are unaffected by this change. (Derived from a suggestion by Arthur David Olson.) Allow POSIX-like TZ strings where daylight saving time is in effect all year. E.g., TZ='WART4WARST,J1/0,J365/25' for Western Argentina Summer Time all year. This supports a more compact way to represent the 2013d data for America/Argentina/San_Luis. Because of the change for San Luis noted above this change does not affect the current data. (Thanks to Andrew Main (Zefram) for suggestions that improved this change.) Where these two TZ changes take effect, there is a minor extension to the tz file format in that it allows new values for the embedded TZ-format string, and the tz file format version number has therefore been increased from 2 to 3 as a precaution. Version-2-based client code should continue to work as before for all timestamps before 2038. Existing version-2-based client code (tzcode, GNU/Linux, Solaris) has been tested on version-3-format files, and typically works in practice even for timestamps after 2037; the only known exception is America/Godthab. Changes affecting timestamps before 1970 Pacific/Johnston is now a link to Pacific/Honolulu. This corrects some errors before 1947. Some zones have been turned into links, when they differ from existing zones only in older data entries that were likely invented or that differ only in LMT or transitions from LMT. These changes affect only timestamps before 1943. The affected zones are: Africa/Juba, America/Anguilla, America/Aruba, America/Dominica, America/Grenada, America/Guadeloupe, America/Marigot, America/Montserrat, America/St_Barthelemy, America/St_Kitts, America/St_Lucia, America/St_Thomas, America/St_Vincent, America/Tortola, and Europe/Vaduz. (Thanks to Alois Treindl for confirming that the old Europe/Vaduz zone was wrong and the new link is better for WWII-era times.) Change Kingston Mean Time from -5:07:12 to -5:07:11. This affects America/Cayman, America/Jamaica and America/Grand_Turk timestamps from 1890 to 1912. Change the UT offset of Bern Mean Time from 0:29:44 to 0:29:46. This affects Europe/Zurich timestamps from 1853 to 1894. (Thanks to Alois Treindl.) Change the date of the circa-1850 Zurich transition from 1849-09-12 to 1853-07-16, overriding Shanks with data from Messerli about postal and telegraph time in Switzerland. Changes affecting time zone abbreviations before 1970 For Asia/Jakarta, use BMT (not JMT) for mean time from 1923 to 1932, as Jakarta was called Batavia back then. Changes affecting API The 'zic' command now outputs a dummy transition when far-future data can't be summarized using a TZ string, and uses a 402-year window rather than a 400-year window. For the current data, this affects only the Asia/Tehran file. It does not affect any of the timestamps that this file represents, so zdump outputs the same information as before. (Thanks to Andrew Main (Zefram).) The 'date' command has a new '-r' option, which lets you specify the integer time to display, a la FreeBSD. The 'tzselect' command has two new options '-c' and '-n', which lets you select a zone based on latitude and longitude. The 'zic' command's '-v' option now warns about constructs that require the new version-3 binary file format. (Thanks to Arthur David Olson for the suggestion.) Support for floating-point time_t has been removed. It was always dicey, and POSIX no longer requires it. (Thanks to Eric Blake for suggesting to the POSIX committee to remove it, and thanks to Alan Barrett, Clive D.W. Feather, Andy Heninger, Arthur David Olson, and Alois Treindl, for reporting bugs and elucidating some of the corners of the old floating-point implementation.) The signatures of 'offtime', 'timeoff', and 'gtime' have been changed back to the old practice of using 'long' to represent UT offsets. This had been inadvertently and mistakenly changed to 'int_fast32_t'. (Thanks to Christos Zoulas.) The code avoids undefined behavior on integer overflow in some more places, including gmtime, localtime, mktime and zdump. Changes affecting the zdump utility zdump now outputs "UT" when referring to Universal Time, not "UTC". "UTC" does not make sense for timestamps that predate the introduction of UTC, whereas "UT", a more generic term, does. (Thanks to Steve Allen for clarifying UT vs UTC.) Data changes affecting behavior of tzselect and similar programs Country code BQ is now called the more common name "Caribbean Netherlands" rather than the more official "Bonaire, St Eustatius & Saba". Remove from zone.tab the names America/Montreal, America/Shiprock, and Antarctica/South_Pole, as they are equivalent to existing same-country-code zones for post-1970 timestamps. The data entries for these names are unchanged, so the names continue to work as before. Changes affecting code internals zic -c now runs way faster on 64-bit hosts when given large numbers. zic now uses vfprintf to avoid allocating and freeing some memory. tzselect now computes the list of continents from the data, rather than have it hard-coded. Minor changes pacify GCC 4.7.3 and GCC 4.8.1. Changes affecting the build procedure The 'leapseconds' file is now generated automatically from a new file 'leap-seconds.list', which is a copy of A new source file 'leapseconds.awk' implements this. The goal is simplification of the future maintenance of 'leapseconds'. When building the 'posix' or 'right' subdirectories, if the subdirectory would be a copy of the default subdirectory, it is now made a symbolic link if that is supported. This saves about 2 MB of file system space. The links America/Shiprock and Antarctica/South_Pole have been moved to the 'backward' file. This affects only nondefault builds that omit 'backward'. Changes affecting version-control only .gitignore now ignores 'date'. Changes affecting documentation and commentary Changes to the 'tzfile' man page It now mentions that the binary file format may be extended in future versions by appending data. It now refers to the 'zdump' and 'zic' man pages. Changes to the 'zic' man page It lists conditions that elicit a warning with '-v'. It says that the behavior is unspecified when duplicate names are given, or if the source of one link is the target of another. Its examples are updated to match the latest data. The definition of white space has been clarified slightly. (Thanks to Michael Deckers.) Changes to the 'Theory' file There is a new section about the accuracy of the tz database, describing the many ways that errors can creep in, and explaining why so many of the pre-1970 timestamps are wrong or misleading (thanks to Steve Allen, Lester Caine, and Garrett Wollman for discussions that contributed to this). The 'Theory' file describes LMT better (this follows a suggestion by Guy Harris). It refers to the 2013 edition of POSIX rather than the 2004 edition. It's mentioned that excluding 'backward' should not affect the other data, and it suggests at least one zone.tab name per inhabited country (thanks to Stephen Colebourne). Some longstanding restrictions on names are documented, e.g., 'America/New_York' precludes 'America/New_York/Bronx'. It gives more reasons for the 1970 cutoff. It now mentions which time_t variants are supported, such as signed integer time_t. (Thanks to Paul Goyette for reporting typos in an experimental version of this change.) (Thanks to Philip Newton for correcting typos in these changes.) Documentation and commentary is more careful to distinguish UT in general from UTC in particular. (Thanks to Steve Allen.) Add a better source for the Zurich 1894 transition. (Thanks to Pierre-Yves Berger.) Update shapefile citations in tz-link.htm. (Thanks to Guy Harris.) Release 2013d - 2013-07-05 07:38:01 -0700 Changes affecting future timestamps: Morocco's midsummer transitions this year are July 7 and August 10, not July 9 and August 8. (Thanks to Andrew Paprocki.) Israel now falls back on the last Sunday of October. (Thanks to Ephraim Silverberg.) Changes affecting past timestamps: Specify Jerusalem's location more precisely; this changes the pre-1880 times by 2 s. Changing affecting metadata only: Fix typos in the entries for country codes BQ and SX. Changes affecting code: Rework the code to fix a bug with handling Australia/Macquarie on 32-bit hosts (thanks to Arthur David Olson). Port to platforms like NetBSD, where time_t can be wider than long. Add support for testing time_t types other than the system's. Run 'make check_time_t_alternatives' to try this out. Currently, the tests fail for unsigned time_t; this should get fixed at some point. Changes affecting documentation and commentary: Deemphasize the significance of national borders. Update the zdump man page. Remove obsolete NOID comment (thanks to Denis Excoffier). Update several URLs and comments in the web pages. Spelling fixes (thanks to Kevin Lyda and Jonathan Leffler). Update URL for CLDR Zone->Tzid table (thanks to Yoshito Umaoka). Release 2013c - 2013-04-19 16:17:40 -0700 Changes affecting current and future timestamps: Palestine observed DST starting March 29, 2013. (Thanks to Steffen Thorsen.) From 2013 on, Gaza and Hebron both observe DST, with the predicted rules being the last Thursday in March at 24:00 to the first Friday on or after September 21 at 01:00. Assume that the recent change to Paraguay's DST rules is permanent, by moving the end of DST to the 4th Sunday in March every year. (Thanks to Carlos Raúl Perasso.) Changes affecting past timestamps: Fix some historical data for Palestine to agree with that of timeanddate.com, as follows: The spring 2008 change in Gaza and Hebron was on 00:00 Mar 28, not 00:00 Apr 1. The fall 2009 change in Gaza and Hebron on Sep 4 was at 01:00, not 02:00. The spring 2010 change in Hebron was 00:00 Mar 26, not 00:01 Mar 27. The spring 2011 change in Gaza was 00:01 Apr 1, not 12:01 Apr 2. The spring 2011 change in Hebron on Apr 1 was at 00:01, not 12:01. The fall 2011 change in Hebron on Sep 30 was at 00:00, not 03:00. Fix times of habitation for Macquarie to agree with the Tasmania Parks & Wildlife Service history, which indicates that permanent habitation was 1899-1919 and 1948 on. Changing affecting metadata only: Macquarie Island is politically part of Australia, not Antarctica. (Thanks to Tobias Conradi.) Sort Macquarie more consistently with other parts of Australia. (Thanks to Tim Parenti.) Release 2013b - 2013-03-10 22:33:40 -0700 Changes affecting current and future timestamps: Haiti uses US daylight-saving rules this year, and presumably future years. This changes timestamps starting today. (Thanks to Steffen Thorsen.) Paraguay will end DST on March 24 this year. (Thanks to Steffen Thorsen.) For now, assume it's just this year. Morocco does not observe DST during Ramadan; try to predict Ramadan in Morocco as best we can. (Thanks to Erik Homoet for the heads-up.) Changes affecting commentary: Update URLs in tz-link page. Add URLs for webOS, BB10, iOS. Update URL for Solaris. Mention Internet RFC 6557. Update Internet RFCs 2445->5545, 2822->5322. Switch from FTP to HTTP for Internet RFCs. Release 2013a - 2013-02-27 09:20:35 -0800 Change affecting binary data format: The zone offset at the end of version-2-format zone files is now allowed to be 24:00, as per POSIX.1-2008. (Thanks to Arthur David Olson.) Changes affecting current and future timestamps: Chile's 2013 rules, and we guess rules for 2014 and later, will be the same as 2012, namely Apr Sun>=23 03:00 UTC to Sep Sun>=2 04:00 UTC. (Thanks to Steffen Thorsen and Robert Elz.) New Zones Asia/Khandyga, Asia/Ust-Nera, Europe/Busingen. (Thanks to Tobias Conradi and Arthur David Olson.) Many changes affect historical timestamps before 1940. These were deduced from: Milne J. Civil time. Geogr J. 1899 Feb;13(2):173-94 . Changes affecting the code: Fix zic bug that mishandled Egypt's 2010 changes (this also affected the data). (Thanks to Arthur David Olson.) Fix localtime bug when time_t is unsigned and data files were generated by a signed time_t system. (Thanks to Doug Bailey for reporting and to Arthur David Olson for fixing.) Allow the email address for bug reports to be set by the packager. The default is tz@iana.org, as before. (Thanks to Joseph S. Myers.) Update HTML checking to be compatible with Ubuntu 12.10. Check that files are a safe subset of ASCII. At some point we may relax this requirement to a safe subset of UTF-8. Without the check, some non-UTF-8 encodings were leaking into the distribution. Commentary changes: Restore a comment about copyright notices that was inadvertently deleted. (Thanks to Arthur David Olson.) Improve the commentary about which districts observe what times in Russia. (Thanks to Oscar van Vlijmen and Arthur David Olson.) Add web page links to tz.js. Add "Run by the Monkeys" to tz-art. (Thanks to Arthur David Olson.) Release 2012j - 2012-11-12 18:34:49 -0800 Libya moved to CET this weekend, but with DST planned next year. (Thanks to Even Scharning, Steffen Thorsen, and Tim Parenti.) Signatures now have the extension .asc, not .sign, as that's more standard. (Thanks to Phil Pennock.) The output of 'zdump --version', and of 'zic --version', now uses a format that is more typical for --version. (Thanks to Joseph S. Myers.) The output of 'tzselect --help', 'zdump --help', and 'zic --help' now uses tz@iana.org rather than the old elsie address. zic -v now complains about abbreviations that are less than 3 or more than 6 characters, as per POSIX. Formerly, it checked for abbreviations that were more than 3. 'make public' no longer puts its temporary directory under /tmp, and uses the just-built zic rather than the system zic. Various fixes to documentation and commentary. Release 2012i - 2012-11-03 12:57:09 -0700 Cuba switches from DST tomorrow at 01:00. (Thanks to Steffen Thorsen.) Linker flags can now be specified via LDFLAGS. AWK now defaults to 'awk', not 'nawk'. The shell in tzselect now defaults to /bin/bash, but this can be overridden by specifying KSHELL. The main web page now mentions the unofficial GitHub repository. (Thanks to Mike Frysinger.) Tarball signatures can now be built by running 'make signatures'. There are also new makefile rules 'tarballs', 'check_public', and separate makefile rules for each tarball and signature file. A few makefile rules are now more portable to strict POSIX. The main web page now lists the canonical IANA URL. Release 2012h - 2012-10-26 22:49:10 -0700 Bahia no longer has DST. (Thanks to Kelley Cook.) Tocantins has DST. (Thanks to Rodrigo Severo.) Israel has new DST rules next year. (Thanks to Ephraim Silverberg.) Jordan stays on DST this winter. (Thanks to Steffen Thorsen.) Web page updates. More C modernization, except that at Arthur David Olson's suggestion the instances of 'register' were kept. Release 2012g - 2012-10-17 20:59:45 -0700 Samoa fall 2012 and later. (Thanks to Nicholas Pereira and Robert Elz.) Palestine fall 2012. (Thanks to Steffen Thorsen.) Assume C89. To attack the version-number problem, this release ships the file 'Makefile' (which contains the release number) in both the tzcode and the tzdata tarballs. The two Makefiles are identical, and should be identical in any matching pair of tarballs, so it shouldn't matter which order you extract the tarballs. Perhaps we can come up with a better version-number scheme at some point; this scheme does have the virtue of not adding more files. Release 2012f - 2012-09-12 23:17:03 -0700 * australasia (Pacific/Fiji): Fiji DST is October 21 through January 20 this year. (Thanks to Steffen Thorsen.) Release 2012e - 2012-08-02 20:44:55 -0700 * australasia (Pacific/Fakaofo): Tokelau is UT +13, not +14. (Thanks to Steffen Thorsen.) * Use a single version number for both code and data. * .gitignore: New file. * Remove trailing white space. Release code2012c-data2012d - 2012-07-19 16:35:33 -0700 Changes for Morocco's timestamps, which take effect in a couple of hours, along with infrastructure changes to accommodate how the tz code and data are released on IANA. Release data2012c - 2012-03-27 12:17:25 -0400 africa Summer time changes for Morocco (to start late April 2012) asia Changes for 2012 for Gaza & the West Bank (Hebron) and Syria northamerica Haiti following US/Canada rules for 2012 (and we're assuming, for now anyway, for the future). Release 2012b - 2012-03-02 12:29:15 +0700 There is just one change to tzcode2012b (compared with 2012a): the Makefile that was accidentally included with 2012a has been replaced with the version that should have been there, which is identical with the previous version (from tzcode2011i). There are just two changes in tzdata2012b compared with 2012a. Most significantly, summer time in Cuba has been delayed 3 weeks (now starts April 1 rather than March 11). Since Mar 11 (the old start date, as listed in 2012a) is just a little over a week away, this change is urgent. Less importantly, an excess tab in one of the changes in zone.tab in 2012a has been removed. Release 2012a - 2012-03-01 18:28:10 +0700 The changes in tzcode2012a (compared to the previous version, 2011i) are entirely to the README and tz-art.htm and tz-link.htm files, if none of those concern you, you can ignore the code update. The changes reflect the changed addresses for the mailing list and the code and data distribution points & methods (and a link to DateTime::TimeZone::Tzfile has been added to tz-link.htm). In tzdata2012a (compared to the previous release, which was 2011n) the major changes are: Chile 2011/2012 and 2012/2013 summer time date adjustments. Falkland Islands onto permanent summer time (we're assuming for the foreseeable future, though 2012 is all we're fairly certain of.) Armenia has abolished Summer Time. Tokelau jumped the International Date Line back last December (just the same as their near neighbour, Samoa). America/Creston is a new zone for a small area of British Columbia There will be a leapsecond 2012-06-30 23:59:60 UTC. Other minor changes are: Corrections to 1918 Canadian summer time end dates. Updated URL for UK time zone history (in comments) A few typos in Le Corre's list of free French place names (comments) Release data2011n - 2011-10-30 14:57:54 +0700 There are three changes of note - most urgently, Cuba (America/Havana) has extended summer time by two weeks, now to end on Nov 13, rather than the (already past) Oct 30. Second, the Pridnestrovian Moldavian Republic (Europe/Tiraspol) decided not to split from the rest of Moldova after all, and consequently that zone has been removed (again) and reinstated in the "backward" file as a link to Europe/Chisinau. And third, the end date for Fiji's summer time this summer was moved forward from the earlier planned Feb 26, to Jan 22. Apart from that, Moldova (MD) returns to a single entry in zone.tab (and the incorrect syntax that was in the 2011m version of that file is so fixed - it would have been fixed in a different way had this change not happened - that's the "missing" sccs version id). Release data2011m - 2011-10-24 21:42:16 +0700 In particular, the typos in comments in the data (2011-11-17 should have been 2011-10-17 as Alan Barrett noted, and spelling of Tiraspol that Tim Parenti noted) have been fixed, and the change for Ukraine has been made in all 4 Ukrainian zones, rather than just Europe/Kiev (again, thanks to Tim Parenti, and also Denys Gavrysh). In addition, I added Europe/Tiraspol to zone.tab. This time, all the files have new version numbers... (including the files otherwise unchanged in 2011m that were changed in 2011l but didn't get new version numbers there...) Release data2011l - 2011-10-10 11:15:43 +0700 There are just 2 changes that cause different generated tzdata files from zic, to Asia/Hebron and Pacific/Fiji - the possible change for Bahia, Brazil is included, but commented out. Compared with the diff I sent out last week, this version also includes attributions for the sources for the changes (in much the same format as ado used, but the html tags have not been checked, verified, or used in any way at all, so if there are errors there, please let me know.) Release data2011k - 2011-09-20 17:54:03 -0400 [not summarized] Release data2011j - 2011-09-12 09:22:49 -0400 (contemporary changes for Samoa; past changes for Kenya, Uganda, and Tanzania); there are also two spelling corrections to comments in the australasia file (with thanks to Christos Zoulas). Release 2011i - 2011-08-29 05:56:32 -0400 [not summarized] Release data2011h - 2011-06-15 18:41:48 -0400 Russia and Curaçao changes Release 2011g - 2011-04-25 09:07:22 -0400 update the rules for Egypt to reflect its abandonment of DST this year Release 2011f - 2011-04-06 17:14:53 -0400 [not summarized] Release 2011e - 2011-03-31 16:04:38 -0400 Morocco, Chile, and tz-link changes Release 2011d - 2011-03-14 09:18:01 -0400 changes that impact present-day timestamps in Cuba, Samoa, and Turkey Release 2011c - 2011-03-07 09:30:09 -0500 These do affect current timestamps in Chile and Annette Island, Canada. Release 2011b - 2011-02-07 08:44:50 -0500 [not summarized] Release 2011a - 2011-01-24 10:30:16 -0500 [not summarized] Release data2010o - 2010-11-01 09:18:23 -0400 change to the end of DST in Fiji in 2011 Release 2010n - 2010-10-25 08:19:17 -0400 [not summarized] Release 2010m - 2010-09-27 09:24:48 -0400 Hong Kong, Vostok, and zic.c changes Release 2010l - 2010-08-16 06:57:25 -0400 [not summarized] Release 2010k - 2010-07-26 10:42:27 -0400 [not summarized] Release 2010j - 2010-05-10 09:07:48 -0400 changes for Bahía de Banderas and for version naming Release data2010i - 2010-04-16 18:50:45 -0400 the end of DST in Morocco on 2010-08-08 Release data2010h - 2010-04-05 09:58:56 -0400 [not summarized] Release data2010g - 2010-03-24 11:14:53 -0400 [not summarized] Release 2010f - 2010-03-22 09:45:46 -0400 [not summarized] Release data2010e - 2010-03-08 14:24:27 -0500 corrects the Dhaka bug found by Danvin Ruangchan Release data2010d - 2010-03-06 07:26:01 -0500 [not summarized] Release 2010c - 2010-03-01 09:20:58 -0500 changes including KRE's suggestion for earlier initialization of "goahead" and "goback" structure elements Release code2010a - 2010-02-16 10:40:04 -0500 [not summarized] Release data2010b - 2010-01-20 12:37:01 -0500 Mexico changes Release data2010a - 2010-01-18 08:30:04 -0500 changes to Dhaka Release data2009u - 2009-12-26 08:32:28 -0500 changes to DST in Bangladesh Release 2009t - 2009-12-21 13:24:27 -0500 [not summarized] Release data2009s - 2009-11-14 10:26:32 -0500 (cosmetic) Antarctica change and the DST-in-Fiji-in-2009-and-2010 change Release 2009r - 2009-11-09 10:10:31 -0500 "antarctica" and "tz-link.htm" changes Release 2009q - 2009-11-02 09:12:40 -0500 with two corrections as reported by Eric Muller and Philip Newton Release data2009p - 2009-10-23 15:05:27 -0400 Argentina (including San Luis) changes (with the correction from Mariano Absatz) Release data2009o - 2009-10-14 16:49:38 -0400 Samoa (commentary only), Pakistan, and Bangladesh changes Release data2009n - 2009-09-22 15:13:38 -0400 added commentary for Argentina and a change to the end of DST in 2009 in Pakistan Release data2009m - 2009-09-03 10:23:43 -0400 Samoa and Palestine changes Release data2009l - 2009-08-14 09:13:07 -0400 Samoa (comments only) and Egypt Release 2009k - 2009-07-20 09:46:08 -0400 [not summarized] Release data2009j - 2009-06-15 06:43:59 -0400 Bangladesh change (with a short turnaround since the DST change is impending) Release 2009i - 2009-06-08 09:21:22 -0400 updating for DST in Bangladesh this year Release 2009h - 2009-05-26 09:19:14 -0400 [not summarized] Release data2009g - 2009-04-20 16:34:07 -0400 Cairo Release data2009f - 2009-04-10 11:00:52 -0400 correct DST in Pakistan Release 2009e - 2009-04-06 09:08:11 -0400 [not summarized] Release 2009d - 2009-03-23 09:38:12 -0400 Morocco, Tunisia, Argentina, and American Astronomical Society changes Release data2009c - 2009-03-16 09:47:51 -0400 change to the start of Cuban DST Release 2009b - 2009-02-09 11:15:22 -0500 [not summarized] Release 2009a - 2009-01-21 10:09:39 -0500 [not summarized] Release data2008i - 2008-10-21 12:10:25 -0400 southamerica and zone.tab files, with Argentina DST rule changes and United States zone reordering and recommenting Release 2008h - 2008-10-13 07:33:56 -0400 [not summarized] Release 2008g - 2008-10-06 09:03:18 -0400 Fix a broken HTML anchor and update Brazil's DST transitions; there's also a slight reordering of information in tz-art.htm. Release data2008f - 2008-09-09 22:33:26 -0400 [not summarized] Release 2008e - 2008-07-28 14:11:17 -0400 changes by Arthur David Olson and Jesper Nørgaard Welen Release data2008d - 2008-07-07 09:51:38 -0400 changes by Arthur David Olson, Paul Eggert, and Rodrigo Severo Release data2008c - 2008-05-19 17:48:03 -0400 Pakistan, Morocco, and Mongolia Release data2008b - 2008-03-24 08:30:59 -0400 including renaming Asia/Calcutta to Asia/Kolkata, with a backward link provided Release 2008a - 2008-03-08 05:42:16 -0500 [not summarized] Release 2007k - 2007-12-31 10:25:22 -0500 most importantly, changes to the "southamerica" file based on Argentina's readoption of daylight saving time Release 2007j - 2007-12-03 09:51:01 -0500 1. eliminate the "P" (parameter) macro; 2. the "noncontroversial" changes circulated on the time zone mailing list (less the changes to "logwtmp.c"); 3. eliminate "too many transition" errors when "min" is used in time zone rules; 4. changes by Paul Eggert (including updated information for Venezuela). Release data2007i - 2007-10-30 10:28:11 -0400 changes for Cuba and Syria Release 2007h - 2007-10-01 10:05:51 -0400 changes by Paul Eggert, as well as an updated link to the ICU project in tz-link.htm Release 2007g - 2007-08-20 10:47:59 -0400 changes by Paul Eggert The "leapseconds" file has been updated to incorporate the most recent International Earth Rotation and Reference Systems Service (IERS) bulletin. There's an addition to tz-art.htm regarding the television show "Medium". Release 2007f - 2007-05-07 10:46:46 -0400 changes by Paul Eggert (including Haiti, Turks and Caicos, and New Zealand) changes to zic.c to allow hour values greater than 24 (along with Paul's improved time value overflow checking) Release 2007e - 2007-04-02 10:11:52 -0400 Syria and Honduras changes by Paul Eggert zic.c variable renaming changes by Arthur David Olson Release 2007d - 2007-03-20 08:48:30 -0400 changes by Paul Eggert the elimination of white space at the ends of lines Release 2007c - 2007-02-26 09:09:37 -0500 changes by Paul Eggert Release 2007b - 2007-02-12 09:34:20 -0500 Paul Eggert's proposed change to the quotation handling logic in zic.c. changes to the commentary in "leapseconds" reflecting the IERS announcement that there is to be no positive leap second at the end of June 2007. Release 2007a - 2007-01-08 12:28:29 -0500 changes by Paul Eggert Derick Rethans's Asmara change Oscar van Vlijmen's Easter Island local mean time change symbolic link changes Release 2006p - 2006-11-27 08:54:27 -0500 changes by Paul Eggert Release 2006o - 2006-11-06 09:18:07 -0500 changes by Paul Eggert Release 2006n - 2006-10-10 11:32:06 -0400 changes by Paul Eggert Release 2006m - 2006-10-02 15:32:35 -0400 changes for Uruguay, Palestine, and Egypt by Paul Eggert (minimalist) changes to zic.8 to clarify "until" information Release data2006l - 2006-09-18 12:58:11 -0400 Paul's best-effort work on this coming weekend's Egypt time change Release 2006k - 2006-08-28 12:19:09 -0400 changes by Paul Eggert Release 2006j - 2006-08-21 09:56:32 -0400 changes by Paul Eggert Release code2006i - 2006-08-07 12:30:55 -0400 localtime.c fixes Ken Pizzini's conversion script Release code2006h - 2006-07-24 09:19:37 -0400 adds public domain notices to four files includes a fix for transition times being off by a second adds a new recording to the "arts" file (information courtesy Colin Bowern) Release 2006g - 2006-05-08 17:18:09 -0400 northamerica changes by Paul Eggert Release 2006f - 2006-05-01 11:46:00 -0400 a missing version number problem is fixed (with thanks to Bradley White for catching the problem) Release 2006d - 2006-04-17 14:33:43 -0400 changes by Paul Eggert added new items to tz-arts.htm that were found by Paul Release 2006c - 2006-04-03 10:09:32 -0400 two sets of data changes by Paul Eggert a fencepost error fix in zic.c changes to zic.c and the "europe" file to minimize differences between output produced by the old 32-bit zic and the new 64-bit version Release 2006b - 2006-02-20 10:08:18 -0500 [tz32code2006b + tz64code2006b + tzdata2006b] 64-bit code All SCCS IDs were bumped to "8.1" for this release. Release 2006a - 2006-01-30 08:59:31 -0500 changes by Paul Eggert (in particular, Indiana time zone moves) an addition to the zic manual page to describe how special-case transitions are handled Release 2005r - 2005-12-27 09:27:13 -0500 Canadian changes by Paul Eggert They also add "
" directives to time zone data files and reflect
  changes to warning message logic in "zdump.c" (but with calls to
  "gettext" kept unbundled at the suggestion of Ken Pizzini).


Release 2005q - 2005-12-13 09:17:09 -0500

  Nothing earth-shaking here:
	1.  Electronic mail addresses have been removed.
	2.  Casts of the return value of exit have been removed.
	3.  Casts of the argument of is.* macros have been added.
	4.  Indentation in one section of zic.c has been fixed.
	5.  References to dead URLs in the data files have been dealt with.


Release 2005p - 2005-12-05 10:30:53 -0500

  "systemv", "tz-link.htm", and "zdump.c" changes
  (less the casts of arguments to the is* macros)


Release 2005o - 2005-11-28 10:55:26 -0500

  Georgia, Cuba, Nicaragua, and Jordan changes by Paul Eggert

  zdump.c lint fixes by Arthur David Olson


Release 2005n - 2005-10-03 09:44:09 -0400

  changes by Paul Eggert (both the Uruguay changes and the Kyrgyzstan
  et al. changes)


Release 2005m - 2005-08-29 12:15:40 -0400

  changes by Paul Eggert (with a small tweak to the tz-art change)

  a declaration of an unused variable has been removed from zdump.c


Release 2005l - 2005-08-22 12:06:39 -0400

  changes by Paul Eggert

  overflow/underflow checks by Arthur David Olson, minus changes to
  the "Theory" file about the pending addition of 64-bit data (I grow
  less confident of the changes being accepted with each passing day,
  and the changes no longer increase the data files nine-fold--there's
  less than a doubling in size by my local Sun's reckoning)


Release 2005k - 2005-07-14 14:14:24 -0400

  The "leapseconds" file has been edited to reflect the recently
  announced leap second at the end of 2005.

  I've also deleted electronic mail addresses from the files as an
  anti-spam measure.


Release 2005j - 2005-06-13 14:34:13 -0400

  These reflect changes to limit the length of time zone abbreviations
  and the characters used in those abbreviations.

  There are also changes to handle POSIX-style "quoted" timezone
  environment variables.

  The changes were circulated on the time zone mailing list; the only
  change since then was the removal of a couple of minimum-length of
  abbreviation checks.


Release data2005i - 2005-04-21 15:04:16 -0400

  changes (most importantly to Nicaragua and Haiti) by Paul Eggert


Release 2005h - 2005-04-04 11:24:47 -0400

  changes by Paul Eggert

  minor changes to Makefile and zdump.c to produce more useful output
  when doing a "make typecheck"


Release 2005g - 2005-03-14 10:11:21 -0500

  changes by Paul Eggert (a change to current DST rules in Uruguay and
  an update to a link to time zone software)


Release 2005f - 2005-03-01 08:45:32 -0500

  data and documentation changes by Paul Eggert


Release 2005e - 2005-02-10 15:59:44 -0500

  [not summarized]


Release code2005d - 2005-01-31 09:21:47 -0500

  make zic complain about links to links if the -v flag is used

  have "make public" do more code checking

  add an include to "localtime.c" for the benefit of gcc systems


Release 2005c - 2005-01-17 18:36:29 -0500

  get better results when mktime runs on a system where time_t is double

  changes to the data files (most importantly to Paraguay)


Release 2005b - 2005-01-10 09:19:54 -0500

  Get localtime and gmtime working on systems with exotic time_t types.

  Update the leap second commentary in the "leapseconds" file.


Release 2005a - 2005-01-01 13:13:44 -0500

  [not summarized]


Release code2004i - 2004-12-14 13:42:58 -0500

  Deal with systems where time_t is unsigned.


Release code2004h - 2004-12-07 11:40:18 -0500

  64-bit-time_t changes


Release 2004g - 2004-11-02 09:06:01 -0500

  update to Cuba (taking effect this weekend)

  other changes by Paul Eggert

  correction of the spelling of Oslo

  changed versions of difftime.c and private.h


Release code2004f - 2004-10-21 10:25:22 -0400

  Cope with wide-ranging tm_year values.


Release 2004e - 2004-10-11 14:47:21 -0400

  Brazil/Argentina/Israel changes by Paul Eggert

  changes to tz-link.htm by Paul

  one small fix to Makefile


Release 2004d - 2004-09-22 08:27:29 -0400

  Avoid overflow problems when TM_YEAR_BASE is added to an integer.


Release 2004c - 2004-08-11 12:06:26 -0400

  asctime-related changes

  (variants of) some of the documentation changes suggested by Paul Eggert


Release 2004b - 2004-07-19 14:33:35 -0400

  data changes by Paul Eggert - most importantly, updates for Argentina


Release 2004a - 2004-05-27 12:00:47 -0400

  changes by Paul Eggert

  Handle DST transitions that occur at the end of a month in some
  years but at the start of the following month in other years.

  Add a copy of the correspondence that's the basis for claims about
  DST in the Navajo Nation.


Release 2003e - 2003-12-15 09:36:47 -0500

  changes by Arthur David Olson (primarily code changes)

  changes by Paul Eggert (primarily data changes)

  minor changes to "Makefile" and "northamerica" (in the latter case,
  optimization of the "Toronto" rules)


Release 2003d - 2003-10-06 09:34:44 -0400

  changes by Paul Eggert


Release 2003c - 2003-09-16 10:47:05 -0400

  Fix bad returns in zic.c's inleap function.
  Thanks to Bradley White for catching the problem!


Release 2003b - 2003-09-16 07:13:44 -0400

  Add a "--version" option (and documentation) to the zic and zdump commands.

  changes to overflow/underflow checking in zic

  a localtime typo fix.

  Update the leapseconds and tz-art.htm files.


Release 2003a - 2003-03-24 09:30:54 -0500

  changes by Paul Eggert

  a few additions and modifications to the tz-art.htm file


Release 2002d - 2002-10-15 13:12:42 -0400

  changes by Paul Eggert, less the "Britain (UK)" change in iso3166.tab

  There's also a new time zone quote in "tz-art.htm".


Release 2002c - 2002-04-04 11:55:20 -0500

  changes by Paul Eggert

  Change zic.c to avoid creating symlinks to files that don't exist.


Release 2002b - 2002-01-28 12:56:03 -0500

  [These change notes are for Release 2002a, which was corrupted.
  2002b was a corrected version of 2002a.]

  changes by Paul Eggert

  Update the "leapseconds" file to note that there'll be no leap
  second at the end of June, 2002.

  Change "zic.c" to deal with a problem in handling the "Asia/Bishkek" zone.

  Change to "difftime.c" to avoid sizeof problems.


Release 2001d - 2001-10-09 13:31:32 -0400

  changes by Paul Eggert


Release 2001c - 2001-06-05 13:59:55 -0400

  changes by Paul Eggert and Andrew Brown


Release 2001b - 2001-04-05 16:44:38 -0400

  changes by Paul Eggert (modulo jnorgard's typo fix)

  tz-art.htm has been HTMLified.


Release 2001a - 2001-03-13 12:57:44 -0500

  changes by Paul Eggert

  An addition to the "leapseconds" file: comments with the text of the
  latest IERS leap second notice.

  Trailing white space has been removed from data file lines, and
  repeated spaces in "Rule Jordan" lines in the "asia" file have been
  converted to tabs.


Release 2000h - 2000-12-14 15:33:38 -0500

  changes by Paul Eggert

  one typo fix in the "art" file

  With providence, this is the last update of the millennium.


Release 2000g - 2000-10-10 11:35:22 -0400

  changes by Paul Eggert

  correction of John Mackin's name submitted by Robert Elz

  Garry Shandling's Daylight Saving Time joke (!?!) from the recent
  Emmy Awards broadcast.


Release 2000f - 2000-08-10 09:31:58 -0400

  changes by Paul Eggert

  Added information in "tz-art.htm" on a Seinfeld reference to DST.

  Error checking and messages in the "yearistype" script have been
  improved.


Release 2000e - 2000-07-31 09:27:54 -0400

  data changes by Paul Eggert

  a change to the default value of the defined constant HAVE_STRERROR

  the addition of a Dave Barry quote on DST to the tz-arts file


Release 2000d - 2000-04-20 15:43:04 -0400

  changes to the documentation and code of strftime for C99 conformance

  a bug fix for date.c

  These are based on (though modified from) changes by Paul Eggert.


Release 2000c - 2000-03-04 10:31:43 -0500

  changes by Paul Eggert


Release 2000b - 2000-02-21 12:16:29 -0500

  changes by Paul Eggert and Joseph Myers

  modest tweaks to the tz-art.htm and tz-link.htm files


Release 2000a - 2000-01-18 09:21:26 -0500

  changes by Paul Eggert

  The two hypertext documents have also been renamed.


Release code1999i-data1999j - 1999-11-15 18:43:22 -0500

  Paul Eggert's changes

  additions to the "zic" manual page and the "Arts.htm" file


Release code1999h-data1999i - 1999-11-08 14:55:21 -0500

  [not summarized]


Release data1999h - 1999-10-07 03:50:29 -0400

  changes by Paul Eggert to "europe" (most importantly, fixing
  Lithuania and Estonia)


Release 1999g - 1999-09-28 11:06:18 -0400

  data changes by Paul Eggert (most importantly, the change for
  Lebanon that buys correctness for this coming Sunday)

  The "code" file contains changes to "Makefile" and "checktab.awk" to
  allow better checking of time zone files before they are published.


Release 1999f - 1999-09-23 09:48:14 -0400

  changes by Arthur David Olson and Paul Eggert


Release 1999e - 1999-08-17 15:20:54 -0400

  changes circulated by Paul Eggert, although the change to handling
  of DST-specifying timezone names has been commented out for now
  (search for "XXX" in "localtime.c" for details).  These files also
  do not make any changes to the start of DST in Brazil.

  In addition to Paul's changes, there are updates to "Arts.htm" and
  cleanups of URLs.


Release 1999d - 1999-03-30 11:31:07 -0500

  changes by Paul Eggert

  The Makefile's "make public" rule has also been changed to do a test
  compile of each individual time zone data file (which should help
  avoid problems such as the one we had with Nicosia).


Release 1999c - 1999-03-25 09:47:47 -0500

  changes by Paul Eggert, most importantly the change for Chile.


Release 1999b - 1999-02-01 17:51:44 -0500

  changes by Paul Eggert

  code changes (suggested by Mani Varadarajan, mani at be.com) for
  correct handling of symbolic links when building using a relative directory

  code changes to generate correct messages for failed links

  updates to the URLs in Arts.htm


Release 1999a - 1999-01-19 16:20:29 -0500

  error message internationalizations and corrections in zic.c and
  zdump.c (as suggested by Vladimir Michl, vladimir.michl at upol.cz,
  to whom thanks!)


Release code1998h-data1998i - 1998-10-01 09:56:10 -0400

  changes for Brazil, Chile, and Germany

  support for use of "24:00" in the input files for the time zone compiler


Release code1998g-data1998h - 1998-09-24 10:50:28 -0400

  changes by Paul Eggert

  correction to a define in the "private.h" file


Release data1998g - 1998-08-11 03:28:35 -0000
  [tzdata1998g.tar.gz is missing!]

  Lithuanian change provided by mgedmin at pub.osf.it

  Move creation of the GMT link with Etc/GMT to "etcetera" (from
  "backward") to ensure that the GMT file is created even where folks
  don't want the "backward" links (as suggested by Paul Eggert).


Release data1998f - 1998-07-20 13:50:00 -0000
  [tzdata1998f.tar.gz is missing!]

  Update the "leapseconds" file to include the newly announced
  insertion at the end of 1998.


Release code1998f - 1998-06-01 10:18:31 -0400

  addition to localtime.c by Guy Harris


Release 1998e - 1998-05-28 09:56:26 -0400

  The Makefile is changed to produce zoneinfo-posix rather than
  zoneinfo/posix, and to produce zoneinfo-leaps rather than
  zoneinfo/right.

  data changes by Paul Eggert

  changes from Guy Harris to provide asctime_r and ctime_r

  A usno1998 file (substantially identical to usno1997) has been added.


Release 1998d - 1998-05-14 11:58:34 -0400

  changes to comments (in particular, elimination of references to CIA maps).
  "Arts.htm", "WWW.htm", "asia", and "australasia" are the only places
  where changes occur.


Release 1998c - 1998-02-28 12:32:26 -0500

  changes by Paul Eggert (save the "French correction," on which I'll
  wait for the dust to settle)

  symlink changes

  changes and additions to Arts.htm


Release 1998b - 1998-01-17 14:31:51 -0500

  URL cleanups and additions


Release 1998a - 1998-01-13 12:37:35 -0500

  changes by Paul Eggert


Release code1997i-data1997k - 1997-12-29 09:53:41 -0500

  changes by Paul Eggert, with minor modifications from Arthur David
  Olson to make the files more browser friendly


Release code1997h-data1997j - 1997-12-18 17:47:35 -0500

  minor changes to put "TZif" at the start of each timezone information file

  a rule has also been added to the Makefile so you can
	make zones
  to just recompile the zone information files (rather than doing a
  full "make install" with its other effects).


Release data1997i - 1997-10-07 08:45:38 -0400

  changes to Africa by Paul Eggert


Release code1997g-data1997h - 1997-09-04 16:56:54 -0400

  corrections for Uruguay (and other locations)

  Arthur David Olson's simple-minded fix allowing mktime to both
  correctly handle leap seconds and correctly handle tm_sec values
  upon which arithmetic has been performed.


Release code1997f-data1997g - 1997-07-19 13:15:02 -0400

  Paul Eggert's updates

  a small change to a function prototype;

  "Music" has been renamed "Arts.htm", HTMLified, and augmented to
  include information on Around the World in Eighty Days.


Release code1997e-data1997f - 1997-05-03 18:52:34 -0400

  fixes to zic's error handling

  changes inspired by the item circulated on Slovenia

  The description of Web resources has been HTMLified for browsing
  convenience.

  A new piece of tz-related music has been added to the "Music" file.


Release code1997d-data1997e - 1997-03-29 12:48:52 -0500

  Paul Eggert's latest suggestions


Release code1997c-data1997d - 1997-03-07 20:37:54 -0500

  changes to "zic.c" to correct performance of the "-s" option

  a new file "usno1997"


Release data1997c - 1997-03-04 09:58:18 -0500

  changes in Israel


Release 1997b - 1997-02-27 18:34:19 -0500

  The data file incorporates the 1997 leap second.

  The code file incorporates Arthur David Olson's take on the
  zic/multiprocessor/directory-creation situation.


Release 1997a - 1997-01-21 09:11:10 -0500

  Paul Eggert's Antarctica (and other changes)

  Arthur David Olson finessed the "getopt" issue by checking against
  both -1 and EOF (regardless of POSIX, SunOS 4.1.1's manual says -1
  is returned while SunOS 5.5's manual says EOF is returned).


Release code1996o-data1996n - 1996-12-27 21:42:05 -0500

  Paul Eggert's latest changes


Release code1996n - 1996-12-16 09:42:02 -0500

  link snapping fix from Bruce Evans (via Garrett Wollman)


Release data1996m - 1996-11-24 02:37:34 -0000
  [tzdata1996m.tar.gz is missing!]

  Paul Eggert's batch of changes


Release code1996m-data1996l - 1996-11-05 14:00:12 -0500

  No functional changes here; the files have simply been changed to
  make more use of ISO style dates in comments. The names of the above
  files now include the year in full.


Release code96l - 1996-09-08 17:12:20 -0400

  tzcode96k was missing a couple of pieces.


Release 96k - 1996-09-08 16:06:22 -0400

  the latest round of changes from Paul Eggert

  the recent Year 2000 material


Release code96j - 1996-07-30 13:18:53 -0400

  Set sp->typecnt as suggested by Timothy Patrick Murphy.


Release code96i - 1996-07-27 20:11:35 -0400

  Paul's suggested patch for strftime %V week numbers


Release data96i - 1996-07-01 18:13:04 -0400

  "northamerica" and "europe" changes by Paul Eggert


Release code96h - 1996-06-05 08:02:21 -0400

  fix for handling transitions specified in Universal Time

  Some "public domain" notices have also been added.


Release code96g - 1996-05-16 14:00:26 -0400

  fix for the simultaneous-DST-and-zone-change challenge


Release data96h - 1996-05-09 17:40:51 -0400

  changes by Paul Eggert


Release code96f-data96g - 1996-05-03 03:09:59 -0000
  [tzcode96f.tar.gz + tzdata96g.tar.gz are both missing!]

  The changes get us some of the way to fixing the problems noted in Paul
  Eggert's letter yesterday (in addition to a few others).  The approach
  has been to make zic a bit smarter about figuring out what time zone
  abbreviations apply just after the time specified in the "UNTIL" part
  of a zone line.  Putting the smarts in zic means avoiding having
  transition times show up in both "Zone" lines and "Rule" lines, which
  in turn avoids multiple transition time entries in time zone files.
  (This also makes the zic input files such as "europe" a bit shorter and
  should ease maintenance.)


Release data96f - 1996-04-19 19:20:03 -0000
  [tzdata96f.tar.gz is missing!]

  The only changes are to the "northamerica" file; the time zone
  abbreviation for Denver is corrected to MST (and MDT), and the
  comments for Mexico have been updated.


Release data96e - 1996-03-19 17:37:26 -0500

  Proposals by Paul Eggert, in particular the Portugal change that
  comes into play at the end of this month.


Release data96d - 1996-03-18 20:49:39 -0500

  [not summarized]


Release code96e - 1996-02-29 15:43:27 -0000
  [tzcode96e.tar.gz is missing!]

  internationalization changes and the fix to the documentation for strftime


Release code96d-data96c - 1996-02-12 11:05:27 -0500

  The "code" file simply updates Bob Kridle's electronic address.

  The "data" file updates rules for Mexico.


Release data96b - 1996-01-27 15:44:42 -0500

  Kiribati change


Release code96c - 1996-01-16 16:58:15 -0500

  leap-year streamlining and binary-search changes

  fix to newctime.3


Release code96b - 1996-01-10 20:42:39 -0500

  fixes and enhancements from Paul Eggert, including code that
  emulates the behavior of recent versions of the SunOS "date"
  command.


Release 96a - 1996-01-06 09:08:24 -0500

  Israel updates

  fixes to strftime.c for correct ISO 8601 week number generation,
  plus support for two new formats ('G' and 'g') to give ISO 8601 year
  numbers (which are not necessarily the same as calendar year numbers)


Release code95i-data95m - 1995-12-21 12:46:47 -0500

  The latest revisions from Paul Eggert are included, the usno1995
  file has been updated, and a new file ("WWW") covering useful URLs
  has been added.


Release code95h-data95l - 1995-12-19 18:10:12 -0500

  A simplification of a macro definition, a change to data for Sudan,
  and (for last minute shoppers) notes in the "Music" file on the CD
  "Old Man Time".


Release code95g-data95k - 1995-10-30 10:32:47 -0500

  (slightly reformatted) 8-bit-clean proposed patch

  minor patch: US/Eastern -> America/New_York

  snapshot of the USNO's latest data ("usno1995")

  some other minor cleanups


Release code95f-data95j - 1995-10-28 21:01:34 -0000
  [tzcode95f.tar.gz + tzdata95j.tar.gz are both missing!]

  European cleanups

  support for 64-bit time_t's

  optimization in localtime.c


Release code95e - 1995-10-13 13:23:57 -0400

  the mktime change to scan from future to past when trying to find time zone
  offsets


Release data95i - 1995-09-26 10:43:26 -0400

  For Canada/Central, guess that the Sun customer's "one week too
  early" was just a approximation, and the true error is one month
  too early.  This is consistent with the rest of Canada.


Release data95h - 1995-09-21 11:26:48 -0400

  latest changes from Paul Eggert


Release code95d - 1995-09-14 11:14:45 -0400

  the addition of a "Music" file, which documents four recorded
  versions of the tune "Save That Time".


Release data95g - 1995-09-01 17:21:36 -0400

  "yearistype" correction


Release data95f - 1995-08-28 20:46:56 -0400

  Paul Eggert's change to the australasia file


Release data95e - 1995-07-08 18:02:34 -0400

  The only change is a leap second at the end of this year.
  Thanks to Bradley White for forwarding news on the leap second.


Release data95d - 1995-07-03 13:26:22 -0400

  Paul Eggert's changes


Release data95c - 1995-07-02 19:19:28 -0400

  changes to "asia", "backward", "europe", and "southamerica"
  (read: northamericacentrics need not apply)


Release code95c - 1995-03-13 14:00:46 -0500

  one-line fix for sign extension problems in detzcode


Release 95b - 1995-03-04 11:22:38 -0500

  Minor changes in both:

  The "code" file contains a workaround for the lack of "unistd.h" in
  Microsoft C++ version 7.

  The "data" file contains a fixed "Link" for America/Shiprock.


Release 94h - 1994-12-10 12:51:14 -0500

  The files:

  *	incorporate the changes to "zdump" and "date" to make changes to
	the "TZ" environment variable permanent;

  *	incorporate the table changes by Paul Eggert;

  *	include (and document) support for universal time specifications in
	data files - but do not (yet) include use of this feature in the
	data files.

  Think of this as "TZ Classic" - the software has been set up not to break if
  universal time shows up in its input, and data entries have been
  left as is so as not to break existing implementations.


Release data94f - 1994-08-20 12:56:09 -0400

  (with thanks!) the latest data updates from Paul Eggert


Release data94e - 1994-06-04 13:13:53 -0400

  [not summarized]


Release code94g - 1994-05-05 12:14:07 -0400

  fix missing "optind.c" and a reference to it in the Makefile


Release code94f - 1994-05-05 13:00:33 -0000
  [tzcode94f.tar.gz is missing!]

  changes to avoid overflow in difftime, as well as changes to cope
  with the 52/53 challenge in strftime


Release code94e - 1994-03-30 23:32:59 -0500

  change for the benefit of PCTS


Release 94d - 1994-02-24 15:42:25 -0500

  Avoid clashes with POSIX semantics for zones such as GMT+4.

  Some other very minor housekeeping is also present.


Release code94c - 1994-02-10 08:52:40 -0500

  Fix bug where mkdirs was broken unless you compile with
  -fwritable-strings (which is generally losing to do).


Release 94b - 1994-02-07 10:04:33 -0500

  work by Paul Eggert who notes:

  I found another book of time zone histories by E W Whitman; it's not
  as extensive as Shanks but has a few goodies of its own.  I used it
  to update the tables.  I also fixed some more as a result of
  correspondence with Adam David and Peter Ilieve, and move some stray
  links from 'europe' to 'backward'.  I corrected some scanning errors
  in usno1989.

  As far as the code goes, I fixed zic to allow years in the range
  INT_MIN to INT_MAX; this fixed a few boundary conditions around 1900.
  And I cleaned up the zic documentation a little bit.


Release data94a - 1994-02-03 08:58:54 -0500

  It simply incorporates the recently announced leap second into the
  "leapseconds" file.


Release 93g - 1993-11-22 17:28:27 -0500

  Paul Eggert has provided a good deal of historic information (based
  on Shanks), and there are some code changes to deal with the buglets
  that crawled out in dealing with the new information.


Release 93f - 1993-10-15 12:27:46 -0400

  Paul Eggert's changes


Release 93e - 1993-09-05 21:21:44 -0400

  This has updated data for Israel, England, and Kwajalein.  There's
  also an update to "zdump" to cope with Kwajalein's 24-hour jump.
  Thanks to Paul Eggert and Peter Ilieve for the changes.


Release 93d - 1993-06-17 23:34:17 -0400

  new fix and new data on Israel


Release 93c - 1993-06-06 19:31:55 -0400

  [not summarized]


Release 93b - 1993-02-02 14:53:58 -0500

  updated "leapseconds" file


Release 93 - 1993-01-08 07:01:06 -0500

  At kre's suggestion, the package has been split in two - a code piece
  (which also includes documentation) that's only of use to folks who
  want to recompile things and a data piece useful to anyone who can
  run "zic".

  The new version has a few changes to the data files, a few
  portability changes, and an off-by-one fix (with thanks to
  Tom Karzes at deshaw.com for providing a description and a
  solution).


Release 92c - 1992-11-21 17:35:36 -0000
  [tz92c.tar.Z is missing!]

  The fallout from the latest round of DST transitions.

  There are changes for Portugal, Saskatchewan, and "Pacific-New";
  there's also a change to "zic.c" that makes it portable to more systems.


Release 92 - 1992-04-25 18:17:03 -0000
  [tz92.tar.Z is missing!]

  By popular demand (well, at any rate, following a request by kre at munnari)


The 1989 update of the time zone package featured:

  *	POSIXization (including interpretation of POSIX-style TZ environment
	variables, provided by Guy Harris),
  *	ANSIfication (including versions of "mktime" and "difftime"),
  *	SVIDulation (an "altzone" variable)
  *	MACHination (the "gtime" function)
  *	corrections to some time zone data (including corrections to the rules
	for Great Britain and New Zealand)
  *	reference data from the United States Naval Observatory for folks who
	want to do additional time zones
  *	and the 1989 data for Saudi Arabia.

  (Since this code will be treated as "part of the implementation" in some
  places and as "part of the application" in others, there's no good way to
  name functions, such as timegm, that are not part of the proposed ANSI C
  standard; such functions have kept their old, underscore-free names in this
  update.)

  And the "dysize" function has disappeared; it was present to allow
  compilation of the "date" command on old BSD systems, and a version of "date"
  is now provided in the package.  The "date" command is not created when you
  "make all" since it may lack options provided by the version distributed with
  your operating system, or may not interact with the system in the same way
  the native version does.

  Since POSIX frowns on correct leap second handling, the default behavior of
  the "zic" command (in the absence of a "-L" option) has been changed to omit
  leap second information from its output files.


-----
Notes

This file contains copies of the part of each release announcement
that talks about the changes in that release.  The text has been
adapted and reformatted for the purposes of this file.

Traditionally a release R consists of a pair of tarball files,
tzcodeR.tar.gz and tzdataR.tar.gz.  However, some releases (e.g.,
code2010a, data2012c) consist of just one or the other tarball, and a
few (e.g., code2012c-data2012d) have tarballs with mixed version
numbers.  Recent releases also come in an experimental format
consisting of a single tarball tzdb-R.tar.lz with extra data.

A release’s timestamp is taken from the release’s commit (for newer,
Git-based releases), from the newest file in the tarball (for older
releases, where this info is available) or from the email announcing
the release (if all else fails; these are marked with a time zone
abbreviation of -0000 and an "is missing!" comment).

Earlier versions of the code and data were not announced on the tz
list and are not summarized here.

This file is in the public domain.

Local Variables:
coding: utf-8
End:
README0000644000000000000000000000462615242710656006704 0ustar00README for the tz distribution

“Where do I set the hands of the clock?” – Les Tremayne as The King
“Oh that – you can set them any place you want.” – Frank Baxter as The Scientist
					(from the Bell System film “About Time”)

The Time Zone Database (called tz, tzdb or zoneinfo) contains code and
data that represent the history of local time for many representative
locations around the globe.  It is updated periodically to reflect
changes made by political bodies to time zone boundaries, UTC offsets,
and daylight-saving rules.

See  or the
file tz-link.html for how to acquire the code and data.

Once acquired, read the leading comments in the file ‘Makefile’
and make any changes needed to make things right for your system,
especially when using a platform other than current GNU/Linux.

Then run the following commands, substituting your desired
installation directory for ‘$HOME/tzdir’:

	make TOPDIR="$HOME/tzdir" install
	"$HOME/tzdir/usr/bin/zdump" -v America/Los_Angeles

See the file tz-how-to.html for examples of how to read the data files.

This database of historical local time information has several goals:

 * Provide a compendium of data about the history of civil time that
   is useful even if not 100% accurate.

 * Give an idea of the variety of local time rules that have existed
   in the past and thus may be expected in the future.

 * Test the generality of the local time rule description system.

The information in the time zone data files is by no means authoritative;
fixes and enhancements are welcome.  Please see the file CONTRIBUTING
for details.

Thanks to these Time Zone Caballeros who’ve made major contributions to the
time conversion package: Keith Bostic; Bob Devine; Paul Eggert; Robert Elz;
Guy Harris; Mark Horton; John Mackin; and Bradley White.  Thanks also to
Michael Bloom, Art Neilson, Stephen Prince, John Sovereign, and Frank Wales
for testing work, and to Gwillim Law for checking local mean time data.
Thanks in particular to Arthur David Olson, the project’s founder and first
maintainer, to whom the time zone community owes the greatest debt of all.
None of them are responsible for remaining errors.

-----

This file is in the public domain, so clarified as of 2009-05-17 by
Arthur David Olson.  The other files in this distribution are either
public domain or BSD licensed; see the file LICENSE for details.
SECURITY0000644000000000000000000000141315054513212007173 0ustar00Please report any sensitive security-related bugs via email to the
tzdb designated coordinators, currently Paul Eggert
 and Tim Parenti .
Put “tzdb security” at the start of your email’s subject line.
We prefer communications to be in English.

You should receive a response within a week. If not, please follow up
via email to make sure we received your original message.

If we confirm the bug, we plan to notify affected third-party services
or software that we know about, prepare an advisory, commit fixes to
the main development branch as quickly as is practical, and finally
publish the advisory on tz@iana.org.  As with all tzdb contributions,
we give credit to security contributors unless they wish to remain
anonymous.
theory.html0000644000000000000000000020427415251274013010215 0ustar00


  Theory and pragmatics of the tz code and data
  
  
  






Theory and pragmatics of the tz code and data

  

  
Scope of the tz database



The tz
database attempts to record the history and predicted future of
civil time scales.
It organizes time zone and daylight saving time
data by partitioning the world into timezones
whose clocks all agree about timestamps that occur after the POSIX Epoch
(1970-01-01 00:00:00 UTC).
Although 1970 is a somewhat-arbitrary cutoff, there are significant
challenges to moving the cutoff earlier even by a decade or two, due
to the wide variety of local practices before computer timekeeping
became prevalent.
Most timezones correspond to a notable location and the database
records all known clock transitions for that location;
some timezones correspond instead to a fixed UTC offset.





Each timezone typically corresponds to a geographical region that is
smaller than a traditional time zone, because clocks in a timezone
all agree after 1970 whereas a traditional time zone merely
specifies current standard time. For example, applications that deal
with current and future timestamps in the traditional North
American mountain time zone can choose from the timezones
America/Denver which observes US-style daylight saving
time (DST),
and America/Phoenix which does not observe DST.
Applications that also deal with past timestamps in the mountain time
zone can choose from over a dozen timezones, such as
America/Boise and
America/Hermosillo, each of which currently uses mountain
time but differs from other timezones for some timestamps after 1970.





Clock transitions before 1970 are recorded for location-based timezones,
because most systems support timestamps before 1970 and could
misbehave if data entries were omitted for pre-1970 transitions.
However, the database is not designed for and does not suffice for
applications requiring accurate handling of all past times everywhere,
as it would take far too much effort and guesswork to record all
details of pre-1970 civil timekeeping.
Although some information outside the scope of the database is
collected in a file backzone that is distributed along
with the database proper, this file is less reliable and does not
necessarily follow database guidelines.





As described below, reference source code for using the
tz database is also available.
The tz code is upwards compatible with POSIX, an international
standard for UNIX-like systems.
As of this writing, the current edition of POSIX is
POSIX.1-2024
(The Open Group Base Specifications Issue 8, IEEE Std 1003.1-2024).
Unlike its predecessors
POSIX.1-1988 through
POSIX.1-2017,
POSIX.1-2024 requires support for the
tz database, which has a
model for describing civil time that is more complex than the
standard and daylight saving times required by earlier POSIX editions.
A tz timezone corresponds to a ruleset that can
have more than two changes per year, these changes need not merely
flip back and forth between two alternatives, and the rules themselves
can change at times.
Whether and when a timezone changes its clock,
and even the timezone’s notional base offset from UTC,
are variable.
It does not always make sense to talk about a timezone’s
“base offset”, which is not necessarily a single number.



  
Timezone identifiers



Each timezone has a name that uniquely identifies the timezone.
Inexperienced users are not expected to select these names unaided.
Distributors should provide documentation and/or a simple selection
interface that explains each name via a map or via descriptive text like
“Czech Republic” instead of the timezone name “Europe/Prague”.
If geolocation information is available, a selection interface can
locate the user on a timezone map or prioritize names that are
geographically close. For an example selection interface, see the
tzselect program in the tz code.
Unicode’s Common Locale Data
Repository (CLDR)
contains data that may be useful for other selection
interfaces; it maps timezone names like Europe/Prague to
locale-dependent strings like “Prague”, “Praha”, “Прага”, and “布拉格”.





The naming conventions attempt to strike a balance
among the following goals:




    
  
  • 
    Uniquely identify every timezone where clocks have agreed since 1970.
    This is essential for the intended use: static clocks keeping local
    civil time.
  
    • 
  
  • 
    Indicate to experts where the timezone’s clocks typically are.
  
    • 
  
  • 
    Be robust in the presence of political changes.
    For example, names are typically not tied to countries, to avoid
    incompatibilities when countries change their name (e.g.,
    Swaziland→Eswatini) or when locations change countries (e.g., Hong
    Kong from UK colony to China).
    There is no requirement that every country or national
    capital must have a timezone name.
  
    • 
  
  • 
    Be portable to a wide variety of implementations.
  
    • 
  
  • 
    Use a consistent naming conventions over the entire world.
  
    • 





Names normally have the format
AREA/LOCATION, where
AREA is a continent or ocean, and
LOCATION is a specific location within the area.
North and South America share the same area, America.
Typical names are Africa/Cairo,
America/New_York, and Pacific/Honolulu.
Some names are further qualified to help avoid confusion; for example,
America/Indiana/Petersburg distinguishes Petersburg,
Indiana from other Petersburgs in America.





Here are the general guidelines used for
choosing timezone names,
in decreasing order of importance:




    
  
  • 
    Use only valid POSIX file name components (i.e., the parts of
    names other than "/").
    Do not use the file name components "." and
    "..".
    Within a file name component, use only ASCII letters,
    ".", "-" and "_".
    Do not use digits, as that might create an ambiguity with POSIX’s
    proleptic TZ strings.
    A file name component must not exceed 14 characters or start with
    "-".
    E.g., prefer America/Noronha to
    America/Fernando_de_Noronha.
    Exceptions: see the discussion of legacy names below.
  
    • 
  
  • 
    A name must not be empty, or contain "//", or
    start or end with "/".
    Also, a name must not be "Etc/Unknown", as
    CLDR uses that string for an unknown or invalid timezone.
  
    • 
  
  • 
    Do not use names that differ only in case.
    Although the reference implementation is case-sensitive, some
    other implementations are not, and they would mishandle names
    differing only in case.
  
    • 
  
  • 
    If one name A is an initial prefix of another
    name AB (ignoring case), then B must not
    start with "/", as a regular file cannot have the
    same name as a directory in POSIX.
    For example, America/New_York precludes
    America/New_York/Bronx.
  
    • 
  
  • 
    Uninhabited regions like the North Pole and Bouvet Island
    do not need locations, since local time is not defined there.
  
    • 
  
  • 
    If all clocks in a region have agreed since 1970,
    give them just one name even if some of the clocks disagreed before 1970,
    or reside in different countries or in notable or faraway locations.
    Otherwise these tables would become annoyingly large.
    For example, do not create a name Indian/Crozet
    as a near-duplicate or alias of Asia/Dubai
    merely because they are different countries or territories,
    or their clocks disagreed before 1970, or the
    Crozet Islands
    are notable in their own right,
    or the Crozet Islands are not adjacent to other locations
    that use Asia/Dubai.
  
    • 
  
  • 
    If boundaries between regions are fluid, such as during a war or
    insurrection, do not bother to create a new timezone merely
    because of yet another boundary change. This helps prevent table
    bloat and simplifies maintenance.
  
    • 
  
  • 
    If a name is ambiguous, use a less ambiguous alternative;
    e.g., many cities are named San José and Georgetown, so
    prefer America/Costa_Rica to
    America/San_Jose and America/Guyana
    to America/Georgetown.
  
    • 
  
  • 
    Keep locations compact.
    Use cities or small islands, not countries or regions, so that any
    future changes do not split individual locations into different
    timezones.
    E.g., prefer Europe/Paris to Europe/France,
    since
    France
    has had multiple time zones.
  
    • 
  
  • 
    Use mainstream English spelling, e.g., prefer
    Europe/Rome to Europa/Roma, and
    prefer Europe/Athens to the Greek
    Ευρώπη/Αθήνα or the Romanized
    Evrópi/Athína.
    The POSIX file name restrictions encourage this guideline.
  
    • 
  
  • 
    Use the most populous among locations in a region,
    e.g., prefer Asia/Shanghai to
    Asia/Beijing.
    Among locations with similar populations, pick the best-known
    location, e.g., prefer Europe/Rome to
    Europe/Milan.
  
    • 
  
  • 
    Use the singular form, e.g., prefer Atlantic/Canary to
    Atlantic/Canaries.
  
    • 
  
  • 
    Omit common suffixes like "_Islands" and
    "_City", unless that would lead to ambiguity.
    E.g., prefer America/Cayman to
    America/Cayman_Islands and
    America/Guatemala to
    America/Guatemala_City, but prefer
    America/Mexico_City to
    America/Mexico
    because the
    country of Mexico has several time zones.
  
    • 
  
  • 
    Use "_" to represent a space.
  
    • 
  
  • 
    Omit "." from abbreviations in names.
    E.g., prefer Atlantic/St_Helena to
    Atlantic/St._Helena.
  
    • 
  
  • 
    Do not change established names if they only marginally violate
    the above guidelines.
    For example, do not change the existing name Europe/Rome to
    Europe/Milan merely because Milan’s population has grown
    to be somewhat greater than Rome’s.
  
    • 
  
  • 
    If a name is changed, put its old spelling in the
    "backward" file as a link to the new spelling.
    This means old spellings will continue to work.
    Ordinarily a name change should occur only in the rare case when
    a location’s consensus English-language spelling changes; for example,
    in 2008 Asia/Calcutta was renamed to Asia/Kolkata
    due to long-time widespread use of the new city name instead of the old.
  
    • 





Guidelines have evolved with time, and names following old versions of
these guidelines might not follow the current version. When guidelines
have changed, old names continue to be supported. Guideline changes
have included the following:




    

  • 
Older versions of this package used a different naming scheme.
See the file "backward" for most of these older names
(e.g., US/Eastern instead of America/New_York).
The other old-fashioned names still supported are
WET, CET, MET, and
EET (see the file "europe").

    • 


  • 
Older versions of this package defined legacy names that are
incompatible with the first guideline of location names, but which are
still supported.
These legacy names are mostly defined in the file
"etcetera".
Also, the file "backward" defines the legacy names
Etc/GMT0, Etc/GMT-0, Etc/GMT+0,
GMT0, GMT-0, GMT+0,
EST5EDT, CST6CDT,
MST7MDT, and PST8PDT.

    • 


  • 
Older versions of these guidelines said that
there should typically be at least one name for each ISO
3166-1 officially assigned two-letter code for an inhabited
country or territory.
This old guideline has been dropped, as it was not needed to handle
timestamps correctly and it increased maintenance burden.

    • 





The file zone1970.tab lists geographical locations used
to name timezones.
It is intended to be an exhaustive list of names for geographic
regions as described above; this is a subset of the timezones in the data.
Although a zone1970.tab location’s
longitude
corresponds to
its local mean
time (LMT) offset with one hour for every 15°
east longitude, this relationship is not exact.
The backward-compatibility file zone.tab is similar
but conforms to the older-version guidelines related to ISO 3166-1;
it lists only one country code per entry and unlike zone1970.tab
it can list names defined in backward.
Applications that process only timestamps from now on can instead use the file
zonenow.tab, which partitions the world more coarsely,
into regions where clocks agree now and in the predicted future;
this file is smaller and simpler than zone1970.tab
and zone.tab.





The database defines each timezone name to be a zone, or a link to a zone.
The source file backward defines links
and a few zones for backward compatibility.
Although backward was originally designed to be optional,
nowadays distributions typically use it
and no great weight should be attached to whether a link or zone
is defined in backward or in some other file.
The source file etcetera defines names that may be useful
on platforms that do not support proleptic TZ strings
like <+08>-8;
no other source file other than backward
contains links to its zones.
One of etcetera’s names is Etc/UTC,
used by functions like gmtime to obtain leap
second information on platforms that support leap seconds.
Another etcetera name, GMT,
is used by older code releases.



  
Time zone abbreviations



When this package is installed, it generates time zone abbreviations
like EST to be compatible with human tradition and POSIX.
Here are the general guidelines used for choosing time zone abbreviations,
in decreasing order of importance:




    
  
  • 
    Use three to six characters that are ASCII alphanumerics or
    "+" or "-".
    Previous editions of this database also used characters like
    space and "?", but these characters have a
    special meaning to the
    UNIX shell
    and cause commands like
    "set
    `date`"
    to have unexpected effects.
    Previous editions of this guideline required upper-case letters, but the
    Congressman who introduced
    Chamorro
    Standard Time preferred “ChST”, so lower-case letters are now allowed.
    Also, POSIX from 2001 on relaxed the rule to allow "-",
    "+", and alphanumeric characters from the portable
    character set in the current locale.
    In practice ASCII alphanumerics and "+" and
    "-" are safe in all locales.

    
    
    In other words, in the C locale the POSIX extended regular
    expression [-+[:alnum:]]{3,6} should match the
    abbreviation.
    This guarantees that all abbreviations could have been specified
    explicitly by a POSIX proleptic TZ string.
    
    
  
    • 
  
  • 
    Use abbreviations that are in common use among English-speakers,
    e.g., “EST” for Eastern Standard Time in North America.
    We assume that applications translate them to other languages
    as part of the normal localization process; for example,
    a French application might translate “EST” to “HNE”.

    
    
    These abbreviations (for standard/daylight/etc. time) are:
      ACST/ACDT Australian Central,
      AST/ADT/APT/AWT/ADDT Atlantic,
      AEST/AEDT Australian Eastern,
      AHST/AHDT Alaska-Hawaii,
      AKST/AKDT Alaska,
      AWST/AWDT Australian Western,
      BST/BDT Bering,
      CAT/CAST Central Africa,
      CET/CEST/CEMT Central European,
      ChST Chamorro,
      CST/CDT/CWT/CPT Central [North America],
      CST/CDT China,
      GMT/BST/IST/BDST Greenwich,
      EAT East Africa,
      EST/EDT/EWT/EPT Eastern [North America],
      EET/EEST Eastern European,
      GST/GDT Guam,
      HST/HDT/HWT/HPT Hawaii,
      HKT/HKST/HKWT Hong Kong,
      IST India,
      IST/GMT Irish,
      IST/IDT/IDDT Israel,
      JST/JDT Japan,
      KST/KDT Korea,
      MET/MEST Middle European (a backward-compatibility alias for
	Central European),
      MSK/MSD Moscow,
      MST/MDT/MWT/MPT Mountain,
      NST/NDT/NWT/NPT/NDDT Newfoundland,
      NST/NDT/NWT/NPT Nome,
      NZMT/NZST New Zealand through 1945,
      NZST/NZDT New Zealand 1946–present,
      PKT/PKST Pakistan,
      PST/PDT/PWT/PPT Pacific,
      PST/PDT Philippine,
      SAST South Africa,
      SST Samoa,
      UTC Universal,
      WAT/WAST West Africa,
      WET/WEST/WEMT Western European,
      WIB Waktu Indonesia Barat,
      WIT Waktu Indonesia Timur,
      WITA Waktu Indonesia Tengah,
      YST/YDT/YWT/YPT/YDDT Yukon.
    
    
  
    • 
  
  • 
    
    
    For times taken from a city’s longitude, use the
    traditional xMT notation.
    The only abbreviation like this in current use is GMT.
    The others are for timestamps before 1960,
    except that Monrovia Mean Time persisted until 1972.
    Typically, numeric abbreviations (e.g., -004430 for
    MMT) would cause trouble here, as the numeric strings would exceed
    the POSIX length limit.
    
    

    
    
    These abbreviations are:
      AMT Asunción, Athens;
      BMT Baghdad, Bangkok, Batavia, Bermuda, Bern, Bogotá,
        Brussels, Bucharest;
      CMT Calamarca, Caracas, Chisinau, Colón, Córdoba;
      DMT Dublin/Dunsink;
      EMT Easter;
      FFMT Fort-de-France;
      FMT Funchal;
      GMT Greenwich;
      HMT Havana, Helsinki, Horta, Howrah;
      IMT Irkutsk, Istanbul;
      JMT Jerusalem;
      KMT Kaunas, Kyiv, Kingston;
      LMT Lima, Lisbon, local;
      MMT Macassar, Madras, Malé, Managua, Minsk, Monrovia, Montevideo,
	Moratuwa, Moscow;
      PLMT Phù Liễn;
      PMT Paramaribo, Paris, Perm, Pontianak, Prague;
      PMMT Port Moresby;
      PPMT Port-au-Prince;
      QMT Quito;
      RMT Rangoon, Riga, Rome;
      SDMT Santo Domingo;
      SJMT San José;
      SMT Santiago, Simferopol, Singapore, Stanley;
      TBMT Tbilisi;
      TMT Tallinn, Tehran;
      WMT Warsaw.
    
    

    
    
    A few abbreviations also follow the pattern that
    GMT/BST established for time in the UK.
    They are:
      BMT/BST for Bermuda 1890–1930,
      CMT/BST for Calamarca Mean Time and Bolivian Summer Time
	1890–1932,
      DMT/IST for Dublin/Dunsink Mean Time and Irish Summer Time
	1880–1916,
      MMT/MST/MDST for Moscow 1880–1919, and
      RMT/LST for Riga Mean Time and Latvian Summer time 1880–1926.
    
    
    
  
    • 
  
  • 
    Use “LMT” for local mean time of locations before the
    introduction of standard time; see “Scope of the
    tz database”.
  
    • 
  
  • 
    If there is no common English abbreviation, use numeric offsets like
    -05 and +0530 that are generated
    by zic’s %z notation.
  
    • 
  
  • 
    Use current abbreviations for older timestamps to avoid confusion.
    For example, in 1910 a common English abbreviation for time
    in central Europe was “MEZ” (short for both “Middle European
    Zone” and for “Mitteleuropäische Zeit” in German).
    Nowadays “CET” (“Central European Time”) is more common in
    English, and the database uses “CET” even for circa-1910
    timestamps as this is less confusing for modern users and avoids
    the need for determining when “CET” supplanted “MEZ” in common
    usage.
  
    • 
  
  • 
    Use a consistent style in a timezone’s history.
    For example, if a history tends to use numeric
    abbreviations and a particular entry could go either way, use a
    numeric abbreviation.
  
    • 
  
  • 
    Use
    Universal Time
    (UT) (with time zone abbreviation -00) for
    locations while uninhabited.
    The leading "-" is a flag that the UT offset is in
    some sense undefined; this notation is derived
    from Internet
    RFC 3339.
    (The abbreviation Z that
    Internet
    RFC 9557 uses for this concept
    would violate the POSIX requirement
    of at least three characters in an abbreviation.)
  
    • 





Application writers should note that these abbreviations are ambiguous
in practice: e.g., CST means one thing in China and something else
in North America, and IST can refer to time in India, Ireland or
Israel.
To avoid ambiguity, use numeric UT offsets like
-0600 instead of time zone abbreviations like CST.



  
Accuracy of the tz database



The tz database is not authoritative, and it
surely has errors.
Corrections are welcome and encouraged; see the file CONTRIBUTING.
Users requiring authoritative data should consult national standards
bodies and the references cited in the database’s comments.





Errors in the tz database arise from many sources:




    
  
  • 
    The tz database predicts future
    timestamps, and current predictions
    will be incorrect after future governments change the rules.
    For example, if today someone schedules a meeting for 13:00 next
    October 1, Dublin time, and tomorrow Ireland changes its
    daylight saving rules, software can mess up after the rule change
    if it blithely relies on conversions made before the change.
  
    • 
  
  • 
    The pre-1970 entries in this database cover only a tiny sliver of how
    clocks actually behaved; the vast majority of the necessary
    information was lost or never recorded.
    Thousands more timezones would be needed if
    the tz database’s scope were extended to
    cover even just the known or guessed history of standard time; for
    example, the current single entry for France would need to split
    into dozens of entries, perhaps hundreds.
    And in most of the world even this approach would be misleading
    due to widespread disagreement or indifference about what times
    should be observed.
    In her 2015 book
    The
    Global Transformation of Time, 1870–1950,
    Vanessa Ogle writes
    “Outside of Europe and North America there was no system of time
    zones at all, often not even a stable landscape of mean times,
    prior to the middle decades of the twentieth century”.
    See: Timothy Shenk, Booked:
      A Global History of Time. Dissent 2015-12-17.
  
    • 
  
  • 
    Most of the pre-1970 data entries come from unreliable sources, often
    astrology books that lack citations and whose compilers evidently
    invented entries when the true facts were unknown, without
    reporting which entries were known and which were invented.
    These books often contradict each other or give implausible entries,
    and on the rare occasions when they are checked they are
    typically found to be incorrect.
  
    • 
  
  • 
    For the UK the tz database relies on
    years of first-class work done by
    Joseph Myers and others; see
    “History of
    legal time in Britain”.
    Other countries are not done nearly as well.
  
    • 
  
  • 
    Sometimes, different people in the same city maintain clocks
    that differ significantly.
    Historically, railway time was used by railroad companies (which
    did not always
    agree with each other), church-clock time was used for birth
    certificates, etc.
    More recently, competing political groups might disagree about
    clock settings. Often this is merely common practice, but
    sometimes it is set by law.
    For example, from 1891 to 1911 the UT offset in France
    was legally UT +00:09:21 outside train stations and
    UT +00:04:21 inside. Other examples include
    Chillicothe in 1920, Palm Springs in 1946/7, and Jerusalem and
    Ürümqi to this day.
  
    • 
  
  • 
    Although a named location in the tz
    database stands for the containing region, its pre-1970 data
    entries are often accurate for only a small subset of that region.
    For example, Europe/London stands for the United
    Kingdom, but its pre-1847 times are valid only for locations that
    have London’s exact meridian, and its 1847 transition
    to GMT is known to be valid only for the L&NW and
    the Caledonian railways.
  
    • 
  
  • 
    The tz database does not record the
    earliest time for which a timezone’s
    data entries are thereafter valid for every location in the region.
    For example, Europe/London is valid for all locations
    in its region after GMT was made the standard time,
    but the date of standardization (1880-08-02) is not in the
    tz database, other than in commentary.
    For many timezones the earliest time of
    validity is unknown.
  
    • 
  
  • 
    The tz database does not record a
    region’s boundaries, and in many cases the boundaries are not known.
    For example, the timezone
    America/Kentucky/Louisville represents a region
    around the city of Louisville, the boundaries of which are
    unclear.
  
    • 
  
  • 
    Changes that are modeled as instantaneous transitions in the
    tz
    database were often spread out over hours, days, or even decades.
  
    • 
  
  • 
    Even if the time is specified by law, locations sometimes
    deliberately flout the law.
  
    • 
  
  • 
    Early timekeeping practices, even assuming perfect clocks, were
    often not specified to the accuracy that the
    tz database requires.
  
    • 
  
  • 
    The tz database cannot represent stopped clocks.
    However, on 1911-03-11 at 00:00, some public-facing French clocks
    were changed by stopping them for a few minutes to effect a transition.
    The tz database models this via a
    backward transition; the relevant French legislation does not
    specify exactly how the transition was to occur.
  
    • 
  
  • 
    Sometimes historical timekeeping was specified more precisely
    than what the tz code can handle.
    For example, from 1880 to 1916 clocks in Ireland observed Dublin Mean
    Time (estimated to be UT
    −00:25:21.1); although the tz
    source data can represent the .1 second, TZif files and the code cannot.
    In practice these old specifications were rarely if ever
    implemented to subsecond precision.
  
    • 
  
  • 
    Even when all the timestamp transitions recorded by the
    tz database are correct, the
    tz rules that generate them may not
    faithfully reflect the historical rules.
    For example, from 1922 until World War II the UK moved clocks
    forward the day following the third Saturday in April unless that
    was Easter, in which case it moved clocks forward the previous
    Sunday.
    Because the tz database has no
    way to specify Easter, these exceptional years are entered as
    separate tz Rule lines, even though the
    legal rules did not change.
    When transitions are known but the historical rules behind them are not,
    the database contains Zone and Rule
    entries that are intended to represent only the generated
    transitions, not any underlying historical rules; however, this
    intent is recorded at best only in commentary.
  
    • 
  
  • 
    The tz database models time
    using the proleptic
    Gregorian calendar with days containing 24 equal-length hours
    numbered 00 through 23, except when clock transitions occur.
    Pre-standard time is modeled as local mean time.
    However, historically many people used other calendars and other timescales.
    For example, the Roman Empire used
    the Julian
    calendar,
    and Roman
    timekeeping had twelve varying-length daytime hours with a
    non-hour-based system at night.
    And even today, some local practices diverge from the Gregorian
    calendar with 24-hour days. These divergences range from
    relatively minor, such as Japanese bars giving times like 24:30 for the
    wee hours of the morning, to more-significant differences such as the
    east African practice of starting the day at dawn, renumbering
    the Western 06:00 to be 12:00. These practices are largely outside
    the scope of the tz code and data, which
    provide only limited support for date and time localization
    such as that required by POSIX.
    If DST is not used a different time zone
    can often do the trick; for example, in Kenya a TZ setting
    like <-03>3 or America/Cayenne starts
    the day six hours later than Africa/Nairobi does.
  
    • 
  
  • 
    Early clocks were less reliable, and data entries do not represent
    clock error.
  
    • 
  
  • 
    The tz database assumes Universal Time
    (UT) as an origin, even though UT is not
    standardized for older timestamps.
    In the tz database commentary,
    UT denotes a family of time standards that includes
    Coordinated Universal Time (UTC) along with other
    variants such as UT1 and GMT,
    with days starting at midnight.
    Although UT equals UTC for modern
    timestamps, UTC was not defined until 1960, so
    commentary uses the more general abbreviation UT for
    timestamps that might predate 1960.
    Since UT, UT1, etc. disagree slightly,
    and since pre-1972 UTC seconds varied in length,
    interpretation of older timestamps can be problematic when
    subsecond accuracy is needed.
  
    • 
  
  • 
    Civil time was not based on atomic time before 1972, and we do not
    know the history of
    earth’s
    rotation accurately enough to map SI seconds to
    historical solar time
    to more than about one-hour accuracy.
    See: Morrison LV, Stephenson FR, Hohenkerk CY, Zawilski M.
    Addendum 2020
    to ‘Measurement of the Earth’s rotation: 720 BC to AD 2015’.
    Proc Royal Soc A. 2021;477:20200776.
    Also see: Espenak F. Uncertainty
    in Delta T (ΔT).
  
    • 
  
  • 
    The relationship between POSIX time (that is, UTC but
    ignoring leap
    seconds) and UTC is not agreed upon.
    This affects time stamps during the leap second era (1972–2035).
    Although the POSIX
    clock officially stops during an inserted leap second, at least one
    proposed standard has it jumping back a second instead; and in
    practice POSIX clocks more typically either progress glacially during
    a leap second, or are slightly slowed while near a leap second.
  
    • 
  
  • 
    The tz database does not represent how
    uncertain its information is.
    Ideally it would contain information about when data entries are
    incomplete or dicey.
    Partial temporal knowledge is a field of active research, though,
    and it is not clear how to apply it here.
  
    • 





In short, many, perhaps most, of the tz
database’s pre-1970 and future timestamps are either wrong or
misleading.
Any attempt to pass the
tz database off as the definition of time
should be unacceptable to anybody who cares about the facts.
In particular, the tz database’s
LMT offsets should not be considered meaningful, and
should not prompt creation of timezones
merely because two locations
differ in LMT or transitioned to standard time at
different dates.



  
Time and date functions



The tz code contains time and date functions
that are upwards compatible with those of POSIX.
Code compatible with this package is already
part of many platforms, where the
primary use of this package is to update obsolete time-related files.
To do this, you may need to compile the time zone compiler
zic supplied with this package instead of using the
system zic, since the format of zic’s
input is occasionally extended, and a platform may still be shipping
an older zic.





In POSIX, time display in a process is controlled by the
environment variable TZ, which can have two forms:



    
  
  • 
    A proleptic TZ value
    like CET-1CEST,M3.5.0,M10.5.0/3 uses a complex
    notation that specifies a single standard time along with daylight
    saving rules that apply to all years past, present, and future.
  
    • 
  
  • 
    A geographical TZ value
    like Europe/Berlin names a location that stands for
    civil time near that location, which can have more than
    one standard time and more than one set of daylight saving rules,
    to record timekeeping practice more accurately.
    These names are defined by the tz database.
  
    • 




POSIX.1-2017 properties and limitations



Some platforms support only the features required by POSIX.1-2017
and earlier editions,
and have not yet upgraded to POSIX.1-2024.
Code intended to be portable to these platforms must deal
with problems that were fixed in later POSIX editions.




    
  
  • 
    POSIX.1-2017 does not require support for geographical TZ,
    and there is no convenient and efficient way to determine
    the UT offset and time zone abbreviation of arbitrary
    timestamps, particularly for timezones
    that do not fit into the POSIX model.
  
    • 
  
  • 
    
    
    The proleptic TZ string,
    which is all that POSIX.1-2017 requires,
    has a format that is hard to describe and is error-prone in practice.
    Also, proleptic TZ strings cannot deal with daylight
    saving time rules not based on the Gregorian calendar (as in
    Palestine), or with situations where more than two time zone
    abbreviations or UT offsets are used in an area.
    
    

    
    
    A proleptic TZ string has the following format:
    
    

    
    
    stdoffset[dst[offset][,date[/time],date[/time]]]
    
    

    
    
    where:
    
    

    
    
      
    std and dst
    
	are 3 or more characters specifying the standard
	and daylight saving time (DST) zone abbreviations.
	Starting with POSIX.1-2001, std and dst
	may also be quoted in angle brackets, like <+09>;
	this allows "+" and "-" in the names.
      
    
      
    offset
    
	is of the form
	[±]hh:[mm[:ss]]
	and specifies the offset west of UT.
	hh may be a single digit;
	0≤hh≤24.
	The default DST offset is one hour ahead of
	standard time.
      
    
      
    date[/time],date[/time]
    
	specifies the beginning and end of DST.
	If this is absent, the system supplies its own ruleset
	for DST, typically	current US
	DST rules.
      
    
      
    time
    
	takes the form
	hh:[mm[:ss]]
	and defaults to 02:00.
	This is the same format as the offset, except that a
	leading "+" or "-" is not allowed.
      
    
      
    date
    
	takes one of the following forms:
	
    
	  
    Jn (1≤n≤365)
    
	    origin-1 day number not counting February 29
	  
    
	  
    n (0≤n≤365)
    
	    origin-0 day number counting February 29 if present
	  
    
	  
    Mm.n.d
	    (0[Sunday]≤d≤6[Saturday], 1≤n≤5,
	    1≤m≤12)
    
	    for the dth day of week n of
	    month m of the year, where week 1 is the first
	    week in which day d appears, and
	    "5" stands for the last week in which
	    day d appears (which may be either the 4th or
	    5th week).
	    Typically, this is the only useful form; the n
	    and Jn forms are rarely used.
	  
    
	
    
      
    
    
    

    
    
    Here is an example proleptic TZ string for New
    Zealand after 2007.
    It says that standard time (NZST) is 12 hours ahead
    of UT, and that daylight saving time
    (NZDT) is observed from September’s last Sunday at
    02:00 until April’s first Sunday at 03:00:
    
    

    
    TZ='NZST-12NZDT,M9.5.0,M4.1.0/3'
    

    
    
    This proleptic TZ string is hard to remember, and
    mishandles some timestamps before 2008.
    With this package you can use a geographical TZ instead:
    
    

    
    TZ='Pacific/Auckland'
    
  
    • 





POSIX.1-2017 also has the limitations of POSIX.1-2024,
discussed in the next section.




POSIX.1-2024 properties and limitations



POSIX.1-2024 extends POSIX.1-2017 in the following significant ways:



    
  
  • 
    POSIX.1-2024 requires support for geographical TZ.
    Earlier POSIX editions require support only for proleptic TZ.
  
    • 
  
  • 
    POSIX.1-2024 requires struct tm
    to have a UT offset member tm_gmtoff
    and a time zone abbreviation member tm_zone.
    Earlier POSIX editions lack this requirement.
  
    • 
  
  • 
    DST transition times can range from −167:59:59
    to 167:59:59 instead of merely from 00:00:00 to 24:59:59.
    This allows for proleptic TZ strings
    like "<-02>2<-01>,M3.5.0/-1,M10.5.0/0"
    where the transition time −1:00 means 23:00 the previous day.
  
    • 




However POSIX.1-2024, like earlier POSIX editions, has some limitations:

    
  
  • 
    The TZ environment variable is process-global, which
    makes it hard to write efficient, thread-safe applications that
    need access to multiple timezones.
  
    • 
  
  • 
    In POSIX, there is no tamper-proof way for a process to learn the
    system’s best idea of local (wall clock) time.
    This is important for applications that an administrator wants
    used only at certain times – without regard to whether the
    user has fiddled the
    TZ environment variable.
    While an administrator can “do everything in UT” to
    get around the problem, doing so is inconvenient and precludes
    handling daylight saving time shifts – as might be required to
    limit phone calls to off-peak hours.
  
    • 
  
  • 
    POSIX requires that time_t clock counts exclude leap
    seconds.
  
    • 
  
  • 
    POSIX does not define the DST transitions
    for settings like TZ='EST5EDT'.
    Traditionally the current US DST rules
    were used to interpret such values, but this meant that the
    US DST rules were compiled into each
    time conversion package, and when
    US time conversion rules changed (as in the United
    States in 1987 and again in 2007), all packages that
    interpreted TZ values had to be updated
    to ensure proper results.
  
    • 




Extensions to POSIX in the
tz code



  The tz code defines some properties
  left unspecified by POSIX, and attempts to support some
  extensions to POSIX.




    
  
  • 
    The tz code attempts to support all the
    time_t implementations allowed by POSIX.
    The time_t type represents a nonnegative count of seconds
    since 1970-01-01 00:00:00 UTC, ignoring leap seconds.
    In practice, time_t is usually a signed 64- or 32-bit
    integer; 32-bit signed time_t values stop working after
    2038-01-19 03:14:07 UTC, so new implementations these
    days typically use a signed 64-bit integer.
    Unsigned 32-bit integers are used on one or two platforms, and 36-bit
    and 40-bit integers are also used occasionally.
    Although earlier POSIX versions allowed time_t to be a
    floating-point type, this was not supported by any practical system,
    and POSIX.1-2013+ and the tz code both
    require time_t to be an integer type.
  
    • 
  
  • 
    
    
    If the TZ environment variable uses the geographical format,
    it is used in generating
    the name of a file from which time-related information is read.
    The file’s format is TZif,
    a timezone information format that contains binary data; see
    Internet
    RFC 9636.
    The daylight saving time rules to be used for a
    particular timezone are encoded in the
    TZif file; the format of the file allows US,
    Australian, and other rules to be encoded, and
    allows for situations where more than two time zone
    abbreviations are used.
    
    
    
    
    When the tz code was developed in the 1980s,
    it was recognized that allowing the TZ environment
    variable to take on values such as America/New_York
    might cause old programs (that expect TZ to have a
    certain format) to operate incorrectly; consideration was given to using
    some other environment variable (for example, TIMEZONE)
    to hold the string used to generate the TZif file’s name.
    In the end, however, it was decided to continue using
    TZ: it is widely used for time zone purposes;
    separately maintaining both TZ
    and TIMEZONE seemed a nuisance; and systems where
    new forms of TZ might cause problems can simply
    use legacy settings such as TZ='EST5EDT' which
    can be used by new programs as well as by old programs that
    assume pre-POSIX TZ values.
    
    
  
    • 
  
  • 
    Functions tzalloc, tzfree,
    localtime_rz, and mktime_z for
    more-efficient thread-safe applications that need to use multiple
    timezones.
    The tzalloc and tzfree functions
    allocate and free objects of type timezone_t,
    and localtime_rz and mktime_z are
    like localtime_r and mktime with an
    extra timezone_t argument.
    The functions were inspired by NetBSD.
  
    • 
  
  • 
    Negative time_t values are supported, on systems
    where time_t is signed.
  
    • 
  
  • 
    These functions can account for leap seconds;
    see Leap seconds below.
  
    • 




POSIX features no longer needed



POSIX and ISO C
define some APIs that are vestigial:
they are not needed, and are relics of a too-simple model that does
not suffice to handle many real-world timestamps.
Although the tz code supports these
vestigial APIs for backwards compatibility, they should
be avoided in portable applications.
The vestigial APIs are:



    
  
  • 
    The POSIX tzname variable does not suffice and is no
    longer needed.
    It is planned to be removed in a future edition of POSIX.
    To get a timestamp’s time zone abbreviation, consult
    the tm_zone member if available; otherwise,
    use strftime’s "%Z" conversion
    specification.
  
    • 
  
  • 
    The POSIX daylight and timezone
    variables do not suffice and are no longer needed.
    They are planned to be removed in a future edition of POSIX.
    To get a timestamp’s UT offset, consult
    the tm_gmtoff member if available; otherwise,
    subtract values returned by localtime
    and gmtime using the rules of the Gregorian calendar,
    or use strftime’s "%z" conversion
    specification if a string like "+0900" suffices.
  
    • 
  
  • 
    The tm_isdst member is almost never needed and most of
    its uses should be discouraged in favor of the
    APIs mentioned above.
    It was intended as an index into the tzname variable,
    but as mentioned previously that usage is obsolete.
    Although it can still be used in arguments to
    mktime to disambiguate timestamps near
    a DST transition when the clock jumps back on
    platforms lacking tm_gmtoff, this
    disambiguation works only for proleptic TZ strings;
    it does not work in general for geographical timezones,
    such as when a location changes to a time zone with a
    lesser UT offset.
  
    • 




Other portability notes


    
  
  • 
    The 7th Edition
    UNIX timezone function is not present in this
    package; it is impossible to reliably map timezone’s
    arguments (a “minutes west of GMT” value and a
    “daylight saving time in effect” flag) to a time zone
    abbreviation, and we refuse to guess.
    Programs that in the past used the timezone function
    may now examine localtime(&clock)->tm_zone
    (if TM_ZONE is defined) or
    use strftime with a %Z conversion specification
    to learn the correct time
    zone abbreviation to use.
  
    • 
  
  • 
    The 4.2BSD
    gettimeofday function is not
    used in this package.
    This formerly let users obtain the current UTC offset
    and DST flag, but this functionality was removed in
    later versions of BSD.
  
    • 
  
  • 
    In SVR2, time conversion fails for near-minimum or
    near-maximum time_t values when doing conversions
    for places that do not use UT.
    This package takes care to do these conversions correctly.
    A comment in the source code tells how to get compatibly wrong
    results.
  
    • 
  
  • 
    The functions that are conditionally compiled
    if STD_INSPIRED is nonzero should, at this point, be
    looked on primarily as food for thought.
    They are not in any sense “standard compatible” – some are
    not, in fact, specified in any standard.
    They do, however, represent responses of various authors to
    standardization proposals.
  
    • 
  
  • 
    Other time conversion proposals, in particular those supported by the
    Time Zone
    Database Parser, offer a wider selection of functions
    that provide capabilities beyond those provided here.
    The absence of such functions from this package is not meant to
    discourage the development, standardization, or use of such
    functions.
    Rather, their absence reflects the decision to make this package
    contain valid extensions to POSIX, to ensure its broad
    acceptability.
    If more powerful time conversion functions can be standardized, so
    much the better.
  
    • 



  
Interface stability



The tz code and data supply the following interfaces:




    
  
  • 
    A set of timezone names as per
      “Timezone identifiers” above.
  
    • 
  
  • 
    Library functions described in “Time and date
      functions” above.
  
    • 
  
  • 
    The programs tzselect, zdump,
    and zic, documented in their man pages.
  
    • 
  
  • 
    The format of zic input files, documented in
    the zic man page.
  
    • 
  
  • 
    The format of zic output files, documented in
    the tzfile man page.
  
    • 
  
  • 
    The format of zone table files, documented in zone1970.tab.
  
    • 
  
  • 
    The format of the country code file, documented in iso3166.tab.
  
    • 
  
  • 
    The version number of the code and data, as the first line of
    the text file "version" in each release.
  
    • 





Interface changes in a release attempt to preserve compatibility with
recent releases.
For example, tz data files typically do not
rely on recently added zic features, so that users can
run older zic versions to process newer data files.
Downloading
the tz database describes how releases
are tagged and distributed.





Interfaces not listed above are less stable.
For example, users should not rely on particular UT
offsets or abbreviations for timestamps, as data entries are often
based on guesswork and these guesses may be corrected or improved.





Timezone boundaries are not part of the stable interface.
For example, even though the Asia/Bangkok timezone
currently includes Chang Mai, Hanoi, and Phnom Penh, this is not part
of the stable interface and the timezone can split at any time.
If a calendar application records a future event in some location other
than Bangkok by putting Asia/Bangkok in the event’s record,
the application should be robust in the presence of timezone splits
between now and the future time.



  
Leap seconds



Leap seconds were introduced in 1972 to accommodate the
difference between atomic time and the less regular rotation of the earth.
Unfortunately they have caused so many problems with civil
timekeeping that there are
plans
to discontinue them by 2035.
Even if these plans come to fruition, a record of leap seconds will still be
needed to resolve timestamps from 1972 through 2035,
and there may also be a need to record whatever mechanism replaces them.





The tz code and data can account for leap seconds,
thanks to code contributed by Bradley White.
However, the leap second support of this package is rarely used directly
because POSIX requires leap seconds to be excluded and many
software packages would mishandle leap seconds if they were present.
Instead, leap seconds are more commonly handled by occasionally adjusting
the operating system kernel clock as described in
Precision timekeeping,
and this package by default installs a leapseconds file
commonly used by
NTP
software that adjusts the kernel clock.
However, kernel-clock twiddling approximates UTC only roughly,
and systems needing more precise UTC can use this package’s leap
second support directly.





The directly supported mechanism assumes that time_t
counts of seconds since the POSIX epoch normally include leap seconds,
as opposed to POSIX time_t counts which exclude leap seconds.
This modified timescale is converted to UTC
at the same point that time zone and DST
adjustments are applied –
namely, at calls to localtime and analogous functions –
and the process is driven by leap second information
stored in alternate versions of the TZif files.
Because a leap second adjustment may be needed even
if no time zone correction is desired,
calls to gmtime-like functions
also need to consult a TZif file,
conventionally named Etc/UTC
(GMT in previous versions),
to see whether leap second corrections are needed.
To convert an application’s time_t timestamps to or from
POSIX time_t timestamps (for use when, say,
embedding or interpreting timestamps in portable
tar
files),
the application can call the utility functions
time2posix and posix2time
included with this package.





If the POSIX-compatible TZif file set is installed
in a directory whose basename is zoneinfo, the
leap-second-aware file set is by default installed in a separate
directory zoneinfo-leaps.
Although each process can have its own time zone by setting
its TZ environment variable, there is no support for some
processes being leap-second aware while other processes are
POSIX-compatible; the leap-second choice is system-wide.
So if you configure your kernel to count leap seconds, you should also
discard zoneinfo and rename zoneinfo-leaps
to zoneinfo.
Alternatively, you can install just one set of TZif files
in the first place; see the REDO variable in this package’s
makefile.



  
Calendrical issues



Calendrical issues are a bit out of scope for a time zone database,
but they indicate the sort of problems that we would run into if we
extended the time zone database further into the past.
An excellent resource in this area is Edward M. Reingold
and Nachum Dershowitz, Calendrical
Calculations: The Ultimate Edition, Cambridge University Press (2018).
Other information and sources are given in the file "calendars"
in the tz distribution.
They sometimes disagree.



  
Time and time zones off Earth



Efforts are underway to establish a reference timescale for the Moon, which has
days roughly equivalent to 29.5 Earth days, and where relativistic
effects cause clocks to tick slightly faster than on Earth.
The European Space Agency began
considering the problem in 2023, and in 2024
NASA
was ordered
to consider the establishment of Coordinated Lunar Time (LTC).
In 2025 a team at China’s
Purple
Mountain Observatory published
Lunar Time Ephemeris
440, the first formal proposal for a standard
for Temps Coordonné Lunaire (TCL).
A draft
resolution of the 13–15 October 2026 General Conference on Weights
and Measures recommends that the Bureau
International des Poids et Mesures collaborate with space agencies
and other organizations on this matter.





Some people’s work schedules have used
Mars time.
Jet Propulsion Laboratory (JPL) coordinators kept Mars time on
and off during the
Mars
Pathfinder mission (1997).
Some of their family members also adapted to Mars time.
Dozens of special Mars watches were built for JPL workers who kept
Mars time during the
Mars
Exploration Rovers (MER) mission (2004–2018).
These timepieces looked like normal Seikos and Citizens but were adjusted
to use Mars seconds rather than terrestrial seconds, although
unfortunately the adjusted watches were unreliable and appear to have
had only limited use.





A Mars solar day is called a “sol” and has a mean period equal to
about 24 hours 39 minutes 35.244 seconds in terrestrial time.
It is divided into a conventional 24-hour clock, so each Mars second
equals about 1.02749125 terrestrial seconds.
(One MER worker noted, “If I am working Mars hours, and Mars hours are
2.5% more than Earth hours, shouldn’t I get an extra 2.5% pay raise?”)





The prime
meridian of Mars goes through the center of the crater
Airy-0, named in
honor of the British astronomer who built the Greenwich telescope that
defines Earth’s prime meridian.
Mean solar time on the Mars prime meridian is
called Mars Coordinated Time (MTC).





Each landed mission on Mars has adopted a different reference for
solar timekeeping, so there is no real standard for Mars time zones.
For example, the MER mission defined two time zones “Local
Solar Time A” and “Local Solar Time B” for its two missions, each zone
designed so that its time equals local true solar time at
approximately the middle of the nominal mission.
The A and B zones differ enough so that an MER worker assigned to
the A zone might suffer “Mars lag” when switching to work in the B zone.
Such a “time zone” is not particularly suited for any application
other than the mission itself.





Many calendars have been proposed for Mars, but none have achieved
wide acceptance.
Astronomers often use Mars Sol Date (MSD) which is a
sequential count of Mars solar days elapsed since about 1873-12-29
12:00 GMT.





In our solar system, Mars is the planet with time and calendar most
like Earth’s.
On other planets, Sun-based time and calendars would work quite
differently.
For example, although Mercury’s
sidereal
rotation period is 58.646 Earth days, Mercury revolves around the
Sun so rapidly that an observer on Mercury’s equator would see a
sunrise only every 175.97 Earth days, i.e., a Mercury year is 0.5 of a
Mercury day.
Venus is more complicated, partly because its rotation is slightly
retrograde:
its year is 1.92 of its days.
Gas giants like Jupiter are trickier still, as their polar and
equatorial regions rotate at different rates, so that the length of a
day depends on latitude.
This effect is most pronounced on Neptune, where the day is about 12
hours at the poles and 18 hours at the equator.





Although the tz database does not support
time on other planets, it is documented here in the hopes that support
will be added eventually.





Sources for time on other planets:













This web page is in the public domain, so clarified as of
2009-05-17 by Arthur David Olson.


Please send corrections to this web page to the
time zone mailing list.
The mailing list and its archives are public,
so please do not send confidential information.




version0000644000000000000000000000000615257052271007416 0ustar002026e
newctime.30000644000000000000000000002060615216527262007717 0ustar00.\" This file is in the public domain, so clarified as of
.\" 2009-05-17 by Arthur David Olson.
.TH newctime 3 "" "Time Zone Database"
.SH NAME
asctime, ctime, difftime, gmtime, localtime, mktime \- convert date and time
.SH SYNOPSIS
.nf
.B "#include "
.PP
.B "struct tm *localtime(time_t const *clock);"
.B "struct tm *localtime_r(time_t const *restrict clock,"
.B "    struct tm *restrict result);"
.B "struct tm *localtime_rz(timezone_t restrict zone,"
.B "    time_t const *restrict clock,"
.B "    struct tm *restrict result);"
.PP
.B "struct tm *gmtime(time_t const *clock);"
.B "struct tm *gmtime_r(time_t const *restrict clock,"
.B "    struct tm *restrict result);"
.PP
.B "time_t mktime(struct tm *tm);"
.B "time_t mktime_z(timezone_t restrict zone,"
.B "    struct tm *restrict tm);"
.PP
.B "double difftime(time_t time1, time_t time0);"
.PP
.B "[[deprecated]] char *asctime(struct tm const *tm);"
.B "[[deprecated]] char *ctime(time_t const *clock);"
.PP
/* Only in POSIX.1-2017 and earlier.  */
.B "char *ctime_r(time_t const *clock, char *buf);"
.B "char *asctime_r(struct tm const *restrict tm,"
.B "    char *restrict result);"
.PP
.BR cc " ... " \-ltz
.fi
.SH DESCRIPTION
.ie '\(en'' .ds en \-
.el .ds en \(en
.ie '\(lq'' .ds lq \&"\"
.el .ds lq \(lq\"
.ie '\(rq'' .ds rq \&"\"
.el .ds rq \(rq\"
.de q
\\$3\*(lq\\$1\*(rq\\$2
..
The
.B localtime
and
.B gmtime
functions
convert an integer, pointed to by
.IR clock ,
and
return pointers to
.q "tm"
structures, described below.
If the integer is out of range for conversion,
these functions return a null pointer.
The
.B localtime
function
corrects for the time zone and any time zone adjustments
(such as Daylight Saving Time in the United States).
The
.B gmtime
function converts to Coordinated Universal Time.
.PP
The
.BI * clock
timestamp represents the time in seconds since 1970-01-01 00:00:00
Coordinated Universal Time (UTC).
The POSIX standard says that timestamps must be nonnegative
and must ignore leap seconds.
Many implementations extend POSIX by allowing negative timestamps,
and can therefore represent timestamps that predate the
introduction of UTC and are some other flavor of Universal Time (UT).
Some implementations support leap seconds, in contradiction to POSIX.
.PP
The
.B mktime
function
converts the broken-down time,
expressed as local time,
in the structure pointed to by
.I tm
into a calendar time value with the same encoding as that of the values
returned by the
.B time
function.
The original values of the
.B tm_wday
and
.B tm_yday
components of the structure are ignored,
and the original values of the other components are not restricted
to their normal ranges.
(A positive or zero value for
.B tm_isdst
causes
.B mktime
to presume initially that daylight saving time
respectively,
is or is not in effect for the specified time.
A negative value for
.B tm_isdst
causes the
.B mktime
function to attempt to divine whether daylight saving time is in effect
for the specified time; in this case it does not use a consistent
rule and may give a different answer when later
presented with the same argument.)
On successful completion, the values of the
.B tm_wday
and
.B tm_yday
components of the structure are set appropriately,
and the other components are set to represent the specified calendar time,
but with their values forced to their normal ranges; the final value of
.B tm_mday
is not set until
.B tm_mon
and
.B tm_year
are determined.
The
.B mktime
function
returns the specified calendar time.
If the calendar time cannot be represented,
it returns \-1 without updating the structure.
To distinguish failure from a valid \-1 return,
you can set
.B tm_wday
or
.B tm_yday
to a negative value before calling
.BR mktime ;
if that value is still negative when
.B mktime
returns, the calendar time could not be represented.
.PP
The
.B difftime
function
returns the difference between two calendar times,
.RI ( time1
\-
.IR time0 ),
expressed in seconds.
.PP
The
.B asctime
function
converts a time value contained in a
.q "tm"
structure to a pointer to a
string of the form
.br
.ce
.eo
Thu Nov 24 18:22:48 1986\n\0
.br
.ec
Years requiring fewer than four characters are padded with leading zeroes.
For years longer than four characters, the string is of the form
.br
.ce
.eo
Thu Nov 24 18:22:48     81986\n\0
.ec
.br
with five spaces before the year.
This unusual format is designed to make it less likely that older
software that expects exactly 26 bytes of output will mistakenly output
misleading values for out-of-range years.
This function is deprecated starting in C23.
Callers can use
.B strftime
instead.
.PP
The
.B ctime
function is equivalent to calling
.B localtime
and then calling
.B asctime
on the result.
Like
.BR asctime ,
this function is deprecated starting in C23.
Callers can use
.B localtime
and
.B strftime
instead.
.PP
The
.BR ctime_r ,
.BR localtime_r ,
.BR gmtime_r ,
and
.B asctime_r
functions
are like their unsuffixed counterparts, except that they accept an
additional argument specifying where to store the result if successful.
The
.B ctime_r
and
.B asctime_r
functions are present only on systems supporting POSIX.1-2017 and earlier,
as they are removed in POSIX.1-2024 and user code can define these
functions with other meanings.
.PP
The
.B localtime_rz
and
.B mktime_z
functions
are like their unsuffixed counterparts, except that they accept an
extra initial
.B zone
argument specifying the timezone to be used for conversion.
If
.B zone
is null, UT is used; otherwise,
.B zone
should be have been allocated by
.B tzalloc
and should not be freed until after all uses (e.g., by calls to
.BR strftime )
of the filled-in
.B tm_zone
fields.
.PP
Declarations of all the functions and externals, and the
.q "tm"
structure,
are in the
.B 
header file.
The structure (of type)
.B "struct tm"
includes the following fields:
.RS
.PP
.nf
.ta 2n +\w'long tm_gmtoff;nn'u
	int tm_sec;	/* seconds (0\*(en60) */
	int tm_min;	/* minutes (0\*(en59) */
	int tm_hour;	/* hours (0\*(en23) */
	int tm_mday;	/* day of month (1\*(en31) */
	int tm_mon;	/* month of year (0\*(en11) */
	int tm_year;	/* year \- 1900 */
	int tm_wday;	/* day of week (Sunday = 0) */
	int tm_yday;	/* day of year (0\*(en365) */
	int tm_isdst;	/* is daylight saving time in effect? */
	char *tm_zone;	/* time zone abbreviation (optional) */
	long tm_gmtoff;	/* offset from UT in seconds (optional) */
.fi
.RE
.PP
The
.B tm_isdst
field
is non-zero if daylight saving time is in effect.
.PP
The
.B tm_gmtoff
field
is the offset (in seconds) of the time represented
from UT, with positive values indicating east
of the Prime Meridian.
The field's name is derived from Greenwich Mean Time, a precursor of UT.
.PP
In platforms conforming to POSIX.1-2024 the
.B "struct tm"
the
.B tm_zone
and
.B tm_gmtoff
fields exist, and are filled in.
For
.B localtime_rz
and
.B mktime_rz
the storage lifetime of the strings addressed by
.B tm_zone
extends until the corresponding
.B timezone_t
object is freed via
.BR tzfree .
For the other functions the lifetime extends until the
.I TZ
environment variable changes state and
.B tzset
is then called.
.PP
As a side effect, the
.BR ctime ,
.B localtime
and
.B mktime
functions also behave as if
.B tzset
were called.
The
.B ctime_r
and
.B localtime_r
functions might (or might not) also behave this way.
This is for compatibility with older platforms, as required by POSIX.
.SH FILES
.ta \w'/usr/share/zoneinfo/GMT\0\0'u
/etc/localtime	local timezone file
.br
/usr/share/zoneinfo	timezone directory
.br
/usr/share/zoneinfo/GMT	for UTC leap seconds
.PP
If /usr/share/zoneinfo/GMT is absent,
UTC leap seconds are loaded from /usr/share/zoneinfo/GMT0 if present.
.SH SEE ALSO
.BR getenv (3),
.BR newstrftime (3),
.BR newtzset (3),
.BR time (2),
.BR tzfile (5).
.SH NOTES
The return values of
.BR asctime ,
.BR ctime ,
.BR gmtime ,
and
.B localtime
point to static data
overwritten by each call.
The remaining functions and data are thread-safe.
.PP
The
.BR asctime ,
.BR asctime_r ,
.BR ctime ,
and
.B ctime_r
functions
behave strangely for years before 1000 or after 9999.
The 1989 and 1999 editions of the C Standard say
that years from \-99 through 999 are converted without
extra spaces, but this conflicts with longstanding
tradition and with this implementation.
The 2011 edition says that the behavior
is undefined if the year is before 1000 or after 9999.
Traditional implementations of these two functions are
restricted to years in the range 1900 through 2099.
To avoid this portability mess, new programs should use
.B strftime
instead.
newstrftime.30000644000000000000000000002506415216527262010456 0ustar00.\" strftime man page
.\"
.\" Based on the UCB file whose corrected copyright information appears below.
.\" Copyright 1989, 1991 The Regents of the University of California.
.\" All rights reserved.
.\"
.\" This code is derived from software contributed to Berkeley by
.\" the American National Standards Committee X3, on Information
.\" Processing Systems.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\"    notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\"    notice, this list of conditions and the following disclaimer in the
.\"    documentation and/or other materials provided with the distribution.
.\" 3. Neither the name of the University nor the names of its contributors
.\"    may be used to endorse or promote products derived from this software
.\"    without specific prior written permission.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS "AS IS" AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\"     from: @(#)strftime.3	5.12 (Berkeley) 6/29/91
.\"	$Id: strftime.3,v 1.4 1993/12/15 20:33:00 jtc Exp $
.\"
.TH newstrftime 3 "" "Time Zone Database"
.SH NAME
strftime \- format date and time
.SH SYNOPSIS
.nf
.B "#include "
.PP
.B "size_t strftime(char *restrict buf, size_t maxsize,"
.B "    char const *restrict format, struct tm const *restrict timeptr);"
.PP
.BR cc " ... " \-ltz
.fi
.SH DESCRIPTION
.ie '\(lq'' .ds lq \&"\"
.el .ds lq \(lq\"
.ie '\(rq'' .ds rq \&"\"
.el .ds rq \(rq\"
.de c
.ie \n(.g \f(CR\\$1\fP\\$2
.el \\$1\\$2
..
.de q
\\$3\*(lq\\$1\*(rq\\$2
..
The
.B strftime
function formats the information from
.BI * timeptr
into the array pointed to by
.I buf
according to the string pointed to by
.IR format .
.PP
The
.I format
string consists of zero or more conversion specifications and
ordinary characters.
All ordinary characters are copied directly into the array.
A conversion specification consists of a percent sign
.q %
and one other character.
.PP
No more than
.I maxsize
bytes are placed into the array.
.PP
Each conversion specification is replaced by the characters as
follows which are then copied into the array.
The characters depend on the values of zero or more members of
.BI * timeptr
as specified by brackets in the description.
If a bracketed member name is followed by
.q + ,
.B strftime
can use the named member even though POSIX.1-2024 does not list it;
if the name is followed by
.q \- ,
.B strftime
ignores the member even though POSIX.1-2024 lists it
which means portable code should set it.
For portability,
.BI * timeptr
should be initialized as if by a successful call to
.BR gmtime ,
.BR localtime ,
.BR mktime ,
.BR timegm ,
or similar functions.
.TP
%A
is replaced by the locale's full weekday name.
.RI [ tm_wday ]
.TP
%a
is replaced by the locale's abbreviated weekday name.
.RI [ tm_wday ]
.TP
%B
is replaced by the locale's full month name.
.RI [ tm_mon ]
.TP
%b or %h
is replaced by the locale's abbreviated month name.
.RI [ tm_mon ]
.TP
%C
is replaced by the century (a year divided by 100 and truncated to an integer)
as a decimal number, with at least two digits by default.
.RI [ tm_year ]
.TP
%c
is replaced by the locale's appropriate date and time representation.
.RI [ tm_year ,
.IR tm_yday ,
.IR tm_mon ,
.IR tm_mday ,
.IR tm_wday ,
.IR tm_hour ,
.IR tm_min ,
.IR tm_sec ,
.IR tm_gmtoff ,
.IR tm_zone ,
.IR tm_isdst \-].
.TP
%D
is equivalent to
.c %m/%d/%y .
Although used in the United States for current dates,
this format is ambiguous elsewhere
and for dates that might involve other centuries.
.RI [ tm_year ,
.IR tm_mon ,
.IR tm_mday ]
.TP
%d
is replaced by the day of the month as a decimal number [01,31].
.RI [ tm_mday ]
.TP
%e
is replaced by the day of month as a decimal number [1,31];
single digits are preceded by a blank.
.RI [ tm_mday ]
.TP
%F
is equivalent to
.c %Y-%m-%d
(the ISO 8601 date format).
.RI [ tm_year ,
.IR tm_mon ,
.IR tm_mday ]
.TP
%G
is replaced by the ISO 8601 year with century as a decimal number.
This is the year that includes the greater part of the week.
(Monday as the first day of a week).
See also the
.c %V
conversion specification.
.RI [ tm_year ,
.IR tm_yday ,
.IR tm_wday ]
.TP
%g
is replaced by the ISO 8601 year without century as a decimal number [00,99].
Since it omits the century, it is ambiguous for dates.
.RI [ tm_year ,
.IR tm_yday ,
.IR tm_wday ]
.TP
%H
is replaced by the hour (24-hour clock) as a decimal number [00,23].
.RI [ tm_hour ]
.TP
%I
is replaced by the hour (12-hour clock) as a decimal number [01,12].
.RI [ tm_hour ]
.TP
%j
is replaced by the day of the year as a decimal number [001,366].
.RI [ tm_yday ]
.TP
%k
is replaced by the hour (24-hour clock) as a decimal number [0,23];
single digits are preceded by a blank.
.RI [ tm_hour ]
.TP
%l
is replaced by the hour (12-hour clock) as a decimal number [1,12];
single digits are preceded by a blank.
.RI [ tm_hour ]
.TP
%M
is replaced by the minute as a decimal number [00,59].
.RI [ tm_min ]
.TP
%m
is replaced by the month as a decimal number [01,12].
.RI [ tm_mon ]
.TP
%n
is replaced by a newline.
.TP
%p
is replaced by the locale's equivalent of either
.q AM
or
.q PM .
.RI [ tm_hour ]
.TP
%R
is replaced by the time in the format
.c %H:%M .
.RI [ tm_hour ,
.IR tm_min ]
.TP
%r
is replaced by the locale's representation of 12-hour clock time
using AM/PM notation.
.RI [ tm_hour ,
.IR tm_min ,
.IR tm_sec ]
.TP
%S
is replaced by the second as a decimal number [00,60].
The range of
seconds is [00,60] instead of [00,59] to allow for the periodic occurrence
of leap seconds.
.RI [ tm_sec ]
.TP
%s
is replaced by the number of seconds since the Epoch (see
.BR ctime (3)).
Although %s is reliable in this implementation,
it can have glitches on other platforms
(notably obsolescent platforms lacking
.I tm_gmtoff
or where
.B time_t
is no wider than int), and POSIX allows
.B strftime
to set
.B errno
to
.B EINVAL
or
.B EOVERFLOW
and return 0 if the number of seconds would be negative or out of range for
.BR time_t .
Portable code should therefore format a
.B time_t
value directly via something like
.B sprintf
instead of via
.B localtime
followed by
.B strftime
with "%s".
.RI [ tm_year ,
.IR tm_mon ,
.IR tm_mday ,
.IR tm_hour ,
.IR tm_min ,
.IR tm_sec ,
.IR tm_gmtoff +,
.IR tm_isdst \-].
.TP
%T
is replaced by the time in the format
.c %H:%M:%S .
.RI [ tm_hour ,
.IR tm_min ,
.IR tm_sec ]
.TP
%t
is replaced by a tab.
.TP
%U
is replaced by the week number of the year (Sunday as the first day of
the week) as a decimal number [00,53].
.RI [ tm_wday ,
.IR tm_yday ,
.IR tm_year \-]
.TP
%u
is replaced by the weekday (Monday as the first day of the week)
as a decimal number [1,7].
.RI [ tm_wday ]
.TP
%V
is replaced by the week number of the year (Monday as the first day of
the week) as a decimal number [01,53].  If the week containing January
1 has four or more days in the new year, then it is week 1; otherwise
it is week 53 of the previous year, and the next week is week 1.
The year is given by the
.c %G
conversion specification.
.RI [ tm_year ,
.IR tm_yday ,
.IR tm_wday ]
.TP
%W
is replaced by the week number of the year (Monday as the first day of
the week) as a decimal number [00,53].
.RI [ tm_yday ,
.IR tm_wday ]
.TP
%w
is replaced by the weekday (Sunday as the first day of the week)
as a decimal number [0,6].
.RI [ tm_year ,
.IR tm_yday ,
.IR tm_wday ]
.TP
%X
is replaced by the locale's appropriate time representation.
.RI [ tm_year \-,
.IR tm_yday \-,
.IR tm_mon \-,
.IR tm_mday \-,
.IR tm_wday \-,
.IR tm_hour ,
.IR tm_min ,
.IR tm_sec ,
.IR tm_gmtoff ,
.IR tm_zone ,
.IR tm_isdst \-].
.TP
%x
is replaced by the locale's appropriate date representation.
This format can be ambiguous for dates, e.g.,
it can generate "01/02/03" in the C locale.
.RI [ tm_year ,
.IR tm_yday ,
.IR tm_mon ,
.IR tm_mday ,
.IR tm_wday ,
.IR tm_hour \-,
.IR tm_min \-,
.IR tm_sec \-,
.IR tm_gmtoff \-,
.IR tm_zone \-,
.IR tm_isdst \-].
.TP
%Y
is replaced by the year with century as a decimal number.
.RI [ tm_year ]
.TP
%y
is replaced by the year without century as a decimal number [00,99].
Since it omits the century, it is ambiguous for dates.
.RI [ tm_year ]
.TP
%Z
is replaced by the time zone abbreviation,
or by the empty string if this is not determinable.
.RI [ tm_zone ,
.IR tm_isdst \-]
.TP
%z
is replaced by the offset from the Prime Meridian
in the format +HHMM or \-HHMM (ISO 8601) as appropriate,
with positive values representing locations east of Greenwich,
or by the empty string if this is not determinable.
The numeric time zone abbreviation \-0000 is used when the time is
Universal Time
but local time is indeterminate; by convention this is used for
locations while uninhabited, and corresponds to a zero offset when the
time zone abbreviation begins with
.q "\-" .
.RI [ tm_gmtoff ,
.IR tm_zone +,
.IR tm_isdst \-]
.TP
%%
is replaced by a single %.
.TP
%+
is replaced by the locale's date and time in
.BR date (1)
format.
.RI [ tm_year ,
.IR tm_yday ,
.IR tm_mon ,
.IR tm_mday ,
.IR tm_wday ,
.IR tm_hour ,
.IR tm_min ,
.IR tm_sec ,
.IR tm_gmtoff ,
.IR tm_zone ]
.PP
As a side effect,
.B strftime
also behaves as if
.B tzset
were called.
This is for compatibility with older platforms, as required by POSIX;
it is not needed for
.BR strftime 's
own use.
.SH "RETURN VALUE"
If the conversion is successful,
.B strftime
returns the number of bytes placed into the array, not counting the
terminating NUL;
.B errno
is unchanged if the returned value is zero.
Otherwise,
.B errno
is set to indicate the error, zero is returned,
and the array contents are unspecified.
.SH ERRORS
This function fails if:
.TP
[ERANGE]
The total number of resulting bytes, including the terminating
NUL character, is more than
.IR maxsize .
.SH SEE ALSO
.BR date (1),
.BR getenv (3),
.BR newctime (3),
.BR newtzset (3),
.BR time (2),
.BR tzfile (5).
.SH BUGS
There is no conversion specification for the phase of the moon.
newtzset.30000644000000000000000000002436015216527262007770 0ustar00.\" This file is in the public domain, so clarified as of
.\" 2009-05-17 by Arthur David Olson.
.TH newtzset 3 "" "Time Zone Database"
.SH NAME
tzset \- initialize time conversion information
.SH SYNOPSIS
.nf
.B "#include "
.PP
.BI "timezone_t tzalloc(char const *" TZ );
.PP
.BI "void tzfree(timezone_t " tz );
.PP
.B "void tzset(void);"
.PP
/* Optional and obsolescent:  */
.br
.B "extern char *tzname[];"
.br
.B "extern long timezone;"
.br
.B "extern int daylight;"
.PP
.BR cc " ... " \-ltz
.fi
.SH DESCRIPTION
.ie '\(en'' .ds en \-
.el .ds en \(en
.ie '\(lq'' .ds lq \&"\"
.el .ds lq \(lq\"
.ie '\(rq'' .ds rq \&"\"
.el .ds rq \(rq\"
.de q
\\$3\*(lq\\$1\*(rq\\$2
..
The
.B tzalloc
function
allocates and returns a timezone object described by
.IR TZ .
.PP
If
.I TZ
is a null pointer,
.B tzalloc
uses the best available approximation to local (wall
clock) time, as specified by the
.BR tzfile (5)-format
file
.B localtime
in the system time conversion information directory.
.PP
If
.I TZ
is the empty string,
.B tzalloc
uses Universal Time (UT), with the abbreviation "UTC"
and without leap second correction; please see
.BR newctime (3)
for more about UT, UTC, and leap seconds.
.PP
If
.I TZ
is nonnull and nonempty:
.IP
if the value begins with a colon, it is used as a pathname of a file
from which to read the time conversion information;
.IP
if the value does not begin with a colon, it is first used as the
pathname of a file from which to read the time conversion information,
and, if that file cannot be read, is used directly as a specification of
the time conversion information.
.PP
When
.I TZ
contents are used as a pathname, a pathname beginning with
.q "/"
is used as-is; otherwise
the pathname is relative to a system time conversion information
directory.
In a privileged program the pathname must be relative.
Relative pathnames must not contain
.q "..\&"
components.
For the purpose of these checks, a file name beginning with
.q "/"
is considered to be relative if it is the
.B localtime
file's name, or if it starts with the
system timezone directory's name followed by one more more slashes.
The file must be a regular file in the format specified in
.BR tzfile (5).
.PP
When
.I TZ
is used directly as a specification of the time conversion information,
it must have the following syntax:
.IP
\fIstd\|offset\fR[\fIdst\fR[\fIoffset\fR][\fB,\fIrule\fR]]
.PP
Where:
.RS
.TP
.IR std " and " dst
Three or more bytes that are the designation for the standard
.RI ( std )
or the alternative
.RI ( dst ,
such as daylight saving time)
time zone.  Only
.I std
is required; if
.I dst
is missing, then daylight saving time does not apply in this locale.
Upper- and lowercase letters are explicitly allowed.  Any characters
except a leading colon
.RB ( : ),
digits, comma
.RB ( , ),
ASCII minus
.RB ( \- ),
ASCII plus
.RB ( + ),
and NUL bytes are allowed.
Alternatively, a designation can be surrounded by angle brackets
.B <
and
.BR > ;
in this case, the designation can contain any characters other than
.B >
and NUL.
.TP
.I offset
Indicates the value one must add to the local time to arrive at
Coordinated Universal Time.  The
.I offset
has the form:
.RS
.IP
\fIhh\fR[\fB:\fImm\fR[\fB:\fIss\fR]]
.RE
.IP
The minutes
.RI ( mm )
and seconds
.RI ( ss )
are optional.  The hour
.RI ( hh )
is required and may be a single digit.  The
.I offset
following
.I std
is required.  If no
.I offset
follows
.IR dst ,
daylight saving time is assumed to be one hour ahead of standard time.  One or
more digits may be used; the value is always interpreted as a decimal
number.  The hour must be between zero and 24, and the minutes (and
seconds) \*(en if present \*(en between zero and 59.  If preceded by a
.q "\-" ,
the time zone shall be east of the Prime Meridian; otherwise it shall be
west (which may be indicated by an optional preceding
.q "+" ).
.TP
.I rule
Indicates when to change to and back from daylight saving time.  The
.I rule
has the form:
.RS
.IP
\fIdate\fB/\fItime\fB,\fIdate\fB/\fItime\fR
.RE
.IP
where the first
.I date
describes when the change from standard to daylight saving time occurs and the
second
.I date
describes when the change back happens.  Each
.I time
field describes when, in current local time, the change to the other
time is made.
Daylight saving is assumed to be in effect
all year if it begins January 1 at 00:00 and ends December 31 at
24:00 plus the difference between daylight saving and standard time,
leaving no room for standard time in the calendar.
.IP
The format of
.I date
is one of the following:
.RS
.TP
.BI J n
The Julian day
.I n
.RI "(1\ \(<=" "\ n\ " "\(<=\ 365).
Leap days are not counted; that is, in all years \*(en including leap
years \*(en February 28 is day 59 and March 1 is day 60.  It is
impossible to explicitly refer to the occasional February 29.
.TP
.I n
The zero-based Julian day
.RI "(0\ \(<=" "\ n\ " "\(<=\ 365).
Leap days are counted, and it is possible to refer to February 29.
.TP
.BI M m . n . d
The
.IR d' th
day
.RI "(0\ \(<=" "\ d\ " "\(<=\ 6)
of week
.I n
of month
.I m
of the year
.RI "(1\ \(<=" "\ n\ " "\(<=\ 5,
.RI "1\ \(<=" "\ m\ " "\(<=\ 12,
where week 5 means
.q "the last \fId\fP day in month \fIm\fP"
which may occur in either the fourth or the fifth week).  Week 1 is the
first week in which the
.IR d' th
day occurs.  Day zero is Sunday.
.RE
.IP
The
.I time
has the same format as
.I offset
except that the hours part of
.I time
can range from \-167 through 167; this allows for unusual rules such
as
.q "the Saturday before the first Sunday of March" .
The default, if
.I time
is not given, is
.BR 02:00:00 .
.RE
.LP
Here are some examples of
.I TZ
values that directly specify the timezone.
.TP
.B EST5
stands for US Eastern Standard
Time (EST), 5 hours behind UT, without daylight saving.
.TP
.B <+12>\-12<+13>,M11.1.0,M1.2.1/147
stands for Fiji time, 12 hours ahead
of UT, springing forward on November's first Sunday at 02:00, and
falling back on January's second Monday at 147:00 (i.e., 03:00 on the
first Sunday on or after January 14).  The abbreviations for standard
and daylight saving time are
.q "+12"
and
.q "+13".
.TP
.B IST\-2IDT,M3.4.4/26,M10.5.0
stands for Israel Standard Time (IST) and Israel Daylight Time (IDT),
2 hours ahead of UT, springing forward on March's fourth
Thursday at 26:00 (i.e., 02:00 on the first Friday on or after March
23), and falling back on October's last Sunday at 02:00.
.TP
.B <\-04>4<\-03>,J1/0,J365/25
stands for permanent daylight saving time, 3 hours behind UT with
abbreviation
.q "\-03".
There is a dummy fall-back transition on December 31 at 25:00 daylight
saving time (i.e., 24:00 standard time, equivalent to January 1 at
00:00 standard time), and a simultaneous spring-forward transition on
January 1 at 00:00 standard time, so daylight saving time is in effect
all year and the initial
.B <\-04>
is a placeholder.
.TP
.B <\-03>3<\-02>,M3.5.0/\-2,M10.5.0/\-1
stands for time in western Greenland, 3 hours behind UT, where clocks
follow the EU rule of
springing forward on March's last Sunday at 01:00 UT (\-02:00 local
time, i.e., 22:00 the previous day) and falling back on October's last
Sunday at 01:00 UT (\-01:00 local time, i.e., 23:00 the previous day).
The abbreviations for standard and daylight saving time are
.q "\-03"
and
.q "\-02".
.PP
If
.I TZ
specifies daylight saving time but does not specify a
.IR rule ,
the rule typically defaults to the current US daylight-saving rule,
although such a default is not guaranteed and
is incorrect for many locations outside the US.
Therefore, if a
.I TZ
string directly specifies a timezone with daylight saving time,
it should specify the daylight saving rule explicitly.
.PP
For compatibility with System V Release 3.1, a semicolon
.RB ( ; )
may be used to separate the
.I rule
from the rest of the specification;
this is an extension to POSIX.
.PP
The
.B tzfree
function
frees a timezone object
.IR tz ,
which should have been successfully allocated by
.BR tzalloc .
This invalidates any
.B tm_zone
pointers that
.I tz
was used to set.
.PP
The
.B tzset
function
acts like
.BR tzalloc(getenv("TZ")) ,
except it saves any resulting timezone object into internal
storage that is accessed by
.BR localtime ,
.BR localtime_r ,
and
.BR mktime .
The anonymous shared timezone object is freed by the next call to
.BR tzset .
If the implied call to
.B getenv
fails,
.B tzset
acts like
.BR tzalloc(nullptr) ;
if the implied call to
.B tzalloc
fails,
.B tzset
falls back on UT.
.PP
As a side effect, the
.B tzset
function sets some external variables if the platform defines them.
It sets
.BR tzname [0]
and
.BR tzname [1]
to pointers to strings that are time zone abbreviations to be used with
standard and daylight saving time, respectively.
It also sets
.B timezone
to be the number of seconds that standard time is west of the Prime Meridian,
and
.B daylight
to be zero if daylight saving time is never in effect, non-zero otherwise.
.SH "RETURN VALUE"
If successful, the
.B tzalloc
function returns a nonnull pointer to the newly allocated object.
Otherwise, it returns a null pointer and sets
.IR errno .
The
.B tzfree
function does not modify
.IR errno .
.SH ERRORS
.TP
.B EOVERFLOW
.I TZ
directly specifies time conversion information,
and contains an integer out of machine range
or a time zone abbreviation that is too long for this platform.
.PP
The
.B tzalloc
function may also fail and set
.I errno
for any of the errors specified for the routines
.BR access (2),
.BR close (2),
.BR malloc (3),
.BR open (2),
and
.BR read (2).
.SH FILES
.ta \w'/usr/share/zoneinfo/GMT\0\0'u
/etc/localtime	local timezone file
.br
/usr/share/zoneinfo	timezone directory
.br
/usr/share/zoneinfo/GMT	for UTC leap seconds
.PP
If /usr/share/zoneinfo/GMT is absent,
UTC leap seconds are loaded from /usr/share/zoneinfo/GMT0 if present.
.SH SEE ALSO
.BR getenv (3),
.BR newctime (3),
.BR newstrftime (3),
.BR time (2),
.BR tzfile (5).
.SH NOTES
Portable code should not rely on the contents of the external variables
.BR tzname ,
.B timezone
and
.B daylight
as their contents are unspecified (and do not make sense in general)
when a geographical TZ is used.
In multithreaded applications behavior is undefined if one thread accesses
one of these variables while another thread invokes
.BR tzset .
A future version of POSIX is planned to remove these variables;
callers can instead use the
.I tm_gmtoff
and
.I tm_zone
members of
.BR "struct tm",
or use
.B strftime
with "%z" or "%Z".
time2posix.30000644000000000000000000000772315216527262010214 0ustar00.\" This file is in the public domain, so clarified as of
.\" 1996-06-05 by Arthur David Olson.
.TH time2posix 3 "" "Time Zone Database"
.SH NAME
time2posix, posix2time \- convert seconds since the Epoch
.SH SYNOPSIS
.nf
.B "#include "
.PP
.B "time_t time2posix(time_t t);"
.PP
.B "time_t posix2time(time_t t);"
.PP
.BR cc " ... " \-ltz
.fi
.SH DESCRIPTION
.ie '\(en'' .ds en \-
.el .ds en \(en
.ie '\(lq'' .ds lq \&"\"
.el .ds lq \(lq\"
.ie '\(rq'' .ds rq \&"\"
.el .ds rq \(rq\"
.de q
\\$3\*(lq\\$1\*(rq\\$2
..
IEEE Standard 1003.1
(POSIX)
says that
.B time_t
values cannot count leap
seconds and,
therefore,
that the system time must be adjusted as each leap occurs.
.PP
If the time package is configured with leap-second support
enabled,
however,
no such adjustment is needed and
.B time_t
values continue to increase over leap events
(as a true
.q "seconds since...\&"
value).
This means that these values will differ from those required by POSIX
by the net number of leap seconds inserted since the Epoch.
.PP
For many C programs this is not a problem as the C standard says that
.B time_t
is
(mostly)
opaque \*(en
.B time_t
values should be obtained from and
passed to functions such as
.BR time(2) ,
.BR localtime(3) ,
.BR mktime(3) ,
and
.BR difftime(3) .
However,
POSIX gives an arithmetic
expression for computing a
.B time_t
value directly from a given Universal date and time,
and the same relationship is assumed by some
applications.
Any programs creating/dissecting
.B time_t
values
using such a relationship will typically not handle intervals
over leap seconds correctly.
.PP
The
.B time2posix
and
.B posix2time
functions address this mismatch by converting
between local
.B time_t
values and their POSIX equivalents.
This is done by accounting for the number of time-base changes that
would have taken place on a POSIX system as leap seconds were inserted
or deleted.
These converted values can then be used
when communicating with POSIX-compliant systems.
.PP
The
.B time2posix
function converts a
.B time_t
value to its POSIX counterpart, if one exists.
The
.B posix2time
function does the reverse but
is less well-behaved:
for a positive leap second hit the result is not unique,
and for a negative leap second hit the corresponding
POSIX
.B time_t
doesn't exist so an adjacent value is returned.
Both of these are indicate problems with the
POSIX representation.
.PP
The following table summarizes the relationship between a time
T and its conversion to,
and back from,
the POSIX representation over the leap second inserted at the end of June,
1993.
In this table, X=time2posix(T), Y=posix2time(X), A=741484816, and B=A\-17
because 17 positive leap seconds preceded this leap second.
.PP
.in +2
.nf
.ta \w'1993-06-30\0'u +\w'23:59:59\0'u +\w'A+0\0'u +\w'B+0\0'u
DATE	TIME	T	X	Y
1993-06-30	23:59:59	A	B	A
1993-06-30	23:59:60	A+1	B+1	A+1 or A+2
1993-07-01	00:00:00	A+2	B+1	A+1 or A+2
1993-07-01	00:00:01	A+3	B+2	A+3
.in
.fi
.PP
A leap second deletion would look like the following, and
posix2time(B+1) would return either A or A+1.
.PP
.in +2
.nf
DATE	TIME	T	X	Y
????-06-30	23:59:58	A	B	A
????-07-01	00:00:00	A+1	B+2	A+1
????-07-01	00:00:01	A+2	B+3	A+2
.fi
.in
.PP
If leap-second support is not enabled,
local
.B time_t
and
POSIX
.B time_t
values are equivalent,
and both
.B time2posix
and
.B posix2time
degenerate to the identity function.
.SH "RETURN VALUE"
If successful, these functions return the resulting timestamp without modifying
.BR errno .
Otherwise, they set
.B errno
and return
.BR "((time_t) -1)" .
.SH ERRORS
These functions fail if:
.TP
[EOVERFLOW]
The resulting value cannot be represented.
This can happen for
.B posix2time
if its argument is close to the maximum
.B time_t
value.
In environments where the
.I TZ
environment variable names a TZif file,
overflow can happen for either function for an argument sufficiently
close to an extreme
.B time_t
value if the TZif file specifies unrealistic leap second corrections.
.SH SEE ALSO
.BR difftime (3),
.BR localtime (3),
.BR mktime (3),
.BR time (2).
tzfile.50000644000000000000000000004765115132622175007410 0ustar00.\" This file is in the public domain, so clarified as of
.\" 1996-06-05 by Arthur David Olson.
.TH tzfile 5 "" "Time Zone Database"
.SH NAME
tzfile \- timezone information
.SH DESCRIPTION
.ie '\(lq'' .ds lq \&"\"
.el .ds lq \(lq\"
.ie '\(rq'' .ds rq \&"\"
.el .ds rq \(rq\"
.de q
\\$3\*(lq\\$1\*(rq\\$2
..
The timezone information files used by
.BR tzset (3)
are typically found under a directory with a name like
.IR /usr/share/zoneinfo .
These files use the format described in Internet RFC 9636.
Each file is a sequence of 8-bit bytes.
In a file, a binary integer is represented by a sequence of one or
more bytes in network order (bigendian, or high-order byte first),
with all bits significant,
a signed binary integer is represented using two's complement,
and a boolean is represented by a one-byte binary integer that is
either 0 (false) or 1 (true).
The format begins with a 44-byte header containing the following fields:
.RS "\w'  'u"
.IP \(bu "\w'\(bu  'u"
The magic four-byte ASCII sequence
.q "TZif"
identifies the file as a timezone information file.
.IP \(bu
A byte identifying the version of the file's format
(as of 2021, either an ASCII NUL,
.q "2",
.q "3",
or
.q "4" ).
.IP \(bu
Fifteen bytes containing zeros reserved for future use.
.IP \(bu
Six four-byte integer values, in the following order:
.RS "\w'\(bu    'u"
.TP "\w'  'u"
.B tzh_ttisutcnt
The number of UT/local indicators stored in the file.
(UT is Universal Time.)
.TP
.B tzh_ttisstdcnt
The number of standard/wall indicators stored in the file.
.TP
.B tzh_leapcnt
The number of leap seconds for which data entries are stored in the file.
.TP
.B tzh_timecnt
The number of transition times for which data entries are stored
in the file.
.TP
.B tzh_typecnt
The number of local time types for which data entries are stored
in the file (must not be zero).
.TP
.B tzh_charcnt
The number of bytes of time zone abbreviation strings
stored in the file.
.RE
.RE
.PP
The above header is followed by the following fields, whose lengths
depend on the contents of the header:
.RS "\w'  'u"
.IP \(bu "\w'\(bu  'u"
.B tzh_timecnt
four-byte signed integer values sorted in ascending order.
These values are written in network byte order.
Each is used as a transition time (as returned by
.BR time (2))
at which the rules for computing local time change.
.IP \(bu
.B tzh_timecnt
one-byte unsigned integer values;
each one but the last tells which of the different types of local time types
described in the file is associated with the time period
starting with the same-indexed transition time
and continuing up to but not including the next transition time.
(The last time type is present only for consistency checking with the
proleptic TZ string described below.)
These values serve as indices into the next field.
.IP \(bu
.B tzh_typecnt
.B ttinfo
entries, each defined as follows:
.in +2
.sp
.nf
.ta \w'\0\0\0\0'u +\w'unsigned char\0'u
struct ttinfo {
	int32_t	tt_utoff;
	unsigned char	tt_isdst;
	unsigned char	tt_desigidx;
};
.in
.fi
.sp
Each structure is written as a four-byte signed integer value for
.BR tt_utoff ,
in network byte order, followed by a one-byte boolean for
.B tt_isdst
and a one-byte value for
.BR tt_desigidx .
In each structure,
.B tt_utoff
gives the number of seconds to be added to UT,
.B tt_isdst
tells whether
.B tm_isdst
should be set by
.BR localtime (3)
and
.B tt_desigidx
serves as an index into the array of time zone abbreviation bytes
that follow the
.B ttinfo
entries in the file; if the designated string is "\-00", the
.B ttinfo
entry is a placeholder indicating that local time is unspecified.
The
.B tt_utoff
value is never equal to \-2**31, to let 32-bit clients negate it without
overflow.
Also, in realistic applications
.B tt_utoff
is in the range [\-89999, 93599] (i.e., more than \-25 hours and less
than 26 hours); this allows easy support by implementations that
already support the POSIX-required range [\-24:59:59, 25:59:59].
.IP \(bu
.B tzh_charcnt
bytes that represent time zone designations,
which are null-terminated byte strings, each indexed by the
.B tt_desigidx
values mentioned above, and each corresponding to a time zone abbreviation.
The byte strings can overlap if one is a suffix of the other.
The encoding of these strings is not specified.
.IP \(bu
.B tzh_leapcnt
pairs of four-byte values, written in network byte order;
the first value of each pair gives the non-negative time
(as returned by
.BR time (2))
at which a leap second occurs or at which the leap second table expires;
the second is a signed integer specifying the correction, which is the
.I total
number of leap seconds to be applied during the time period
starting at the given time.
The pairs of values are sorted in strictly ascending order by time.
Each pair denotes one leap second, either positive or negative,
except that if the last pair has the same correction as the previous one,
the last pair denotes the leap second table's expiration time.
Each leap second is at the end of a UTC calendar month.
The first leap second has a non-negative occurrence time,
and is a positive leap second if and only if its correction is positive;
the correction for each leap second after the first differs
from the previous leap second by either 1 for a positive leap second,
or \-1 for a negative leap second.
If the leap second table is empty, the leap-second correction is zero
for all timestamps;
otherwise, for timestamps before the first occurrence time,
the leap-second correction is zero if the first pair's correction is 1 or \-1,
and is unspecified otherwise (which can happen only in files
truncated at the start).
.IP \(bu
.B tzh_ttisstdcnt
standard/wall indicators, each stored as a one-byte boolean;
they tell whether the transition times associated with local time types
were specified as standard time or local (wall clock) time.
.IP \(bu
.B tzh_ttisutcnt
UT/local indicators, each stored as a one-byte boolean;
they tell whether the transition times associated with local time types
were specified as UT or local time.
If a UT/local indicator is set, the corresponding standard/wall indicator
must also be set.
.RE
.PP
The standard/wall and UT/local indicators were designed for
transforming a TZif file's transition times into transitions appropriate
for another time zone specified via
a proleptic TZ string that lacks rules.
For example, when TZ="EET\-2EEST" and there is no TZif file "EET\-2EEST",
the idea was to adapt the transition times from a TZif file with the
well-known name "posixrules" that was present only for this purpose and
was a copy of the file "Europe/Brussels", a file with a different UT offset.
POSIX does not specify the details of this obsolete transformational behavior,
the default rules are installation-dependent, and no implementation
is known to support this feature for timestamps past 2037,
so users desiring (say) Greek time should instead specify
TZ="Europe/Athens" for better historical coverage, falling back on
TZ="EET\-2EEST,M3.5.0/3,M10.5.0/4"
if conformance to POSIX.1-2017 or earlier is required
and older timestamps need not be handled accurately.
.PP
The
.BR localtime (3)
function
normally uses the first
.B ttinfo
structure in the file
if either
.B tzh_timecnt
is zero or the time argument is less than the first transition time recorded
in the file.
.SS Version 2 format
For version-2-format timezone files,
the above header and data are followed by a second header and data,
identical in format except that
eight bytes are used for each transition time or leap second time.
(Leap second counts remain four bytes.)
After the second header and data comes a newline-enclosed string
in the style of the contents of a proleptic TZ,
for use in handling instants
after the last transition time stored in the file
or for all instants if the file has no transitions.
The TZ string is empty (i.e., nothing between the newlines)
if there is no proleptic representation for such instants.
.PP
If non-empty, the TZ string must agree with the local time
type after the last transition time if present in the eight-byte data;
for example, given the string
.q "WET0WEST,M3.5.0/1,M10.5.0"
then if a last transition time is in July, the transition's local time
type must specify a daylight-saving time abbreviated
.q "WEST"
that is one hour east of UT.
.PP
The TZ string can contain time zone abbreviations and UT offsets that
do not appear elsewhere in the TZif file.
.PP
Also, if there is at least one transition, time type 0 is associated
with the time period from the indefinite past up to but not including
the earliest transition time.
.SS Version 3 format
For version-3-format timezone files, a TZ string (see
.BR newtzset (3))
may use the following POSIX.1-2024 extensions to POSIX.1-2017:
First, as in TZ="<\-02>2<\-01>,M3.5.0/\-1,M10.5.0/0",
the hours part of its transition times may be signed and range from
\-167 through 167 instead of being limited to unsigned values
from 0 through 24.
Second, as in TZ="XXX3EDT4,0/0,J365/23", DST is in effect all year if it starts
January 1 at 00:00 and ends December 31 at 24:00 plus the difference
between daylight saving and standard time.
.SS Version 4 format
For version-4-format TZif files,
the first leap second record can have a correction that is neither
+1 nor \-1, to represent truncation of the TZif file at the start.
Also, if two or more leap second transitions are present and the last
entry's correction equals the previous one, the last entry
denotes the expiration of the leap second table instead of a leap second;
timestamps after this expiration are unreliable in that future
releases will likely add leap second entries after the expiration, and
the added leap seconds will change how post-expiration timestamps are treated.
.SS Interoperability considerations
Future changes to the format may append more data.
.PP
Version 1 files are considered a legacy format and
should not be generated, as they do not support transition
times after the year 2038.
Readers that understand only Version 1 must ignore
any data that extends beyond the calculated end of the version
1 data block.
.PP
Other than version 1, writers should generate
the lowest version number needed by a file's data.
For example, a writer should generate a version 4 file
only if its leap second table either expires or is truncated at the start.
Likewise, a writer not generating a version 4 file
should generate a version 3 file only if
TZ string extensions are necessary to accurately
model transition times.
.PP
The sequence of time changes defined by the version 1
header and data block should be a contiguous sub-sequence
of the time changes defined by the version 2+ header and data
block, and by the footer.
This guideline helps obsolescent version 1 readers
agree with current readers about timestamps within the
contiguous sub-sequence.  It also lets writers not
supporting obsolescent readers use a
.B tzh_timecnt
of zero
in the version 1 data block to save space.
.PP
When a TZif file contains a leap second table expiration
time, TZif readers should either refuse to process
post-expiration timestamps, or process them as if the expiration
time did not exist (possibly with an error indication).
.PP
Time zone abbreviations should consist of at least three (3)
and no more than six (6) ASCII characters from the set of
alphanumerics,
.q "\-",
and
.q "+".
This is for compatibility with POSIX requirements for
time zone abbreviations.
.PP
A numeric time zone abbreviation should match the UT offset.
For example, "+0530" should be used only if the UT offset is 5.5 hours
ahead of UT, and "\-00" should be used only if the UT offset is zero.
.PP
When reading a version 2 or higher file, readers
should ignore the version 1 header and data block except for
the purpose of skipping over them.
.PP
Readers should calculate the total lengths of the
headers and data blocks and check that they all fit within
the actual file size, as part of a validity check for the file.
.PP
When a positive leap second occurs, readers should append an extra
second to the local minute containing the second just before the leap
second.  If this occurs when the UTC offset is not a multiple of 60
seconds, the leap second occurs earlier than the last second of the
local minute and the minute's remaining local seconds are numbered
through 60 instead of the usual 59; the UTC offset is unaffected.
.SS Common interoperability issues
This section documents common problems in reading or writing TZif files.
Most of these are problems in generating TZif files for use by
older readers.
The goals of this section are to help:
.RS "\w'  'u"
.IP \(bu "\w'\(bu  'u"
TZif writers output files that avoid common
pitfalls in older or buggy TZif readers,
.IP \(bu
TZif readers avoid common pitfalls when reading
files generated by future TZif writers, and
.IP \(bu
any future specification authors see what sort of
problems arise when the TZif format is changed.
.RE
.PP
When new versions of the TZif format have been defined, a
design goal has been that a reader can successfully use a TZif
file even if the file is of a later TZif version than what the
reader was designed for.
When complete compatibility was not achieved, an attempt was
made to limit glitches to rarely used timestamps and allow
simple partial workarounds in writers designed to generate
newer-version data useful even for older-version readers.
This section attempts to document these compatibility issues and
workarounds as well as documenting other common bugs in
readers.
.PP
Interoperability problems with TZif include the following:
.RS "\w'  'u"
.IP \(bu "\w'\(bu  'u"
Some readers examine only version 1 data.
As a partial workaround, a writer can output as much version 1
data as possible.
However, a reader should ignore version 1 data, and should use
version 2+ data even if the reader's native timestamps have only
32 bits.
.IP \(bu
Some readers designed for version 2 might mishandle
timestamps after a version 3 or higher file's last transition, because
they cannot parse the POSIX.1-2024 extensions to POSIX.1-2017
in the proleptic TZ string.
As a partial workaround, a writer can output more transitions
than necessary, so that only far-future timestamps are
mishandled by version 2 readers.
.IP \(bu
Some readers might mishandle timestamps after a file's last transition,
because they require that all abbreviations or UT offsets
in the proleptic TZ string must also occur somewhere in the file's tables
of time zone designations and local time type records.
As a workaround, a writer can output more transitions than necessary,
so that the other tables contain duplicates of the proleptic TZ string's
abbreviations and offsets.
.IP \(bu
Some readers designed for version 2 do not support
permanent daylight saving time with transitions after 24:00
\(en e.g., a TZ string
.q "EST5EDT,0/0,J365/25"
denoting permanent Eastern Daylight Time
(\-04).
As a workaround, a writer can substitute standard time
for two time zones east, e.g.,
.q "XXX3EDT4,0/0,J365/23"
for a time zone with a never-used standard time (XXX, \-03)
and negative daylight saving time (EDT, \-04) all year.
Alternatively,
as a partial workaround, a writer can substitute standard time
for the next time zone east \(en e.g.,
.q "AST4"
for permanent
Atlantic Standard Time (\-04).
.IP \(bu
Some readers designed for version 2 or 3 and that require strict
conformance to RFC 9636 reject version 4 files whose leap second
tables are truncated at the start or end in expiration times.
.IP \(bu
Some readers ignore the footer, and instead predict future
timestamps from the time type of the last transition.
As a partial workaround, a writer can output more transitions
than necessary.
.IP \(bu
Some stripped-down readers ignore everything but the footer,
and use its proleptic TZ string to calculate all timestamps.
Although this approach often works for current and future timestamps,
it obviously has problems with past timestamps,
and even for current timestamps it can fail for settings like
TZ="Africa/Casablanca".  This corresponds to a TZif file
containing explicit transitions through the year 2087,
followed by a footer containing the TZ string
.q <+01>\-1 ,
which should be used only for timestamps after the last
explicit transition.
.IP \(bu
Some readers do not use time type 0 for timestamps before
the first transition, in that they infer a time type using a
heuristic that does not always select time type 0.
As a partial workaround, a writer can output a dummy (no-op)
first transition at an early time.
.IP \(bu
Some readers mishandle timestamps before the first
transition that has a timestamp that is not less than \-2**31.
Readers that support only 32-bit timestamps are likely to be
more prone to this problem, for example, when they process
64-bit transitions only some of which are representable in 32
bits.
As a partial workaround, a writer can output a dummy
transition at timestamp \-2**31.
.IP \(bu
Some readers mishandle a transition if its timestamp has
the minimum possible signed 64-bit value.
Timestamps less than \-2**59 are not recommended.
.IP \(bu
Some readers mishandle proleptic TZ strings that
contain
.q "<"
or
.q ">".
As a partial workaround, a writer can avoid using
.q "<"
or
.q ">"
for time zone abbreviations containing only alphabetic
characters.
.IP \(bu
Many readers mishandle time zone abbreviations that contain
non-ASCII characters.
These characters are not recommended.
.IP \(bu
Some readers may mishandle time zone abbreviations that
contain fewer than 3 or more than 6 characters or that
contain ASCII characters other than alphanumerics,
.q "\-",
and
.q "+".
These abbreviations are not recommended.
.IP \(bu
Some readers mishandle TZif files that specify
daylight-saving time UT offsets that are less than the UT
offsets for the corresponding standard time.
These readers do not support locations like Ireland, which
uses the equivalent of the TZ string
.q "IST\-1GMT0,M10.5.0,M3.5.0/1",
observing standard time
(IST, +01) in summer and daylight saving time (GMT, +00) in winter.
As a partial workaround, a writer can output data for the
equivalent of the TZ string
.q "GMT0IST,M3.5.0/1,M10.5.0",
thus swapping standard and daylight saving time.
Although this workaround misidentifies which part of the year
uses daylight saving time, it records UT offsets and time zone
abbreviations correctly.
.IP \(bu
Some readers generate ambiguous timestamps for positive leap seconds
that occur when the UTC offset is not a multiple of 60 seconds.
For example, with UTC offset +01:23:45 and
a positive leap second 78796801 (1972-06-30 23:59:60 UTC), some readers will
map both 78796800 and 78796801 to 01:23:45 local time the next day
instead of mapping the latter to 01:23:46, and they will map 78796815 to
01:23:59 instead of to 01:23:60.
This has not yet been a practical problem, since no civil authority
has observed such UTC offsets since leap seconds were
introduced in 1972.
.RE
.PP
Some interoperability problems are reader bugs that
are listed here mostly as warnings to developers of readers.
.RS "\w'  'u"
.IP \(bu "\w'\(bu  'u"
Some readers do not support negative timestamps.
Developers of distributed applications should keep this
in mind if they need to deal with pre-1970 data.
.IP \(bu
Some readers mishandle timestamps before the first
transition that has a non-negative timestamp.
Readers that do not support negative timestamps are likely to
be more prone to this problem.
.IP \(bu
Some readers mishandle time zone abbreviations like
.q "\-08"
that contain
.q "+",
.q "\-",
or digits.
.IP \(bu
Some readers mishandle UT offsets that are out of the
traditional range of \-12 through +12 hours, and so do not
support locations like Kiritimati that are outside this
range.
.IP \(bu
Some readers mishandle UT offsets in the range [\-3599, \-1]
seconds from UT because they integer-divide the offset by
3600 to get 0 and then display the hour part as
.q "+00".
.IP \(bu
Some readers mishandle UT offsets that are not a multiple
of one hour, or of 15 minutes, or of 1 minute.
.RE
.SH SEE ALSO
.BR time (2),
.BR localtime (3),
.BR tzset (3),
.BR tzselect (8),
.BR zdump (8),
.BR zic (8).
.PP
Olson A, Eggert P, Murchison K. The Time Zone Information Format (TZif).
October 2024.
.UR https://\:www.rfc-editor.org/\:rfc/\:rfc9636
Internet RFC 9636
.UE
.UR https://\:doi.org/\:10.17487/\:RFC9636
doi:10.17487/RFC9636
.UE .
tzselect.80000644000000000000000000000573414737431435007756 0ustar00.\" This file is in the public domain, so clarified as of
.\" 2009-05-17 by Arthur David Olson.
.TH tzselect 8 "" "Time Zone Database"
.SH NAME
tzselect \- select a timezone
.SH SYNOPSIS
.ds d " degrees
.ds m " minutes
.ds s " seconds
.ds _ " \&
.if t \{\
. if \n(.g .if c \(de .if c \(fm .if c \(sd \{\
.  ds d \(de
.  ds m \(fm
.  ds s \(sd
.  ds _ \|
. \}
.\}
.B tzselect
[
.B \-c
.I coord
] [
.B \-n
.I limit
] [
.B \-\-help
] [
.B \-\-version
]
.SH DESCRIPTION
The
.B tzselect
program asks the user for information about the current location,
and outputs the resulting timezone to standard output.
The output is suitable as a value for the TZ environment variable.
.PP
All interaction with the user is done via standard input and standard error.
.SH OPTIONS
.TP
.BI "\-c " coord
Instead of asking for continent and then country and then city,
ask for selection from time zones whose largest cities
are closest to the location with geographical coordinates
.I coord.
Use ISO 6709 notation for
.I coord,
that is, a latitude immediately followed by a longitude.  The latitude
and longitude should be signed integers followed by an optional
decimal point and fraction: positive numbers represent north and east,
negative south and west.  Latitudes with two and longitudes with three
integer digits are treated as degrees; latitudes with four or six and
longitudes with five or seven integer digits are treated as
.I "DDMM, DDDMM, DDMMSS,"
or
.I DDDMMSS
representing
.I DD
or
.I DDD
degrees,
.I MM
minutes,
and zero or
.I SS
seconds, with any trailing fractions represent fractional minutes or
(if
.I SS
is present) seconds.  The decimal point is that of the current locale.
For example, in the (default) C locale,
.B "\-c\ +40.689\-074.045"
specifies 40.689\*d\*_N, 74.045\*d\*_W,
.B "\-c\ +4041.4\-07402.7"
specifies 40\*d\*_41.4\*m\*_N, 74\*d\*_2.7\*m\*_W, and
.B "\-c\ +404121\-0740240"
specifies 40\*d\*_41\*m\*_21\*s\*_N, 74\*d\*_2\*m\*_40\*s\*_W.
If
.I coord
is not one of the documented forms, the resulting behavior is unspecified.
.TP
.BI "\-n " limit
When
.B \-c
is used, display the closest
.I limit
locations (default 10).
.TP
.B "\-\-help"
Output help information and exit.
.TP
.B "\-\-version"
Output version information and exit.
.SH "ENVIRONMENT VARIABLES"
.TP
\f3AWK\fP
Name of a POSIX-compliant
.B awk
program (default:
.BR awk ).
.TP
\f3TZDIR\fP
Name of the directory containing timezone data files (default:
.BR /usr/share/zoneinfo ).
.SH FILES
.TP
\f2TZDIR\fP\f3/iso3166.tab\fP
Table of ISO 3166 2-letter country codes and country names.
.TP
\f2TZDIR\fP\f3/zone1970.tab\fP
Table of country codes, latitude and longitude, timezones, and
descriptive comments.
.TP
\f2TZDIR\fP\f3/\fP\f2TZ\fP
Timezone data file for timezone \f2TZ\fP.
.SH "EXIT STATUS"
The exit status is zero if a timezone was successfully obtained from the user,
nonzero otherwise.
.SH "SEE ALSO"
newctime(3), tzfile(5), zdump(8), zic(8)
.SH NOTES
Applications should not assume that
.BR tzselect 's
output matches the user's political preferences.
zic.80000644000000000000000000006406715246336075006712 0ustar00.\" This file is in the public domain, so clarified as of
.\" 2009-05-17 by Arthur David Olson.
.TH zic 8 "" "Time Zone Database"
.SH NAME
zic \- timezone compiler
.SH SYNOPSIS
.B zic
[
.I option
\&... ] [
.I filename
\&... ]
.SH DESCRIPTION
.ie '\(lq'' .ds lq \&"\"
.el .ds lq \(lq\"
.ie '\(rq'' .ds rq \&"\"
.el .ds rq \(rq\"
.de q
\\$3\*(lq\\$1\*(rq\\$2
..
.ie '\(la'' .ds < <
.el .ds < \(la
.ie '\(ra'' .ds > >
.el .ds > \(ra
.ds d " degrees
.ds m " minutes
.ds s " seconds
.ds _ " \&
.if t \{\
. if \n(.g .if c \(de .if c \(fm .if c \(sd \{\
.  ds d \(de
.  ds m \(fm
.  ds s \(sd
.  ds _ \|
. \}
.\}
The
.B zic
program reads text from the file(s) named on the command line
and creates the timezone information format (TZif) files
specified in this input.
If a
.I filename
is
.q "\-" ,
standard input is read.
.SH OPTIONS
.TP
.B "\-\-version"
Output version information and exit.
.TP
.B \-\-help
Output short usage message and exit.
.TP
.BI "\-b " bloat
Output backward-compatibility data as specified by
.IR bloat .
If
.I bloat
is
.BR fat ,
generate additional data entries that work around potential bugs or
incompatibilities in older software, such as software that mishandles
a TZif file's 64-bit data or proleptic TZ string.
If
.I bloat
is
.BR slim ,
keep the output files small; this can help check for the bugs
and incompatibilities.
The default is
.BR slim ,
as the
.B fat
workarounds are typically good only until the year 2038 anyway.
Also see the
.B \-r
option for another way to alter output size.
.TP
.BI \-D
Do not create ancestor directories of output files,
For example, for a zone named America/Los_Angeles
the directory America should already exist.
By default, the directory and its ancestors are created
if they do not already exist.
.TP
.BI "\-d " directory
Create time conversion information files in the named directory rather than
in the standard directory named below.
.TP
.BI "\-l " timezone
Use
.I timezone
as local time.
.B zic
will act as if the input contained a link line of the form
.sp
.ti +2
.ta \w'Link\0\0'u  +\w'\fItimezone\fP\0\0'u
Link	\fItimezone\fP		localtime
.sp
If
.I timezone
is
.BR \- ,
any already-existing link is removed.
.TP
.BI "\-L " leapsecondfilename
Read leap second information from the file with the given name.
If this option is not used,
no leap second information appears in output files;
this is required by some TZif readers.
.TP
.BI "\-p " timezone
Act as if the input contained a link line of the form
.sp
.ti +2
Link	\fItimezone\fP		posixrules
.sp
If
.I timezone
is
.q "\-"
(the default), any already-existing link is removed.
.sp
Unless
.I timezone
is
.q "\-" ,
this option is obsolete and is no longer supported by most runtimes.
Among other things it should not be used for timestamps after the year 2037,
and it should not be combined with
.B "\-b slim"
if
.IR timezone 's
transitions are at standard time or Universal Time (UT) instead of local time.
The option is present only to support obsolete runtimes that used
.IR timezone 's
rules when handling obsolescent
TZ strings like "AST4ADT" that lack transition rules;
modern runtimes that support these TZ strings
typically just use current US rules
as the TZ strings were mainly used in the US.
Similarly, any Zone or Link named "posixrules" is obsolete and problematic.
.TP
.BI "\-m " mode
Create TZif files with the given file mode bits.
By default the files are created with mode 644 as modified by the umask.
With this option they are created with the given mode instead.
For portability the mode should be an unsigned octal integer,
typically 644 or 444;
some platforms also support
.BR chmod (1)-style
symbolic modes.
This option does not affect created ancestor directories,
which have mode 755 as modified by the umask.
The option is ignored on platforms lacking the notion of file mode bits.
.TP
.BR "\-r " "[\fB@\fP\fIlo\fP][\fB/@\fP\fIhi\fP]"
Limit the applicability of output files
to timestamps in the range from
.I lo
(inclusive) to
.I hi
(exclusive), where
.I lo
and
.I hi
are possibly signed decimal counts of seconds since the Epoch
(1970-01-01 00:00:00 UTC).
Omitted counts default to extreme values.
The output files use UT offset 0 and abbreviation
.q "\-00"
in place of the omitted timestamp data.
For example,
.q "zic \-r @0"
omits data intended for negative timestamps (i.e., before the Epoch), and
.q "zic \-r @0/@2147483648"
outputs data intended only for nonnegative timestamps that fit into
31-bit signed integers.
On platforms with GNU
.BR date ,
.q "zic \-r @$(date +%s)"
omits data intended for past timestamps.
Although this option typically reduces the output file's size,
the size can increase due to the need to represent the timestamp range
boundaries, particularly if
.I hi
causes a TZif file to contain explicit entries for
.RI pre- hi
transitions rather than concisely representing them
with a proleptic TZ string.
Also see the
.B "\-b slim"
option for another way to shrink output size.
.TP
.BI "\-R @" hi
Generate redundant trailing explicit transitions for timestamps
that occur less than
.I hi
seconds since the Epoch, even though the transitions could be
more concisely represented via the proleptic TZ string.
This option does not affect the represented timestamps.
Although it accommodates nonstandard TZif readers
that ignore the proleptic TZ string,
it increases the size of the altered output files.
.TP
.BI "\-t " file
When creating local time information, put the configuration link in
the named file rather than in the standard location.
.TP
.BI "\-u " owner\fR[:\fPgroup\fR]\fP
Change the output regular files' owner and group to those specified.
The
.I owner
is either a user name, or an unsigned decimal integer user ID,
or an empty string meaning no change to the owner.
The
.I group
is similar for group names and IDs.
This option does not affect directories or hard or symbolic links.
It typically needs special privileges to change ownership,
and is ignored on platforms that lack the notions of owners and groups.
.TP
.B \-v
Be more verbose, and complain about the following situations:
.RS
.PP
The input specifies a link to a link,
something not supported by some older parsers, including
.B zic
itself through release 2022e.
.PP
A year that appears in a data file is outside the range
of representable years.
.PP
A time of 24:00 or more appears in the input.
Pre-1998 versions of
.B zic
prohibit 24:00, and pre-2007 versions prohibit times greater than 24:00.
.PP
A rule goes past the start or end of the month.
Pre-2004 versions of
.B zic
prohibit this.
.PP
A time zone abbreviation uses a
.B %z
format.
Pre-2015 versions of
.B zic
do not support this.
.PP
A timestamp contains fractional seconds.
Pre-2018 versions of
.B zic
do not support this.
.PP
The input contains abbreviations that are mishandled by pre-2018 versions of
.B zic
due to a longstanding coding bug.
These abbreviations include
.q L
for
.q Link ,
.q mi
for
.q min ,
.q Sa
for
.q Sat ,
and
.q Su
for
.q Sun .
.PP
The output file does not contain all the information about the
long-term future of a timezone, because the future cannot be summarized as
a proleptic TZ string.  For example, as of 2026 this problem
occurs for Palestine's daylight-saving rules, as these rules are based
on predictions for when Ramadan will be observed, something that
a proleptic TZ string cannot represent.
.PP
The output contains data that may not be handled properly by client
code designed for older
.B zic
output formats.  These compatibility issues affect only timestamps
before 1970 or after the start of 2038.
.PP
The output contains a truncated leap second table,
which can cause some older TZif readers to misbehave.
This can occur if the
.B "\-L"
option is used, and either an Expires line is present or
the
.B "\-r"
option is also used.
.PP
The output file contains more than 1200 transitions,
which may be mishandled by some clients.
The current reference client supports at most 2000 transitions;
pre-2014 versions of the reference client support at most 1200
transitions.
.PP
A time zone abbreviation has fewer than 3 or more than 6 characters.
POSIX requires at least 3, and requires implementations to support
at least 6.
.PP
An output file name contains a byte that is not an ASCII letter,
.q "\-" ,
.q "/" ,
or
.q "_" ;
or it contains a file name component that contains more than 14 bytes
or that starts with
.q "\-" .
.RE
.SH FILES
Input files use the format described in this section; output files use
.BR tzfile (5)
format.
.PP
Input files should be text files, that is, they should be a series of
zero or more lines, each ending in a newline byte and containing at
most 2048 bytes counting the newline, and without any NUL bytes.
The input text's encoding
is typically UTF-8 or ASCII; it should have a unibyte representation
for the
.UR https://\:pubs\:.opengroup\:.org/\:onlinepubs/\:9799919799/\:basedefs/\:V1_chap06\:.html
POSIX Portable Character Set (PPCS)
.UE
and the encoding's non-unibyte characters should consist entirely of
non-PPCS bytes.
Non-PPCS characters typically occur only in comments.
.PP
Input lines are made up of fields.
Fields are separated from one another by one or more white space characters.
The white space characters are space, form feed, carriage return, newline,
tab, and vertical tab.
Leading and trailing white space on input lines is ignored.
An unquoted sharp character (#) in the input introduces a comment which extends
to the end of the line the sharp character appears on.
White space characters and sharp characters may be enclosed in double quotes
(") if they're to be used as part of a field.
Any line that is blank (after comment stripping) is ignored.
Nonblank lines are expected to be of one of three types:
rule lines, zone lines, and link lines.
.PP
Keywords are in English and are case insensitive.
They are not reserved words.
They appear in several contexts, and include month and weekday names
and other words such as
.BR "maximum" ,
.BR "only" ,
.BR "Rolling" ,
and
.BR "Zone" .
A keyword can be abbreviated by omitting all but an initial prefix; any
abbreviation must be unambiguous in context.
.PP
A rule line has the form
.nf
.ti +2
.ta \w'Rule\0\0'u +\w'NAME\0\0'u +\w'FROM\0\0'u +\w'1973\0\0'u +\w'\-\0\0'u +\w'Apr\0\0'u +\w'lastSun\0\0'u +\w'2:00w\0\0'u +\w'1:00d\0\0'u
.sp
Rule	NAME	FROM	TO	\-	IN	ON	AT	SAVE	LETTER/S
.sp
For example:
.ti +2
.sp
Rule	US	1967	1973	\-	Apr	lastSun	2:00w	1:00d	D
.sp
.fi
The fields that make up a rule line are:
.TP
.B NAME
Gives the name of the rule set that contains this line.
The name must start with a character that is neither
an ASCII digit nor
.q \-
nor
.q + .
To allow for future extensions,
an unquoted name should not contain characters from the set
.ie \n(.g .q \f(CR!$%&\(aq()*,/:;<=>?@[\e]\(ha\(ga{|}\(ti\fP .
.el .ie t .q \f(CW!$%&'()*,/:;<=>?@[\e]^\(ga{|}~\fP .
.el .q !$%&'()*,/:;<=>?@[\e]^`{|}~ .
.TP
.B FROM
Gives the first year in which the rule applies.
Any signed integer year can be supplied; the proleptic Gregorian calendar
is assumed, with year 0 preceding year 1.
Rules can describe times that are not representable as time values,
with the unrepresentable times ignored; this allows rules to be portable
among hosts with differing time value types.
.TP
.B TO
Gives the final year in which the rule applies.
The word
.B maximum
(or an abbreviation) means the indefinite future, and the word
.B only
(or an abbreviation)
may be used to repeat the value of the
.B FROM
field.
.TP
.B \-
Is a reserved field and should always contain
.q \-
for compatibility with older versions of
.BR zic .
It was previously known as the
.B TYPE
field, which could contain values to allow a
separate script to further restrict in which
.q types
of years the rule would apply.
.TP
.B IN
Names the month in which the rule takes effect.
Month names may be abbreviated as mentioned previously;
for example, January can appear as
.q January ,
.q JANU
or
.q Ja ,
but not as
.q j
which would be ambiguous with both June and July.
.TP
.B ON
Gives the day on which the rule takes effect.
Recognized forms include:
.nf
.in +2
.sp
.ta \w'Sun<=25\0\0'u
5	the fifth of the month
lastSun	the last Sunday in the month
lastMon	the last Monday in the month
Sun>=8	first Sunday on or after the eighth
Sun<=25	last Sunday on or before the 25th
.fi
.in
.sp
A weekday name (e.g.,
.BR "Sunday" )
or a weekday name preceded by
.q "last"
(e.g.,
.BR "lastSunday" )
may be abbreviated as mentioned previously,
e.g.,
.q Su
for Sunday and
.q lastsa
for the last Saturday.
There must be no white space characters within the
.B ON
field.
The
.q <=
and
.q >=
constructs can result in a day in the neighboring month;
for example, the IN-ON combination
.q "Oct Sun>=31"
stands for the first Sunday on or after October 31,
even if that Sunday occurs in November.
.TP
.B AT
Gives the time of day at which the rule takes effect,
relative to 00:00, the start of a calendar day.
Recognized forms include:
.nf
.in +2
.sp
.ta \w'00:19:32.13\0\0'u
2	time in hours
2:00	time in hours and minutes
01:28:14	time in hours, minutes, and seconds
00:19:32.13	time with fractional seconds
12:00	midday, 12 hours after 00:00
15:00	3 PM, 15 hours after 00:00
24:00	end of day, 24 hours after 00:00
260:00	260 hours after 00:00
\-2:30	2.5 hours before 00:00
\-	equivalent to 0
.fi
.in
.sp
Although
.B zic
rounds times to the nearest integer second
(breaking ties to the even integer), the fractions may be useful
to other applications requiring greater precision.
The source format does not specify any maximum precision.
Any of these forms may be followed by the letter
.B w
if the given time is local or
.q "wall clock"
time,
.B s
if the given time is standard time without any adjustment for daylight saving,
or
.B u
(or
.B g
or
.BR z )
if the given time is universal time;
in the absence of an indicator,
local (wall clock) time is assumed.
These forms ignore leap seconds; for example,
if a leap second occurs at 00:59:60 local time,
.q "1:00"
stands for 3601 seconds after local midnight instead of the usual 3600 seconds.
The intent is that a rule line describes the instants when a
clock/calendar set to the type of time specified in the
.B AT
field would show the specified date and time of day.
.TP
.B SAVE
Gives the amount of time to be added to local standard time when the rule is in
effect, and whether the resulting time is standard or daylight saving.
This field has the same format as the
.B AT
field
except with a different set of suffix letters:
.B s
for standard time and
.B d
for daylight saving time.
The suffix letter is typically omitted, and defaults to
.B s
if the offset is zero and to
.B d
otherwise.
Negative offsets are allowed; in Ireland, for example, daylight saving
time is observed in winter and has a negative offset relative to
Irish Standard Time.
The offset is merely added to standard time; for example,
.B zic
does not distinguish a 10:30 standard time plus an 0:30
.B SAVE
from a 10:00 standard time plus a 1:00
.BR SAVE .
.TP
.B LETTER/S
Gives the
.q "variable part"
(for example, the
.q "S"
or
.q "D"
in
.q "EST"
or
.q "EDT" )
of time zone abbreviations to be used when this rule is in effect.
If this field is
.q \- ,
the variable part is null.
.PP
A zone line has the form
.sp
.nf
.ti +2
.ta \w'Zone\0\0'u +\w'Asia/Amman\0\0'u +\w'STDOFF\0\0'u +\w'Jordan\0\0'u +\w'FORMAT\0\0'u
Zone	NAME	STDOFF	RULES	FORMAT	[UNTIL]
.sp
For example:
.sp
.ti +2
Zone	Asia/Amman	2:00	Jordan	EE%sT	2017 Oct 27 01:00
.sp
.fi
The fields that make up a zone line are:
.TP
.B NAME
The name of the timezone.
This is the name used in creating the time conversion information file for the
timezone.
It should not contain a file name component
that is empty or is
.q ".\&"
or
.q ".." ;
a file name component is a maximal substring that does not contain
.q "/" .
Names can be further restricted by the operating system or file system.
Most systems reject a name
.IB A / B
if there is another name
.IR A ,
some reject names differing only in case,
and some reject names containing special characters like
.q ":"
and
.q "\e" .
Other software will work better if names are
limited to the restricted syntax described under the
.B \-v
option.
.TP
.B STDOFF
The amount of time to add to UT to get standard time,
without any adjustment for daylight saving.
This field has the same format as the
.B AT
and
.B SAVE
fields of rule lines, except without suffix letters;
begin the field with a minus sign if time must be subtracted from UT.
.TP
.B RULES
The name of the rules that apply in the timezone or,
alternatively, a field in the same format as a rule-line
.B SAVE
field,
giving the amount of time to be added to local standard time
and whether the resulting time is standard or daylight saving.
Standard time applies if this field is
.B \-
or for timestamps occurring before any rule takes effect.
When an amount of time is given, only the sum of standard time and
this amount matters.
.TP
.B FORMAT
The format for time zone abbreviations.
The pair of characters
.B %s
shows where to put the time zone abbreviation's variable part,
which is taken from the
.B LETTER/S
field of the corresponding rule;
any timestamps that precede the earliest rule use the
.B LETTER/S
of the earliest standard-time rule (which in this case must exist).
Alternatively, a format can use the pair of characters
.B %z
to stand for the UT offset in the form
.RI \(+- hh ,
.RI \(+- hhmm ,
or
.RI \(+- hhmmss ,
using the shortest form that does not lose information, where
.IR hh ,
.IR mm ,
and
.I ss
are the hours, minutes, and seconds east (+) or west (\-) of UT.
Alternatively,
a slash (/)
separates standard and daylight abbreviations.
To conform to POSIX, a time zone abbreviation should contain only
alphanumeric ASCII characters,
.q "+"
and
.q "\-".
By convention, the time zone abbreviation
.q "\-00"
is a placeholder that means local time is unspecified.
.TP
.B UNTIL
The time at which the UT offset or the rule(s) change for a location.
It takes the form of one to four fields YEAR [MONTH [DAY [TIME]]].
If this is specified,
the time zone information is generated from the given UT offset
and rule change until the time specified, which is interpreted using
the rules in effect just before the transition.
The month, day, and time of day have the same format as the IN, ON, and AT
fields of a rule; trailing fields can be omitted, and default to the
earliest possible value for the missing fields.
.IP
The next line must be a
.q "continuation"
line; this has the same form as a zone line except that the
string
.q "Zone"
and the name are omitted, as the continuation line will
place information starting at the time specified as the
.q "until"
information in the previous line in the file used by the previous line.
Continuation lines may contain
.q "until"
information, just as zone lines do, indicating that the next line is a further
continuation.
.PP
If a zone changes at the same instant that a rule would otherwise take
effect in the earlier zone or continuation line, the rule is ignored.
A zone or continuation line
.I L
with a named rule set starts with standard time by default:
that is, any of
.IR L 's
timestamps preceding
.IR L 's
earliest rule use the rule in effect after
.IR L 's
first transition into standard time.
In a single zone it is an error if two rules take effect at the same
instant, or if two zone changes take effect at the same instant.
.PP
If a continuation line subtracts
.I N
seconds from the UT offset after a transition that would be
interpreted to be later if using the continuation line's UT offset and
rules, the
.q "until"
time of the previous zone or continuation line is interpreted
according to the continuation line's UT offset and rules, and any rule
that would otherwise take effect in the next
.I N
seconds is instead assumed to take effect simultaneously.
For example:
.br
.ne 7
.nf
.in +2
.ta \w'# Rule\0\0'u +\w'NAME\0\0'u +\w'FROM\0\0'u +\w'2006\0\0'u +\w'\-\0\0'u +\w'Oct\0\0'u +\w'lastSun\0\0'u +\w'2:00\0\0'u +\w'SAVE\0\0'u
.sp
# Rule	NAME	FROM	TO	\-	IN	ON	AT	SAVE	LETTER/S
Rule	US	1967	2006	-	Oct	lastSun	2:00	0	S
Rule	US	1967	1973	-	Apr	lastSun	2:00	1:00	D
.ta \w'# Zone\0\0'u +\w'America/Menominee\0\0'u +\w'STDOFF\0\0'u +\w'RULES\0\0'u +\w'FORMAT\0\0'u
# Zone	NAME	STDOFF	RULES	FORMAT	[UNTIL]
Zone	America/Menominee	\-5:00	\-	EST	1973 Apr 29 2:00
		\-6:00	US	C%sT
.sp
.in
.fi
Here, an incorrect reading would be there were two clock changes on 1973-04-29,
the first from 02:00 EST (\-05) to 01:00 CST (\-06) according to the
.q "until"
value in the zone line,
and the second an hour later from 02:00 CST (\-06) to 03:00 CDT (\-05)
according to the values in the April rule line.
However,
.B zic
interprets this more sensibly as a single transition from 02:00 EST (\-05) to
02:00 CDT (\-05).
.PP
A link line has the form
.sp
.nf
.ti +2
.ta \w'Link\0\0'u +\w'Europe/Istanbul\0\0'u
Link	TARGET	LINK-NAME
.sp
.fi
where
.B TARGET
and
.B LINK-NAME
both have the same syntax as a zone line's
.BR NAME .
For example:
.sp
.nf
.ti +2
Link	Europe/Istanbul	Asia/Istanbul
.sp
.fi
The
.B TARGET
field should appear as the
.B NAME
field in some zone line or as the
.B LINK-NAME
field in some link line.
The
.B LINK-NAME
field is an alternative name for that zone.
Links can chain together, although the behavior is unspecified if a
chain of one or more links does not terminate in a Zone name.
A link line can appear before the line that defines the link target.
For example:
.sp
.ne 3
.nf
.in +2
.ta \w'Zone\0\0'u +\w'Greenwich\0\0'u
Link	Greenwich	G_M_T
Link	Etc/GMT	Greenwich
Zone	Etc/GMT\0\00\0\0\-\0\0GMT
.sp
.in
.fi
The two links are chained together, and G_M_T, Greenwich, and Etc/GMT
all name the same zone.
.PP
Except for continuation lines,
lines may appear in any order in the input.
However, the behavior is unspecified if multiple zone or link lines
define the same name.
.PP
The file that describes leap seconds can have leap lines and an
expiration line.
Leap lines have the following form:
.nf
.ti +2
.ta \w'Leap\0\0'u +\w'YEAR\0\0'u +\w'MONTH\0\0'u +\w'DAY\0\0'u +\w'HH:MM:SS\0\0'u +\w'CORR\0\0'u
.sp
Leap	YEAR	MONTH	DAY	HH:MM:SS	CORR	R/S
.sp
For example:
.ti +2
.sp
Leap	2016	Dec	31	23:59:60	+	S
.sp
.fi
The
.BR YEAR ,
.BR MONTH ,
.BR DAY ,
and
.B HH:MM:SS
fields tell when the leap second happened.
The
.B CORR
field
should be
.q "+"
if a second was added
or
.q "\-"
if a second was skipped.
The
.B R/S
field
should be (an abbreviation of)
.q "Stationary"
if the leap second time given by the other fields should be interpreted as UTC
or
(an abbreviation of)
.q "Rolling"
if the leap second time given by the other fields should be interpreted as
local (wall clock) time.
.PP
Rolling leap seconds would let one see
Times Square ball drops where there'd be a
.q "3... 2... 1... leap... Happy New Year"
countdown, placing the leap second at
midnight New York time rather than midnight UTC.
Although stationary leap seconds are the common practice,
rolling leap seconds can be useful in specialized applications
like SMPTE timecodes that may prefer to put leap second
discontinuities at the end of a local broadcast day.
However, rolling leap seconds are not supported if the
.B \-r
option is used.
.PP
The expiration line, if present, has the form:
.nf
.ti +2
.ta \w'Expires\0\0'u +\w'YEAR\0\0'u +\w'MONTH\0\0'u +\w'DAY\0\0'u
.sp
Expires	YEAR	MONTH	DAY	HH:MM:SS
.sp
For example:
.ti +2
.sp
Expires	2020	Dec	28	00:00:00
.sp
.fi
The
.BR YEAR ,
.BR MONTH ,
.BR DAY ,
and
.B HH:MM:SS
fields give the expiration timestamp in UTC for the leap second table.
.br
.ne 22
.SH "EXTENDED EXAMPLE"
Here is an extended example of
.B zic
input, intended to illustrate many of its features.
.nf
.in +2
.ta \w'# Rule\0\0'u +\w'NAME\0\0'u +\w'FROM\0\0'u +\w'1973\0\0'u +\w'\-\0\0'u +\w'Apr\0\0'u +\w'lastSun\0\0'u +\w'2:00\0\0'u +\w'SAVE\0\0'u
.sp
# Rule	NAME	FROM	TO	\-	IN	ON	AT	SAVE	LETTER/S
Rule	Swiss	1941	1942	\-	May	Mon>=1	1:00	1:00	S
Rule	Swiss	1941	1942	\-	Oct	Mon>=1	2:00	0	\-
.sp .5
Rule	EU	1977	1980	\-	Apr	Sun>=1	1:00u	1:00	S
Rule	EU	1977	only	\-	Sep	lastSun	1:00u	0	\-
Rule	EU	1978	only	\-	Oct	 1	1:00u	0	\-
Rule	EU	1979	1995	\-	Sep	lastSun	1:00u	0	\-
Rule	EU	1981	max	\-	Mar	lastSun	1:00u	1:00	S
Rule	EU	1996	max	\-	Oct	lastSun	1:00u	0	\-
.sp
.ta \w'# Zone\0\0'u +\w'Europe/Zurich\0\0'u +\w'0:29:45.50\0\0'u +\w'RULES\0\0'u +\w'FORMAT\0\0'u
# Zone	NAME	STDOFF	RULES	FORMAT	[UNTIL]
Zone	Europe/Zurich	0:34:08	\-	LMT	1853 Jul 16
		0:29:45.50	\-	BMT	1894 Jun
		1:00	Swiss	CE%sT	1981
		1:00	EU	CE%sT
.sp
Link	Europe/Zurich	Europe/Vaduz
.sp
.in
.fi
In this example, the EU rules are for the European Union
and for its predecessor organization, the European Communities.
The timezone is named Europe/Zurich and it has the alias Europe/Vaduz.
This example says that Zurich was 34 minutes and 8
seconds east of UT until 1853-07-16 at 00:00, when the legal offset
was changed to
7\*d\*_26\*m\*_22.50\*s,
which works out to 0:29:45.50;
.B zic
treats this by rounding it to 0:29:46.
After 1894-06-01 at 00:00 the UT offset became one hour
and Swiss daylight saving rules (defined with lines beginning with
.q "Rule Swiss")
apply.  From 1981 to the present, EU daylight saving rules have
applied, and the UTC offset has remained at one hour.
.PP
In 1941 and 1942, daylight saving time applied from the first Monday
in May at 01:00 to the first Monday in October at 02:00.
The pre-1981 EU daylight-saving rules have no effect
here, but are included for completeness.  Since 1981, daylight
saving has begun on the last Sunday in March at 01:00 UTC.
Until 1995 it ended the last Sunday in September at 01:00 UTC,
but this changed to the last Sunday in October starting in 1996.
.PP
For purposes of display,
.q "LMT"
and
.q "BMT"
were initially used, respectively.  Since
Swiss rules and later EU rules were applied, the time zone abbreviation
has been CET for standard time and CEST for daylight saving
time.
.SH FILES
.TP
.I /etc/localtime
Default local timezone file.
.TP
.I /usr/share/zoneinfo
Default timezone information directory.
.SH NOTES
For areas with more than two types of local time,
you may need to use local standard time in the
.B AT
field of the earliest transition time's rule to ensure that
the earliest transition time recorded in the compiled file is correct.
.PP
If,
for a particular timezone,
a clock advance caused by the start of daylight saving
coincides with and is equal to
a clock retreat caused by a change in UT offset,
.B zic
produces a single transition to daylight saving at the new UT offset
without any change in local (wall clock) time.
To get separate transitions
use multiple zone continuation lines
specifying transition instants using universal time.
.SH SEE ALSO
.BR tzfile (5),
.BR zdump (8)
zdump.80000644000000000000000000001526614737431435007261 0ustar00.\" This file is in the public domain, so clarified as of
.\" 2009-05-17 by Arthur David Olson.
.TH zdump 8 "" "Time Zone Database"
.SH NAME
zdump \- timezone dumper
.SH SYNOPSIS
.B zdump
[
.I option
\&... ] [
.I timezone
\&... ]
.SH DESCRIPTION
.ie '\(lq'' .ds lq \&"\"
.el .ds lq \(lq\"
.ie '\(rq'' .ds rq \&"\"
.el .ds rq \(rq\"
.de q
\\$3\*(lq\\$1\*(rq\\$2
..
The
.B zdump
program prints the current time in each
.I timezone
named on the command line.
A
.I timezone
of
.B \-
is treated as if it were /dev/stdin;
this can be used to pipe TZif data into
.BR zdump .
.SH OPTIONS
.TP
.B \-\-version
Output version information and exit.
.TP
.B \-\-help
Output short usage message and exit.
.TP
.B \-i
Output a description of time intervals.  For each
.I timezone
on the command line, output an interval-format description of the
timezone.  See
.q "INTERVAL FORMAT"
below.
.TP
.B \-v
Output a verbose description of time intervals.
For each
.I timezone
on the command line,
print the times at the two extreme time values,
the times (if present) at and just beyond the boundaries of years that
.BR localtime (3)
and
.BR gmtime (3)
can represent, and
the times both one second before and exactly at
each detected time discontinuity.
Each line is followed by
.BI isdst= D
where
.I D
is positive, zero, or negative depending on whether
the given time is daylight saving time, standard time,
or an unknown time type, respectively.
Each line is also followed by
.BI gmtoff= N
if the given local time is known to be
.I N
seconds east of Greenwich.
.TP
.B \-V
Like
.BR \-v ,
except omit output concerning extreme time and year values.
This generates output that is easier to compare to that of
implementations with different time representations.
.TP
.BI "\-c " \fR[\fIloyear , \fR]\fIhiyear
Cut off interval output at the given year(s).
Cutoff times are computed using the proleptic Gregorian calendar with year 0
and with Universal Time (UT) ignoring leap seconds.
Cutoffs are at the start of each year, where the lower-bound
timestamp is inclusive and the upper is exclusive; for example,
.B "\-c 1970,2070"
selects transitions on or after 1970-01-01 00:00:00 UTC
and before 2070-01-01 00:00:00 UTC.
The default cutoff is
.BR \-500,2500 .
.TP
.BI "\-t " \fR[\fIlotime , \fR]\fIhitime
Cut off interval output at the given time(s),
given in decimal seconds since 1970-01-01 00:00:00
Coordinated Universal Time (UTC).
The
.I timezone
determines whether the count includes leap seconds.
As with
.BR \-c ,
the cutoff's lower bound is inclusive and its upper bound is exclusive.
.SH "INTERVAL FORMAT"
The interval format is a compact text representation that is intended
to be both human- and machine-readable.  It consists of an empty line,
then a line
.q "TZ=\fIstring\fP"
where
.I string
is a double-quoted string giving the timezone, a second line
.q "\- \- \fIinterval\fP"
describing the time interval before the first transition if any, and
zero or more following lines
.q "\fIdate time interval\fP",
one line for each transition time and following interval.  Fields are
separated by single tabs.
.PP
Dates are in
.IR yyyy - mm - dd
format and times are in 24-hour
.IR hh : mm : ss
format where
.IR hh <24.
Times are in local time immediately after the transition.  A
time interval description consists of a UT offset in signed
.RI \(+- hhmmss
format, a time zone abbreviation, and an isdst flag.  An abbreviation
that equals the UT offset is omitted; other abbreviations are
double-quoted strings unless they consist of one or more alphabetic
characters.  An isdst flag is omitted for standard time, and otherwise
is a decimal integer that is unsigned and positive (typically 1) for
daylight saving time and negative for unknown.
.PP
In times and in UT offsets with absolute value less than 100 hours,
the seconds are omitted if they are zero, and
the minutes are also omitted if they are also zero.  Positive UT
offsets are east of Greenwich.  The UT offset \-00 denotes a UT
placeholder in areas where the actual offset is unspecified; by
convention, this occurs when the UT offset is zero and the time zone
abbreviation begins with
.q "\-"
or is
.q "zzz".
.PP
In double-quoted strings, escape sequences represent unusual
characters.  The escape sequences are \es for space, and \e", \e\e,
\ef, \en, \er, \et, and \ev with their usual meaning in the C
programming language.  E.g., the double-quoted string
\*(lq"CET\es\e"\e\e"\*(rq represents the character sequence \*(lqCET
"\e\*(rq.\""
.PP
.ne 9
Here is an example of the output, with the leading empty line omitted.
(This example is shown with tab stops set far enough apart so that the
tabbed columns line up.)
.nf
.sp
.if \n(.g .ft CR
.if t .in +.5i
.if n .in +2
.nr w \w'1896-01-13  'u+\n(.i
.ta \w'1896-01-13\0\0'u +\w'12:01:26\0\0'u +\w'-103126\0\0'u +\w'HWT\0\0'u
TZ="Pacific/Honolulu"
-	-	-103126	LMT
1896-01-13	12:01:26	-1030	HST
1933-04-30	03	-0930	HDT	1
1933-05-21	11	-1030	HST
1942-02-09	03	-0930	HWT	1
1945-08-14	13:30	-0930	HPT	1
1945-09-30	01	-1030	HST
1947-06-08	02:30	-10	HST
.in
.if \n(.g .ft
.sp
.fi
Here, local time begins 10 hours, 31 minutes and 26 seconds west of
UT, and is a standard time abbreviated LMT.  Immediately after the
first transition, the date is 1896-01-13 and the time is 12:01:26, and
the following time interval is 10.5 hours west of UT, a standard time
abbreviated HST.  Immediately after the second transition, the date is
1933-04-30 and the time is 03:00:00 and the following time interval is
9.5 hours west of UT, is abbreviated HDT, and is daylight saving time.
Immediately after the last transition the date is 1947-06-08 and the
time is 02:30:00, and the following time interval is 10 hours west of
UT, a standard time abbreviated HST.
.PP
.ne 10
Here are excerpts from another example:
.nf
.sp
.if \n(.g .ft CR
.if t .in +.5i
.if n .in +2
TZ="Europe/Astrakhan"
-	-	+031212	LMT
1924-04-30	23:47:48	+03
1930-06-21	01	+04
1981-04-01	01	+05		1
1981-09-30	23	+04
\&...
2014-10-26	01	+03
2016-03-27	03	+04
.in
.if \n(.g .ft
.sp
.fi
This time zone is east of UT, so its UT offsets are positive.  Also,
many of its time zone abbreviations are omitted since they duplicate
the text of the UT offset.
.SH LIMITATIONS
Time discontinuities are found by sampling the results returned by
.BR localtime (3)
at twelve-hour intervals.
This works in all real-world cases;
one can construct artificial time zones for which this fails.
.PP
In the
.B \-v
and
.B \-V
output,
.q "UT"
denotes the value returned by
.BR gmtime (3),
which uses UTC for modern timestamps and some other UT flavor for
timestamps that predate the introduction of UTC.
No attempt is currently made to have the output use
.q "UTC"
for newer and
.q "UT"
for older timestamps, partly because the exact date of the
introduction of UTC is problematic.
.SH SEE ALSO
.BR tzfile (5),
.BR zic (8)
date.10000644000000000000000000000342115134527373007015 0ustar00.\" This file is in the public domain, so clarified as of
.\" 2009-05-17 by Arthur David Olson.
.TH date 1 "" "Time Zone Database"
.SH NAME
date \- show and set date and time
.SH SYNOPSIS
.B date
[
.B \-u
] [
.B \-c
] [
.B \-r
.I seconds
] [
.BI + format
] [
\fR[\fIyyyy\fR]\fImmddhhmm\fR[\fIyy\fR][\fB.\fIss\fR]
]
.SH DESCRIPTION
.ie '\(lq'' .ds lq \&"\"
.el .ds lq \(lq\"
.ie '\(rq'' .ds rq \&"\"
.el .ds rq \(rq\"
.de q
\\$3\*(lq\\$1\*(rq\\$2
..
The
.B date
command
without arguments writes the date and time to the standard output in
the form
.ce 1
Sat Mar  8 14:54:40 EST 2025
.br
with
.B EST
replaced by the local time zone's abbreviation
(or by the abbreviation for the time zone specified in the
.B TZ
environment variable if set).
The exact output format depends on the locale.
.PP
If a command-line argument starts with a plus sign (\c
.q "\fB+\fP" ),
the rest of the argument is used as a
.I format
that is processed by
.BR strftime (3)
to determine what to output;
a newline character is appended.
For example, the shell command:
.ce 1
date +"%Y\-%m\-%d %H:%M:%S %z"
.br
outputs a line like
.q "2025\-03\-08 14:54:40 \-0500"
instead.
.PP
These options are available:
.TP
.BR \-u " or " \-c
Use Universal Time when setting and showing the date and time.
.TP
.BI "\-r " seconds
Output the date that corresponds to
.I seconds
past the epoch of 1970-01-01 00:00:00 UTC, where
.I seconds
should be an integer, either decimal, octal (leading 0), or
hexadecimal (leading 0x), preceded by an optional sign.
.SH FILES
.ta \w'/usr/share/zoneinfo/Etc/UTC\0\0'u
/etc/localtime	local timezone file
.br
/usr/lib/locale/\f2L\fP/LC_TIME	description of time locale \f2L\fP
.br
/usr/share/zoneinfo	timezone directory
.br
/usr/share/zoneinfo/Etc/UTC	for UTC leap seconds, if supported
.SH SEE ALSO
.BR strftime (3).
newctime.3.txt0000644000000000000000000002140415216527263010533 0ustar00newctime(3)                Library Functions Manual                newctime(3)

NAME
       asctime, ctime, difftime, gmtime, localtime, mktime - convert date and
       time

SYNOPSIS
       #include 

       struct tm *localtime(time_t const *clock);
       struct tm *localtime_r(time_t const *restrict clock,
           struct tm *restrict result);
       struct tm *localtime_rz(timezone_t restrict zone,
           time_t const *restrict clock,
           struct tm *restrict result);

       struct tm *gmtime(time_t const *clock);
       struct tm *gmtime_r(time_t const *restrict clock,
           struct tm *restrict result);

       time_t mktime(struct tm *tm);
       time_t mktime_z(timezone_t restrict zone,
           struct tm *restrict tm);

       double difftime(time_t time1, time_t time0);

       [[deprecated]] char *asctime(struct tm const *tm);
       [[deprecated]] char *ctime(time_t const *clock);

       /* Only in POSIX.1-2017 and earlier.  */
       char *ctime_r(time_t const *clock, char *buf);
       char *asctime_r(struct tm const *restrict tm,
           char *restrict result);

       cc ... -ltz

DESCRIPTION
       The localtime and gmtime functions convert an integer, pointed to by
       clock, and return pointers to “tm” structures, described below.  If the
       integer is out of range for conversion, these functions return a null
       pointer.  The localtime function corrects for the time zone and any
       time zone adjustments (such as Daylight Saving Time in the United
       States).  The gmtime function converts to Coordinated Universal Time.

       The *clock timestamp represents the time in seconds since 1970-01-01
       00:00:00 Coordinated Universal Time (UTC).  The POSIX standard says
       that timestamps must be nonnegative and must ignore leap seconds.  Many
       implementations extend POSIX by allowing negative timestamps, and can
       therefore represent timestamps that predate the introduction of UTC and
       are some other flavor of Universal Time (UT).  Some implementations
       support leap seconds, in contradiction to POSIX.

       The mktime function converts the broken-down time, expressed as local
       time, in the structure pointed to by tm into a calendar time value with
       the same encoding as that of the values returned by the time function.
       The original values of the tm_wday and tm_yday components of the
       structure are ignored, and the original values of the other components
       are not restricted to their normal ranges.  (A positive or zero value
       for tm_isdst causes mktime to presume initially that daylight saving
       time respectively, is or is not in effect for the specified time.  A
       negative value for tm_isdst causes the mktime function to attempt to
       divine whether daylight saving time is in effect for the specified
       time; in this case it does not use a consistent rule and may give a
       different answer when later presented with the same argument.)  On
       successful completion, the values of the tm_wday and tm_yday components
       of the structure are set appropriately, and the other components are
       set to represent the specified calendar time, but with their values
       forced to their normal ranges; the final value of tm_mday is not set
       until tm_mon and tm_year are determined.  The mktime function returns
       the specified calendar time.  If the calendar time cannot be
       represented, it returns -1 without updating the structure.  To
       distinguish failure from a valid -1 return, you can set tm_wday or
       tm_yday to a negative value before calling mktime; if that value is
       still negative when mktime returns, the calendar time could not be
       represented.

       The difftime function returns the difference between two calendar
       times, (time1 - time0), expressed in seconds.

       The asctime function converts a time value contained in a “tm”
       structure to a pointer to a string of the form
                            Thu Nov 24 18:22:48 1986\n\0
       Years requiring fewer than four characters are padded with leading
       zeroes.  For years longer than four characters, the string is of the
       form
                          Thu Nov 24 18:22:48     81986\n\0
       with five spaces before the year.  This unusual format is designed to
       make it less likely that older software that expects exactly 26 bytes
       of output will mistakenly output misleading values for out-of-range
       years.  This function is deprecated starting in C23.  Callers can use
       strftime instead.

       The ctime function is equivalent to calling localtime and then calling
       asctime on the result.  Like asctime, this function is deprecated
       starting in C23.  Callers can use localtime and strftime instead.

       The ctime_r, localtime_r, gmtime_r, and asctime_r functions are like
       their unsuffixed counterparts, except that they accept an additional
       argument specifying where to store the result if successful.  The
       ctime_r and asctime_r functions are present only on systems supporting
       POSIX.1-2017 and earlier, as they are removed in POSIX.1-2024 and user
       code can define these functions with other meanings.

       The localtime_rz and mktime_z functions are like their unsuffixed
       counterparts, except that they accept an extra initial zone argument
       specifying the timezone to be used for conversion.  If zone is null, UT
       is used; otherwise, zone should be have been allocated by tzalloc and
       should not be freed until after all uses (e.g., by calls to strftime)
       of the filled-in tm_zone fields.

       Declarations of all the functions and externals, and the “tm”
       structure, are in the  header file.  The structure (of type)
       struct tm includes the following fields:

                int tm_sec;      /* seconds (0–60) */
                int tm_min;      /* minutes (0–59) */
                int tm_hour;     /* hours (0–23) */
                int tm_mday;     /* day of month (1–31) */
                int tm_mon;      /* month of year (0–11) */
                int tm_year;     /* year - 1900 */
                int tm_wday;     /* day of week (Sunday = 0) */
                int tm_yday;     /* day of year (0–365) */
                int tm_isdst;    /* is daylight saving time in effect? */
                char *tm_zone;   /* time zone abbreviation (optional) */
                long tm_gmtoff;  /* offset from UT in seconds (optional) */

       The tm_isdst field is non-zero if daylight saving time is in effect.

       The tm_gmtoff field is the offset (in seconds) of the time represented
       from UT, with positive values indicating east of the Prime Meridian.
       The field's name is derived from Greenwich Mean Time, a precursor of
       UT.

       In platforms conforming to POSIX.1-2024 the struct tm the tm_zone and
       tm_gmtoff fields exist, and are filled in.  For localtime_rz and
       mktime_rz the storage lifetime of the strings addressed by tm_zone
       extends until the corresponding timezone_t object is freed via tzfree.
       For the other functions the lifetime extends until the TZ environment
       variable changes state and tzset is then called.

       As a side effect, the ctime, localtime and mktime functions also behave
       as if tzset were called.  The ctime_r and localtime_r functions might
       (or might not) also behave this way.  This is for compatibility with
       older platforms, as required by POSIX.

FILES
       /etc/localtime           local timezone file
       /usr/share/zoneinfo      timezone directory
       /usr/share/zoneinfo/GMT  for UTC leap seconds

       If /usr/share/zoneinfo/GMT is absent, UTC leap seconds are loaded from
       /usr/share/zoneinfo/GMT0 if present.

SEE ALSO
       getenv(3), newstrftime(3), newtzset(3), time(2), tzfile(5).

NOTES
       The return values of asctime, ctime, gmtime, and localtime point to
       static data overwritten by each call.  The remaining functions and data
       are thread-safe.

       The asctime, asctime_r, ctime, and ctime_r functions behave strangely
       for years before 1000 or after 9999.  The 1989 and 1999 editions of the
       C Standard say that years from -99 through 999 are converted without
       extra spaces, but this conflicts with longstanding tradition and with
       this implementation.  The 2011 edition says that the behavior is
       undefined if the year is before 1000 or after 9999.  Traditional
       implementations of these two functions are restricted to years in the
       range 1900 through 2099.  To avoid this portability mess, new programs
       should use strftime instead.

Time Zone Database                                                 newctime(3)
newstrftime.3.txt0000644000000000000000000002242115216527263011267 0ustar00newstrftime(3)             Library Functions Manual             newstrftime(3)

NAME
       strftime - format date and time

SYNOPSIS
       #include 

       size_t strftime(char *restrict buf, size_t maxsize,
           char const *restrict format, struct tm const *restrict timeptr);

       cc ... -ltz

DESCRIPTION
       The strftime function formats the information from *timeptr into the
       array pointed to by buf according to the string pointed to by format.

       The format string consists of zero or more conversion specifications
       and ordinary characters.  All ordinary characters are copied directly
       into the array.  A conversion specification consists of a percent sign
       “%” and one other character.

       No more than maxsize bytes are placed into the array.

       Each conversion specification is replaced by the characters as follows
       which are then copied into the array.  The characters depend on the
       values of zero or more members of *timeptr as specified by brackets in
       the description.  If a bracketed member name is followed by “+”,
       strftime can use the named member even though POSIX.1-2024 does not
       list it; if the name is followed by “-”, strftime ignores the member
       even though POSIX.1-2024 lists it which means portable code should set
       it.  For portability, *timeptr should be initialized as if by a
       successful call to gmtime, localtime, mktime, timegm, or similar
       functions.

       %A     is replaced by the locale's full weekday name.  [tm_wday]

       %a     is replaced by the locale's abbreviated weekday name.  [tm_wday]

       %B     is replaced by the locale's full month name.  [tm_mon]

       %b or %h
              is replaced by the locale's abbreviated month name.  [tm_mon]

       %C     is replaced by the century (a year divided by 100 and truncated
              to an integer) as a decimal number, with at least two digits by
              default.  [tm_year]

       %c     is replaced by the locale's appropriate date and time
              representation.  [tm_year, tm_yday, tm_mon, tm_mday, tm_wday,
              tm_hour, tm_min, tm_sec, tm_gmtoff, tm_zone, tm_isdst-].

       %D     is equivalent to %m/%d/%y.  Although used in the United States
              for current dates, this format is ambiguous elsewhere and for
              dates that might involve other centuries.  [tm_year, tm_mon,
              tm_mday]

       %d     is replaced by the day of the month as a decimal number [01,31].
              [tm_mday]

       %e     is replaced by the day of month as a decimal number [1,31];
              single digits are preceded by a blank.  [tm_mday]

       %F     is equivalent to %Y-%m-%d (the ISO 8601 date format).  [tm_year,
              tm_mon, tm_mday]

       %G     is replaced by the ISO 8601 year with century as a decimal
              number.  This is the year that includes the greater part of the
              week.  (Monday as the first day of a week).  See also the %V
              conversion specification.  [tm_year, tm_yday, tm_wday]

       %g     is replaced by the ISO 8601 year without century as a decimal
              number [00,99].  Since it omits the century, it is ambiguous for
              dates.  [tm_year, tm_yday, tm_wday]

       %H     is replaced by the hour (24-hour clock) as a decimal number
              [00,23].  [tm_hour]

       %I     is replaced by the hour (12-hour clock) as a decimal number
              [01,12].  [tm_hour]

       %j     is replaced by the day of the year as a decimal number
              [001,366].  [tm_yday]

       %k     is replaced by the hour (24-hour clock) as a decimal number
              [0,23]; single digits are preceded by a blank.  [tm_hour]

       %l     is replaced by the hour (12-hour clock) as a decimal number
              [1,12]; single digits are preceded by a blank.  [tm_hour]

       %M     is replaced by the minute as a decimal number [00,59].  [tm_min]

       %m     is replaced by the month as a decimal number [01,12].  [tm_mon]

       %n     is replaced by a newline.

       %p     is replaced by the locale's equivalent of either “AM” or “PM”.
              [tm_hour]

       %R     is replaced by the time in the format %H:%M.  [tm_hour, tm_min]

       %r     is replaced by the locale's representation of 12-hour clock time
              using AM/PM notation.  [tm_hour, tm_min, tm_sec]

       %S     is replaced by the second as a decimal number [00,60].  The
              range of seconds is [00,60] instead of [00,59] to allow for the
              periodic occurrence of leap seconds.  [tm_sec]

       %s     is replaced by the number of seconds since the Epoch (see
              ctime(3)).  Although %s is reliable in this implementation, it
              can have glitches on other platforms (notably obsolescent
              platforms lacking tm_gmtoff or where time_t is no wider than
              int), and POSIX allows strftime to set errno to EINVAL or
              EOVERFLOW and return 0 if the number of seconds would be
              negative or out of range for time_t.  Portable code should
              therefore format a time_t value directly via something like
              sprintf instead of via localtime followed by strftime with "%s".
              [tm_year, tm_mon, tm_mday, tm_hour, tm_min, tm_sec, tm_gmtoff+,
              tm_isdst-].

       %T     is replaced by the time in the format %H:%M:%S.  [tm_hour,
              tm_min, tm_sec]

       %t     is replaced by a tab.

       %U     is replaced by the week number of the year (Sunday as the first
              day of the week) as a decimal number [00,53].  [tm_wday,
              tm_yday, tm_year-]

       %u     is replaced by the weekday (Monday as the first day of the week)
              as a decimal number [1,7].  [tm_wday]

       %V     is replaced by the week number of the year (Monday as the first
              day of the week) as a decimal number [01,53].  If the week
              containing January 1 has four or more days in the new year, then
              it is week 1; otherwise it is week 53 of the previous year, and
              the next week is week 1.  The year is given by the %G conversion
              specification.  [tm_year, tm_yday, tm_wday]

       %W     is replaced by the week number of the year (Monday as the first
              day of the week) as a decimal number [00,53].  [tm_yday,
              tm_wday]

       %w     is replaced by the weekday (Sunday as the first day of the week)
              as a decimal number [0,6].  [tm_year, tm_yday, tm_wday]

       %X     is replaced by the locale's appropriate time representation.
              [tm_year-, tm_yday-, tm_mon-, tm_mday-, tm_wday-, tm_hour,
              tm_min, tm_sec, tm_gmtoff, tm_zone, tm_isdst-].

       %x     is replaced by the locale's appropriate date representation.
              This format can be ambiguous for dates, e.g., it can generate
              "01/02/03" in the C locale.  [tm_year, tm_yday, tm_mon, tm_mday,
              tm_wday, tm_hour-, tm_min-, tm_sec-, tm_gmtoff-, tm_zone-,
              tm_isdst-].

       %Y     is replaced by the year with century as a decimal number.
              [tm_year]

       %y     is replaced by the year without century as a decimal number
              [00,99].  Since it omits the century, it is ambiguous for dates.
              [tm_year]

       %Z     is replaced by the time zone abbreviation, or by the empty
              string if this is not determinable.  [tm_zone, tm_isdst-]

       %z     is replaced by the offset from the Prime Meridian in the format
              +HHMM or -HHMM (ISO 8601) as appropriate, with positive values
              representing locations east of Greenwich, or by the empty string
              if this is not determinable.  The numeric time zone abbreviation
              -0000 is used when the time is Universal Time but local time is
              indeterminate; by convention this is used for locations while
              uninhabited, and corresponds to a zero offset when the time zone
              abbreviation begins with “-”.  [tm_gmtoff, tm_zone+, tm_isdst-]

       %%     is replaced by a single %.

       %+     is replaced by the locale's date and time in date(1) format.
              [tm_year, tm_yday, tm_mon, tm_mday, tm_wday, tm_hour, tm_min,
              tm_sec, tm_gmtoff, tm_zone]

       As a side effect, strftime also behaves as if tzset were called.  This
       is for compatibility with older platforms, as required by POSIX; it is
       not needed for strftime's own use.

RETURN VALUE
       If the conversion is successful, strftime returns the number of bytes
       placed into the array, not counting the terminating NUL; errno is
       unchanged if the returned value is zero.  Otherwise, errno is set to
       indicate the error, zero is returned, and the array contents are
       unspecified.

ERRORS
       This function fails if:

       [ERANGE]
              The total number of resulting bytes, including the terminating
              NUL character, is more than maxsize.

SEE ALSO
       date(1), getenv(3), newctime(3), newtzset(3), time(2), tzfile(5).

BUGS
       There is no conversion specification for the phase of the moon.

Time Zone Database                                              newstrftime(3)
newtzset.3.txt0000644000000000000000000002705715216527263010615 0ustar00newtzset(3)                Library Functions Manual                newtzset(3)

NAME
       tzset - initialize time conversion information

SYNOPSIS
       #include 

       timezone_t tzalloc(char const *TZ);

       void tzfree(timezone_t tz);

       void tzset(void);

       /* Optional and obsolescent:  */
       extern char *tzname[];
       extern long timezone;
       extern int daylight;

       cc ... -ltz

DESCRIPTION
       The tzalloc function allocates and returns a timezone object described
       by TZ.

       If TZ is a null pointer, tzalloc uses the best available approximation
       to local (wall clock) time, as specified by the tzfile(5)-format file
       localtime in the system time conversion information directory.

       If TZ is the empty string, tzalloc uses Universal Time (UT), with the
       abbreviation "UTC" and without leap second correction; please see
       newctime(3) for more about UT, UTC, and leap seconds.

       If TZ is nonnull and nonempty:

              if the value begins with a colon, it is used as a pathname of a
              file from which to read the time conversion information;

              if the value does not begin with a colon, it is first used as
              the pathname of a file from which to read the time conversion
              information, and, if that file cannot be read, is used directly
              as a specification of the time conversion information.

       When TZ contents are used as a pathname, a pathname beginning with “/”
       is used as-is; otherwise the pathname is relative to a system time
       conversion information directory.  In a privileged program the pathname
       must be relative.  Relative pathnames must not contain “..” components.
       For the purpose of these checks, a file name beginning with “/” is
       considered to be relative if it is the localtime file's name, or if it
       starts with the system timezone directory's name followed by one more
       more slashes.  The file must be a regular file in the format specified
       in tzfile(5).

       When TZ is used directly as a specification of the time conversion
       information, it must have the following syntax:

              stdoffset[dst[offset][,rule]]

       Where:

              std and dst
                     Three or more bytes that are the designation for the
                     standard (std) or the alternative (dst, such as daylight
                     saving time) time zone.  Only std is required; if dst is
                     missing, then daylight saving time does not apply in this
                     locale.  Upper- and lowercase letters are explicitly
                     allowed.  Any characters except a leading colon (:),
                     digits, comma (,), ASCII minus (-), ASCII plus (+), and
                     NUL bytes are allowed.  Alternatively, a designation can
                     be surrounded by angle brackets < and >; in this case,
                     the designation can contain any characters other than >
                     and NUL.

              offset Indicates the value one must add to the local time to
                     arrive at Coordinated Universal Time.  The offset has the
                     form:

                            hh[:mm[:ss]]

                     The minutes (mm) and seconds (ss) are optional.  The hour
                     (hh) is required and may be a single digit.  The offset
                     following std is required.  If no offset follows dst,
                     daylight saving time is assumed to be one hour ahead of
                     standard time.  One or more digits may be used; the value
                     is always interpreted as a decimal number.  The hour must
                     be between zero and 24, and the minutes (and seconds) –
                     if present – between zero and 59.  If preceded by a “-”,
                     the time zone shall be east of the Prime Meridian;
                     otherwise it shall be west (which may be indicated by an
                     optional preceding “+”).

              rule   Indicates when to change to and back from daylight saving
                     time.  The rule has the form:

                            date/time,date/time

                     where the first date describes when the change from
                     standard to daylight saving time occurs and the second
                     date describes when the change back happens.  Each time
                     field describes when, in current local time, the change
                     to the other time is made.  Daylight saving is assumed to
                     be in effect all year if it begins January 1 at 00:00 and
                     ends December 31 at 24:00 plus the difference between
                     daylight saving and standard time, leaving no room for
                     standard time in the calendar.

                     The format of date is one of the following:

                     Jn     The Julian day n (1 ≤ n ≤ 365).  Leap days are not
                            counted; that is, in all years – including leap
                            years – February 28 is day 59 and March 1 is day
                            60.  It is impossible to explicitly refer to the
                            occasional February 29.

                     n      The zero-based Julian day (0 ≤ n ≤ 365).  Leap
                            days are counted, and it is possible to refer to
                            February 29.

                     Mm.n.d The d'th day (0 ≤ d ≤ 6) of week n of month m of
                            the year (1 ≤ n ≤ 5, 1 ≤ m ≤ 12, where week 5
                            means “the last d day in month m” which may occur
                            in either the fourth or the fifth week).  Week 1
                            is the first week in which the d'th day occurs.
                            Day zero is Sunday.

                     The time has the same format as offset except that the
                     hours part of time can range from -167 through 167; this
                     allows for unusual rules such as “the Saturday before the
                     first Sunday of March”.  The default, if time is not
                     given, is 02:00:00.

       Here are some examples of TZ values that directly specify the timezone.

       EST5   stands for US Eastern Standard Time (EST), 5 hours behind UT,
              without daylight saving.

       <+12>-12<+13>,M11.1.0,M1.2.1/147
              stands for Fiji time, 12 hours ahead of UT, springing forward on
              November's first Sunday at 02:00, and falling back on January's
              second Monday at 147:00 (i.e., 03:00 on the first Sunday on or
              after January 14).  The abbreviations for standard and daylight
              saving time are “+12” and “+13”.

       IST-2IDT,M3.4.4/26,M10.5.0
              stands for Israel Standard Time (IST) and Israel Daylight Time
              (IDT), 2 hours ahead of UT, springing forward on March's fourth
              Thursday at 26:00 (i.e., 02:00 on the first Friday on or after
              March 23), and falling back on October's last Sunday at 02:00.

       <-04>4<-03>,J1/0,J365/25
              stands for permanent daylight saving time, 3 hours behind UT
              with abbreviation “-03”.  There is a dummy fall-back transition
              on December 31 at 25:00 daylight saving time (i.e., 24:00
              standard time, equivalent to January 1 at 00:00 standard time),
              and a simultaneous spring-forward transition on January 1 at
              00:00 standard time, so daylight saving time is in effect all
              year and the initial <-04> is a placeholder.

       <-03>3<-02>,M3.5.0/-2,M10.5.0/-1
              stands for time in western Greenland, 3 hours behind UT, where
              clocks follow the EU rule of springing forward on March's last
              Sunday at 01:00 UT (-02:00 local time, i.e., 22:00 the previous
              day) and falling back on October's last Sunday at 01:00 UT
              (-01:00 local time, i.e., 23:00 the previous day).  The
              abbreviations for standard and daylight saving time are “-03”
              and “-02”.

       If TZ specifies daylight saving time but does not specify a rule, the
       rule typically defaults to the current US daylight-saving rule,
       although such a default is not guaranteed and is incorrect for many
       locations outside the US.  Therefore, if a TZ string directly specifies
       a timezone with daylight saving time, it should specify the daylight
       saving rule explicitly.

       For compatibility with System V Release 3.1, a semicolon (;) may be
       used to separate the rule from the rest of the specification; this is
       an extension to POSIX.

       The tzfree function frees a timezone object tz, which should have been
       successfully allocated by tzalloc.  This invalidates any tm_zone
       pointers that tz was used to set.

       The tzset function acts like tzalloc(getenv("TZ")), except it saves any
       resulting timezone object into internal storage that is accessed by
       localtime, localtime_r, and mktime.  The anonymous shared timezone
       object is freed by the next call to tzset.  If the implied call to
       getenv fails, tzset acts like tzalloc(nullptr); if the implied call to
       tzalloc fails, tzset falls back on UT.

       As a side effect, the tzset function sets some external variables if
       the platform defines them.  It sets tzname[0] and tzname[1] to pointers
       to strings that are time zone abbreviations to be used with standard
       and daylight saving time, respectively.  It also sets timezone to be
       the number of seconds that standard time is west of the Prime Meridian,
       and daylight to be zero if daylight saving time is never in effect,
       non-zero otherwise.

RETURN VALUE
       If successful, the tzalloc function returns a nonnull pointer to the
       newly allocated object.  Otherwise, it returns a null pointer and sets
       errno.  The tzfree function does not modify errno.

ERRORS
       EOVERFLOW
              TZ directly specifies time conversion information, and contains
              an integer out of machine range or a time zone abbreviation that
              is too long for this platform.

       The tzalloc function may also fail and set errno for any of the errors
       specified for the routines access(2), close(2), malloc(3), open(2), and
       read(2).

FILES
       /etc/localtime           local timezone file
       /usr/share/zoneinfo      timezone directory
       /usr/share/zoneinfo/GMT  for UTC leap seconds

       If /usr/share/zoneinfo/GMT is absent, UTC leap seconds are loaded from
       /usr/share/zoneinfo/GMT0 if present.

SEE ALSO
       getenv(3), newctime(3), newstrftime(3), time(2), tzfile(5).

NOTES
       Portable code should not rely on the contents of the external variables
       tzname, timezone and daylight as their contents are unspecified (and do
       not make sense in general) when a geographical TZ is used.  In
       multithreaded applications behavior is undefined if one thread accesses
       one of these variables while another thread invokes tzset.  A future
       version of POSIX is planned to remove these variables; callers can
       instead use the tm_gmtoff and tm_zone members of struct tm, or use
       strftime with "%z" or "%Z".

Time Zone Database                                                 newtzset(3)
time2posix.3.txt0000644000000000000000000001006415216527263011023 0ustar00time2posix(3)              Library Functions Manual              time2posix(3)

NAME
       time2posix, posix2time - convert seconds since the Epoch

SYNOPSIS
       #include 

       time_t time2posix(time_t t);

       time_t posix2time(time_t t);

       cc ... -ltz

DESCRIPTION
       IEEE Standard 1003.1 (POSIX) says that time_t values cannot count leap
       seconds and, therefore, that the system time must be adjusted as each
       leap occurs.

       If the time package is configured with leap-second support enabled,
       however, no such adjustment is needed and time_t values continue to
       increase over leap events (as a true “seconds since...” value).  This
       means that these values will differ from those required by POSIX by the
       net number of leap seconds inserted since the Epoch.

       For many C programs this is not a problem as the C standard says that
       time_t is (mostly) opaque – time_t values should be obtained from and
       passed to functions such as time(2), localtime(3), mktime(3), and
       difftime(3).  However, POSIX gives an arithmetic expression for
       computing a time_t value directly from a given Universal date and time,
       and the same relationship is assumed by some applications.  Any
       programs creating/dissecting time_t values using such a relationship
       will typically not handle intervals over leap seconds correctly.

       The time2posix and posix2time functions address this mismatch by
       converting between local time_t values and their POSIX equivalents.
       This is done by accounting for the number of time-base changes that
       would have taken place on a POSIX system as leap seconds were inserted
       or deleted.  These converted values can then be used when communicating
       with POSIX-compliant systems.

       The time2posix function converts a time_t value to its POSIX
       counterpart, if one exists.  The posix2time function does the reverse
       but is less well-behaved: for a positive leap second hit the result is
       not unique, and for a negative leap second hit the corresponding POSIX
       time_t doesn't exist so an adjacent value is returned.  Both of these
       are indicate problems with the POSIX representation.

       The following table summarizes the relationship between a time T and
       its conversion to, and back from, the POSIX representation over the
       leap second inserted at the end of June, 1993.  In this table,
       X=time2posix(T), Y=posix2time(X), A=741484816, and B=A-17 because 17
       positive leap seconds preceded this leap second.

         DATE       TIME     T   X   Y
         1993-06-30 23:59:59 A   B   A
         1993-06-30 23:59:60 A+1 B+1 A+1 or A+2
         1993-07-01 00:00:00 A+2 B+1 A+1 or A+2
         1993-07-01 00:00:01 A+3 B+2 A+3

       A leap second deletion would look like the following, and
       posix2time(B+1) would return either A or A+1.

         DATE       TIME     T   X   Y
         ????-06-30 23:59:58 A   B   A
         ????-07-01 00:00:00 A+1 B+2 A+1
         ????-07-01 00:00:01 A+2 B+3 A+2

       If leap-second support is not enabled, local time_t and POSIX time_t
       values are equivalent, and both time2posix and posix2time degenerate to
       the identity function.

RETURN VALUE
       If successful, these functions return the resulting timestamp without
       modifying errno.  Otherwise, they set errno and return ((time_t) -1).

ERRORS
       These functions fail if:

       [EOVERFLOW]
              The resulting value cannot be represented.  This can happen for
              posix2time if its argument is close to the maximum time_t value.
              In environments where the TZ environment variable names a TZif
              file, overflow can happen for either function for an argument
              sufficiently close to an extreme time_t value if the TZif file
              specifies unrealistic leap second corrections.

SEE ALSO
       difftime(3), localtime(3), mktime(3), time(2).

Time Zone Database                                               time2posix(3)
tzfile.5.txt0000644000000000000000000005456415132622176010230 0ustar00tzfile(5)                     File Formats Manual                    tzfile(5)

NAME
       tzfile - timezone information

DESCRIPTION
       The timezone information files used by tzset(3) are typically found
       under a directory with a name like /usr/share/zoneinfo.  These files
       use the format described in Internet RFC 9636.  Each file is a sequence
       of 8-bit bytes.  In a file, a binary integer is represented by a
       sequence of one or more bytes in network order (bigendian, or high-
       order byte first), with all bits significant, a signed binary integer
       is represented using two's complement, and a boolean is represented by
       a one-byte binary integer that is either 0 (false) or 1 (true).  The
       format begins with a 44-byte header containing the following fields:

         •  The magic four-byte ASCII sequence “TZif” identifies the file as a
            timezone information file.

         •  A byte identifying the version of the file's format (as of 2021,
            either an ASCII NUL, “2”, “3”, or “4”).

         •  Fifteen bytes containing zeros reserved for future use.

         •  Six four-byte integer values, in the following order:

              tzh_ttisutcnt
                The number of UT/local indicators stored in the file.  (UT is
                Universal Time.)

              tzh_ttisstdcnt
                The number of standard/wall indicators stored in the file.

              tzh_leapcnt
                The number of leap seconds for which data entries are stored
                in the file.

              tzh_timecnt
                The number of transition times for which data entries are
                stored in the file.

              tzh_typecnt
                The number of local time types for which data entries are
                stored in the file (must not be zero).

              tzh_charcnt
                The number of bytes of time zone abbreviation strings stored
                in the file.

       The above header is followed by the following fields, whose lengths
       depend on the contents of the header:

         •  tzh_timecnt four-byte signed integer values sorted in ascending
            order.  These values are written in network byte order.  Each is
            used as a transition time (as returned by time(2)) at which the
            rules for computing local time change.

         •  tzh_timecnt one-byte unsigned integer values; each one but the
            last tells which of the different types of local time types
            described in the file is associated with the time period starting
            with the same-indexed transition time and continuing up to but not
            including the next transition time.  (The last time type is
            present only for consistency checking with the proleptic TZ string
            described below.)  These values serve as indices into the next
            field.

         •  tzh_typecnt ttinfo entries, each defined as follows:

              struct ttinfo {
                  int32_t       tt_utoff;
                  unsigned char tt_isdst;
                  unsigned char tt_desigidx;
              };

            Each structure is written as a four-byte signed integer value for
            tt_utoff, in network byte order, followed by a one-byte boolean
            for tt_isdst and a one-byte value for tt_desigidx.  In each
            structure, tt_utoff gives the number of seconds to be added to UT,
            tt_isdst tells whether tm_isdst should be set by localtime(3) and
            tt_desigidx serves as an index into the array of time zone
            abbreviation bytes that follow the ttinfo entries in the file; if
            the designated string is "-00", the ttinfo entry is a placeholder
            indicating that local time is unspecified.  The tt_utoff value is
            never equal to -2**31, to let 32-bit clients negate it without
            overflow.  Also, in realistic applications tt_utoff is in the
            range [-89999, 93599] (i.e., more than -25 hours and less than 26
            hours); this allows easy support by implementations that already
            support the POSIX-required range [-24:59:59, 25:59:59].

         •  tzh_charcnt bytes that represent time zone designations, which are
            null-terminated byte strings, each indexed by the tt_desigidx
            values mentioned above, and each corresponding to a time zone
            abbreviation.  The byte strings can overlap if one is a suffix of
            the other.  The encoding of these strings is not specified.

         •  tzh_leapcnt pairs of four-byte values, written in network byte
            order; the first value of each pair gives the non-negative time
            (as returned by time(2)) at which a leap second occurs or at which
            the leap second table expires; the second is a signed integer
            specifying the correction, which is the total number of leap
            seconds to be applied during the time period starting at the given
            time.  The pairs of values are sorted in strictly ascending order
            by time.  Each pair denotes one leap second, either positive or
            negative, except that if the last pair has the same correction as
            the previous one, the last pair denotes the leap second table's
            expiration time.  Each leap second is at the end of a UTC calendar
            month.  The first leap second has a non-negative occurrence time,
            and is a positive leap second if and only if its correction is
            positive; the correction for each leap second after the first
            differs from the previous leap second by either 1 for a positive
            leap second, or -1 for a negative leap second.  If the leap second
            table is empty, the leap-second correction is zero for all
            timestamps; otherwise, for timestamps before the first occurrence
            time, the leap-second correction is zero if the first pair's
            correction is 1 or -1, and is unspecified otherwise (which can
            happen only in files truncated at the start).

         •  tzh_ttisstdcnt standard/wall indicators, each stored as a one-byte
            boolean; they tell whether the transition times associated with
            local time types were specified as standard time or local (wall
            clock) time.

         •  tzh_ttisutcnt UT/local indicators, each stored as a one-byte
            boolean; they tell whether the transition times associated with
            local time types were specified as UT or local time.  If a
            UT/local indicator is set, the corresponding standard/wall
            indicator must also be set.

       The standard/wall and UT/local indicators were designed for
       transforming a TZif file's transition times into transitions
       appropriate for another time zone specified via a proleptic TZ string
       that lacks rules.  For example, when TZ="EET-2EEST" and there is no
       TZif file "EET-2EEST", the idea was to adapt the transition times from
       a TZif file with the well-known name "posixrules" that was present only
       for this purpose and was a copy of the file "Europe/Brussels", a file
       with a different UT offset.  POSIX does not specify the details of this
       obsolete transformational behavior, the default rules are installation-
       dependent, and no implementation is known to support this feature for
       timestamps past 2037, so users desiring (say) Greek time should instead
       specify TZ="Europe/Athens" for better historical coverage, falling back
       on TZ="EET-2EEST,M3.5.0/3,M10.5.0/4" if conformance to POSIX.1-2017 or
       earlier is required and older timestamps need not be handled
       accurately.

       The localtime(3) function normally uses the first ttinfo structure in
       the file if either tzh_timecnt is zero or the time argument is less
       than the first transition time recorded in the file.

   Version 2 format
       For version-2-format timezone files, the above header and data are
       followed by a second header and data, identical in format except that
       eight bytes are used for each transition time or leap second time.
       (Leap second counts remain four bytes.)  After the second header and
       data comes a newline-enclosed string in the style of the contents of a
       proleptic TZ, for use in handling instants after the last transition
       time stored in the file or for all instants if the file has no
       transitions.  The TZ string is empty (i.e., nothing between the
       newlines) if there is no proleptic representation for such instants.

       If non-empty, the TZ string must agree with the local time type after
       the last transition time if present in the eight-byte data; for
       example, given the string “WET0WEST,M3.5.0/1,M10.5.0” then if a last
       transition time is in July, the transition's local time type must
       specify a daylight-saving time abbreviated “WEST” that is one hour east
       of UT.

       The TZ string can contain time zone abbreviations and UT offsets that
       do not appear elsewhere in the TZif file.

       Also, if there is at least one transition, time type 0 is associated
       with the time period from the indefinite past up to but not including
       the earliest transition time.

   Version 3 format
       For version-3-format timezone files, a TZ string (see newtzset(3)) may
       use the following POSIX.1-2024 extensions to POSIX.1-2017: First, as in
       TZ="<-02>2<-01>,M3.5.0/-1,M10.5.0/0", the hours part of its transition
       times may be signed and range from -167 through 167 instead of being
       limited to unsigned values from 0 through 24.  Second, as in
       TZ="XXX3EDT4,0/0,J365/23", DST is in effect all year if it starts
       January 1 at 00:00 and ends December 31 at 24:00 plus the difference
       between daylight saving and standard time.

   Version 4 format
       For version-4-format TZif files, the first leap second record can have
       a correction that is neither +1 nor -1, to represent truncation of the
       TZif file at the start.  Also, if two or more leap second transitions
       are present and the last entry's correction equals the previous one,
       the last entry denotes the expiration of the leap second table instead
       of a leap second; timestamps after this expiration are unreliable in
       that future releases will likely add leap second entries after the
       expiration, and the added leap seconds will change how post-expiration
       timestamps are treated.

   Interoperability considerations
       Future changes to the format may append more data.

       Version 1 files are considered a legacy format and should not be
       generated, as they do not support transition times after the year 2038.
       Readers that understand only Version 1 must ignore any data that
       extends beyond the calculated end of the version 1 data block.

       Other than version 1, writers should generate the lowest version number
       needed by a file's data.  For example, a writer should generate a
       version 4 file only if its leap second table either expires or is
       truncated at the start.  Likewise, a writer not generating a version 4
       file should generate a version 3 file only if TZ string extensions are
       necessary to accurately model transition times.

       The sequence of time changes defined by the version 1 header and data
       block should be a contiguous sub-sequence of the time changes defined
       by the version 2+ header and data block, and by the footer.  This
       guideline helps obsolescent version 1 readers agree with current
       readers about timestamps within the contiguous sub-sequence.  It also
       lets writers not supporting obsolescent readers use a tzh_timecnt of
       zero in the version 1 data block to save space.

       When a TZif file contains a leap second table expiration time, TZif
       readers should either refuse to process post-expiration timestamps, or
       process them as if the expiration time did not exist (possibly with an
       error indication).

       Time zone abbreviations should consist of at least three (3) and no
       more than six (6) ASCII characters from the set of alphanumerics, “-”,
       and “+”.  This is for compatibility with POSIX requirements for time
       zone abbreviations.

       A numeric time zone abbreviation should match the UT offset.  For
       example, "+0530" should be used only if the UT offset is 5.5 hours
       ahead of UT, and "-00" should be used only if the UT offset is zero.

       When reading a version 2 or higher file, readers should ignore the
       version 1 header and data block except for the purpose of skipping over
       them.

       Readers should calculate the total lengths of the headers and data
       blocks and check that they all fit within the actual file size, as part
       of a validity check for the file.

       When a positive leap second occurs, readers should append an extra
       second to the local minute containing the second just before the leap
       second.  If this occurs when the UTC offset is not a multiple of 60
       seconds, the leap second occurs earlier than the last second of the
       local minute and the minute's remaining local seconds are numbered
       through 60 instead of the usual 59; the UTC offset is unaffected.

   Common interoperability issues
       This section documents common problems in reading or writing TZif
       files.  Most of these are problems in generating TZif files for use by
       older readers.  The goals of this section are to help:

         •  TZif writers output files that avoid common pitfalls in older or
            buggy TZif readers,

         •  TZif readers avoid common pitfalls when reading files generated by
            future TZif writers, and

         •  any future specification authors see what sort of problems arise
            when the TZif format is changed.

       When new versions of the TZif format have been defined, a design goal
       has been that a reader can successfully use a TZif file even if the
       file is of a later TZif version than what the reader was designed for.
       When complete compatibility was not achieved, an attempt was made to
       limit glitches to rarely used timestamps and allow simple partial
       workarounds in writers designed to generate newer-version data useful
       even for older-version readers.  This section attempts to document
       these compatibility issues and workarounds as well as documenting other
       common bugs in readers.

       Interoperability problems with TZif include the following:

         •  Some readers examine only version 1 data.  As a partial
            workaround, a writer can output as much version 1 data as
            possible.  However, a reader should ignore version 1 data, and
            should use version 2+ data even if the reader's native timestamps
            have only 32 bits.

         •  Some readers designed for version 2 might mishandle timestamps
            after a version 3 or higher file's last transition, because they
            cannot parse the POSIX.1-2024 extensions to POSIX.1-2017 in the
            proleptic TZ string.  As a partial workaround, a writer can output
            more transitions than necessary, so that only far-future
            timestamps are mishandled by version 2 readers.

         •  Some readers might mishandle timestamps after a file's last
            transition, because they require that all abbreviations or UT
            offsets in the proleptic TZ string must also occur somewhere in
            the file's tables of time zone designations and local time type
            records.  As a workaround, a writer can output more transitions
            than necessary, so that the other tables contain duplicates of the
            proleptic TZ string's abbreviations and offsets.

         •  Some readers designed for version 2 do not support permanent
            daylight saving time with transitions after 24:00 – e.g., a TZ
            string “EST5EDT,0/0,J365/25” denoting permanent Eastern Daylight
            Time (-04).  As a workaround, a writer can substitute standard
            time for two time zones east, e.g., “XXX3EDT4,0/0,J365/23” for a
            time zone with a never-used standard time (XXX, -03) and negative
            daylight saving time (EDT, -04) all year.  Alternatively, as a
            partial workaround, a writer can substitute standard time for the
            next time zone east – e.g., “AST4” for permanent Atlantic Standard
            Time (-04).

         •  Some readers designed for version 2 or 3 and that require strict
            conformance to RFC 9636 reject version 4 files whose leap second
            tables are truncated at the start or end in expiration times.

         •  Some readers ignore the footer, and instead predict future
            timestamps from the time type of the last transition.  As a
            partial workaround, a writer can output more transitions than
            necessary.

         •  Some stripped-down readers ignore everything but the footer, and
            use its proleptic TZ string to calculate all timestamps.  Although
            this approach often works for current and future timestamps, it
            obviously has problems with past timestamps, and even for current
            timestamps it can fail for settings like TZ="Africa/Casablanca".
            This corresponds to a TZif file containing explicit transitions
            through the year 2087, followed by a footer containing the TZ
            string “<+01>-1”, which should be used only for timestamps after
            the last explicit transition.

         •  Some readers do not use time type 0 for timestamps before the
            first transition, in that they infer a time type using a heuristic
            that does not always select time type 0.  As a partial workaround,
            a writer can output a dummy (no-op) first transition at an early
            time.

         •  Some readers mishandle timestamps before the first transition that
            has a timestamp that is not less than -2**31.  Readers that
            support only 32-bit timestamps are likely to be more prone to this
            problem, for example, when they process 64-bit transitions only
            some of which are representable in 32 bits.  As a partial
            workaround, a writer can output a dummy transition at timestamp
            -2**31.

         •  Some readers mishandle a transition if its timestamp has the
            minimum possible signed 64-bit value.  Timestamps less than -2**59
            are not recommended.

         •  Some readers mishandle proleptic TZ strings that contain “<” or
            “>”.  As a partial workaround, a writer can avoid using “<” or “>”
            for time zone abbreviations containing only alphabetic characters.

         •  Many readers mishandle time zone abbreviations that contain non-
            ASCII characters.  These characters are not recommended.

         •  Some readers may mishandle time zone abbreviations that contain
            fewer than 3 or more than 6 characters or that contain ASCII
            characters other than alphanumerics, “-”, and “+”.  These
            abbreviations are not recommended.

         •  Some readers mishandle TZif files that specify daylight-saving
            time UT offsets that are less than the UT offsets for the
            corresponding standard time.  These readers do not support
            locations like Ireland, which uses the equivalent of the TZ string
            “IST-1GMT0,M10.5.0,M3.5.0/1”, observing standard time (IST, +01)
            in summer and daylight saving time (GMT, +00) in winter.  As a
            partial workaround, a writer can output data for the equivalent of
            the TZ string “GMT0IST,M3.5.0/1,M10.5.0”, thus swapping standard
            and daylight saving time.  Although this workaround misidentifies
            which part of the year uses daylight saving time, it records UT
            offsets and time zone abbreviations correctly.

         •  Some readers generate ambiguous timestamps for positive leap
            seconds that occur when the UTC offset is not a multiple of 60
            seconds.  For example, with UTC offset +01:23:45 and a positive
            leap second 78796801 (1972-06-30 23:59:60 UTC), some readers will
            map both 78796800 and 78796801 to 01:23:45 local time the next day
            instead of mapping the latter to 01:23:46, and they will map
            78796815 to 01:23:59 instead of to 01:23:60.  This has not yet
            been a practical problem, since no civil authority has observed
            such UTC offsets since leap seconds were introduced in 1972.

       Some interoperability problems are reader bugs that are listed here
       mostly as warnings to developers of readers.

         •  Some readers do not support negative timestamps.  Developers of
            distributed applications should keep this in mind if they need to
            deal with pre-1970 data.

         •  Some readers mishandle timestamps before the first transition that
            has a non-negative timestamp.  Readers that do not support
            negative timestamps are likely to be more prone to this problem.

         •  Some readers mishandle time zone abbreviations like “-08” that
            contain “+”, “-”, or digits.

         •  Some readers mishandle UT offsets that are out of the traditional
            range of -12 through +12 hours, and so do not support locations
            like Kiritimati that are outside this range.

         •  Some readers mishandle UT offsets in the range [-3599, -1] seconds
            from UT because they integer-divide the offset by 3600 to get 0
            and then display the hour part as “+00”.

         •  Some readers mishandle UT offsets that are not a multiple of one
            hour, or of 15 minutes, or of 1 minute.

SEE ALSO
       time(2), localtime(3), tzset(3), tzselect(8), zdump(8), zic(8).

       Olson A, Eggert P, Murchison K. The Time Zone Information Format
       (TZif).  October 2024.  Internet RFC 9636 doi:10.17487/RFC9636.

Time Zone Database                                                   tzfile(5)
tzselect.8.txt0000644000000000000000000000617615105202465010562 0ustar00tzselect(8)                 System Manager's Manual                tzselect(8)

NAME
       tzselect - select a timezone

SYNOPSIS
       tzselect [ -c coord ] [ -n limit ] [ --help ] [ --version ]

DESCRIPTION
       The tzselect program asks the user for information about the current
       location, and outputs the resulting timezone to standard output.  The
       output is suitable as a value for the TZ environment variable.

       All interaction with the user is done via standard input and standard
       error.

OPTIONS
       -c coord
              Instead of asking for continent and then country and then city,
              ask for selection from time zones whose largest cities are
              closest to the location with geographical coordinates coord.
              Use ISO 6709 notation for coord, that is, a latitude immediately
              followed by a longitude.  The latitude and longitude should be
              signed integers followed by an optional decimal point and
              fraction: positive numbers represent north and east, negative
              south and west.  Latitudes with two and longitudes with three
              integer digits are treated as degrees; latitudes with four or
              six and longitudes with five or seven integer digits are treated
              as DDMM, DDDMM, DDMMSS, or DDDMMSS representing DD or DDD
              degrees, MM minutes, and zero or SS seconds, with any trailing
              fractions represent fractional minutes or (if SS is present)
              seconds.  The decimal point is that of the current locale.  For
              example, in the (default) C locale, -c +40.689-074.045 specifies
              40.689 degrees N, 74.045 degrees W, -c +4041.4-07402.7 specifies
              40 degrees 41.4 minutes N, 74 degrees 2.7 minutes W, and
              -c +404121-0740240 specifies 40 degrees 41 minutes 21 seconds N,
              74 degrees 2 minutes 40 seconds W.  If coord is not one of the
              documented forms, the resulting behavior is unspecified.

       -n limit
              When -c is used, display the closest limit locations (default
              10).

       --help Output help information and exit.

       --version
              Output version information and exit.

ENVIRONMENT VARIABLES
       AWK    Name of a POSIX-compliant awk program (default: awk).

       TZDIR  Name of the directory containing timezone data files (default:
              /usr/share/zoneinfo).

FILES
       TZDIR/iso3166.tab
              Table of ISO 3166 2-letter country codes and country names.

       TZDIR/zone1970.tab
              Table of country codes, latitude and longitude, timezones, and
              descriptive comments.

       TZDIR/TZ
              Timezone data file for timezone TZ.

EXIT STATUS
       The exit status is zero if a timezone was successfully obtained from
       the user, nonzero otherwise.

SEE ALSO
       newctime(3), tzfile(5), zdump(8), zic(8)

NOTES
       Applications should not assume that tzselect's output matches the
       user's political preferences.

Time Zone Database                                                 tzselect(8)
zic.8.txt0000644000000000000000000007141415246336076007523 0ustar00zic(8)                      System Manager's Manual                     zic(8)

NAME
       zic - timezone compiler

SYNOPSIS
       zic [ option ... ] [ filename ... ]

DESCRIPTION
       The zic program reads text from the file(s) named on the command line
       and creates the timezone information format (TZif) files specified in
       this input.  If a filename is “-”, standard input is read.

OPTIONS
       --version
              Output version information and exit.

       --help Output short usage message and exit.

       -b bloat
              Output backward-compatibility data as specified by bloat.  If
              bloat is fat, generate additional data entries that work around
              potential bugs or incompatibilities in older software, such as
              software that mishandles a TZif file's 64-bit data or proleptic
              TZ string.  If bloat is slim, keep the output files small; this
              can help check for the bugs and incompatibilities.  The default
              is slim, as the fat workarounds are typically good only until
              the year 2038 anyway.  Also see the -r option for another way to
              alter output size.

       -D     Do not create ancestor directories of output files, For example,
              for a zone named America/Los_Angeles the directory America
              should already exist.  By default, the directory and its
              ancestors are created if they do not already exist.

       -d directory
              Create time conversion information files in the named directory
              rather than in the standard directory named below.

       -l timezone
              Use timezone as local time.  zic will act as if the input
              contained a link line of the form

                Link  timezone  localtime

              If timezone is -, any already-existing link is removed.

       -L leapsecondfilename
              Read leap second information from the file with the given name.
              If this option is not used, no leap second information appears
              in output files; this is required by some TZif readers.

       -p timezone
              Act as if the input contained a link line of the form

                Link  timezone  posixrules

              If timezone is “-” (the default), any already-existing link is
              removed.

              Unless timezone is “-”, this option is obsolete and is no longer
              supported by most runtimes.  Among other things it should not be
              used for timestamps after the year 2037, and it should not be
              combined with -b slim if timezone's transitions are at standard
              time or Universal Time (UT) instead of local time.  The option
              is present only to support obsolete runtimes that used
              timezone's rules when handling obsolescent TZ strings like
              "AST4ADT" that lack transition rules; modern runtimes that
              support these TZ strings typically just use current US rules as
              the TZ strings were mainly used in the US.  Similarly, any Zone
              or Link named "posixrules" is obsolete and problematic.

       -m mode
              Create TZif files with the given file mode bits.  By default the
              files are created with mode 644 as modified by the umask.  With
              this option they are created with the given mode instead.  For
              portability the mode should be an unsigned octal integer,
              typically 644 or 444; some platforms also support chmod(1)-style
              symbolic modes.  This option does not affect created ancestor
              directories, which have mode 755 as modified by the umask.  The
              option is ignored on platforms lacking the notion of file mode
              bits.

       -r [@lo][/@hi]
              Limit the applicability of output files to timestamps in the
              range from lo (inclusive) to hi (exclusive), where lo and hi are
              possibly signed decimal counts of seconds since the Epoch
              (1970-01-01 00:00:00 UTC).  Omitted counts default to extreme
              values.  The output files use UT offset 0 and abbreviation “-00”
              in place of the omitted timestamp data.  For example, “zic -r
              @0” omits data intended for negative timestamps (i.e., before
              the Epoch), and “zic -r @0/@2147483648” outputs data intended
              only for nonnegative timestamps that fit into 31-bit signed
              integers.  On platforms with GNU date, “zic -r @$(date +%s)”
              omits data intended for past timestamps.  Although this option
              typically reduces the output file's size, the size can increase
              due to the need to represent the timestamp range boundaries,
              particularly if hi causes a TZif file to contain explicit
              entries for pre-hi transitions rather than concisely
              representing them with a proleptic TZ string.  Also see the -b
              slim option for another way to shrink output size.

       -R @hi Generate redundant trailing explicit transitions for timestamps
              that occur less than hi seconds since the Epoch, even though the
              transitions could be more concisely represented via the
              proleptic TZ string.  This option does not affect the
              represented timestamps.  Although it accommodates nonstandard
              TZif readers that ignore the proleptic TZ string, it increases
              the size of the altered output files.

       -t file
              When creating local time information, put the configuration link
              in the named file rather than in the standard location.

       -u owner[:group]
              Change the output regular files' owner and group to those
              specified.  The owner is either a user name, or an unsigned
              decimal integer user ID, or an empty string meaning no change to
              the owner.  The group is similar for group names and IDs.  This
              option does not affect directories or hard or symbolic links.
              It typically needs special privileges to change ownership, and
              is ignored on platforms that lack the notions of owners and
              groups.

       -v     Be more verbose, and complain about the following situations:

              The input specifies a link to a link, something not supported by
              some older parsers, including zic itself through release 2022e.

              A year that appears in a data file is outside the range of
              representable years.

              A time of 24:00 or more appears in the input.  Pre-1998 versions
              of zic prohibit 24:00, and pre-2007 versions prohibit times
              greater than 24:00.

              A rule goes past the start or end of the month.  Pre-2004
              versions of zic prohibit this.

              A time zone abbreviation uses a %z format.  Pre-2015 versions of
              zic do not support this.

              A timestamp contains fractional seconds.  Pre-2018 versions of
              zic do not support this.

              The input contains abbreviations that are mishandled by pre-2018
              versions of zic due to a longstanding coding bug.  These
              abbreviations include “L” for “Link”, “mi” for “min”, “Sa” for
              “Sat”, and “Su” for “Sun”.

              The output file does not contain all the information about the
              long-term future of a timezone, because the future cannot be
              summarized as a proleptic TZ string.  For example, as of 2026
              this problem occurs for Palestine's daylight-saving rules, as
              these rules are based on predictions for when Ramadan will be
              observed, something that a proleptic TZ string cannot represent.

              The output contains data that may not be handled properly by
              client code designed for older zic output formats.  These
              compatibility issues affect only timestamps before 1970 or after
              the start of 2038.

              The output contains a truncated leap second table, which can
              cause some older TZif readers to misbehave.  This can occur if
              the -L option is used, and either an Expires line is present or
              the -r option is also used.

              The output file contains more than 1200 transitions, which may
              be mishandled by some clients.  The current reference client
              supports at most 2000 transitions; pre-2014 versions of the
              reference client support at most 1200 transitions.

              A time zone abbreviation has fewer than 3 or more than 6
              characters.  POSIX requires at least 3, and requires
              implementations to support at least 6.

              An output file name contains a byte that is not an ASCII letter,
              “-”, “/”, or “_”; or it contains a file name component that
              contains more than 14 bytes or that starts with “-”.

FILES
       Input files use the format described in this section; output files use
       tzfile(5) format.

       Input files should be text files, that is, they should be a series of
       zero or more lines, each ending in a newline byte and containing at
       most 2048 bytes counting the newline, and without any NUL bytes.  The
       input text's encoding is typically UTF-8 or ASCII; it should have a
       unibyte representation for the POSIX Portable Character Set (PPCS) and
       the encoding's non-unibyte characters should consist entirely of non-
       PPCS bytes.  Non-PPCS characters typically occur only in comments.

       Input lines are made up of fields.  Fields are separated from one
       another by one or more white space characters.  The white space
       characters are space, form feed, carriage return, newline, tab, and
       vertical tab.  Leading and trailing white space on input lines is
       ignored.  An unquoted sharp character (#) in the input introduces a
       comment which extends to the end of the line the sharp character
       appears on.  White space characters and sharp characters may be
       enclosed in double quotes (") if they're to be used as part of a field.
       Any line that is blank (after comment stripping) is ignored.  Nonblank
       lines are expected to be of one of three types: rule lines, zone lines,
       and link lines.

       Keywords are in English and are case insensitive.  They are not
       reserved words.  They appear in several contexts, and include month and
       weekday names and other words such as maximum, only, Rolling, and Zone.
       A keyword can be abbreviated by omitting all but an initial prefix; any
       abbreviation must be unambiguous in context.

       A rule line has the form

         Rule  NAME  FROM  TO    -  IN   ON       AT     SAVE   LETTER/S

       For example:

         Rule  US    1967  1973  -  Apr  lastSun  2:00w  1:00d  D

       The fields that make up a rule line are:

       NAME   Gives the name of the rule set that contains this line.  The
              name must start with a character that is neither an ASCII digit
              nor “-” nor “+”.  To allow for future extensions, an unquoted
              name should not contain characters from the set
              “!$%&'()*,/:;<=>?@[\]^`{|}~”.

       FROM   Gives the first year in which the rule applies.  Any signed
              integer year can be supplied; the proleptic Gregorian calendar
              is assumed, with year 0 preceding year 1.  Rules can describe
              times that are not representable as time values, with the
              unrepresentable times ignored; this allows rules to be portable
              among hosts with differing time value types.

       TO     Gives the final year in which the rule applies.  The word
              maximum (or an abbreviation) means the indefinite future, and
              the word only (or an abbreviation) may be used to repeat the
              value of the FROM field.

       -      Is a reserved field and should always contain “-” for
              compatibility with older versions of zic.  It was previously
              known as the TYPE field, which could contain values to allow a
              separate script to further restrict in which “types” of years
              the rule would apply.

       IN     Names the month in which the rule takes effect.  Month names may
              be abbreviated as mentioned previously; for example, January can
              appear as “January”, “JANU” or “Ja”, but not as “j” which would
              be ambiguous with both June and July.

       ON     Gives the day on which the rule takes effect.  Recognized forms
              include:

                5        the fifth of the month
                lastSun  the last Sunday in the month
                lastMon  the last Monday in the month
                Sun>=8   first Sunday on or after the eighth
                Sun<=25  last Sunday on or before the 25th

              A weekday name (e.g., Sunday) or a weekday name preceded by
              “last” (e.g., lastSunday) may be abbreviated as mentioned
              previously, e.g., “Su” for Sunday and “lastsa” for the last
              Saturday.  There must be no white space characters within the ON
              field.  The “<=” and “>=” constructs can result in a day in the
              neighboring month; for example, the IN-ON combination “Oct
              Sun>=31” stands for the first Sunday on or after October 31,
              even if that Sunday occurs in November.

       AT     Gives the time of day at which the rule takes effect, relative
              to 00:00, the start of a calendar day.  Recognized forms
              include:

                2            time in hours
                2:00         time in hours and minutes
                01:28:14     time in hours, minutes, and seconds
                00:19:32.13  time with fractional seconds
                12:00        midday, 12 hours after 00:00
                15:00        3 PM, 15 hours after 00:00
                24:00        end of day, 24 hours after 00:00
                260:00       260 hours after 00:00
                -2:30        2.5 hours before 00:00
                -            equivalent to 0

              Although zic rounds times to the nearest integer second
              (breaking ties to the even integer), the fractions may be useful
              to other applications requiring greater precision.  The source
              format does not specify any maximum precision.  Any of these
              forms may be followed by the letter w if the given time is local
              or “wall clock” time, s if the given time is standard time
              without any adjustment for daylight saving, or u (or g or z) if
              the given time is universal time; in the absence of an
              indicator, local (wall clock) time is assumed.  These forms
              ignore leap seconds; for example, if a leap second occurs at
              00:59:60 local time, “1:00” stands for 3601 seconds after local
              midnight instead of the usual 3600 seconds.  The intent is that
              a rule line describes the instants when a clock/calendar set to
              the type of time specified in the AT field would show the
              specified date and time of day.

       SAVE   Gives the amount of time to be added to local standard time when
              the rule is in effect, and whether the resulting time is
              standard or daylight saving.  This field has the same format as
              the AT field except with a different set of suffix letters: s
              for standard time and d for daylight saving time.  The suffix
              letter is typically omitted, and defaults to s if the offset is
              zero and to d otherwise.  Negative offsets are allowed; in
              Ireland, for example, daylight saving time is observed in winter
              and has a negative offset relative to Irish Standard Time.  The
              offset is merely added to standard time; for example, zic does
              not distinguish a 10:30 standard time plus an 0:30 SAVE from a
              10:00 standard time plus a 1:00 SAVE.

       LETTER/S
              Gives the “variable part” (for example, the “S” or “D” in “EST”
              or “EDT”) of time zone abbreviations to be used when this rule
              is in effect.  If this field is “-”, the variable part is null.

       A zone line has the form

         Zone  NAME        STDOFF  RULES   FORMAT  [UNTIL]

       For example:

         Zone  Asia/Amman  2:00    Jordan  EE%sT   2017 Oct 27 01:00

       The fields that make up a zone line are:

       NAME   The name of the timezone.  This is the name used in creating the
              time conversion information file for the timezone.  It should
              not contain a file name component that is empty or is “.” or
              “..”; a file name component is a maximal substring that does not
              contain “/”.  Names can be further restricted by the operating
              system or file system.  Most systems reject a name A/B if there
              is another name A, some reject names differing only in case, and
              some reject names containing special characters like “:” and
              “\”.  Other software will work better if names are limited to
              the restricted syntax described under the -v option.

       STDOFF The amount of time to add to UT to get standard time, without
              any adjustment for daylight saving.  This field has the same
              format as the AT and SAVE fields of rule lines, except without
              suffix letters; begin the field with a minus sign if time must
              be subtracted from UT.

       RULES  The name of the rules that apply in the timezone or,
              alternatively, a field in the same format as a rule-line SAVE
              field, giving the amount of time to be added to local standard
              time and whether the resulting time is standard or daylight
              saving.  Standard time applies if this field is - or for
              timestamps occurring before any rule takes effect.  When an
              amount of time is given, only the sum of standard time and this
              amount matters.

       FORMAT The format for time zone abbreviations.  The pair of characters
              %s shows where to put the time zone abbreviation's variable
              part, which is taken from the LETTER/S field of the
              corresponding rule; any timestamps that precede the earliest
              rule use the LETTER/S of the earliest standard-time rule (which
              in this case must exist).  Alternatively, a format can use the
              pair of characters %z to stand for the UT offset in the form
              ±hh, ±hhmm, or ±hhmmss, using the shortest form that does not
              lose information, where hh, mm, and ss are the hours, minutes,
              and seconds east (+) or west (-) of UT.  Alternatively, a slash
              (/) separates standard and daylight abbreviations.  To conform
              to POSIX, a time zone abbreviation should contain only
              alphanumeric ASCII characters, “+” and “-”.  By convention, the
              time zone abbreviation “-00” is a placeholder that means local
              time is unspecified.

       UNTIL  The time at which the UT offset or the rule(s) change for a
              location.  It takes the form of one to four fields YEAR [MONTH
              [DAY [TIME]]].  If this is specified, the time zone information
              is generated from the given UT offset and rule change until the
              time specified, which is interpreted using the rules in effect
              just before the transition.  The month, day, and time of day
              have the same format as the IN, ON, and AT fields of a rule;
              trailing fields can be omitted, and default to the earliest
              possible value for the missing fields.

              The next line must be a “continuation” line; this has the same
              form as a zone line except that the string “Zone” and the name
              are omitted, as the continuation line will place information
              starting at the time specified as the “until” information in the
              previous line in the file used by the previous line.
              Continuation lines may contain “until” information, just as zone
              lines do, indicating that the next line is a further
              continuation.

       If a zone changes at the same instant that a rule would otherwise take
       effect in the earlier zone or continuation line, the rule is ignored.
       A zone or continuation line L with a named rule set starts with
       standard time by default: that is, any of L's timestamps preceding L's
       earliest rule use the rule in effect after L's first transition into
       standard time.  In a single zone it is an error if two rules take
       effect at the same instant, or if two zone changes take effect at the
       same instant.

       If a continuation line subtracts N seconds from the UT offset after a
       transition that would be interpreted to be later if using the
       continuation line's UT offset and rules, the “until” time of the
       previous zone or continuation line is interpreted according to the
       continuation line's UT offset and rules, and any rule that would
       otherwise take effect in the next N seconds is instead assumed to take
       effect simultaneously.  For example:

         # Rule  NAME  FROM  TO    -  IN   ON       AT    SAVE  LETTER/S
         Rule    US    1967  2006  -  Oct  lastSun  2:00  0     S
         Rule    US    1967  1973  -  Apr  lastSun  2:00  1:00  D
         # Zone  NAME               STDOFF  RULES  FORMAT  [UNTIL]
         Zone    America/Menominee  -5:00   -      EST     1973 Apr 29 2:00
                                    -6:00   US     C%sT

       Here, an incorrect reading would be there were two clock changes on
       1973-04-29, the first from 02:00 EST (-05) to 01:00 CST (-06) according
       to the “until” value in the zone line, and the second an hour later
       from 02:00 CST (-06) to 03:00 CDT (-05) according to the values in the
       April rule line.  However, zic interprets this more sensibly as a
       single transition from 02:00 EST (-05) to 02:00 CDT (-05).

       A link line has the form

         Link  TARGET           LINK-NAME

       where TARGET and LINK-NAME both have the same syntax as a zone line's
       NAME.  For example:

         Link  Europe/Istanbul  Asia/Istanbul

       The TARGET field should appear as the NAME field in some zone line or
       as the LINK-NAME field in some link line.  The LINK-NAME field is an
       alternative name for that zone.  Links can chain together, although the
       behavior is unspecified if a chain of one or more links does not
       terminate in a Zone name.  A link line can appear before the line that
       defines the link target.  For example:

         Link  Greenwich  G_M_T
         Link  Etc/GMT    Greenwich
         Zone  Etc/GMT  0  -  GMT

       The two links are chained together, and G_M_T, Greenwich, and Etc/GMT
       all name the same zone.

       Except for continuation lines, lines may appear in any order in the
       input.  However, the behavior is unspecified if multiple zone or link
       lines define the same name.

       The file that describes leap seconds can have leap lines and an
       expiration line.  Leap lines have the following form:

         Leap  YEAR  MONTH  DAY  HH:MM:SS  CORR  R/S

       For example:

         Leap  2016  Dec    31   23:59:60  +     S

       The YEAR, MONTH, DAY, and HH:MM:SS fields tell when the leap second
       happened.  The CORR field should be “+” if a second was added or “-” if
       a second was skipped.  The R/S field should be (an abbreviation of)
       “Stationary” if the leap second time given by the other fields should
       be interpreted as UTC or (an abbreviation of) “Rolling” if the leap
       second time given by the other fields should be interpreted as local
       (wall clock) time.

       Rolling leap seconds would let one see Times Square ball drops where
       there'd be a “3... 2... 1... leap... Happy New Year” countdown, placing
       the leap second at midnight New York time rather than midnight UTC.
       Although stationary leap seconds are the common practice, rolling leap
       seconds can be useful in specialized applications like SMPTE timecodes
       that may prefer to put leap second discontinuities at the end of a
       local broadcast day.  However, rolling leap seconds are not supported
       if the -r option is used.

       The expiration line, if present, has the form:

         Expires  YEAR  MONTH  DAY  HH:MM:SS

       For example:

         Expires  2020  Dec    28   00:00:00

       The YEAR, MONTH, DAY, and HH:MM:SS fields give the expiration timestamp
       in UTC for the leap second table.

EXTENDED EXAMPLE
       Here is an extended example of zic input, intended to illustrate many
       of its features.

         # Rule  NAME  FROM  TO    -  IN   ON       AT    SAVE  LETTER/S
         Rule    Swiss 1941  1942  -  May  Mon>=1   1:00  1:00  S
         Rule    Swiss 1941  1942  -  Oct  Mon>=1   2:00  0     -
         Rule    EU    1977  1980  -  Apr  Sun>=1   1:00u 1:00  S
         Rule    EU    1977  only  -  Sep  lastSun  1:00u 0     -
         Rule    EU    1978  only  -  Oct   1       1:00u 0     -
         Rule    EU    1979  1995  -  Sep  lastSun  1:00u 0     -
         Rule    EU    1981  max   -  Mar  lastSun  1:00u 1:00  S
         Rule    EU    1996  max   -  Oct  lastSun  1:00u 0     -

         # Zone  NAME           STDOFF      RULES  FORMAT  [UNTIL]
         Zone    Europe/Zurich  0:34:08     -      LMT     1853 Jul 16
                                0:29:45.50  -      BMT     1894 Jun
                                1:00        Swiss  CE%sT   1981
                                1:00        EU     CE%sT

         Link    Europe/Zurich  Europe/Vaduz

       In this example, the EU rules are for the European Union and for its
       predecessor organization, the European Communities.  The timezone is
       named Europe/Zurich and it has the alias Europe/Vaduz.  This example
       says that Zurich was 34 minutes and 8 seconds east of UT until
       1853-07-16 at 00:00, when the legal offset was changed to 7 degrees 26
       minutes 22.50 seconds, which works out to 0:29:45.50; zic treats this
       by rounding it to 0:29:46.  After 1894-06-01 at 00:00 the UT offset
       became one hour and Swiss daylight saving rules (defined with lines
       beginning with “Rule Swiss”) apply.  From 1981 to the present, EU
       daylight saving rules have applied, and the UTC offset has remained at
       one hour.

       In 1941 and 1942, daylight saving time applied from the first Monday in
       May at 01:00 to the first Monday in October at 02:00.  The pre-1981 EU
       daylight-saving rules have no effect here, but are included for
       completeness.  Since 1981, daylight saving has begun on the last Sunday
       in March at 01:00 UTC.  Until 1995 it ended the last Sunday in
       September at 01:00 UTC, but this changed to the last Sunday in October
       starting in 1996.

       For purposes of display, “LMT” and “BMT” were initially used,
       respectively.  Since Swiss rules and later EU rules were applied, the
       time zone abbreviation has been CET for standard time and CEST for
       daylight saving time.

FILES
       /etc/localtime
              Default local timezone file.

       /usr/share/zoneinfo
              Default timezone information directory.

NOTES
       For areas with more than two types of local time, you may need to use
       local standard time in the AT field of the earliest transition time's
       rule to ensure that the earliest transition time recorded in the
       compiled file is correct.

       If, for a particular timezone, a clock advance caused by the start of
       daylight saving coincides with and is equal to a clock retreat caused
       by a change in UT offset, zic produces a single transition to daylight
       saving at the new UT offset without any change in local (wall clock)
       time.  To get separate transitions use multiple zone continuation lines
       specifying transition instants using universal time.

SEE ALSO
       tzfile(5), zdump(8)

Time Zone Database                                                      zic(8)
zdump.8.txt0000644000000000000000000001627315105202465010063 0ustar00zdump(8)                    System Manager's Manual                   zdump(8)

NAME
       zdump - timezone dumper

SYNOPSIS
       zdump [ option ... ] [ timezone ... ]

DESCRIPTION
       The zdump program prints the current time in each timezone named on the
       command line.  A timezone of - is treated as if it were /dev/stdin;
       this can be used to pipe TZif data into zdump.

OPTIONS
       --version
              Output version information and exit.

       --help Output short usage message and exit.

       -i     Output a description of time intervals.  For each timezone on
              the command line, output an interval-format description of the
              timezone.  See “INTERVAL FORMAT” below.

       -v     Output a verbose description of time intervals.  For each
              timezone on the command line, print the times at the two extreme
              time values, the times (if present) at and just beyond the
              boundaries of years that localtime(3) and gmtime(3) can
              represent, and the times both one second before and exactly at
              each detected time discontinuity.  Each line is followed by
              isdst=D where D is positive, zero, or negative depending on
              whether the given time is daylight saving time, standard time,
              or an unknown time type, respectively.  Each line is also
              followed by gmtoff=N if the given local time is known to be N
              seconds east of Greenwich.

       -V     Like -v, except omit output concerning extreme time and year
              values.  This generates output that is easier to compare to that
              of implementations with different time representations.

       -c [loyear,]hiyear
              Cut off interval output at the given year(s).  Cutoff times are
              computed using the proleptic Gregorian calendar with year 0 and
              with Universal Time (UT) ignoring leap seconds.  Cutoffs are at
              the start of each year, where the lower-bound timestamp is
              inclusive and the upper is exclusive; for example, -c 1970,2070
              selects transitions on or after 1970-01-01 00:00:00 UTC and
              before 2070-01-01 00:00:00 UTC.  The default cutoff is
              -500,2500.

       -t [lotime,]hitime
              Cut off interval output at the given time(s), given in decimal
              seconds since 1970-01-01 00:00:00 Coordinated Universal Time
              (UTC).  The timezone determines whether the count includes leap
              seconds.  As with -c, the cutoff's lower bound is inclusive and
              its upper bound is exclusive.

INTERVAL FORMAT
       The interval format is a compact text representation that is intended
       to be both human- and machine-readable.  It consists of an empty line,
       then a line “TZ=string” where string is a double-quoted string giving
       the timezone, a second line “- - interval” describing the time interval
       before the first transition if any, and zero or more following lines
       “date time interval”, one line for each transition time and following
       interval.  Fields are separated by single tabs.

       Dates are in yyyy-mm-dd format and times are in 24-hour hh:mm:ss format
       where hh<24.  Times are in local time immediately after the transition.
       A time interval description consists of a UT offset in signed ±hhmmss
       format, a time zone abbreviation, and an isdst flag.  An abbreviation
       that equals the UT offset is omitted; other abbreviations are double-
       quoted strings unless they consist of one or more alphabetic
       characters.  An isdst flag is omitted for standard time, and otherwise
       is a decimal integer that is unsigned and positive (typically 1) for
       daylight saving time and negative for unknown.

       In times and in UT offsets with absolute value less than 100 hours, the
       seconds are omitted if they are zero, and the minutes are also omitted
       if they are also zero.  Positive UT offsets are east of Greenwich.  The
       UT offset -00 denotes a UT placeholder in areas where the actual offset
       is unspecified; by convention, this occurs when the UT offset is zero
       and the time zone abbreviation begins with “-” or is “zzz”.

       In double-quoted strings, escape sequences represent unusual
       characters.  The escape sequences are \s for space, and \", \\, \f, \n,
       \r, \t, and \v with their usual meaning in the C programming language.
       E.g., the double-quoted string “"CET\s\"\\"” represents the character
       sequence “CET "\”.

       Here is an example of the output, with the leading empty line omitted.
       (This example is shown with tab stops set far enough apart so that the
       tabbed columns line up.)

         TZ="Pacific/Honolulu"
         -           -         -103126  LMT
         1896-01-13  12:01:26  -1030    HST
         1933-04-30  03        -0930    HDT  1
         1933-05-21  11        -1030    HST
         1942-02-09  03        -0930    HWT  1
         1945-08-14  13:30     -0930    HPT  1
         1945-09-30  01        -1030    HST
         1947-06-08  02:30     -10      HST

       Here, local time begins 10 hours, 31 minutes and 26 seconds west of UT,
       and is a standard time abbreviated LMT.  Immediately after the first
       transition, the date is 1896-01-13 and the time is 12:01:26, and the
       following time interval is 10.5 hours west of UT, a standard time
       abbreviated HST.  Immediately after the second transition, the date is
       1933-04-30 and the time is 03:00:00 and the following time interval is
       9.5 hours west of UT, is abbreviated HDT, and is daylight saving time.
       Immediately after the last transition the date is 1947-06-08 and the
       time is 02:30:00, and the following time interval is 10 hours west of
       UT, a standard time abbreviated HST.

       Here are excerpts from another example:

         TZ="Europe/Astrakhan"
         -           -         +031212  LMT
         1924-04-30  23:47:48  +03
         1930-06-21  01        +04
         1981-04-01  01        +05           1
         1981-09-30  23        +04
         ...
         2014-10-26  01        +03
         2016-03-27  03        +04

       This time zone is east of UT, so its UT offsets are positive.  Also,
       many of its time zone abbreviations are omitted since they duplicate
       the text of the UT offset.

LIMITATIONS
       Time discontinuities are found by sampling the results returned by
       localtime(3) at twelve-hour intervals.  This works in all real-world
       cases; one can construct artificial time zones for which this fails.

       In the -v and -V output, “UT” denotes the value returned by gmtime(3),
       which uses UTC for modern timestamps and some other UT flavor for
       timestamps that predate the introduction of UTC.  No attempt is
       currently made to have the output use “UTC” for newer and “UT” for
       older timestamps, partly because the exact date of the introduction of
       UTC is problematic.

SEE ALSO
       tzfile(5), zic(8)

Time Zone Database                                                    zdump(8)
date.1.txt0000644000000000000000000000340415134527374007635 0ustar00date(1)                     General Commands Manual                    date(1)

NAME
       date - show and set date and time

SYNOPSIS
       date [ -u ] [ -c ] [ -r seconds ] [ +format ] [ [yyyy]mmddhhmm[yy][.ss]
       ]

DESCRIPTION
       The date command without arguments writes the date and time to the
       standard output in the form
                            Sat Mar  8 14:54:40 EST 2025
       with EST replaced by the local time zone's abbreviation (or by the
       abbreviation for the time zone specified in the TZ environment variable
       if set).  The exact output format depends on the locale.

       If a command-line argument starts with a plus sign (“+”), the rest of
       the argument is used as a format that is processed by strftime(3) to
       determine what to output; a newline character is appended.  For
       example, the shell command:
                            date +"%Y-%m-%d %H:%M:%S %z"
       outputs a line like “2025-03-08 14:54:40 -0500” instead.

       These options are available:

       -u or -c
              Use Universal Time when setting and showing the date and time.

       -r seconds
              Output the date that corresponds to seconds past the epoch of
              1970-01-01 00:00:00 UTC, where seconds should be an integer,
              either decimal, octal (leading 0), or hexadecimal (leading 0x),
              preceded by an optional sign.

FILES
       /etc/localtime               local timezone file
       /usr/lib/locale/L/LC_TIME    description of time locale L
       /usr/share/zoneinfo          timezone directory
       /usr/share/zoneinfo/Etc/UTC  for UTC leap seconds, if supported

SEE ALSO
       strftime(3).

Time Zone Database                                                     date(1)
tz-art.html0000644000000000000000000006066315243630342010130 0ustar00





Time and the arts





Time and the arts


Documentaries



Movies


    

  • 
In the 1946 movie A Matter of Life and Death
(U.S. title Stairway to Heaven)
there is a reference to British Double Summer Time.
The time does not play a large part in the plot;
it’s just a passing reference to the time when one of the
characters was supposed to have died (but didn’t).
(IMDb entry.)
(Dave Cantor)

    • 

  • 
The 1953 railway comedy movie The Titfield Thunderbolt includes a
play on words on British Double Summer Time. Valentine’s wife wants
him to leave the pub and asks him, “Do you know what time it is?”
And he, happy where he is, replies: “Yes, my love. Summer double time.”
(IMDb entry.)
(Mark Brader, 2009-10-02)

    • 

  • 
The premise of the 1999 caper movie Entrapment involves computers
in an international banking network being shut down briefly at
midnight in each time zone to avoid any problems at the transition
from the year 1999 to 2000 in that zone. (Hmmmm.) If this shutdown
is extended by 10 seconds, it will create a one-time opportunity for
a gigantic computerized theft. To achieve this, at one location the
crooks interfere with the microwave system supplying time signals to
the computer, advancing the time by 0.1 second each minute over the
last hour of 1999. (So this movie teaches us that 0.1 × 60 = 10.)
(IMDb entry.)
(Mark Brader, 2009-10-02)

    • 

  • 
One mustn’t forget the
trailer
(2014; 2:23) for the movie Daylight Saving.

    • 



TV episodes


    

  • 
An episode of The Adventures of Superman entitled “The Mysterious
Cube”, first aired 1958-02-24, had Superman convincing the controllers
of the Arlington Time Signal to broadcast ahead of actual time;
doing so got a crook trying to be declared dead to
emerge a bit too early from the titular enclosure.
(IMDb entry.)

    • 

  • 
“The Chimes
of Big Ben”, The Prisoner, episode 2, ITC, 1967-10-06.
Our protagonist tumbles to
the fraudulent nature of a Poland-to-England escape upon hearing Big
Ben chiming on Polish local time.
(IMDb entry.)

    • 

  • 
“The Susie”, Seinfeld, season 8, episode 15, NBC, 1997-02-13.
Kramer decides that daylight saving time
isn’t coming fast enough, so he sets his watch ahead an hour.

    • 

  • 
“20 Hours in America”, The West Wing, season 4, episodes 1–2,
2002-09-25, contained a scene that
saw White House staffers stranded in Indiana; they thought they had time to
catch Air Force One but were done in by intra-Indiana local time changes.

    • 

  • 
“In what time zone would you find New York City?” was a $200 question on
the 1999-11-13 United States airing of Who Wants to Be a Millionaire?,
and “In 1883, what industry led the movement to divide the U.S. into four time
zones?” was a $32,000 question on the 2001-05-23 United States airing of
the same show. At this rate, the million-dollar time-zone
question should have been asked 2002-06-04.

    • 

  • 
A private jet’s mid-flight change of time zones distorts Alison Dubois’
premonition in the “We Had a Dream” episode of Medium
(originally aired 2007-02-28).

    • 

  • 
A criminal’s failure to account for the start of daylight saving is pivotal
in “Mr. Monk
and the Rapper” (first aired 2007-07-20).

    • 

  • 
In the 30 Rock episode “Anna Howard Shaw Day”
(first broadcast 2010-02-11),
Jack Donaghy’s date realizes that a Geneva-to-New-York business phone call
received in the evening must be fake given the difference in local times.

    • 

  • 
In the “Run by the Monkeys” episode of Da Vinci’s Inquest
(first broadcast 2002-11-17),
a witness in a five-year-old fire case realizes they may not have set
their clock back when daylight saving ended on the day of the fire,
introducing the possibility of an hour when arson might have occurred.

    • 

  • 
In “The Todd Couple” episode of Outsourced (first aired 2011-02-10),
Manmeet sets up Valentine’s Day teledates for 6:00 and 9:00pm;
since one is with a New Yorker and the other with a San Franciscan,
hilarity ensues.
(Never mind that this should be 7:30am in Mumbai, yet for some reason the show
proceeds as though it’s also mid-evening there.)

    • 

  • 
In the “14 Days to Go”/“T Minus...” episode of
You, Me and the Apocalypse
(first aired 2015-11-11 in the UK, 2016-03-10 in the US),
the success of a mission to deal with a comet
hinges on whether or not Russia observes daylight saving time.
(In the US,
the episode first aired in the week before the switch to DST.)

    • 

  • 
“The Lost Hour”, Eerie, Indiana, episode 10, NBC, 1991-12-01.
Despite Indiana’s then-lack of DST,
Marshall changes his clock with unusual consequences.
See “Eerie,
Indiana was a few dimensions ahead of its time”.

    • 

  • 
“Time Tunnel”, The Adventures of Pete & Pete,
season 2, episode 5, Nickelodeon, 1994-10-23.
The two Petes travel back in time an hour
on the day that DST ends.

    • 

  • 
“King-Size Homer”, The Simpsons, episode 135, Fox, 1995-11-05.
Homer, working from home, remarks “8:58, first
time I’ve ever been early for work. Except for all those daylight
savings days. Lousy farmers.”

    • 

  • 
Last Week Tonight with John Oliver, season 2, episode 5, 2015-03-08,
asked, “Daylight
Saving Time – How Is This Still A Thing?”

    • 

  • 
“Tracks”, The Good Wife, season 7, episode 12,
CBS, 2016-01-17.
The applicability of a contract hinges on the
time zone associated with a video timestamp.

    • 

  • 
“Justice”, Veep, season 6, episode 4, HBO, 2017-05-07.
Jonah’s inability to understand DST ends up impressing a wealthy
backer who sets him up for a 2020 presidential run.

    • 

  • 
“In the Zone”, The Magic School Bus Rides Again, season 3,
episode 3, Netflix, 2020-12-26.
During a New Year’s Eve journey, a software update glitch causes the
Magic School Bus and its passengers to transform every time they cross
time zone boundaries. The visit to the South Pole is especially tricky.

    • 



Books, plays, and magazines


    

  • 
Jules Verne, Around the World in Eighty Days
(Le tour du monde en quatre-vingts jours), 1873.
Wall-clock time plays a central role in the plot.
European readers of the 1870s clearly held the U.S. press in
deep contempt; the protagonists cross the U.S. without once
reading a paper.
Available versions include
an English
translation, and
the original French
“with illustrations from the original 1873 French-language edition”.

    • 

  • 
Nick Enright, Daylight Saving, 1989.
A fast-paced comedy about love and loneliness as the clocks turn back.

    • 

  • 
Umberto Eco,
The
Island of the Day Before
(L’isola del giorno prima), 1994.
“...the story of a 17th century Italian nobleman trapped near an island
on the International Date Line. Time and time zones play an integral
part in the novel.” (Paul Eggert, 2006-04-22)

    • 

  • 
John Dunning, Two
O’Clock, Eastern Wartime, 2001.
Mystery, history, daylight saving time, and old-time radio.

    • 

  • 
Surrealist artist Guy Billout’s work “Date Line”
appeared on page 103 of the 1999-11 Atlantic Monthly.

    • 

  • 
“Gloom, Gloom, Go Away” by Walter Kirn appeared on page 106 of Time
magazine’s 2002-11-11 issue; among other things, it proposed
year-round DST as a way of lessening wintertime despair.

    • 

  • 
Cory Doctorow, Eastern Standard Tribe,
2004. The world splinters into tribes characterized by their timezones.

    • 



Music


    

  • 
Recordings of “Save That Time”, Russ Long, Serrob Publishing, BMI:

      

    • 
Karrin Allyson, I Didn’t Know About You (1993), track 11, 3:44.
Concord Jazz CCD-4543.
Karrin Allyson, vocal;
Russ Long, piano;
Gerald Spaits, bass;
Todd Strait, drums.
CD notes “additional lyric by Karrin Allyson;
arranged by Russ Long and Karrin Allyson”.
ADO ★,
AMG
★★★★, Penguin ★★★⯪.

      • 

    • 
Kevin Mahogany, Double Rainbow (1993), track 3, 6:27. Enja ENJ-7097 2.
Kevin Mahogany, vocal;
Kenny Barron, piano;
Ray Drummond, bass;
Ralph Moore, tenor saxophone;
Lewis Nash, drums.
ADO ★⯪,
AMG
★★★, Penguin ★★★.

      • 

    • 
Joe Williams, Here’s to Life (1994), track 7, 3:58.
Telarc Jazz CD-83357.
Joe Williams, vocal; The Robert Farnon [39 piece] Orchestra.
Also in a 3-CD package “Triple Play”, Telarc CD-83461.
ADO •,
AMG
★★, Penguin ★★★.

      • 

    • 
Charles Fambrough, Keeper of the Spirit (1995), track 7, 7:07.
AudioQuest AQ-CD1033.
Charles Fambrough, bass;
Joel Levine, tenor recorder;
Edward Simon, piano;
Lenny White, drums;
Marion Simon, percussion.
ADO ★,
AMG
unrated, Penguin ★★★.

    

    • 

  • 
Holly Cole Trio, Blame It On My Youth (1992). Manhattan CDP 7 97349 2, 37:45.
Holly Cole, voice;
Aaron Davis, piano;
David Piltch, string bass.
Lyrical reference to “Eastern Standard Time” in
Tom Waits’s “Purple Avenue”.
ADO ★★⯪,
AMG
★★★, Penguin unrated.

    • 

  • 
Milt Hinton,
Old
Man Time (1990).
Chiaroscuro CR(D) 310, 149:38 (two CDs).
Milt Hinton, bass;
Doc Cheatham, Dizzy Gillespie, Clark Terry, trumpet;
Al Grey, trombone;
Eddie Barefield, Joe Camel (Flip Phillips), Buddy Tate,
clarinet and saxophone;
John Bunch, Red Richards, Norman Simmons, Derek Smith,
Ralph Sutton, piano;
Danny Barker, Al Casey, guitar;
Gus Johnson, Gerryck King, Bob Rosengarden, Jackie Williams,
drums;
Lionel Hampton, vibraphone;
Cab Calloway, Joe Williams, vocal;
Buck Clayton, arrangements.
Tunes include “Old Man Time”, “Time After Time”,
“Sometimes I’m Happy”,
“A Hot Time in the Old Town Tonight”,
“Four or Five Times”, “Now’s the Time”,
“Time on My Hands”, “This Time It’s Us”,
and “Good Time Charlie”.
ADO ★★★,
AMG
★★★★⯪, Penguin ★★★.

    • 

  • 
Alan Broadbent, Pacific Standard Time (1995).
Concord Jazz CCD-4664, 62:42.
Alan Broadbent, piano;
Putter Smith, Bass;
Frank Gibson, Jr., drums.
The CD cover features an analemma for equation-of-time fans.
ADO ★,
AMG
★★★★, Penguin ★★★⯪.

    • 

  • 
Anthony Braxton/Richard Teitelbaum, Silence/Time Zones (1996).
Black Lion BLCD 760221, 72:58.
Anthony Braxton, sopranino and alto saxophones,
contrebasse clarinet, miscellaneous instruments;
Leo Smith, trumpet and miscellaneous instruments;
Leroy Jenkins, violin and miscellaneous instruments;
Richard Teitelbaum, modular moog and micromoog synthesizer.
ADO •,
AMG
★★★★.

    • 

  • 
Charles Gayle, Time Zones (2006). Tompkins Square TSQ2839, 49:06.
Charles Gayle, piano.
ADO ★,
AMG
★★★★⯪.

    • 

  • 
The Get Up Kids, Eudora (2001). Vagrant 357, 65:12.
Includes the song “Central Standard Time”.
Thanks to Colin Bowern for this information.
AMG
★★⯪.

    • 

  • 
Coldplay, “Clocks” (2003).
Capitol 52608, 4:13.
Won the 2004 Record of the Year honor at the
Grammy Awards. Co-written and performed by Chris Martin,
great-great-grandson of DST inventor William Willett.
The song’s first line is “Lights go out and I can’t be saved”.

    • 

  • 
Jaime Guevara, “Qué
hora es” (1993), 3:04.
The song protested “Sixto Hour” in Ecuador
(1992–3). Its lyrics include “Amanecía en mitad de la noche, los
guaguas iban a clase sin sol” (“It was dawning in the middle of the
night, the buses went to class without sun”).

    • 

  • 
Irving Kahal and Harry Richman,
“There Ought to be a Moonlight Saving Time” (1931).
This musical standard was a No. 1 hit for Guy Lombardo
in 1931, and was also performed by Maurice Chevalier, Blossom Dearie
and many others. The phrase “Moonlight saving time” also appears in
the 1995 country song “Not Enough Hours in the Night” written by Aaron
Barker, Kim Williams and Rob Harbin and performed by Doug
Supernaw.

    • 

  • 
The Microscopic Septet, Lobster Leaps In (2008).
Cuneiform 272, 73:05.
Includes the song “Twilight Time Zone”.
ADO ★★,
AMG
★★★⯪.

    • 

  • 
Bob Dylan, The Times They Are a-Changin’ (1964).
Columbia CK-8905, 45:36.
ADO ★⯪,
AMG
★★★★⯪.
The title song is also available on “Bob Dylan’s Greatest Hits”
and “The Essential Bob Dylan”.

    • 

  • 
Luciana Souza, Tide (2009). Universal Jazz France B0012688-02, 42:31.
ADO ★★⯪,
AMG
★★★⯪.
Includes the song “Fire and Wood” with the lyric
“The clocks were turned back you remember/Think it’s still November.”

    • 

  • 
Ken Nordine, You’re Getting Better: The Word Jazz Dot Masters (2005).
Geffen B0005171-02, 156:22.
ADO ★,
AMG
★★★★⯪.
Includes the piece “What Time Is It”
(“He knew what time it was everywhere...that counted”).

    • 

  • 
Chicago, Chicago Transit Authority (1969). Columbia 64409, 1:16:20.
AMG ★★★★.
Includes the song “Does Anybody Really Know What Time It Is?”.

    • 

  • 
Emanuele Arciuli,
The Time Curve Preludes (2023).
Neuma 174, 44:46.
The title piece, composed by
William
Duckworth, is the first work of postminimal music.
Unlike minimalism, it does not assume that the listener has plenty of time.

    • 



Comics



Jokes


    

  • 
The idea behind daylight saving time was first proposed as a joke by
Benjamin Franklin. To enforce it, he suggested, “Every
morning, as soon as the sun rises, let all the bells in every church
be set ringing; and if that is not sufficient, let cannon be fired in
every street, to wake the sluggards effectually, and make them open
their eyes to see their true interest. All the difficulty will be in
the first two or three days: after which the reformation will be as
natural and easy as the present irregularity; for, ce n’est que le
premier pas qui coûte.”
Franklin’s
joke was first published on 1784-04-26 by the
Journal de Paris as an
anonymous letter translated into French.

    • 

  • 
“We’ve been using the five-cent nickel in this country since 1492.
Now that’s pretty near 100 years, daylight saving.”
(Groucho Marx as Captain Spaulding in Animal Crackers, 1930,
as noted by Will Fitzgerald)

    • 

  • 
BRADY. ...[Bishop Usher] determined that the Lord began the Creation
on the 23rd of October in the Year 4,004 B.C. at – uh, 9 A.M.!

    
DRUMMOND. That Eastern Standard Time? (Laughter.) Or Rocky Mountain
Time? (More laughter.) It wasn’t daylight-saving time, was it? Because
the Lord didn’t make the sun until the fourth day!

    
(From the play Inherit the Wind by Jerome Lawrence and Robert E. Lee,
filmed in 1960 with Spencer Tracy as Drummond and Fredric March as
Brady, and several other times. Thanks to Mark Brader.)

    • 

  • 
“Good news.”
“What did they do? Extend Daylight Saving Time year round?”
(Professional tanner George Hamilton, in dialog from a
May, 1999 episode of the syndicated television series Baywatch)

    • 

  • 
“A fundamental belief held by Americans is that if you are on land, you
cannot be killed by a fish...So most Americans remain on land, believing
they’re safe. Unfortunately, this belief – like so many myths, such as that
there’s a reason for ‘Daylight Saving Time’ – is false.”
(Dave Barry column, 2000-07-02)

    • 

  • 
“I once had sex for an hour and five minutes, but that was on the day
when you turn the clocks ahead.”
(Garry Shandling, 52nd Annual Emmys, 2000-09-10)

    • 

  • 
“Would it impress you if I told you I invented Daylight Savings Time?”
(“Sahjhan” to “Lilah” in dialog from the “Loyalty” episode of Angel,
originally aired 2002-02-25)

    • 

  • 
“I thought you said Tulsa was a three-hour flight.”
“Well, you’re forgetting about the time difference.”
(“Joey” and “Chandler” in dialog from the
episode of Friends entitled “The One With
Rachel’s Phone Number”, originally aired 2002-12-05)

    • 

  • 
“Is that a pertinent fact,
or are you just trying to dazzle me with your command of time zones?”
(Kelsey Grammer as “Frasier Crane” to “Roz”
from the episode of Frasier entitled “The Kid”,
originally aired 1997-11-04)

    • 

  • 
“I put myself and my staff through this crazy, huge ordeal, all because
I refused to go on at midnight, okay? And so I work, you know, and
then I get this job at eleven, supposed to be a big deal. Then
yesterday daylight [saving] time ended. Right now it’s basically midnight.”
(Conan O’Brien on the 2010-11-08 premiere of Conan)

    • 

  • 
“The best method, I told folks, was to hang a large clock high on a
barn wall where all the cows could see it. If you have Holsteins, you
will need to use an analog clock.” (Jerry Nelson, “How
to adjust dairy cows to daylight saving time”,
Successful Farming, 2017-10-09)

    • 

  • 
“And now, driving to California, I find that I must enter a password
in order to change the time zone on my laptop clock. Evidently,
someone is out to mess up my schedule and my clock must be secured.”
(Garrison Keillor,
“We’ve
never been here before”, 2017-08-22)

    • 

  • 
“Well, in my time zone that’s all the time I have,
but maybe in your time zone I haven’t finished yet. So stay tuned!”
(Goldie Hawn, Rowan & Martin’s Laugh-In No. 65, 1970-03-09)

    • 











This web page is in the public domain, so clarified as of
2009-05-17 by Arthur David Olson.


Please send corrections to this web page to the
time zone mailing list.
The mailing list and its archives are public,
so please do not send confidential information.




tz-how-to.html0000644000000000000000000005552615243644723010571 0ustar00


How to Read the tz Database








How to Read the tz
Database Source Files


by Bill Seymour


This guide uses the America/Chicago and
Pacific/Honolulu zones as examples of how to infer
times of day from the tz database
source files. It might be helpful, but not absolutely necessary,
for the reader to have already downloaded the
latest release of the database and become familiar with the basic layout
of the data files. The format is explained in the “man page”
for the zic compiler, zic.8.txt, in
the code subdirectory.
Although this guide covers many of the common cases, it is not a
complete summary of what zic accepts; the man page is the
authoritative reference.



We’ll begin by talking about the rules for changing between standard
and daylight saving time since we’ll need that information when we talk
about the zones.



First, let’s consider the special daylight saving time rules
for Chicago (from the northamerica file in
the data subdirectory):



#Rule NAME    FROM TO    -   IN  ON      AT   SAVE LETTER
Rule  Chicago 1920 only  -   Jun 13      2:00 1:00 D
Rule  Chicago 1920 1921  -   Oct lastSun 2:00 0    S
Rule  Chicago 1921 only  -   Mar lastSun 2:00 1:00 D
Rule  Chicago 1922 1966  -   Apr lastSun 2:00 1:00 D
Rule  Chicago 1922 1954  -   Sep lastSun 2:00 0    S
Rule  Chicago 1955 1966  -   Oct lastSun 2:00 0    S






  Reformatted a Bit




  From
  To
  On
  At
  Action




  1920 only
  June 13th
  02:00 local
  go to daylight saving time




  1920
  1921
  last Sunday
  in October
  return to standard time




  1921 only
  in March
  go to daylight saving time




  1922
  1966
  in April




  1954
  in September
  return to standard time




  1955
  1966
  in October






The FROM and TO columns, respectively, specify the
first and last calendar years defining a contiguous range over which a specific
Rule line is to apply.  The keyword only can be used in the
TO field to repeat the value of the FROM field in the
event that a rule should only apply to a single year.  Often, the keyword
max is used to extend a rule’s application into the
indefinite future; it is a platform-agnostic stand-in for the largest
representable year.


The next column, -, is reserved; for compatibility with earlier
releases, it always contains a hyphen, which acts as a kind of null value.
Prior to the 2020b release, it was called the TYPE field, though
it had not been used in the main data since the 2000e release.
An obsolescent supplementary file used the
field as a proof-of-concept to allow zic to apply a given Rule
line only to certain “types” of years within the specified range as
dictated by the output of a separate script, such as: only years which would
have a US presidential election, or only years which wouldn’t.


The SAVE column contains the local (wall clock) offset from
local standard time.
This is usually either zero for standard time or one hour for daylight
saving time; but there’s no reason, in principle, why it can’t
take on other values.


The LETTER (sometimes called LETTER/S)
column can contain a variable
part of the usual abbreviation of the time zone’s name, or it can just
be a hyphen if there’s no variable part. For example, the abbreviation
used in the central time zone will be either “CST” or “CDT”.
The variable part is ‘S’ or ‘D’;
and, sure enough, that’s just what we find in
the LETTER column
in the Chicago rules. More about this when we talk about
“Zone” lines.


One important thing to notice is that “Rule” lines
want at once to be both transitions and steady states:

    

  • On the one hand, they represent transitions between standard and
daylight saving time; and any number of Rule lines can be in effect
during a given period (which will always be a non-empty set of
contiguous calendar years).
    • 

  • On the other hand, the SAVE and LETTER
columns contain state that exists between transitions. More about this
when we talk about the US rules.
    • 




In the example above, the transition to daylight saving time
happened on the 13th of June in 1920, and on
the last Sunday in March in 1921; but the return to standard time
happened on the last Sunday in October in both of those
years. Similarly, the rule for changing to daylight saving time was
the same from 1922 to 1966; but the rule for returning to standard
time changed in 1955. Got it?



OK, now for the somewhat more interesting “US” rules:



#Rule NAME FROM TO    -   IN  ON        AT   SAVE LETTER/S
Rule  US   1918 1919  -   Mar lastSun  2:00  1:00 D
Rule  US   1918 1919  -   Oct lastSun  2:00  0    S
Rule  US   1942 only  -   Feb 9        2:00  1:00 W # War
Rule  US   1945 only  -   Aug 14      23:00u 1:00 P # Peace
Rule  US   1945 only  -   Sep 30       2:00  0    S
Rule  US   1967 2006  -   Oct lastSun  2:00  0    S
Rule  US   1967 1973  -   Apr lastSun  2:00  1:00 D
Rule  US   1974 only  -   Jan 6        2:00  1:00 D
Rule  US   1975 only  -   Feb 23       2:00  1:00 D
Rule  US   1976 1986  -   Apr lastSun  2:00  1:00 D
Rule  US   1987 2006  -   Apr Sun>=1   2:00  1:00 D
Rule  US   2007 max   -   Mar Sun>=8   2:00  1:00 D
Rule  US   2007 max   -   Nov Sun>=1   2:00  0    S






  Reformatted a Bit




  From
  To
  On
  At
  Action




  1918
  1919
  last Sunday
  in March
  02:00 local
  go to daylight saving time




  in October
  return to standard time




  1942 only
  February 9th
  go to “war time”




  1945 only
  August 14th
  23:00 UT
  
    rename “war time” to “peace
time;”
    clocks don’t change
  		




  September 30th
  02:00 local
  return to standard time




  1967
  2006
  last Sunday
  in October




  1973
  in April
  go to daylight saving time




  1974 only
  January 6th




  1975 only
  February 23rd




  1976
  1986
  last Sunday
  in April




  1987
  2006
  first Sunday




  2007
  present
  second Sunday in March




  first Sunday in November
  return to standard time






There are two interesting things to note here.



First, the time that something happens (in the AT
column) is not necessarily the local (wall clock) time. The time can be
suffixed with ‘s’ (for “standard”) to mean
local standard time, different from local (wall clock) time when observing
daylight saving time; or it can be suffixed with ‘g’,
‘u’, or ‘z’, all three of which mean the
standard time at the
prime meridian.
‘g’ stands for “GMT”;
‘u’ stands for “UT” or “UTC”
(whichever was official at the time); ‘z’ stands for the
nautical time zone
Z (a.k.a. “Zulu” which, in turn, stands for ‘Z’).
The time can also be suffixed with ‘w’ meaning local (wall
clock) time; but it usually isn’t because that’s the
default.



Second, the day in the ON column, in addition to
“lastSun” or a particular day of the month,
can have the form, “Sun>=x” or
“Sun<=x,” where x is a day
of the month. For example, “Sun>=8” means
“the first Sunday on or after the eighth of the month,” in
other words, the second Sunday of the month. Furthermore, although
there are no examples above, the weekday needn’t be
“Sun” in either form, but can be the usual
three-character English abbreviation for any day of the week.



And the US rules give us more examples of a couple of things
already mentioned:



    

  • The rules for changing to and from daylight saving time are
actually different sets of rules; and the two sets can change
independently. Consider, for example, that the rule for the return to
standard time stayed the same from 1967 to 2006; but the rule for the
transition to daylight saving time changed several times in the same
period.  There can also be periods, 1946 to 1966 for example, when no
rule from this group is in effect, and so either no transition
happened in those years, or some other rule is in effect (perhaps a
state or other more local rule).
    • 


  • The SAVE and LETTER columns
contain steady state, not transitions. Consider, for example,
the transition from “war time” to “peace time”
that happened on August 14, 1945. The “1:00” in
the SAVE column is not an instruction to advance
the clock an hour. It means that clocks should be one hour
ahead of standard time, which they already are because of the previous
rule, so there should be no change.
    • 





OK, now let’s look at a Zone record:



#Zone       NAME      STDOFF   RULES FORMAT [UNTIL]
Zone  America/Chicago -5:50:36 -       LMT  1883 Nov 18 12:09:24
                      -6:00    US      C%sT 1920
                      -6:00    Chicago C%sT 1936 Mar  1  2:00
                      -5:00    -       EST  1936 Nov 15  2:00
                      -6:00    Chicago C%sT 1942
                      -6:00    US      C%sT 1946
                      -6:00    Chicago C%sT 1967
                      -6:00    US      C%sT






  Columns Renamed




  Standard Offset

    from Prime
    Meridian		
  Daylight
Saving Time		
  Abbreviation(s)
  Ending at Local Time




  Date
  Time




  −5:50:36
  not observed
  LMT
  1883-11-18
  12:09:24




  −6:00:00
  US rules
  CST or CDT
  1920-01-01
  00:00:00




  Chicago rules
  1936-03-01
  02:00:00




  −5:00:00
  not observed
  EST
  1936-11-15




  −6:00:00
  Chicago rules
  CST or CDT
  1942-01-01
  00:00:00




  US rules
  CST, CWT or CPT
  1946-01-01




  Chicago rules
  CST or CDT
  1967-01-01




  US rules
  –






There are a couple of interesting differences between Zones and Rules.



First, and somewhat trivially, whereas Rules are considered to
contain one or more records, a Zone is considered to be a single
record with zero or more continuation lines. Thus, the keyword,
“Zone,” and the zone name are not
repeated. The last line is the one without anything in
the [UNTIL] column.



Second, and more fundamentally, each line of a Zone represents a
steady state, not a transition between states. The state exists from
the date and time in the previous line’s [UNTIL]
column up to the date and time in the current
line’s [UNTIL] column. In other words, the date and
time in the [UNTIL] column is the instant that separates
this state from the next. Where that would be ambiguous because
we’re setting our clocks back, the [UNTIL] column
specifies the first occurrence of the instant. The state specified by
the last line, the one without anything in the [UNTIL]
column, continues to the present.



The first line typically specifies the mean solar time observed
before the introduction of standard time. Since there’s no line before
that, it has no beginning. 8-)  For some places near the International
Date Line, the first two lines will show solar times
differing by 24 hours; this corresponds to a movement of the Date
Line.  For example:



#Zone NAME          STDOFF   RULES FORMAT [UNTIL]
Zone America/Juneau 15:02:19 -     LMT    1867 Oct 19  0:31:13u
                    -8:57:41 -     LMT    ...




When Alaska was purchased from Russia in 1867, the Date Line moved
from the Alaska/Canada border to the Bering Strait; and the time in
Alaska was then 24 hours earlier than it had
been. <aside>(This also means that, although the sun had set
on Friday 6 October and risen on Saturday 7 October in the Julian calendar,
which Russia was still using then for religious reasons, upon the transfer
Saturday afternoon became Friday afternoon, and then the sun set and rose once
more on second instances of the same days with different names,
Friday 18 and Saturday 19 October in the Gregorian calendar. Isn’t civil time
wonderful? 8-))</aside>



The abbreviation, “LMT” stands for “local mean
time”, which is an invention of
the tz
database and was probably never actually used during the
period. Furthermore, the value is almost certainly wrong except in the
archetypal place after which the zone is named. (The tz database
usually doesn’t provide a separate Zone record for places where
nothing significant happened after 1970.)



The RULES column tells us whether daylight saving time is being observed:

    

  • A hyphen, a kind of null value, means that we have not set our
clocks ahead of standard time.
    • 


  • An amount of time (usually but not necessarily “1:00”
meaning one hour) means that we have set our clocks ahead by that
amount.
    • 


  • Some alphabetic string means that we might have set our
clocks ahead; and we need to check the rule the name of which is the
given alphabetic string.
    • 




An example of a specific amount of time is:


#Zone NAME            STDOFF RULES FORMAT [UNTIL]
Zone Pacific/Honolulu ...                 1933 Apr 30  2:00
                      -10:30 1:00  HDT    1933 May 21 12:00
                      ...




Hawaii tried daylight saving time for three weeks in 1933 and
decided they didn’t like it. 8-) Note that
the STDOFF column always contains the standard time
offset, so the local (wall clock) time during this period was GMT −
10:30 + 1:00 = GMT − 9:30.



The FORMAT column specifies the usual abbreviation of
the time zone name. It should have one of four forms:


    


  • a time zone abbreviation that is a string of three or more
characters that are either ASCII alphanumerics,
“+”, or “-”
    • 


  • the string “%z”, in which case the
“%z” will be replaced by a numeric time zone
abbreviation
    • 


  • a pair of time zone abbreviations separated by a slash
(‘/’), in which case the first string is the
abbreviation for the standard time name and the second string is the
abbreviation for the daylight saving time name
    • 


  • a string containing “%s”, in which case
the “%s” will be replaced by the text in the
appropriate Rule’s LETTER column, and the resulting
string should be a time zone abbreviation
    • 




The last two make sense only if there’s a named rule in effect.



An example of a slash is:


#Zone NAME          STDOFF RULES FORMAT  [UNTIL]
Zone  Europe/London ...                  1996
                    0:00   EU    GMT/BST




The current time in the UK is called either Greenwich mean time or
British summer time.



One wrinkle, not fully explained in zic.8.txt, is what
happens when switching to a named rule. To what values should
the SAVE and LETTER data be initialized?



    

  • If at least one transition has happened, use
the SAVE and LETTER data from the most
recent.
    • 


  • If switching to a named rule before any transition has happened,
assume standard time (SAVE zero), and use
the LETTER data from the earliest transition with
a SAVE of zero.





And three last things about the FORMAT column:


    


  • The tz
database gives abbreviations for time zones
in popular English-language usage. For
example, the last line in
Zone Pacific/Honolulu (shown below) gives
“HST” for “Hawaii standard time” even though the
legal
name for that time zone is “Hawaii–Aleutian standard time”.
This author has read that there are also some places in Australia where
popular time zone names differ from the legal ones.


  • No attempt is made to localize
the abbreviations. They are intended to be the values returned through the
"%Z" format specifier to
C’s
strftime
function in the
“C” locale.


  • If there is no generally accepted abbreviation for a time zone,
a numeric offset is used instead, e.g., +07 for 7 hours
ahead of Greenwich. By convention, -00 is used in a
zone while uninhabited, where the offset is zero but in some sense
the true offset is undefined.




As a final example, here’s the complete history for Hawaii:



# Relevant Excerpts from the US Rules
#Rule NAME FROM TO   -    IN  ON      AT     SAVE LETTER/S
Rule  US   1918 1919 -    Oct lastSun  2:00  0    S
Rule  US   1942 only -    Feb  9       2:00  1:00 W # War
Rule  US   1945 only -    Aug 14      23:00u 1:00 P # Peace
Rule  US   1945 only -    Sep lastSun  2:00  0    S

# The Zone Record
#Zone NAME            STDOFF    RULES FORMAT [UNTIL]
Zone Pacific/Honolulu -10:31:26 -     LMT    1896 Jan 13 12:00
                      -10:30    -     HST    1933 Apr 30  2:00
                      -10:30    1:00  HDT    1933 May 21  2:00
                      -10:30    US    H%sT   1947 Jun  8  2:00
                      -10:00    -     HST






  What We Infer




  Wall-Clock
Offset from
Prime Meridian		
  Adjust
Clocks		
  Time Zone
  Ending at Local Time




  Abbrv.
  Name
  Date
  Time




  −10:31:26
  –
  LMT
  local mean time
  1896-01-13
  12:00




  −10:30
  +0:01:26
  HST
  Hawaii standard time
  1933-04-30
  02:00




  −9:30
  +1:00
  HDT
  Hawaii daylight time
  1933-05-21
  12:00




  −10:30¹
  −1:00¹
  HST¹
  Hawaii standard time
  1942-02-09
  02:00




  −9:30
  +1:00
  HWT
  Hawaii war time
  1945-08-14
  13:30²




  0
  HPT
  Hawaii peace time
  1945-09-30
  02:00




  −10:30
  −1:00
  HST
  Hawaii standard time
  1947-06-08




  −10:00³
  +0:30³
  –




  
    ¹Switching to US rules...most recent transition (in 1919) was to standard time
  




  
    ²23:00 UT
    + (−9:30) = 13:30 local
  




  
    ³Since 1947-06-08T12:30Z,
    the civil time in Hawaii has been
    UT/UTC
    −10:00 year-round.
  






There will be a short quiz later. 8-)








This web page is in the public domain, so clarified as of
2015-10-20 by Bill Seymour.


All suggestions and corrections will be welcome; all flames will be amusing.
Mail to was at pobox dot com.




tz-link.html0000644000000000000000000017753415243101024010273 0ustar00


Time zone and daylight saving time data








Time zone and daylight saving time data



Time zone and
daylight-saving
rules are controlled by individual
governments. They are sometimes changed with little notice, and their
histories and planned futures are often recorded only fitfully. Here
is a summary of attempts to organize and record relevant data in this
area.


  


The tz database



The public-domain
time zone database contains code and data
that represent the history of local time
for many representative locations around the globe.
It is updated periodically to reflect changes made by political bodies
to time zone boundaries and daylight saving rules.
This database (known as tz,
tzdb, or zoneinfo)
is used by several implementations,
including
the
GNU
C Library (used in
GNU/Linux),
Android,
FreeBSD,
NetBSD,
OpenBSD,
ChromiumOS,
OpenHarmony,
Cygwin,
MariaDB,
musl libc,
MySQL,
webOS,
AIX,
iOS,
macOS,
Microsoft Windows,
OpenVMS,
Oracle Database,
Oracle Solaris,
and QNX.



Each main entry in the database represents a timezone
for a set of civil-time clocks that have all agreed since 1970.
Timezones are typically identified by continent or ocean and then by the
name of the largest city within the region containing the clocks.
For example, America/New_York
represents most of the US eastern time zone;
America/Phoenix represents most of Arizona, which
uses mountain time without daylight saving time (DST);
America/Detroit represents most of Michigan, which uses
eastern time but with different DST rules in 1975;
and other entries represent smaller regions like Starke County,
Indiana, which switched from central to eastern time in 1991
and switched back in 2006.
To use the database on a POSIX.1-2024
implementation set the TZ
environment variable to the location’s full name,
e.g., TZ="America/New_York".



Associated with each timezone is a history of offsets from
Universal
Time (UT), which is Greenwich Mean
Time (GMT) with days beginning at midnight;
for timestamps after 1960 this is more precisely Coordinated
Universal Time (UTC).
The database also records when daylight saving time was in use,
along with some time zone abbreviations such as EST
for Eastern Standard Time in the US.



Downloading the tz database



The following shell commands download
the latest release’s two
tarballs
to a GNU/Linux or similar host.


mkdir tzdb
cd tzdb
wget https://data.iana.org/time-zones/tzcode-latest.tar.gz
wget https://data.iana.org/time-zones/tzdata-latest.tar.gz
gzip -dc tzcode-latest.tar.gz | tar -xf -
gzip -dc tzdata-latest.tar.gz | tar -xf -



Alternatively, the following shell commands download the same
release in a single-tarball format containing extra data
useful for regression testing:


wget https://data.iana.org/time-zones/tzdb-latest.tar.lz
lzip -dc tzdb-latest.tar.lz | tar -xf -



These commands use convenience links to the latest release
of the tz database hosted by the
Time Zone Database website
of the Internet Assigned Numbers
Authority (IANA).
Older releases are in files named
tzcodeV.tar.gz,
tzdataV.tar.gz, and
tzdb-V.tar.lz,
where V is the version.
Since 1996, each version has been a four-digit year followed by
lower-case letter (a through z,
then za through zz, then zza
through zzz, and so on).
Since version 2022a, each release has been distributed in
POSIX
ustar interchange format, compressed as described above;
older releases use a nearly compatible format.
Since version 2016h, each release has contained a text file named
“version” whose first (and currently only) line is the version.
Older archived releases are
available via
HTTPS,
rsync, and
FTP.


Alternatively, a development repository of code and data can be
retrieved from GitHub via the shell
command:


git clone https://github.com/eggert/tz




Since version 2012e, each release has been tagged in development repositories.
Untagged commits are less well tested and probably contain
more errors.



After obtaining the code and data files, see the
README file for what to do next.
The code lets you compile the tz source files into
machine-readable binary files, one for each location.
The binary file format is specified by
The
Time Zone Information Format (TZif)
(Internet RFC 9636).
The code also lets
you read a TZif file and interpret timestamps for that
location.



Changes to the tz database



The tz code and data
are by no means authoritative. If you find errors, please
email changes to tz@iana.org,
the time zone mailing list.
The mailing list and its archives are public,
so please do not send confidential information.
See
the mailing
list’s main page to subscribe or to browse its archive of old messages.
Metadata for mailing list
discussions and corresponding data changes can be
generated automatically.




Changes to the tz code and data are often
propagated to clients via operating system updates, so
client tz data can often be corrected by
applying these updates. With GNU/Linux and similar systems, if your
maintenance provider has not yet adopted the
latest tz data, you can often short-circuit
the process by tailoring the generic instructions in
the tz README file and installing the latest
data yourself. System-specific instructions for installing the
latest tz data have also been published
for AIX
(and
older
AIX),
Android,
ICU,
IBM
JDK,
Joda-Time, MySQL,
Noda Time, and OpenJDK/Oracle JDK.



Since version 2013a,
sources for the tz database have been
UTF-8
text files
with lines terminated by LF,
which can be modified by common text editors such
as GNU Emacs,
gedit, and
vim.
Specialized source-file editing can be done via the
Sublime
zoneinfo package for Sublime Text and the VSCode
zoneinfo extension for Visual
Studio Code.




For further information about updates, please see
Procedures for
Maintaining the Time Zone Database (Internet RFC 6557).
More detail can be
found in Theory and pragmatics of the
tz code and data.
A0 TimeZone Migration
displays changes between recent tzdb versions.




Coordinating with governments and distributors



As discussed in
“How
Time Zones Are Coordinated”, the time zone database relies on
collaboration among governments, the time zone database volunteer
community, and data distributors downstream.



If your government plans to change its time zone boundaries or
daylight saving rules, please send email as described in
“Changes to the tz database”.
Do this well in advance,
as this will lessen confusion and will coordinate updates to many cell phones,
computers, and other devices around the world.
In your email, please cite the legislation or regulation that specifies
the change, so that it can be checked for details such as the exact times
when clock transitions occur.
It is OK if a rule change is planned to affect clocks
far into the future, as a long-planned change can easily be reverted
or otherwise altered with a year’s notice before the change would have
affected clocks.



There is no fixed schedule for tzdb releases.
However, typically a release occurs every few months.
Many downstream timezone data distributors wait for
a tzdb release before they produce an update
to time zone behavior in consumer devices and software products.
After a release, various parties must integrate, test,
and roll out an update before end users see changes.
These updates can be expensive, for both the quality
assurance process and the overall cost of shipping and installing
updates to each device’s copy of tzdb.
Updates may be batched with other updates and may take substantial
time to reach end users after a release.
Older devices may no longer be supported and thus may never be updated,
which means they will continue to use out-of-date rules.



For these reasons any rule change should be promulgated at least a
year before it affects how clocks operate; otherwise, there is a good
chance that many clocks will be wrong due to delays in propagating updates,
and that residents will be confused or even actively resist the change.
The shorter the notice, the more likely clock problems will arise; see “On
the Timing of Time Zone Changes” for examples.




Commentary on the tz database




Web sites using recent versions of the
tz database



These are listed roughly in ascending order of complexity and fanciness.





Network protocols for tz data


    

  • Time Zone
Data Distribution Service (Internet RFC 7808)
is associated with
CalDAV
(Internet RFC 7809), a calendar access protocol for
transferring time zone data by reference.
    • 

  • The iCalendar format
(Internet RFC 5545)
covers time zone
data; see its VTIMEZONE calendar component.
The iCalendar format requires specialized parsers and generators; a
variant xCal
(Internet RFC 6321) uses
XML format, and a variant
jCal
(Internet RFC 7265)
uses JSON format.
    • 




Other tz compilers


Although some of these do not fully support
tz data, in recent tzdb
distributions you can generally work around compatibility problems by
running the command make rearguard_tarballs and compiling
from the resulting tarballs instead.




Other TZif readers


    

  • The GNU C
Library
has a thread-safe TZif file reader.
This library is freely available under the LGPL
and is widely used in GNU/Linux systems.
    • 

  • GNOME’s
GLib has
a TZif file reader written in C that
creates a GTimeZone object representing sets
of UT offsets.
It is freely available under the LGPL.
    • 

  • The musl C library
also has an TZif reader.
It is freely available under the MIT license.
    • 

  • The Time Zone Database Parser for C++, mentioned above,
can be configured to read TZif binary files.
A similar approach is taken in
LLVM
std::chrono::time_zone, which is available
under the Apache License with
LLVM
Exceptions.
    • 

  • The
BDE Standard Library’s
baltzo::TimeZoneUtil component contains a C++
implementation of a TZif file reader. It is freely available under
the Apache License.
    • 

  • CCTZ is a simple C++
library that translates between UT and civil time and
can read TZif files. It is freely available under the Apache
License.
    • 

  • The Go programming language
has a TZif file reader LoadLocationFromTZData.
    • 

  • The
posix_tz_db
package contains Python code
to generate CSV and JSON tables that map
tz settings to proleptic TZ approximations.
For example, it maps "Africa/Cairo"
to "EET-2EEST,M4.5.5/0,M10.5.4/24",
an approximation valid for Cairo timestamps from 2023 on.
This can help porting to platforms that support only proleptic TZ.
The package is freely available under the MIT license.
    • 

  • Timelib is a C
library that reads TZif files and converts
timestamps from one time zone or format to another.
It is used by PHP,
HHVM,
and MongoDB.
It is freely available under the MIT license.
    • 

  • Tcl, mentioned above, also contains a
TZif file reader.
    • 

  • 
DateTime::TimeZone::Tzfile
is a TZif file reader written in Perl.
It is freely available under the same terms as Perl
(dual GPL and Artistic license).
    • 

  • Python has a zoneinfo.ZoneInfo
class that reads TZif data and creates objects
that represent tzdb timezones.
Python is freely available under the
Python Software Foundation
License.
A companion PyPI module
tzdata
supplies TZif data if the underlying system data cannot be found;
it is freely available under the Apache License.
    • 

  • The
public-domain tz.js
library contains a Python tool that
converts TZif data into
JSON-format data suitable for use
in its JavaScript library for time zone conversion. Dates before 1970
are not supported.
    • 

  • The jiff datetime library
for Rust either compiles
tz source into Rust,
or consults TZif data at runtime.
It is in the public domain.
It intends to supplant the soft-deprecated
Chrono-TZ.
    • 

  • The timezone-olson
package contains Haskell code that
parses and uses TZif data. It is freely
available under a BSD-style license.
    • 




Other tz-based time zone software




Other time zone databases




Maps




Time zone boundaries


Geographical boundaries between timezones are available
from several Internet
geolocation
services and other sources.




Civil time concepts and history




National histories of legal time




Australia


The Bureau of Meteorology publishes a list of Implementation
Dates of Daylight Savings Time within Australia.


Belgium


The Royal Observatory of Belgium maintains a table of time in
Belgium (in
Dutch and French).


Brazil


The Time Service Department of the National Observatory
records Brazil’s daylight saving time decrees (in
Portuguese).


Canada


National Research Council Canada publishes current
and some older information about time
zones and daylight saving time.


Chile


The Hydrographic and Oceanographic Service of the Chilean Navy publishes a
history of
Chile’s official time (in Spanish).


China


The Hong Kong Observatory maintains a
history of
 summer time in Hong Kong,
and Macau’s Meteorological and Geophysical Bureau maintains a similar
history for Macau.
Unfortunately the latter is incomplete and has errors.


Czech Republic


When daylight saving time starts and ends (in Czech)
summarizes and cites historical DST regulations.


Germany


The National Institute for Science and Technology maintains the Realisation
of Legal Time in Germany.


Israel


Israel Timezone Files
lists official time-change announcements and laws since 1940,
almost all in Hebrew.


Malaysia


See Singapore below.


Mexico


The Investigation and Analysis Service of the Mexican Library of
Congress has published a history of Mexican local time (in Spanish).


Netherlands


Legal time in the Netherlands (in Dutch)
covers the history of local time in the Netherlands from ancient times.


New Zealand


The Department of Internal Affairs maintains a brief History
of Daylight Saving in NZ.


Palestine


The Ministry of Telecom and Digital Economy publishes a history of clock changes (in Arabic).


Portugal


The Lisbon Astronomical Observatory publishes a
history of
legal time (in Portuguese).


Singapore


Why
is Singapore in the “Wrong” Time Zone? details the
history of legal time in Singapore and Malaysia.


United Kingdom


History of
legal time in Britain discusses in detail the country
with perhaps the best-documented history of clock adjustments.


United States


The Department of Transportation’s Recent
Time Zone Proceedings lists changes to
official written time zone boundaries, and its Time
Zones dataset maps current boundaries.
These boundaries are only for standard time, so the current map puts
all of Arizona in one time zone even though part of Arizona
observes DST and part does not.


Uruguay


The Oceanography, Hydrography, and Meteorology Service of the Uruguayan
Navy (SOHMA) publishes an annual almanac
(in Spanish).





Costs and benefits of time shifts


Various sources argue for and against daylight saving time and time
zone shifts, and many scientific studies have been conducted. This
section summarizes reviews and position statements based on
scientific literature in the area.


    

  • Carey RN, Sarma KM.
Impact of
daylight saving time on road traffic collision risk: a systematic
review.
BMJ Open. 2017;7(6):e014319. doi:10.1136/bmjopen-2016-014319.
This reviews research literature and concludes that the evidence
neither supports nor refutes road safety benefits from
shifts in time zones.
    • 

  • Havranek T, Herman D, Irsova D.
Does daylight saving save electricity? A meta-analysis.
Energy J. 2018;39(2):35–61.
doi:10.5547/01956574.39.2.thav.
This analyzes research literature and concludes, “Electricity savings
are larger for countries farther away from the equator, while
subtropical regions consume more electricity because of DST.”
    • 

  • Neumann P, von Blanckenburg K. What
time will it be? A comprehensive literature review on daylight saving time.
Time Soc. 2025;34(4):684–745.
doi:10.1177/0961463X241310562.
This reviews DST’s effects on electricity, health, crime, road
safety, and the economy, focusing on research since 2010, and concludes that
year-round standard time is preferable overall.
    • 

  • Romigi A, Franco V, Scoditti E et al.
The effects of daylight saving time and clock time transitions on sleep and
sleepiness: a systematic review. Sleep Med Rev. 2025;84:102161. doi:10.1016/j.smrv.2025.102161.
This reviews DST and DST transitions,
and concludes that they both harm sleep, health and behavior.
    • 




The following medical societies have taken positions on the
advisability of clock shifts:





Precision timekeeping


    

  • The
Science of Timekeeping is a thorough introduction
to the theory and practice of precision timekeeping.
    • 

  • The Science of
Time 2016 contains several freely readable papers.
    • 

  • NTP: The Network
Time Protocol (Internet RFC 5905)
discusses how to synchronize clocks of
Internet hosts.
    • 

  • The Huygens
family of software algorithms can achieve accuracy to a few tens of
nanoseconds in scalable server farms without special hardware.
    • 

  • The Precision
Time Protocol (IEEE 1588)
can achieve submicrosecond clock accuracy on a local area network
with special-purpose hardware.
    • 

  • Timezone
Options for DHCP
(Internet RFC 4833)
specifies a DHCP
option for a server to configure
a client’s time zone and daylight saving settings automatically.
    • 

  • Time
Scales describes astronomical time scales like
TDT,
TCG, and
TDB.
    • 

  • The IAU’s SOFA
collection contains C and Fortran
code for converting among time scales like
TAI,
TDB, TDT and
UTC. It is freely available under the
SOFA license.
    • 

  • Mars24 Sunclock
– Time on Mars describes Airy Mean Time (AMT) and the
diverse local time
scales used by each landed mission on Mars.
    • 

  • LeapSecond.com is
dedicated not only to leap seconds but to precise time and frequency
in general. It covers the state of the art in amateur timekeeping, and
how the art has progressed over the past few decades.
    • 

  • The rules for leap seconds are specified in Annex 1 (Time scales) of Standard-frequency
and time-signal emissions, International Telecommunication Union –
Radiocommunication Sector (ITU-R) Recommendation TF.460-6 (02/2002).
    • 

  • IERS
Bulletins contains official publications of the International
Earth Rotation and Reference Systems Service, which decides when leap
seconds occur.
The tz code and data support leap seconds
via an optional "right" configuration where a computer’s internal
time_t integer clock counts every TAI second,
as opposed to the default "posix" configuration
where the internal clock ignores leap seconds.
The two configurations agree for timestamps starting with 1972-01-01 00:00:00
UTC (time_t 63 072 000) and diverge for
timestamps starting with time_t 78 796 800,
which corresponds to the first leap second
1972-06-30 23:59:60 UTC in the "right" configuration,
and to
1972-07-01 00:00:00 UTC in the "posix" configuration.
In practice the two configurations also agree for timestamps before
1972 even though the historical situation is messy, partly because
neither UTC nor TAI
is well-defined for sufficiently old timestamps.
    • 

  • The
NTP Leap Second File covers the text file
leap-seconds.list, which lists the currently known leap seconds.
The IERS maintains this file, and a copy is distributed by
tzdb for use by NTP implementations like
classic
ntpd
and NTPsec.
The tz database also distributes leap second
information in a differently-formatted leapseconds text file,
as well as in the "right" configuration in binary form; for
example, right/UTC can be used
by chrony,
another NTP implementation.
    • 

  • Leap Smear
discusses how to gradually adjust POSIX clocks near a
leap second so that they disagree with UTC by at most a
half second, even though every POSIX minute has exactly
sixty seconds. This approach works with the default tz
"posix" configuration, is supported
by the NTP implementations mentioned above, supports conversion between
UTC and smeared POSIX timestamps, and is used by major
cloud service providers. However, according to
§3.7.1 of
Network Time Protocol Best Current Practices
(Internet RFC 8633), leap smearing is not suitable for
applications requiring accurate UTC or civil time,
and is intended for use only in single, well-controlled environments.
    • 

  • The LEAPSECS List covers McCarthy
and Klepczynski’s 1999 proposal to discontinue leap seconds,
discussed further in
The
leap second: its history and possible future.
UTC
might be redefined
without Leap Seconds gives pointers on this
contentious issue.
The General Conference on Weights and Measures (CGPM)
decided in 2022
to discontinue the use of leap seconds by 2035, and requested that no
discontinuous adjustments be made to UTC for at least a century.
The World Radiocommunication Conference resolved
in 2023 to cooperate with this process. A draft Resolution
C to make continuous UTC effective on 2027-05-20,
and thereby discontinue leap seconds,
has been scheduled for the 28th CGPM starting 2026-10-13 in Paris.
One proposal to implement this
would replace leap seconds with seven 13-second leap smears occurring once per
decade until 2100, with leap smears after that gradually increasing in size.
See:

However, there is still no consensus on whether this is the best way
to replace leap seconds.

    • 




Time notation


    

  • The Unicode Common Locale Data
Repository (CLDR) Project has localizations for time
zone names, abbreviations, identifiers, and formats. For example, it
contains French translations for “Eastern European Summer Time”,
“EEST”, and “Bucharest”. Its
by-type
charts show these values for many locales. Data values are available in
both LDML
(an XML format) and JSON.
    • 

  • 
A summary of
the international standard date and time notation covers
ISO
8601-1:2019 – Date and time – Representations for information
interchange – Part 1: Basic rules.
    • 

  • 
XML
Schema: Datatypes – dateTime specifies a format inspired by
ISO 8601 that is in common use in XML data.
    • 

  • §3.3 of
Internet Message Format (Internet RFC 5322)
specifies the time notation used in email and HTTP
headers.
    • 

  • 
Date and Time
on the Internet: Timestamps (Internet RFC 3339)
specifies an ISO 8601 profile for use in new Internet protocols.
An extension, Date
and Time on the Internet: Timestamps with Additional Information
(Internet RFC 9557) extends this profile
to let you specify the tzdb timezone of a timestamp
via suffixes like [Asia/Tokyo].
    • 

  • 
Date & Time
Formats on the Web surveys web- and Internet-oriented date and time
formats.
    • 

  • Alphabetic time zone abbreviations should not be used as unique
identifiers for UT offsets as they are ambiguous in
practice. For example, in English-speaking North America
“CST” denotes 6 hours behind UT,
but in China it denotes 8 hours ahead of UT,
and French-speaking North Americans prefer
“HNC” to
“CST”. The tz
database contains English abbreviations for many timestamps;
unfortunately some of these abbreviations were merely the database maintainers’
inventions, and these have been removed when possible.
    • 

  • Numeric time zone abbreviations typically count hours east of
UT, e.g., +09 for Japan and
−10 for Hawaii. However, POSIX proleptic
TZ settings use the opposite convention.
For example, one might use TZ="JST-9" and
TZ="HST10"
for Japan and Hawaii, respectively. If the
tz database is available, it is usually better to use
settings like TZ="Asia/Tokyo" and
TZ="Pacific/Honolulu" instead, as this should avoid
confusion, handle old timestamps better, and insulate you better from
any future changes to the rules. One should never set
POSIX TZ to a value like
"GMT-9", though, since this would incorrectly imply that
local time is nine hours ahead of UT and the time zone
is called “GMT”.
    • 











This web page is in the public domain, so clarified as of
2009-05-17 by Arthur David Olson.


Please send corrections to this web page to the
time zone mailing list.
The mailing list and its archives are public,
so please do not send confidential information.




tzfile.h0000644000000000000000000001014215135256172007457 0ustar00/* Layout and location of TZif files.  */

#ifndef TZFILE_H

#define TZFILE_H

/*
** This file is in the public domain, so clarified as of
** 1996-06-05 by Arthur David Olson.
*/

/*
** This header is for use ONLY with the time conversion code.
** There is no guarantee that it will remain unchanged,
** or that it will remain at all.
** Do NOT copy it to any system include directory.
** Thank you!
*/

/* Information about time zone files.
   See Internet RFC 9636 for more details about the following format.  */

/*
** Each file begins with. . .
*/

#define TZ_MAGIC "TZif"

struct tzhead {
	char	tzh_magic[4];		/* TZ_MAGIC */
	char	tzh_version[1];		/* '\0' or '2'-'4' as of 2021 */
	char	tzh_reserved[15];	/* reserved; must be zero */
	char	tzh_ttisutcnt[4];	/* coded number of trans. time flags */
	char	tzh_ttisstdcnt[4];	/* coded number of trans. time flags */
	char	tzh_leapcnt[4];		/* coded number of leap seconds */
	char	tzh_timecnt[4];		/* coded number of transition times */
	char	tzh_typecnt[4];		/* coded number of local time types */
	char	tzh_charcnt[4];		/* coded number of abbr. chars */
};

/*
** . . .followed by. . .
**
**	tzh_timecnt (char [4])s		coded transition times a la time(2)
**	tzh_timecnt (unsigned char)s	types of local time starting at above
**	tzh_typecnt repetitions of
**		one (char [4])		coded UT offset in seconds
**		one (unsigned char)	used to set tm_isdst
**		one (unsigned char)	that's an abbreviation list index
**	tzh_charcnt (char)s		'\0'-terminated zone abbreviations
**	tzh_leapcnt repetitions of
**		one (char [4])		coded leap second transition times
**		one (char [4])		total correction after above
**	tzh_ttisstdcnt (char)s		indexed by type; if 1, transition
**					time is standard time, if 0,
**					transition time is local (wall clock)
**					time; if absent, transition times are
**					assumed to be local time
**	tzh_ttisutcnt (char)s		indexed by type; if 1, transition
**					time is UT, if 0, transition time is
**					local time; if absent, transition
**					times are assumed to be local time.
**					When this is 1, the corresponding
**					std/wall indicator must also be 1.
*/

/*
** If tzh_version is '2' or greater, the above is followed by a second instance
** of tzhead and a second instance of the data in which each coded transition
** time uses 8 rather than 4 chars,
** then a POSIX.1-2017 proleptic TZ string for use in handling
** instants after the last transition time stored in the file
** (with nothing between the newlines if there is no POSIX.1-2017
** representation for such instants).
**
** If tz_version is '3' or greater, the TZ string can be any POSIX.1-2024
** proleptic TZ string, which means the above is extended as follows.
** First, the TZ string's hour offset may range from -167
** through 167 as compared to the range 0 through 24 required
** by POSIX.1-2017 and earlier.
** Second, its DST start time may be January 1 at 00:00 and its stop
** time December 31 at 24:00 plus the difference between DST and
** standard time, indicating DST all year.
*/

/*
** In the current implementation, "tzset()" refuses to deal with files that
** exceed any of the limits below.
*/

#ifndef TZ_MAX_TIMES
/* The following limit applies to localtime.c; zic has no such limit.
   The limit must be at least 310 for Asia/Hebron with 'zic -b fat'.  */
# define TZ_MAX_TIMES 2000
#endif /* !defined TZ_MAX_TIMES */

#ifndef TZ_MAX_TYPES
/* This must be at least 18 for Europe/Vilnius with 'zic -b fat'.  */
# define TZ_MAX_TYPES 256 /* Limited to 256 by Internet RFC 9636.  */
#endif /* !defined TZ_MAX_TYPES */

#ifndef TZ_MAX_CHARS
/* This must be at least 40 for America/Anchorage.  */
# define TZ_MAX_CHARS 256	/* Maximum number of abbreviation characters */
				/* (limited to 256 by Internet RFC 9636) */
#endif /* !defined TZ_MAX_CHARS */

#ifndef TZ_MAX_LEAPS
/* The following limit applies to localtime.c; zic has no such limit.
   The limit must be at least 27 for leap seconds from 1972 through mid-2023.
   There's a plan to discontinue leap seconds by 2035.  */
# define TZ_MAX_LEAPS 50	/* Maximum number of leap second corrections */
#endif /* !defined TZ_MAX_LEAPS */

#endif /* !defined TZFILE_H */
private.h0000644000000000000000000007676515244743474007672 0ustar00/* Private header for tzdb code.  */

#ifndef PRIVATE_H

#define PRIVATE_H

/*
** This file is in the public domain, so clarified as of
** 1996-06-05 by Arthur David Olson.
*/

/*
** This header is for use ONLY with the time conversion code.
** There is no guarantee that it will remain unchanged,
** or that it will remain at all.
** Do NOT copy it to any system include directory.
** Thank you!
*/

/* PORT_TO_C89 means the code should work even if the underlying
   compiler and library support only C89 plus C99's 'long long'
   and perhaps a few other extensions to C89.

   This macro is obsolescent, and the plan is to remove it along with
   associated code.  A good time to do that might be in the year 2029
   because RHEL 7 (whose GCC defaults to C89) extended life cycle
   support (ELS) is scheduled to end on 2028-06-30.  */
#ifndef PORT_TO_C89
# define PORT_TO_C89 0
#endif

/* SUPPORT_C89 means the tzcode library should support C89 callers
   in addition to the usual support for C99-and-later callers.
   This defaults to 1 as POSIX requires, even though that can trigger
   latent bugs in callers.  */
#ifndef SUPPORT_C89
# define SUPPORT_C89 1
#endif


/* The following feature-test macros should be defined before
   any #include of a system header.  */

/* Enable tm_gmtoff, tm_zone, and environ on GNUish systems.  */
#define _GNU_SOURCE 1
/* Fix asctime_r on Solaris 11.  */
#define _POSIX_PTHREAD_SEMANTICS 1
/* Enable strtoimax on pre-C99 Solaris 11.  */
#define __EXTENSIONS__ 1
/* Cause MS-Windows headers to define POSIX names.  */
#define _CRT_DECLARE_NONSTDC_NAMES 1
/* Prevent MS-Windows headers from defining min and max.  */
#define NOMINMAX 1

/* On GNUish systems where time_t might be 32 or 64 bits, use 64.
   On these platforms _FILE_OFFSET_BITS must also be 64; otherwise
   setting _TIME_BITS to 64 does not work.  The code does not
   otherwise rely on _FILE_OFFSET_BITS being 64, since it does not
   use off_t or functions like 'stat' that depend on off_t.  */
#ifndef _TIME_BITS
# ifndef _FILE_OFFSET_BITS
#  define _FILE_OFFSET_BITS 64
# endif
# if _FILE_OFFSET_BITS == 64
#  define _TIME_BITS 64
# endif
#endif

/* End of feature-test macro definitions.  */


#ifndef __STDC_VERSION__
# define __STDC_VERSION__ 0
#endif

/* Define true, false and bool if they don't work out of the box.  */
#if PORT_TO_C89 && __STDC_VERSION__ < 199901
# define true 1
# define false 0
# define bool int
#elif __STDC_VERSION__ < 202311
# include 
#endif

/* For pre-C23 compilers, a substitute for static_assert.
   Some of these compilers may warn if it is used outside the top level.  */
#if __STDC_VERSION__ < 202311 && !defined static_assert
# define static_assert(cond) extern int static_assert_check[(cond) ? 1 : -1]
#endif

/*
** zdump has been made independent of the rest of the time
** conversion package to increase confidence in the verification it provides.
** You can use zdump to help in verifying other implementations.
** To do this, compile with -DUSE_LTZ=0 and link without the tz library.
*/
#ifndef USE_LTZ
# define USE_LTZ 1
#endif

/* This string was in the Factory zone through version 2016f.  */
#ifndef GRANDPARENTED
# define GRANDPARENTED	"Local time zone must be set--see zic manual page"
#endif

/*
** Defaults for preprocessor symbols.
** You can override these in your C compiler options, e.g. '-DHAVE_GETTEXT=1'.
*/

#if !defined HAVE__GENERIC && defined __has_extension
# if !__has_extension(c_generic_selections)
#  define HAVE__GENERIC 0
# endif
#endif
/* _Generic is buggy in pre-4.9 GCC.  */
#if !defined HAVE__GENERIC && defined __GNUC__ && !defined __STRICT_ANSI__
# define HAVE__GENERIC (4 < __GNUC__ + (9 <= __GNUC_MINOR__))
#endif
#ifndef HAVE__GENERIC
# define HAVE__GENERIC (201112 <= __STDC_VERSION__)
#endif

#ifndef HAVE_GETEUID
# define HAVE_GETEUID 1
#endif

#ifndef HAVE_GETRESUID
# define HAVE_GETRESUID 1
#endif

#if !defined HAVE_GETTEXT && defined __has_include
# if __has_include()
#  define HAVE_GETTEXT 1
# endif
#endif
#ifndef HAVE_GETTEXT
# define HAVE_GETTEXT 0
#endif

#ifndef HAVE_INCOMPATIBLE_CTIME_R
# define HAVE_INCOMPATIBLE_CTIME_R 0
#endif

#ifndef HAVE_LINK
# define HAVE_LINK 1
#endif /* !defined HAVE_LINK */

#ifndef HAVE_MALLOC_ERRNO
# define HAVE_MALLOC_ERRNO 1
#endif

#ifndef HAVE_POSIX_DECLS
# define HAVE_POSIX_DECLS 1
#endif

#ifndef HAVE_SETENV
# define HAVE_SETENV 1
#endif

#ifndef HAVE_STRDUP
# define HAVE_STRDUP 1
#endif

#ifndef HAVE_SYMLINK
# define HAVE_SYMLINK 1
#endif /* !defined HAVE_SYMLINK */

#if !defined HAVE_SYS_STAT_H && defined __has_include
# if !__has_include()
#  define HAVE_SYS_STAT_H 0
# endif
#endif
#ifndef HAVE_SYS_STAT_H
# define HAVE_SYS_STAT_H 1
#endif

#if !defined HAVE_UNISTD_H && defined __has_include
# if !__has_include()
#  define HAVE_UNISTD_H 0
# endif
#endif
#ifndef HAVE_UNISTD_H
# define HAVE_UNISTD_H 1
#endif

#ifndef NETBSD_INSPIRED
# define NETBSD_INSPIRED 1
#endif

#if HAVE_INCOMPATIBLE_CTIME_R
# define asctime_r _incompatible_asctime_r
# define ctime_r _incompatible_ctime_r
#endif /* HAVE_INCOMPATIBLE_CTIME_R */

#ifndef TZ_RUNTIME_LEAPS
# define TZ_RUNTIME_LEAPS 1
#endif

/*
** Nested includes
*/

#include 

/* If defining the 'timezone' variable a la POSIX, avoid clashing with the old
   'timezone' function of FreeBSD <= 14, by renaming the latter's declaration.
   This hack can be removed after 2028-11-30, FreeBSD 14's expected EOL.  */
#if (defined __FreeBSD__  && __FreeBSD__ < 15 && defined __BSD_VISIBLE \
     && defined USG_COMPAT  && USG_COMPAT == 2)
# define timezone sys_timezone
# define timezone_defined
#endif

#include 

#ifdef timezone_defined
# undef timezone
# undef timezone_defined
#endif

#include 
#if defined HAVE_STRNLEN && !HAVE_STRNLEN
static size_t
strnlen (char const *s, size_t maxlen)
{
  size_t i;
  for (i = 0; i < maxlen && s[i]; i++)
    continue;
  return i;
}
#endif

#if !PORT_TO_C89
# include 
#endif
#include 	/* for CHAR_BIT et al. */
#include 

#include 

#ifndef EINVAL
# define EINVAL ERANGE
#endif

#ifndef EFTYPE
# define EFTYPE EINVAL
#endif
#ifndef ELOOP
# define ELOOP EINVAL
#endif
#ifndef ENAMETOOLONG
# define ENAMETOOLONG EINVAL
#endif
#ifndef ENOMEM
# define ENOMEM EINVAL
#endif
#ifndef ENOSYS
# define ENOSYS EINVAL
#endif
#ifndef ENOTCAPABLE
# define ENOTCAPABLE EINVAL
#endif
#ifndef ENOTSUP
# define ENOTSUP EINVAL
#endif
#ifndef EOVERFLOW
# define EOVERFLOW EINVAL
#endif
#ifndef EPERM
# define EPERM EINVAL
#endif

#if HAVE_GETTEXT
# include 
#endif /* HAVE_GETTEXT */

#if HAVE_UNISTD_H
# include 
#else
/* Assume getopt.o or equivalent is linked via Makefile configuration.  */
int getopt(int, char *const[], char const *);
extern char *optarg;
extern int optind;
#endif /* HAVE_UNISTD_H */

/* SUPPORT_POSIX2008 means the tzcode library should support
   POSIX.1-2017-and-earlier callers in addition to the usual support for
   POSIX.1-2024-and-later callers; however, this can be
   incompatible with POSIX.1-2024-and-later callers.
   This macro is obsolescent, and the plan is to remove it
   along with any code needed only when it is nonzero.
   A good time to do that might be in the year 2034.
   This macro's name is SUPPORT_POSIX2008 because _POSIX_VERSION == 200809
   in POSIX.1-2017, a minor revision of POSIX.1-2008.  */
#ifndef SUPPORT_POSIX2008
# if defined _POSIX_VERSION && _POSIX_VERSION <= 200809
#  define SUPPORT_POSIX2008 1
# else
#  define SUPPORT_POSIX2008 0
# endif
#endif

#ifndef HAVE_DECL_ASCTIME_R
# if SUPPORT_POSIX2008
#  define HAVE_DECL_ASCTIME_R 1
# else
#  define HAVE_DECL_ASCTIME_R 0
# endif
#endif

#ifndef HAVE_ISSETUGID
# if (defined __FreeBSD__ || defined __NetBSD__ || defined __OpenBSD__ \
      || defined __DragonFly__ \
      || (defined __linux__ && !defined __GLIBC__) /* Android, musl, etc. */ \
      || (defined __APPLE__ && defined __MACH__) || defined __sun)
#  define HAVE_ISSETUGID 1
# else
#  define HAVE_ISSETUGID 0
# endif
#endif

#ifndef HAVE_SNPRINTF
# define HAVE_SNPRINTF (!PORT_TO_C89 || 199901 <= __STDC_VERSION__)
#endif

#ifndef HAVE_STRFTIME_L
# if _POSIX_VERSION < 200809
#  define HAVE_STRFTIME_L 0
# else
#  define HAVE_STRFTIME_L 1
# endif
#endif

#ifndef USG_COMPAT
# ifndef _XOPEN_VERSION
#  define USG_COMPAT 0
# else
#  define USG_COMPAT 1
# endif
#endif

#ifndef HAVE_TZNAME
# if _POSIX_VERSION < 198808 && !USG_COMPAT
#  define HAVE_TZNAME 0
# else
#  define HAVE_TZNAME 1
# endif
#endif

#ifndef ALTZONE
# if defined __sun || defined _M_XENIX
#  define ALTZONE 1
# else
#  define ALTZONE 0
# endif
#endif

#if PORT_TO_C89

/*
** Define HAVE_STDINT_H's default value here, rather than at the
** start, since __GLIBC__ and INTMAX_MAX's values depend on
** previously included files.  glibc 2.1 and Solaris 10 and later have
** stdint.h, even with pre-C99 compilers.
*/
# if !defined HAVE_STDINT_H && defined __has_include
#  define HAVE_STDINT_H 1 /* C23 __has_include implies C99 stdint.h.  */
# endif
# ifndef HAVE_STDINT_H
#  define HAVE_STDINT_H \
    (199901 <= __STDC_VERSION__ \
     || 2 < __GLIBC__ + (1 <= __GLIBC_MINOR__) \
     || __CYGWIN__ || INTMAX_MAX)
# endif /* !defined HAVE_STDINT_H */

# if HAVE_STDINT_H
#  include 
# endif /* !HAVE_STDINT_H */

# ifndef HAVE_INTTYPES_H
#  define HAVE_INTTYPES_H HAVE_STDINT_H
# endif
# if HAVE_INTTYPES_H
#  include 
# endif

/* Pre-C99 GCC compilers define __LONG_LONG_MAX__ instead of LLONG_MAX.  */
# if defined __LONG_LONG_MAX__ && !defined __STRICT_ANSI__
#  ifndef LLONG_MAX
#   define LLONG_MAX __LONG_LONG_MAX__
#  endif
#  ifndef LLONG_MIN
#   define LLONG_MIN (-1 - LLONG_MAX)
#  endif
#  ifndef ULLONG_MAX
#   define ULLONG_MAX (LLONG_MAX * 2ull + 1)
#  endif
# endif

# ifndef INT_FAST64_MAX
#  if 1 <= LONG_MAX >> 31 >> 31
typedef long int_fast64_t;
#   define INT_FAST64_MIN LONG_MIN
#   define INT_FAST64_MAX LONG_MAX
#  else
/* If this fails, compile with -DHAVE_STDINT_H or with a better compiler.  */
typedef long long int_fast64_t;
#   define INT_FAST64_MIN LLONG_MIN
#   define INT_FAST64_MAX LLONG_MAX
#  endif
# endif

# ifndef PRIdFAST64
#  if INT_FAST64_MAX == LONG_MAX
#   define PRIdFAST64 "ld"
#  else
#   define PRIdFAST64 "lld"
#  endif
# endif

# ifndef SCNdFAST64
#  define SCNdFAST64 PRIdFAST64
# endif

# ifndef INT_FAST32_MAX
typedef long int_fast32_t;
#  define INT_FAST32_MAX LONG_MAX
#  define INT_FAST32_MIN LONG_MIN
# endif

# ifndef INT_LEAST32_MAX
typedef int_fast32_t int_least32_t;
# endif

# ifndef INTMAX_MAX
#  ifdef LLONG_MAX
typedef long long intmax_t;
#   ifndef HAVE_STRTOLL
#    define HAVE_STRTOLL 1
#   endif
#   if HAVE_STRTOLL
#    define strtoimax strtoll
#   endif
#   define INTMAX_MAX LLONG_MAX
#   define INTMAX_MIN LLONG_MIN
#  else
typedef long intmax_t;
#   define INTMAX_MAX LONG_MAX
#   define INTMAX_MIN LONG_MIN
#  endif
#  ifndef strtoimax
#   define strtoimax strtol
#  endif
# endif

# ifndef PRIdMAX
#  if INTMAX_MAX == LLONG_MAX
#   define PRIdMAX "lld"
#  else
#   define PRIdMAX "ld"
#  endif
# endif

# ifndef PTRDIFF_MAX
#  define PTRDIFF_MAX MAXVAL(ptrdiff_t, TYPE_BIT(ptrdiff_t))
# endif

# ifndef UINT_FAST32_MAX
typedef unsigned long uint_fast32_t;
# endif

# ifndef UINT_FAST64_MAX
#  if 3 <= ULONG_MAX >> 31 >> 31
typedef unsigned long uint_fast64_t;
#   define UINT_FAST64_MAX ULONG_MAX
#  else
/* If this fails, compile with -DHAVE_STDINT_H or with a better compiler.  */
typedef unsigned long long uint_fast64_t;
#   define UINT_FAST64_MAX ULLONG_MAX
#  endif
# endif

# ifndef UINTMAX_MAX
#  ifdef ULLONG_MAX
typedef unsigned long long uintmax_t;
#   define UINTMAX_MAX ULLONG_MAX
#  else
typedef unsigned long uintmax_t;
#   define UINTMAX_MAX ULONG_MAX
#  endif
# endif

# ifndef PRIuMAX
#  ifdef ULLONG_MAX
#   define PRIuMAX "llu"
#  else
#   define PRIuMAX "lu"
#  endif
# endif

# ifndef SIZE_MAX
#  define SIZE_MAX ((size_t) -1)
# endif

#endif /* PORT_TO_C89 */

/* The maximum size of any created object, as a signed integer.
   Although the C standard does not outright prohibit larger objects,
   behavior is undefined if the result of pointer subtraction does not
   fit into ptrdiff_t, and the code assumes in several places that
   pointer subtraction works.  As a practical matter it's OK to not
   support objects larger than this.  */
#define INDEX_MAX ((ptrdiff_t) min(PTRDIFF_MAX, SIZE_MAX))

/* Support ckd_add, ckd_sub, ckd_mul on C23 or recent-enough GCC-like
   hosts, unless compiled with -DHAVE_STDCKDINT_H=0 or with pre-C23 EDG.  */
#if !defined HAVE_STDCKDINT_H && defined __has_include
# if __has_include()
#  define HAVE_STDCKDINT_H 1
# endif
#endif
#ifdef HAVE_STDCKDINT_H
# if HAVE_STDCKDINT_H
#  include 
# endif
#elif defined __EDG__
/* Do nothing, to work around EDG bug .  */
#elif defined __has_builtin
# if __has_builtin(__builtin_add_overflow)
#  define ckd_add(r, a, b) __builtin_add_overflow(a, b, r)
# endif
# if __has_builtin(__builtin_sub_overflow)
#  define ckd_sub(r, a, b) __builtin_sub_overflow(a, b, r)
# endif
# if __has_builtin(__builtin_mul_overflow)
#  define ckd_mul(r, a, b) __builtin_mul_overflow(a, b, r)
# endif
#elif 7 <= __GNUC__
# define ckd_add(r, a, b) __builtin_add_overflow(a, b, r)
# define ckd_sub(r, a, b) __builtin_sub_overflow(a, b, r)
# define ckd_mul(r, a, b) __builtin_mul_overflow(a, b, r)
#endif

#if (defined __has_c_attribute \
     && (202311 <= __STDC_VERSION__ || !defined __STRICT_ANSI__))
# define HAVE___HAS_C_ATTRIBUTE true
#else
# define HAVE___HAS_C_ATTRIBUTE false
#endif

#ifdef __has_attribute
# if __has_attribute (nonstring)
#  define ATTRIBUTE_NONSTRING __attribute__((__nonstring__))
# endif
#endif
#ifndef ATTRIBUTE_NONSTRING
# define ATTRIBUTE_NONSTRING
#endif

#if HAVE___HAS_C_ATTRIBUTE
# if __has_c_attribute(deprecated)
#  define ATTRIBUTE_DEPRECATED [[deprecated]]
# endif
#endif
#ifndef ATTRIBUTE_DEPRECATED
# if 3 < __GNUC__ + (2 <= __GNUC_MINOR__)
#  define ATTRIBUTE_DEPRECATED __attribute__((deprecated))
# else
#  define ATTRIBUTE_DEPRECATED /* empty */
# endif
#endif

#if HAVE___HAS_C_ATTRIBUTE
# if __has_c_attribute(fallthrough)
#  define ATTRIBUTE_FALLTHROUGH [[fallthrough]]
# endif
#endif
#ifndef ATTRIBUTE_FALLTHROUGH
# if 7 <= __GNUC__
#  define ATTRIBUTE_FALLTHROUGH __attribute__((fallthrough))
# else
#  define ATTRIBUTE_FALLTHROUGH ((void) 0)
# endif
#endif

#if HAVE___HAS_C_ATTRIBUTE
# if __has_c_attribute(maybe_unused)
#  define ATTRIBUTE_MAYBE_UNUSED [[maybe_unused]]
# endif
#endif
#ifndef ATTRIBUTE_MAYBE_UNUSED
# if 2 < __GNUC__ + (7 <= __GNUC_MINOR__)
#  define ATTRIBUTE_MAYBE_UNUSED __attribute__((unused))
# else
#  define ATTRIBUTE_MAYBE_UNUSED /* empty */
# endif
#endif

#if HAVE___HAS_C_ATTRIBUTE
# if __has_c_attribute(noreturn)
#  define ATTRIBUTE_NORETURN [[noreturn]]
# endif
#endif
#ifndef ATTRIBUTE_NORETURN
# if 201112 <= __STDC_VERSION__
#  define ATTRIBUTE_NORETURN _Noreturn
# elif 2 < __GNUC__ + (8 <= __GNUC_MINOR__)
#  define ATTRIBUTE_NORETURN __attribute__((noreturn))
# else
#  define ATTRIBUTE_NORETURN /* empty */
# endif
#endif

#if HAVE___HAS_C_ATTRIBUTE
# if __has_c_attribute(reproducible)
#  define ATTRIBUTE_REPRODUCIBLE [[reproducible]]
# endif
#endif
#ifndef ATTRIBUTE_REPRODUCIBLE
# define ATTRIBUTE_REPRODUCIBLE /* empty */
#endif

#if HAVE___HAS_C_ATTRIBUTE
# if __has_c_attribute(unsequenced)
#  define ATTRIBUTE_UNSEQUENCED [[unsequenced]]
# endif
#endif
#ifndef ATTRIBUTE_UNSEQUENCED
# define ATTRIBUTE_UNSEQUENCED /* empty */
#endif

/* GNU C attributes that are useful in tzcode.
   Although neither __attribute__((const)) nor __attribute__((pure)) are
   stricter than their C23 counterparts [[unsequenced]] and [[reproducible]],
   the C23 attributes happen to work in each tzcode use of ATTRIBUTE_CONST
   and ATTRIBUTE_PURE.  (This might not work outside of tzcode!)  */
#if __GNUC__ < 3
# define ATTRIBUTE_CONST ATTRIBUTE_UNSEQUENCED
# define ATTRIBUTE_FORMAT(spec) /* empty */
# define ATTRIBUTE_PURE ATTRIBUTE_REPRODUCIBLE
#else
# define ATTRIBUTE_CONST __attribute__((const))
# define ATTRIBUTE_FORMAT(spec) __attribute__((format spec))
# define ATTRIBUTE_PURE __attribute__((pure))
#endif

/* Avoid GCC bug 114833 .
   Remove this macro and its uses when the bug is fixed in a GCC release,
   because only the latest GCC matters for $(GCC_DEBUG_FLAGS).  */
#ifdef GCC_LINT
# define ATTRIBUTE_PURE_114833 ATTRIBUTE_PURE
#else
# define ATTRIBUTE_PURE_114833 /* empty */
#endif
/* GCC_LINT hack to pacify GCC bug 114833 even though the attribute is
   not strictly correct, as the function might not return whereas pure
   functions are supposed to return exactly once.  This hack is not
   known to generate wrong code for tzcode on any platform.
   Remove this macro and its uses when the bug is fixed in a GCC release.  */
#define ATTRIBUTE_PURE_114833_HACK ATTRIBUTE_PURE_114833

#if (__STDC_VERSION__ < 199901 && !defined restrict \
     && (PORT_TO_C89 || defined _MSC_VER))
# define restrict /* empty */
#endif

/*
** Workarounds for compilers/systems.
*/

#ifndef EPOCH_LOCAL
# define EPOCH_LOCAL 0
#endif
#ifndef EPOCH_OFFSET
# define EPOCH_OFFSET 0
#endif
#ifndef RESERVE_STD_EXT_IDS
# define RESERVE_STD_EXT_IDS 0
#endif

#ifdef time_tz
# define defined_time_tz true
#else
# define defined_time_tz false
#endif

/* If standard C identifiers with external linkage (e.g., localtime)
   are reserved and are not already being renamed anyway, rename them
   as if compiling with '-Dtime_tz=time_t'.  */
#if !defined time_tz && RESERVE_STD_EXT_IDS && USE_LTZ
# define time_tz time_t
#endif

/*
** Compile with -Dtime_tz=T to build the tz package with a private
** time_t type equivalent to T rather than the system-supplied time_t.
** This debugging feature can test unusual design decisions
** (e.g., time_t wider than 'long', or unsigned time_t) even on
** typical platforms.
*/
#if defined time_tz || EPOCH_LOCAL || EPOCH_OFFSET != 0
# define TZ_TIME_T true
#else
# define TZ_TIME_T false
#endif

#if defined LOCALTIME_IMPLEMENTATION && TZ_TIME_T
static time_t sys_time(time_t *x) { return time(x); }
#endif

#if TZ_TIME_T

typedef time_tz tz_time_t;

# undef  asctime
# define asctime tz_asctime
# undef  ctime
# define ctime tz_ctime
# undef  difftime
# define difftime tz_difftime
# undef  gmtime
# define gmtime tz_gmtime
# undef  gmtime_r
# define gmtime_r tz_gmtime_r
# undef  localtime
# define localtime tz_localtime
# undef  localtime_r
# define localtime_r tz_localtime_r
# undef  localtime_rz
# define localtime_rz tz_localtime_rz
# undef  mktime
# define mktime tz_mktime
# undef  mktime_z
# define mktime_z tz_mktime_z
# undef  offtime
# define offtime tz_offtime
# undef  offtime_r
# define offtime_r tz_offtime_r
# undef  posix2time
# define posix2time tz_posix2time
# undef  posix2time_z
# define posix2time_z tz_posix2time_z
# undef  strftime
# define strftime tz_strftime
# undef  time
# define time tz_time
# undef  time2posix
# define time2posix tz_time2posix
# undef  time2posix_z
# define time2posix_z tz_time2posix_z
# undef  time_t
# define time_t tz_time_t
# undef  timegm
# define timegm tz_timegm
# undef  timelocal
# define timelocal tz_timelocal
# undef  timeoff
# define timeoff tz_timeoff
# undef  tzalloc
# define tzalloc tz_tzalloc
# undef  tzfree
# define tzfree tz_tzfree
# undef  tzset
# define tzset tz_tzset
# if SUPPORT_POSIX2008
#  undef  asctime_r
#  define asctime_r tz_asctime_r
#  undef  ctime_r
#  define ctime_r tz_ctime_r
# endif
# if HAVE_STRFTIME_L
#  undef  strftime_l
#  define strftime_l tz_strftime_l
# endif
# if HAVE_TZNAME
#  undef  tzname
#  define tzname tz_tzname
# endif
# if USG_COMPAT
#  undef  daylight
#  define daylight tz_daylight
#  undef  timezone
#  define timezone tz_timezone
# endif
# if ALTZONE
#  undef  altzone
#  define altzone tz_altzone
# endif

# if __STDC_VERSION__ < 202311
#  define DEPRECATED_IN_C23 /* empty */
# else
#  define DEPRECATED_IN_C23 ATTRIBUTE_DEPRECATED
# endif
DEPRECATED_IN_C23 char *asctime(struct tm const *);
DEPRECATED_IN_C23 char *ctime(time_t const *);
# if SUPPORT_POSIX2008
char *asctime_r(struct tm const *restrict, char *restrict);
char *ctime_r(time_t const *, char *);
# endif
ATTRIBUTE_CONST double difftime(time_t, time_t);
size_t strftime(char *restrict, size_t, char const *restrict,
		struct tm const *restrict);
# if HAVE_STRFTIME_L
size_t strftime_l(char *restrict, size_t, char const *restrict,
		  struct tm const *restrict, locale_t);
# endif
struct tm *gmtime(time_t const *);
struct tm *gmtime_r(time_t const *restrict, struct tm *restrict);
struct tm *localtime(time_t const *);
struct tm *localtime_r(time_t const *restrict, struct tm *restrict);
time_t mktime(struct tm *);
time_t time(time_t *);
time_t timegm(struct tm *);
void tzset(void);
#endif

#ifndef HAVE_DECL_TIMEGM
# if (202311 <= __STDC_VERSION__ \
      || defined __GLIBC__ || defined __tm_zone /* musl */ \
      || defined __FreeBSD__ || defined __NetBSD__ || defined __OpenBSD__ \
      || defined __DragonFly__ || defined __HAIKU__ \
      || (defined __APPLE__ && defined __MACH__))
#  define HAVE_DECL_TIMEGM 1
# else
#  define HAVE_DECL_TIMEGM 0
# endif
#endif
#if !HAVE_DECL_TIMEGM && !defined timegm
time_t timegm(struct tm *);
#endif

#if !HAVE_DECL_ASCTIME_R && !defined asctime_r && SUPPORT_POSIX2008
extern char *asctime_r(struct tm const *restrict, char *restrict);
#endif

#ifndef HAVE_DECL_ENVIRON
# if defined environ || defined __USE_GNU
#  define HAVE_DECL_ENVIRON 1
# else
#  define HAVE_DECL_ENVIRON 0
# endif
#endif

#if !HAVE_DECL_ENVIRON
extern char **environ;
#endif

#ifndef HAVE_MEMPCPY
# if (defined mempcpy \
      || defined __FreeBSD__ || defined __NetBSD__ || defined __DragonFly__ \
      || defined __linux__)
#  define HAVE_MEMPCPY 1
# else
#  define HAVE_MEMPCPY 0
# endif
#endif
#if !HAVE_MEMPCPY
static void *
mempcpy(void *restrict s1, void const *restrict s2, size_t n)
{
  char *p = memcpy(s1, s2, n);
  return p + n;
}
#endif

#if 2 <= HAVE_TZNAME + (TZ_TIME_T || !HAVE_POSIX_DECLS)
extern char *tzname[];
#endif
#if 2 <= USG_COMPAT + (TZ_TIME_T || !HAVE_POSIX_DECLS)
extern long timezone;
extern int daylight;
#endif
#if 2 <= ALTZONE + (TZ_TIME_T || !HAVE_POSIX_DECLS)
extern long altzone;
#endif

/*
** The STD_INSPIRED functions are similar, but most also need
** declarations if time_tz is defined.
*/

#ifndef STD_INSPIRED
# ifdef __NetBSD__
#  define STD_INSPIRED 1
# else
#  define STD_INSPIRED 0
# endif
#endif
#if STD_INSPIRED
# if TZ_TIME_T || !defined offtime
struct tm *offtime(time_t const *, long);
# endif
# if TZ_TIME_T || !defined offtime_r
struct tm *offtime_r(time_t const *restrict, long, struct tm *restrict);
# endif
# if TZ_TIME_T || !defined timelocal
time_t timelocal(struct tm *);
# endif
# if TZ_TIME_T || !defined timeoff
#  define EXTERN_TIMEOFF
# endif
# if TZ_TIME_T || !defined time2posix
time_t time2posix(time_t);
# endif
# if TZ_TIME_T || !defined posix2time
time_t posix2time(time_t);
# endif
#endif

/* Infer TM_ZONE on systems where this information is known, but suppress
   guessing if NO_TM_ZONE is defined.  Similarly for TM_GMTOFF.  */
#if (200809 < _POSIX_VERSION \
     || defined __GLIBC__ \
     || defined __tm_zone /* musl */ \
     || defined __FreeBSD__ || defined __NetBSD__ || defined __OpenBSD__ \
     || defined __DragonFly__ || defined __HAIKU__ \
     || (defined __APPLE__ && defined __MACH__))
# if !defined TM_GMTOFF && !defined NO_TM_GMTOFF
#  define TM_GMTOFF tm_gmtoff
# endif
# if !defined TM_ZONE && !defined NO_TM_ZONE
#  define TM_ZONE tm_zone
# endif
#endif

/*
** Define functions that are ABI compatible with NetBSD.
** What we call 'struct state' NetBSD calls
** 'struct __state', but this is a private name so it doesn't matter.
*/
#if NETBSD_INSPIRED
# ifdef _NETBSD_SOURCE
#  define state __state
# else
typedef struct state *timezone_t;
# endif
struct tm *localtime_rz(timezone_t restrict, time_t const *restrict,
			struct tm *restrict);
time_t mktime_z(timezone_t restrict, struct tm *restrict);
timezone_t tzalloc(char const *);
void tzfree(timezone_t);
# if STD_INSPIRED
#  if TZ_RUNTIME_LEAPS
#   define ATTRIBUTE_POSIX2TIME ATTRIBUTE_PURE
#  else
#   define ATTRIBUTE_POSIX2TIME ATTRIBUTE_CONST
#  endif
#  if TZ_TIME_T || !defined posix2time_z
ATTRIBUTE_POSIX2TIME time_t posix2time_z(timezone_t, time_t);
#  endif
#  if TZ_TIME_T || !defined time2posix_z
ATTRIBUTE_POSIX2TIME time_t time2posix_z(timezone_t, time_t);
#  endif
# endif
#endif

/*
** Finally, some convenience items.
*/

#define TYPE_BIT(type) (CHAR_BIT * (ptrdiff_t) sizeof(type))
#define TYPE_SIGNED(type) (((type) -1) < 0)
#define TWOS_COMPLEMENT(type) (TYPE_SIGNED (type) && (! ~ (type) -1))

/* Yield the value of the arithmetic expression X after integer promotion.
   This is safer than a cast, which in general would accept even
   pointers and which might trap or yield a value not equal to X.  */
#define INT_PROMOTE(x) (+(x))

/* Minimum and maximum of two values.  Use lower case to avoid
   naming clashes with standard include files.  */
#define max(a, b) ((a) > (b) ? (a) : (b))
#define min(a, b) ((a) < (b) ? (a) : (b))

/* Max and min values of the integer type T, of which only the bottom
   B bits are used, and where the highest-order used bit is considered
   to be a sign bit if T is signed.  */
#define MAXVAL(t, b)						\
  ((t) (((t) 1 << ((b) - 1 - TYPE_SIGNED(t)))			\
	- 1 + ((t) 1 << ((b) - 1 - TYPE_SIGNED(t)))))
#define MINVAL(t, b)						\
  ((t) (TYPE_SIGNED(t) ? - TWOS_COMPLEMENT(t) - MAXVAL(t, b) : 0))

/* The extreme time values, assuming no padding.  */
#define TIME_T_MIN_NO_PADDING MINVAL(time_t, TYPE_BIT(time_t))
#define TIME_T_MAX_NO_PADDING MAXVAL(time_t, TYPE_BIT(time_t))

/* The extreme time values.  These are macros, not constants, so that
   any portability problems occur only when compiling .c files that use
   the macros, which is safer for applications that need only zdump and zic.
   This implementation assumes no padding if time_t is signed and
   either the compiler lacks support for _Generic or time_t is not one
   of the standard signed integer types.  */
#if HAVE__GENERIC
# define TIME_T_MIN \
    _Generic((time_t) 0, \
	     signed char: SCHAR_MIN, short: SHRT_MIN, \
	     int: INT_MIN, long: LONG_MIN, long long: LLONG_MIN, \
	     default: TIME_T_MIN_NO_PADDING)
# define TIME_T_MAX \
    (TYPE_SIGNED(time_t) \
     ? _Generic((time_t) 0, \
		signed char: SCHAR_MAX, short: SHRT_MAX, \
		int: INT_MAX, long: LONG_MAX, long long: LLONG_MAX, \
		default: TIME_T_MAX_NO_PADDING)			    \
     : (time_t) -1)
enum { SIGNED_PADDING_CHECK_NEEDED
	 = _Generic((time_t) 0,
		    signed char: false, short: false,
		    int: false, long: false, long long: false,
		    default: true) };
#else
# define TIME_T_MIN TIME_T_MIN_NO_PADDING
# define TIME_T_MAX TIME_T_MAX_NO_PADDING
enum { SIGNED_PADDING_CHECK_NEEDED = true };
#endif
/* Try to check the padding assumptions.  Although TIME_T_MAX and the
   following check can both have undefined behavior on oddball
   platforms due to shifts exceeding widths of signed integers, these
   platforms' compilers are likely to diagnose these issues in integer
   constant expressions, so it shouldn't hurt to check statically.  */
static_assert(! TYPE_SIGNED(time_t) || ! SIGNED_PADDING_CHECK_NEEDED
	      || TIME_T_MAX >> (TYPE_BIT(time_t) - 2) == 1);

/*
** 302 / 1000 is log10(2.0) rounded up.
** Subtract one for the sign bit if the type is signed;
** add one for integer division truncation;
** add one more for a minus sign if the type is signed.
*/
#define INT_STRLEN_MAXIMUM(type) \
	((TYPE_BIT(type) - TYPE_SIGNED(type)) * 302 / 1000 + \
	1 + TYPE_SIGNED(type))

/*
** INITIALIZE(x)
*/

#ifdef GCC_LINT
# define INITIALIZE(x)	((x) = 0)
#else
# define INITIALIZE(x)
#endif

/* Whether memory access must strictly follow the C standard.
   If 0, it's OK to read uninitialized storage so long as the value is
   not relied upon.  Defining it to 0 lets mktime access parts of
   struct tm that might be uninitialized, as a heuristic when the
   standard doesn't say what to return and when tm_gmtoff can help
   mktime likely infer a better value.  */
#ifndef UNINIT_TRAP
# define UNINIT_TRAP 0
#endif

/* strftime.c sometimes needs access to timeoff if it is not already public.
   tz_private_timeoff should be used only by localtime.c and strftime.c.  */
#if (!defined EXTERN_TIMEOFF \
     && defined TM_GMTOFF && (200809 < _POSIX_VERSION || ! UNINIT_TRAP))
# ifndef timeoff
#  define timeoff tz_private_timeoff
# endif
# define EXTERN_TIMEOFF
#endif
#ifdef EXTERN_TIMEOFF
time_t timeoff(struct tm *, long);
#endif

#ifdef DEBUG
# undef unreachable
# define unreachable() abort()
#elif !defined unreachable
# ifdef __has_builtin
#  if __has_builtin(__builtin_unreachable)
#   define unreachable() __builtin_unreachable()
#  endif
# elif 4 < __GNUC__ + (5 <= __GNUC_MINOR__)
#  define unreachable() __builtin_unreachable()
# endif
# ifndef unreachable
#  define unreachable() ((void) 0)
# endif
#endif

/*
** For the benefit of GNU folk...
** '_(MSGID)' uses the current locale's message library string for MSGID.
** The default is to use gettext if available, and use MSGID otherwise.
*/

#ifndef _
# if HAVE_GETTEXT
#  define _(msgid) gettext(msgid)
# else
#  define _(msgid) (msgid)
# endif
#endif
#ifndef N_
# define N_(msgid) (msgid)
#endif

#if !defined TZ_DOMAIN && defined HAVE_GETTEXT
# define TZ_DOMAIN "tz"
#endif

#if HAVE_INCOMPATIBLE_CTIME_R
# undef asctime_r
# undef ctime_r
char *asctime_r(struct tm const *restrict, char *restrict);
char *ctime_r(time_t const *, char *);
#endif /* HAVE_INCOMPATIBLE_CTIME_R */

/* Handy constants that are independent of tzfile implementation.  */

/* 2**31 - 1 as a signed integer, and usable in #if.  */
#define TWO_31_MINUS_1 2147483647

enum {
  SECSPERMIN = 60,
  MINSPERHOUR = 60,
  SECSPERHOUR = SECSPERMIN * MINSPERHOUR,
  HOURSPERDAY = 24,
  DAYSPERWEEK = 7,
  DAYSPERNYEAR = 365,
  DAYSPERLYEAR = DAYSPERNYEAR + 1,
  MONSPERYEAR = 12,
  YEARSPERREPEAT = 400	/* years before a Gregorian repeat */
};

#define SECSPERDAY	((int_fast32_t) SECSPERHOUR * HOURSPERDAY)

#define DAYSPERREPEAT		((int_fast32_t) 400 * 365 + 100 - 4 + 1)
#define SECSPERREPEAT		((int_fast64_t) DAYSPERREPEAT * SECSPERDAY)
#define AVGSECSPERYEAR		(SECSPERREPEAT / YEARSPERREPEAT)

/* How many years to generate (in zic.c) or search through (in localtime.c).
   This is two years larger than the obvious 400, to avoid edge cases.
   E.g., suppose a rule applies from 2012 on with transitions
   in March and September, plus one-off transitions in November 2013,
   and suppose the rule cannot be expressed as a proleptic TZ string.
   If zic looked only at the last 400 years, it would set max_year=2413,
   with the intent that the 400 years 2014 through 2413 will be repeated.
   The last transition listed in the tzfile would be in 2413-09,
   less than 400 years after the last one-off transition in 2013-11.
   Two years is not overkill for localtime.c, as a one-year bump
   would mishandle 2023d's America/Ciudad_Juarez for November 2422.  */
enum { years_of_observations = YEARSPERREPEAT + 2 };

enum {
  TM_SUNDAY,
  TM_MONDAY,
  TM_TUESDAY,
  TM_WEDNESDAY,
  TM_THURSDAY,
  TM_FRIDAY,
  TM_SATURDAY
};

enum {
  TM_JANUARY,
  TM_FEBRUARY,
  TM_MARCH,
  TM_APRIL,
  TM_MAY,
  TM_JUNE,
  TM_JULY,
  TM_AUGUST,
  TM_SEPTEMBER,
  TM_OCTOBER,
  TM_NOVEMBER,
  TM_DECEMBER
};

enum {
  TM_YEAR_BASE = 1900,
  TM_WDAY_BASE = TM_MONDAY,
  EPOCH_YEAR = 1970,
  EPOCH_WDAY = TM_THURSDAY
};

#define isleap(y) (((y) % 4) == 0 && (((y) % 100) != 0 || ((y) % 400) == 0))
#define year_days(y) (DAYSPERNYEAR + isleap(y))

/*
** Since everything in isleap is modulo 400 (or a factor of 400), we know that
**	isleap(y) == isleap(y % 400)
** and so
**	isleap(a + b) == isleap((a + b) % 400)
** or
**	isleap(a + b) == isleap(a % 400 + b % 400)
** This is true even if % means modulo rather than Fortran remainder
** (which is allowed by C89 but not by C99 or later).
** We use this to avoid addition overflow problems.
*/

#define year_sum_days(a, b) (DAYSPERNYEAR + isleap((a) % 400 + (b) % 400))

#endif /* !defined PRIVATE_H */
localtime.c0000644000000000000000000024142515251675224010142 0ustar00/* Convert timestamp from time_t to struct tm.  */

/*
** This file is in the public domain, so clarified as of
** 1996-06-05 by Arthur David Olson.
*/

/*
** Leap second handling from Bradley White.
** POSIX.1-1988 style TZ environment variable handling from Guy Harris.
*/

/*LINTLIBRARY*/

#define LOCALTIME_IMPLEMENTATION
#include "private.h"

#include "tzdir.h"
#include "tzfile.h"
#include 

#if HAVE_SYS_STAT_H
# include 
# ifndef S_ISREG
#  define S_ISREG(mode) (((mode) & S_IFMT) == S_IFREG) /* Ancient UNIX.  */
# endif
#else
struct stat { char st_ctime, st_dev, st_ino; };
# define dev_t char
# define ino_t char
# define fstat(fd, st) (memset(st, 0, sizeof *(st)), 0)
# define stat(name, st) fstat(0, st)
# define S_ISREG(mode) 1
#endif

#ifndef HAVE_STRUCT_STAT_ST_CTIM
# define HAVE_STRUCT_STAT_ST_CTIM 1
#endif
#if !defined st_ctim && defined __APPLE__ && defined __MACH__
# define st_ctim st_ctimespec
#endif

#ifndef THREAD_SAFE
# define THREAD_SAFE 0
#endif

#ifndef THREAD_RWLOCK
# define THREAD_RWLOCK 0
#endif

#ifndef THREAD_TM_MULTI
# define THREAD_TM_MULTI 0
#endif

#ifndef USE_TIMEX_T
# define USE_TIMEX_T false
#endif

#if THREAD_SAFE
# include 

# ifndef THREAD_PREFER_SINGLE
#  define THREAD_PREFER_SINGLE 0
# endif
# if THREAD_PREFER_SINGLE
#  ifndef HAVE___ISTHREADED
#   if defined __FreeBSD__ || defined __OpenBSD__ || defined __DragonFly__
#    define HAVE___ISTHREADED 1
#   else
#    define HAVE___ISTHREADED 0
#   endif
#  endif
#  if HAVE___ISTHREADED
extern int __isthreaded;
#  else
#   if !defined HAVE_SYS_SINGLE_THREADED_H && defined __has_include
#    if __has_include()
#     define HAVE_SYS_SINGLE_THREADED_H 1
#    else
#     define HAVE_SYS_SINGLE_THREADED_H 0
#    endif
#   endif
#   ifndef HAVE_SYS_SINGLE_THREADED_H
#    if defined __GLIBC__ && 2 < __GLIBC__ + (32 <= __GLIBC_MINOR__)
#     define HAVE_SYS_SINGLE_THREADED_H 1
#    else
#     define HAVE_SYS_SINGLE_THREADED_H 0
#    endif
#   endif
#   if HAVE_SYS_SINGLE_THREADED_H
#    include 
#   endif
#  endif
# endif
#endif

#if !defined TM_GMTOFF || !USE_TIMEX_T
# if THREAD_SAFE

/* True if the current process might be multi-threaded,
   false if it is definitely single-threaded.
   If false, it will be false the next time it is called
   unless the caller creates a thread in the meantime.
   If true, it might become false the next time it is called
   if all other threads exit in the meantime.  */
static bool
is_threaded(void)
{
#  if THREAD_PREFER_SINGLE && HAVE___ISTHREADED
  return !!__isthreaded;
#  elif THREAD_PREFER_SINGLE && HAVE_SYS_SINGLE_THREADED_H
  return !__libc_single_threaded;
#  else
  return true;
#  endif
}

#  if THREAD_RWLOCK
static pthread_rwlock_t locallock = PTHREAD_RWLOCK_INITIALIZER;
static int dolock(void) { return pthread_rwlock_rdlock(&locallock); }
static void dounlock(void) { pthread_rwlock_unlock(&locallock); }
#  else
static pthread_mutex_t locallock = PTHREAD_MUTEX_INITIALIZER;
static int dolock(void) { return pthread_mutex_lock(&locallock); }
static void dounlock(void) { pthread_mutex_unlock(&locallock); }
#  endif

/* Get a lock.  Return 0 on success, a positive errno value on failure,
   negative if known to be single-threaded so no lock is needed.  */
static int
lock(void)
{
  if (!is_threaded())
    return -1;
  return dolock();
}
static void
unlock(bool threaded)
{
  if (threaded)
    dounlock();
}
# else
static int lock(void) { return -1; }
static void unlock(ATTRIBUTE_MAYBE_UNUSED bool threaded) { }
# endif
#endif

#if THREAD_SAFE
typedef pthread_once_t once_t;
# define ONCE_INIT PTHREAD_ONCE_INIT
#else
typedef bool once_t;
# define ONCE_INIT false
#endif

static void
once(once_t *once_control, void init_routine(void))
{
#if THREAD_SAFE
  pthread_once(once_control, init_routine);
#else
  if (!*once_control) {
    *once_control = true;
    init_routine();
  }
#endif
}

enum tm_multi { LOCALTIME_TM_MULTI, GMTIME_TM_MULTI, OFFTIME_TM_MULTI };

#if THREAD_SAFE && THREAD_TM_MULTI

enum { N_TM_MULTI = OFFTIME_TM_MULTI + 1 };
static pthread_key_t tm_multi_key;
static int tm_multi_key_err;

static void
tm_multi_key_init(void)
{
  tm_multi_key_err = pthread_key_create(&tm_multi_key, free);
}

#endif

/* Unless intptr_t is missing, pacify gcc -Wcast-qual on char const * exprs.
   Use this carefully, as the casts disable type checking.
   This is a macro so that it can be used in static initializers.  */
#ifdef INTPTR_MAX
# define UNCONST(a) ((char *) (intptr_t) (a))
#else
# define UNCONST(a) ((char *) (a))
#endif

/* A signed type wider than int, so that we can add 1900 + tm_mon/12 to tm_year
   without overflow.  The static_assert checks that it is indeed wider
   than int; if this fails on your platform please let us know.  */
#if INT_MAX < LONG_MAX
typedef long iinntt;
# define IINNTT_MIN LONG_MIN
# define IINNTT_MAX LONG_MAX
#elif INT_MAX < LLONG_MAX
typedef long long iinntt;
# define IINNTT_MIN LLONG_MIN
# define IINNTT_MAX LLONG_MAX
#else
typedef intmax_t iinntt;
# define IINNTT_MIN INTMAX_MIN
# define IINNTT_MAX INTMAX_MAX
#endif
static_assert(IINNTT_MIN < INT_MIN && INT_MAX < IINNTT_MAX);

#ifndef HAVE_STRUCT_TIMESPEC
# define HAVE_STRUCT_TIMESPEC 1
#endif
#if !HAVE_STRUCT_TIMESPEC
struct timespec { time_t tv_sec; long tv_nsec; };
#endif

#if !defined CLOCK_MONOTONIC_COARSE && defined CLOCK_MONOTONIC
# define CLOCK_MONOTONIC_COARSE CLOCK_MONOTONIC
#endif
#ifndef CLOCK_MONOTONIC_COARSE
# undef clock_gettime
# define clock_gettime(id, t) ((t)->tv_sec = time(NULL), (t)->tv_nsec = 0, 0)
#endif

/* How many seconds to wait before checking the default TZif file again.
   Negative means no checking.  Default to 61 if DETECT_TZ_CHANGES
   (as FreeBSD optionally builds its localtime.c with -DDETECT_TZ_CHANGES),
   and to -1 otherwise.  */
#ifndef TZ_CHANGE_INTERVAL
# ifdef DETECT_TZ_CHANGES
#  define TZ_CHANGE_INTERVAL 61
# else
#  define TZ_CHANGE_INTERVAL (-1)
# endif
#endif
static_assert(TZ_CHANGE_INTERVAL < 0 || HAVE_SYS_STAT_H);

/* The change detection interval.  */
#if TZ_CHANGE_INTERVAL < 0 || !defined __FreeBSD__
enum { tz_change_interval = TZ_CHANGE_INTERVAL };
#else
/* FreeBSD uses this private-but-extern var in its internal test suite.  */
int __tz_change_interval = TZ_CHANGE_INTERVAL;
# define tz_change_interval __tz_change_interval
#endif

/* The type of monotonic times.
   This is the system time_t, even if USE_TIMEX_T #defines time_t below.  */
typedef time_t monotime_t;

/* On platforms where offtime or mktime might overflow,
   strftime.c defines USE_TIMEX_T to be true and includes us.
   This tells us to #define time_t to an internal type timex_t that is
   wide enough so that strftime %s never suffers from integer overflow,
   and to #define offtime (if TM_GMTOFF is defined) or mktime (otherwise)
   to a static function that returns the redefined time_t.
   It also tells us to define only data and code needed
   to support the offtime or mktime variant.  */
#if USE_TIMEX_T
# undef TIME_T_MIN
# undef TIME_T_MAX
# undef time_t
# define time_t timex_t
# if MKTIME_FITS_IN(LONG_MIN, LONG_MAX)
typedef long timex_t;
#  define TIME_T_MIN LONG_MIN
#  define TIME_T_MAX LONG_MAX
# elif MKTIME_FITS_IN(LLONG_MIN, LLONG_MAX)
typedef long long timex_t;
#  define TIME_T_MIN LLONG_MIN
#  define TIME_T_MAX LLONG_MAX
# else
typedef intmax_t timex_t;
#  define TIME_T_MIN INTMAX_MIN
#  define TIME_T_MAX INTMAX_MAX
# endif

# ifdef TM_GMTOFF
#  undef timeoff
#  define timeoff timex_timeoff
#  undef EXTERN_TIMEOFF
# else
#  undef mktime
#  define mktime timex_mktime
# endif
#endif

/* Placeholders for platforms lacking AT_FCWD, openat, and fstatat.  */
#ifndef AT_FDCWD
# define AT_FDCWD (-1) /* any negative value will do */
static int openat(int dd, char const *path, int oflag) { unreachable (); }
static int fstatat(int dd, char const *path, struct stat *st, int flags)
{ unreachable(); }
#endif

/* Port to platforms that lack some O_* flags.  Unless otherwise
   specified, the flags are standardized by POSIX.  */

#ifndef O_BINARY
# define O_BINARY 0 /* MS-Windows */
#endif
#ifndef O_CLOEXEC
# define O_CLOEXEC 0
#endif
#ifndef O_CLOFORK
# define O_CLOFORK 0
#endif
#ifndef O_DIRECTORY
# define O_DIRECTORY 0
#endif
#ifndef O_IGNORE_CTTY
# define O_IGNORE_CTTY 0 /* GNU/Hurd */
#endif
#ifndef O_NOCTTY
# define O_NOCTTY 0
#endif
#ifndef O_PATH
# define O_PATH 0
#endif
#ifndef O_REGULAR
# define O_REGULAR 0
#endif
#ifndef O_RESOLVE_BENEATH
# define O_RESOLVE_BENEATH 0
#endif
#ifndef O_SEARCH
# define O_SEARCH 0
#endif

#if !HAVE_ISSETUGID

# if !defined HAVE_SYS_AUXV_H && defined __has_include
#  if __has_include()
#   define HAVE_SYS_AUXV_H 1
#  endif
# endif
# ifndef HAVE_SYS_AUXV_H
#  if defined __GLIBC__ && 2 < __GLIBC__ + (19 <= __GLIBC_MINOR__)
#   define HAVE_SYS_AUXV_H 1
#  else
#   define HAVE_SYS_AUXV_H 0
#  endif
# endif
# if HAVE_SYS_AUXV_H
#  include 
# endif

/* Avoid clash if headers declare but libraries do not define issetugid.  */
# undef issetugid
# define issetugid localtime_issetugid

/* Return 1 if the process is privileged, 0 otherwise.  */
static int
issetugid(void)
{
# if HAVE_SYS_AUXV_H && defined AT_SECURE
  unsigned long val;
  errno = 0;
  val = getauxval(AT_SECURE);
  if (val || errno != ENOENT)
    return !!val;
# endif
# if HAVE_GETRESUID
  {
    uid_t ruid, euid, suid;
    gid_t rgid, egid, sgid;
    if (0 <= getresuid (&ruid, &euid, &suid)) {
      if ((ruid ^ euid) | (ruid ^ suid))
	return 1;
      if (0 <= getresgid (&rgid, &egid, &sgid))
	return !!((rgid ^ egid) | (rgid ^ sgid));
    }
  }
# endif
# if HAVE_GETEUID
  return geteuid() != getuid() || getegid() != getgid();
# else
  return 0;
# endif
}
#endif

#ifndef WILDABBR
/*
** Someone might make incorrect use of a time zone abbreviation:
**	1.	They might reference tzname[0] before calling tzset (explicitly
**		or implicitly).
**	2.	They might reference tzname[1] before calling tzset (explicitly
**		or implicitly).
**	3.	They might reference tzname[1] after setting to a time zone
**		in which Daylight Saving Time is never observed.
**	4.	They might reference tzname[0] after setting to a time zone
**		in which Standard Time is never observed.
**	5.	They might reference tm.TM_ZONE after calling offtime.
** What's best to do in the above cases is open to debate;
** for now, we just set things up so that in any of the five cases
** WILDABBR is used. Another possibility: initialize tzname[0] to the
** string "tzname[0] used before set", and similarly for the other cases.
** And another: initialize tzname[0] to "ERA", with an explanation in the
** manual page of what this "time zone abbreviation" means (doing this so
** that tzname[0] has the "normal" length of three characters).
*/
# define WILDABBR "   "
#endif /* !defined WILDABBR */

static const char	wildabbr[] = WILDABBR;

static char const etc_utc[] = "Etc/UTC";

#if !USE_TIMEX_T || defined TM_ZONE || !defined TM_GMTOFF
static char const *utc = etc_utc + sizeof "Etc/" - 1;
#endif

/*
** The DST rules to use if TZ has no rules.
** Default to US rules as of 2017-05-07.
** POSIX does not specify the default DST rules;
** for historical reasons, US rules are a common default.
*/
#ifndef TZDEFRULESTRING
# define TZDEFRULESTRING ",M3.2.0,M11.1.0"
#endif

/* If compiled with -DOPENAT_TZDIR, then when accessing a relative
   name like "America/Los_Angeles", first open TZDIR (default
   "/usr/share/zoneinfo") as a directory and then use the result in
   openat with "America/Los_Angeles", rather than the traditional
   approach of opening "/usr/share/zoneinfo/America/Los_Angeles".
   Although the OPENAT_TZDIR approach is less efficient, suffers from
   spurious EMFILE and ENFILE failures, and is no more secure in practice,
   bleeding edge FreeBSD started doing it this way in August 2025.  */
#ifndef OPENAT_TZDIR
# define OPENAT_TZDIR 0
#endif

/* If compiled with -DSUPPRESS_TZDIR, do not prepend TZDIR to relative TZ.
   This is intended for specialized applications only, due to its
   security implications.  */
#ifndef SUPPRESS_TZDIR
# define SUPPRESS_TZDIR 0
#endif

/* Limit to time zone abbreviation length in proleptic TZ strings.
   This is distinct from TZ_MAX_CHARS, which limits TZif file contents.
   It defaults to 254, not 255, so that desigidx_type can be an unsigned char.
   unsigned char suffices for TZif files, so the only reason to increase
   TZNAME_MAXIMUM is to support TZ strings specifying abbreviations
   longer than 254 bytes.  There is little reason to do that, though,
   as strings that long are hardly "abbreviations".  */
#ifndef TZNAME_MAXIMUM
# define TZNAME_MAXIMUM 254
#endif

#if TZNAME_MAXIMUM < UCHAR_MAX
typedef unsigned char desigidx_type;
#elif TZNAME_MAXIMUM < INT_MAX
typedef int desigidx_type;
#elif TZNAME_MAXIMUM < PTRDIFF_MAX
typedef ptrdiff_t desigidx_type;
#else
# error "TZNAME_MAXIMUM too large"
#endif

/* A type that can represent any 32-bit two's complement integer,
   i.e., any integer in the range -2**31 .. 2**31 - 1.
   Ordinarily this is int_fast32_t, but on non-C23 hosts
   that are not two's complement it is int_fast64_t.  */
#if INT_FAST32_MIN < -TWO_31_MINUS_1
typedef int_fast32_t int_fast32_2s;
#else
typedef int_fast64_t int_fast32_2s;
#endif

struct ttinfo {				/* time type information */
	int_least32_t	tt_utoff;	/* UT offset in seconds; in the range
					   -2**31 + 1 .. 2**31 - 1  */
	desigidx_type	tt_desigidx;	/* abbreviation list index */
	bool		tt_isdst;	/* used to set tm_isdst */
};

struct lsinfo {				/* leap second information */
	time_t		ls_trans;	/* transition time (positive) */
	int_fast32_2s	ls_corr;	/* correction to apply */
};

/* This abbreviation means local time is unspecified.  */
static char const UNSPEC[] = "-00";

/* How many extra bytes are needed at the end of struct state's chars array.
   This needs to be at least 1 for null termination in case the input
   data isn't properly terminated, and it also needs to be big enough
   for ttunspecified to work without crashing.  */
enum { CHARS_EXTRA = max(sizeof UNSPEC, 2) - 1 };

/* A representation of the contents of a TZif file.  Ideally this
   would have no size limits; the following sizes should suffice for
   practical use.  This struct should not be too large, as instances
   are put on the stack and stacks are relatively small on some platforms.
   See tzfile.h for more about the sizes.  */
struct state {
#if TZ_RUNTIME_LEAPS
	int		leapcnt;
#endif
	int		timecnt;
	int		typecnt;
	int		charcnt;
	bool		goahead;
	time_t		ats[TZ_MAX_TIMES];
	unsigned char	types[TZ_MAX_TIMES];
	struct ttinfo	ttis[TZ_MAX_TYPES];
	char chars[max(max(TZ_MAX_CHARS + CHARS_EXTRA, sizeof "UTC"),
		       2 * (TZNAME_MAXIMUM + 1))];
#if TZ_RUNTIME_LEAPS
	struct lsinfo	lsis[TZ_MAX_LEAPS];
#endif
};

static int
leapcount(ATTRIBUTE_MAYBE_UNUSED struct state const *sp)
{
#if TZ_RUNTIME_LEAPS
  return sp->leapcnt;
#else
  return 0;
#endif
}
static void
set_leapcount(ATTRIBUTE_MAYBE_UNUSED struct state *sp,
	    ATTRIBUTE_MAYBE_UNUSED int leapcnt)
{
#if TZ_RUNTIME_LEAPS
  sp->leapcnt = leapcnt;
#endif
}
static struct lsinfo
lsinfo(ATTRIBUTE_MAYBE_UNUSED struct state const *sp,
       ATTRIBUTE_MAYBE_UNUSED int i)
{
#if TZ_RUNTIME_LEAPS
  return sp->lsis[i];
#else
  unreachable();
#endif
}
static void
set_lsinfo(ATTRIBUTE_MAYBE_UNUSED struct state *sp,
	   ATTRIBUTE_MAYBE_UNUSED int i,
	   ATTRIBUTE_MAYBE_UNUSED struct lsinfo lsinfo)
{
#if TZ_RUNTIME_LEAPS
  sp->lsis[i] = lsinfo;
#endif
}

enum r_type {
  JULIAN_DAY,		/* Jn = Julian day */
  DAY_OF_YEAR,		/* n = day of year */
  MONTH_NTH_DAY_OF_WEEK	/* Mm.n.d = month, week, day of week */
};

struct rule {
	enum r_type	r_type;		/* type of rule */
	int		r_day;		/* day number of rule */
	int		r_week;		/* week number of rule */
	int		r_mon;		/* month number of rule */
	int_fast32_t	r_time;		/* transition time of rule */
};

static struct tm *gmtsub(struct state const *, time_t const *, int_fast32_t,
			 struct tm *, time_t *);
static bool increment_overflow_time(time_t *, int_fast32_2s);
static int_fast32_2s leapcorr(struct state const *, time_t);
static struct tm *timesub(time_t const *, int_fast32_t, struct state const *,
			  struct tm *, time_t *);
static bool tzparse(char const *, struct state *, struct state const *);

#ifndef ALL_STATE
# define ALL_STATE 0
#endif

#if ALL_STATE
static struct state *	lclptr;
static struct state *	gmtptr;
#else
static struct state	lclmem;
static struct state	gmtmem;
static struct state *const lclptr = &lclmem;
static struct state *const gmtptr = &gmtmem;
#endif /* State Farm */

/* Maximum number of bytes in an efficiently-handled TZ string.
   Longer strings work, albeit less efficiently.  */
#ifndef TZ_STRLEN_MAX
# define TZ_STRLEN_MAX 255
#endif /* !defined TZ_STRLEN_MAX */

#if !USE_TIMEX_T || !defined TM_GMTOFF
static char		lcl_TZname[TZ_STRLEN_MAX + 1];
static int		lcl_is_set;
#endif

/*
** Section 4.12.3 of X3.159-1989 requires that
**	Except for the strftime function, these functions [asctime,
**	ctime, gmtime, localtime] return values in one of two static
**	objects: a broken-down time structure and an array of char.
** Thanks to Paul Eggert for noting this.
**
** Although this requirement was removed in C99 it is still present in POSIX.
** Follow the requirement if SUPPORT_C89, even though this is more likely to
** trigger latent bugs in programs.
*/

#if !USE_TIMEX_T

# if SUPPORT_C89
static struct tm	tm;
# endif

# if 2 <= HAVE_TZNAME + TZ_TIME_T
char *tzname[2] = { UNCONST(wildabbr), UNCONST(wildabbr) };
# endif
# if 2 <= USG_COMPAT + TZ_TIME_T
long			timezone;
int			daylight;
# endif
# if 2 <= ALTZONE + TZ_TIME_T
long			altzone;
# endif

#endif

/* Initialize *S to a value based on UTOFF, ISDST, and DESIGIDX.  */
static void
init_ttinfo(struct ttinfo *s, int_fast32_t utoff, bool isdst,
	    desigidx_type desigidx)
{
  s->tt_utoff = utoff;
  s->tt_isdst = isdst;
  s->tt_desigidx = desigidx;
}

/* Return true if SP's time type I does not specify local time.  */
static bool
ttunspecified(struct state const *sp, int i)
{
  char const *abbr = &sp->chars[sp->ttis[i].tt_desigidx];
  /* memcmp is likely faster than strcmp, and is safe due to CHARS_EXTRA.  */
  return memcmp(abbr, UNSPEC, sizeof UNSPEC) == 0;
}

static int_fast32_2s
detzcode(const char *const codep)
{
	register int		i;
	int_fast32_2s
	  maxval = TWO_31_MINUS_1,
	  minval = -1 - maxval,
	  result;

	result = codep[0] & 0x7f;
	for (i = 1; i < 4; ++i)
		result = (result << 8) | (codep[i] & 0xff);

	if (codep[0] & 0x80) {
	  /* Do two's-complement negation even on non-two's-complement machines.
	     This cannot overflow, as int_fast32_2s is wide enough.  */
	  result += minval;
	}
	return result;
}

static int_fast64_t
detzcode64(const char *const codep)
{
	register int_fast64_t result;
	register int	i;
	int_fast64_t one = 1;
	int_fast64_t halfmaxval = one << (64 - 2);
	int_fast64_t maxval = halfmaxval - 1 + halfmaxval;
	int_fast64_t minval = -TWOS_COMPLEMENT(int_fast64_t) - maxval;

	result = codep[0] & 0x7f;
	for (i = 1; i < 8; ++i)
		result = (result << 8) | (codep[i] & 0xff);

	if (codep[0] & 0x80) {
	  /* Do two's-complement negation even on non-two's-complement machines.
	     If the result would be minval - 1, return minval.  */
	  result -= !TWOS_COMPLEMENT(int_fast64_t) && result != 0;
	  result += minval;
	}
	return result;
}

#if !USE_TIMEX_T || !defined TM_GMTOFF

static void
update_tzname_etc(struct state const *sp, struct ttinfo const *ttisp)
{
# if HAVE_TZNAME
  tzname[ttisp->tt_isdst] = UNCONST(&sp->chars[ttisp->tt_desigidx]);
# endif
# if USG_COMPAT
  if (!ttisp->tt_isdst)
    timezone = - ttisp->tt_utoff;
# endif
# if ALTZONE
  if (ttisp->tt_isdst)
    altzone = - ttisp->tt_utoff;
# endif
}

/* If STDDST_MASK indicates that SP's TYPE provides useful info,
   update tzname, timezone, and/or altzone and return STDDST_MASK,
   diminished by the provided info if it is a specified local time.
   Otherwise, return STDDST_MASK.  See settzname for STDDST_MASK.  */
static int
may_update_tzname_etc(int stddst_mask, struct state *sp, int type)
{
  struct ttinfo *ttisp = &sp->ttis[type];
  int this_bit = 1 << ttisp->tt_isdst;
  if (stddst_mask & this_bit) {
    update_tzname_etc(sp, ttisp);
    if (!ttunspecified(sp, type))
      return stddst_mask & ~this_bit;
  }
  return stddst_mask;
}

static void
settzname(void)
{
	register struct state * const	sp = lclptr;
	register int			i;

	/* If STDDST_MASK & 1 we need info about a standard time.
	   If STDDST_MASK & 2 we need info about a daylight saving time.
	   When STDDST_MASK becomes zero we can stop looking.  */
	int stddst_mask = 0;

# if HAVE_TZNAME
	tzname[0] = tzname[1] = UNCONST(sp ? wildabbr : utc);
	stddst_mask = 3;
# endif
# if USG_COMPAT
	timezone = 0;
	stddst_mask = 3;
# endif
# if ALTZONE
	altzone = 0;
	stddst_mask |= 2;
# endif
	/*
	** And to get the latest time zone abbreviations into tzname. . .
	*/
	if (sp) {
	  for (i = sp->timecnt - 1; stddst_mask && 0 <= i; i--)
	    stddst_mask = may_update_tzname_etc(stddst_mask, sp, sp->types[i]);
	  for (i = sp->typecnt - 1; stddst_mask && 0 <= i; i--)
	    stddst_mask = may_update_tzname_etc(stddst_mask, sp, i);
	}
# if USG_COMPAT
	daylight = stddst_mask >> 1 ^ 1;
# endif
}

/* Replace bogus characters in time zone abbreviations.
   Return 0 on success, an errno value if a time zone abbreviation is
   too long.  */
static int
scrub_abbrs(struct state *sp)
{
	int i;

	/* Reject overlong abbreviations.  */
	for (i = 0; i < sp->charcnt - (TZNAME_MAXIMUM + 1); ) {
	  int len = strnlen(&sp->chars[i], TZNAME_MAXIMUM + 1);
	  if (TZNAME_MAXIMUM < len)
	    return EOVERFLOW;
	  i += len + 1;
	}

	/* Replace bogus characters.  */
	for (i = 0; i < sp->charcnt; ++i)
	  switch (sp->chars[i]) {
	  case '\0':
	  case '+': case '-': case '.':
	  case '0': case '1': case '2': case '3': case '4':
	  case '5': case '6': case '7': case '8': case '9':
	  case ':':
	  case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
	  case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
	  case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
	  case 'V': case 'W': case 'X': case 'Y': case 'Z':
	  case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
	  case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
	  case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
	  case 'v': case 'w': case 'x': case 'y': case 'z':
	    break;

	  default:
	    sp->chars[i] = '_';
	    break;
	  }

	return 0;
}

#endif

/* Return true if the TZif file with descriptor FD changed,
   or may have changed, since the last time we were called.
   Return false if it did not change.
   If *ST is valid it is the file's current status;
   otherwise, update *ST to the status if possible.  */
static bool
tzfile_changed(int fd, struct stat *st)
{
  /* If old_ctim.tv_sec, these variables hold the corresponding part
     of the file's metadata the last time this function was called.  */
  static struct timespec old_ctim;
  static dev_t old_dev;
  static ino_t old_ino;

  if (!st->st_ctime && fstat(fd, st) < 0) {
    /* We do not know the file's state, so reset.  */
    old_ctim.tv_sec = 0;
    return true;
  } else {
    /* Use the change time, as it changes more reliably; mod time can
       be set back with futimens etc.  Use subsecond timestamp
       resolution if available, as this can help distinguish files on
       non-POSIX platforms where st_dev and st_ino are unreliable.  */
    struct timespec ctim;
    /* Copy via members, as AIX 7.3 defaults to an incompatible st_ctim.  */
    ctim.tv_sec = st->st_ctime;
#if HAVE_STRUCT_STAT_ST_CTIM
    ctim.tv_nsec = st->st_ctim.tv_nsec;
#else
    ctim.tv_nsec = 0;
#endif

    if ((ctim.tv_sec ^ old_ctim.tv_sec) | (ctim.tv_nsec ^ old_ctim.tv_nsec)
	| (st->st_dev ^ old_dev) | (st->st_ino ^ old_ino)) {
      old_ctim = ctim;
      old_dev = st->st_dev;
      old_ino = st->st_ino;
      return true;
    }

    return false;
  }
}

/* Input buffer for data read from a compiled tz file.  */
union input_buffer {
  /* The first part of the buffer, interpreted as a header.  */
  struct tzhead tzhead;

  /* The entire buffer.  Ideally this would have no size limits;
     the following should suffice for practical use.  */
  char buf[2 * sizeof(struct tzhead) + 2 * sizeof(struct state)
	   + 4 * TZ_MAX_TIMES];
};

/* TZDIR with a trailing '/'.  It is null-terminated if OPENAT_TZDIR.  */
#if !OPENAT_TZDIR
ATTRIBUTE_NONSTRING
#endif
static char const tzdirslash[sizeof TZDIR + OPENAT_TZDIR] = TZDIR "/";
enum { tzdirslashlen = sizeof TZDIR };
#ifdef PATH_MAX
static_assert(tzdirslashlen <= PATH_MAX);  /* Sanity check; assumed below.  */
#endif

/* Local storage needed for 'tzloadbody'.  */
union local_storage {
  /* The results of analyzing the file's contents after it is opened.  */
  struct file_analysis {
    /* The input buffer.  */
    union input_buffer u;

    /* A temporary state used for parsing a TZ string in the file.  */
    struct state st;
  } u;

#if defined PATH_MAX && !OPENAT_TZDIR && !SUPPRESS_TZDIR
  /* The name of the file to be opened.  */
  char fullname[PATH_MAX];
#endif
};

/* These tzload flags can be ORed together, and fit into 'char'.  */

/* TZLOAD_FROMENV means the TZ string is a process-wide setting either
   taken from the TZ environment variable or inferred from TZ being unset.
   When 0 <= tz_change_interval, TZLOAD_FROMENV lets us skip reanalysis
   of a TZif file that did not change during the interval.  */
enum { TZLOAD_FROMENV = 1 };

/* Read any newline-surrounded TZ string.  */
enum { TZLOAD_TZSTRING = 2 };

/* TZ should be a file under TZDIR.  */
enum { TZLOAD_TZDIR_SUB = 4 };


/* Load tz data from the file named NAME into *SP.  Respect TZLOADFLAGS.
   Use **LSPP for temporary storage.  Return 0 on
   success, an errno value on failure.  */
static int
tzloadbody(char const *name, struct state *sp, char tzloadflags,
	   union local_storage **lspp)
{
	register int			i;
	register int			fid;
	register int			stored;
	register ssize_t		nread;
	char const *relname;
	union local_storage *lsp = *lspp;
	union input_buffer *up;
	register int tzheadsize = sizeof(struct tzhead);
	int dd = AT_FDCWD;
	int oflags = (O_RDONLY | O_BINARY | O_CLOEXEC | O_CLOFORK
		      | O_IGNORE_CTTY | O_NOCTTY | O_REGULAR);
	bool might_escape = false;
	int err;
	struct stat st;
	st.st_ctime = 0;

	sp->goahead = false;

	if (! name) {
		name = TZDEFAULT;
		if (! name)
		  return EINVAL;
		/* Omit "tzloadflags &= ~TZLOAD_FROMENV;" here, as that
		   would hurt performance by rereading and reanalyzing the
		   TZDEFAULT file even when it is unchanged.  */
	}

	if (name[0] == ':')
		++name;

	relname = name;

	/* If the program is privileged, NAME is TZDEFAULT or
	   subsidiary to TZDIR.  Also, NAME is not a device.  */
	if (name[0] == '/' && strcmp(name, TZDEFAULT) != 0) {
	  if (!SUPPRESS_TZDIR
	      && strncmp(relname, tzdirslash, tzdirslashlen) == 0)
	    for (relname += tzdirslashlen; *relname == '/'; relname++)
	      continue;
	  else if (issetugid())
	    return ENOTCAPABLE;
	  else
	    might_escape = true;
	}

	if (relname[0] != '/') {
	  if (!OPENAT_TZDIR || !O_RESOLVE_BENEATH) {
	    /* Fail if a relative name contains a non-terminal ".." component,
	       as such a name could read a non-directory outside TZDIR
	       when AT_FDCWD and O_RESOLVE_BENEATH are not available.  */
	    char const *component;
	    for (component = relname; component[0]; component++)
	      if (component[0] == '.' && component[1] == '.'
		  && component[2] == '/'
		  && (component == relname || component[-1] == '/')) {
		if (issetugid())
		  return ENOTCAPABLE;
		might_escape = true;
		break;
	      }
	  }

	  if (OPENAT_TZDIR && !SUPPRESS_TZDIR) {
	    /* Prefer O_SEARCH or O_PATH if available;
	       O_RDONLY should be OK too, as TZDIR is invariably readable.
	       O_DIRECTORY should be redundant but might help
	       on old platforms that mishandle trailing '/'.  */
	    dd = open(tzdirslash,
		      ((O_SEARCH ? O_SEARCH : O_PATH ? O_PATH : O_RDONLY)
		       | O_BINARY | O_CLOEXEC | O_CLOFORK | O_DIRECTORY));
	    if (dd < 0)
	      return errno;
	    if (O_RESOLVE_BENEATH && issetugid()) {
	      oflags |= O_RESOLVE_BENEATH;
	      might_escape = false;
	    }
	  }
	}

	if (!OPENAT_TZDIR && !SUPPRESS_TZDIR && name[0] != '/') {
		char *cp;
		size_t fullnamesize;
#ifdef PATH_MAX
		size_t namesizemax = PATH_MAX - tzdirslashlen;
		size_t namelen = strnlen (name, namesizemax);
		if (namesizemax <= namelen)
		  return ENAMETOOLONG;
#else
		size_t namelen = strlen (name);
#endif
		fullnamesize = tzdirslashlen + namelen + 1;

		/* Create a string "TZDIR/NAME".  Using sprintf here
		   would pull in stdio (and would fail if the
		   resulting string length exceeded INT_MAX!).  */
		if (ALL_STATE || sizeof *lsp < fullnamesize) {
		  lsp = malloc(max(sizeof *lsp, fullnamesize));
		  if (!lsp)
		    return HAVE_MALLOC_ERRNO ? errno : ENOMEM;
		  *lspp = lsp;
		}
		cp = mempcpy(lsp, tzdirslash, tzdirslashlen);
		cp = mempcpy(cp, name, namelen);
		*cp = '\0';
#if defined PATH_MAX && !OPENAT_TZDIR && !SUPPRESS_TZDIR
		name = lsp->fullname;
#else
		name = (char *) lsp;
#endif
	}

	/* For a platform that lacks O_REGULAR and a file that might
	   be outside TZDIR, check that it is a regular file,
	   as merely opening a device could have unwanted side effects.
	   Though racy, there is no portable way to fix the race.  */
	if (!O_REGULAR && might_escape) {
	  /* (oflags & O_RESOLVE_BENEATH) must be zero here.  */
	  if ((OPENAT_TZDIR ? fstatat(dd, relname, &st, 0) : stat(name, &st))
	      < 0)
	    return errno;
	  if (!S_ISREG(st.st_mode))
	    return EFTYPE;
	}
	fid = OPENAT_TZDIR ? openat(dd, relname, oflags) : open(name, oflags);
	err = errno;
	if (0 <= dd)
	  close(dd);
	if (fid < 0)
	  return err;

	/* If detecting changes to the the primary TZif file's state and
	   the file's status is unchanged, save time by returning now.
	   Otherwise read the file's contents.  Close the file either way.  */
	if (0 <= tz_change_interval && (tzloadflags & TZLOAD_FROMENV)
	    && !tzfile_changed(fid, &st))
	  err = -1;
	else {
	  if (ALL_STATE && !lsp) {
	    lsp = malloc(sizeof *lsp);
	    if (!lsp)
	      return HAVE_MALLOC_ERRNO ? errno : ENOMEM;
	    *lspp = lsp;
	  }
	  up = &lsp->u.u;
	  nread = read(fid, up->buf, sizeof up->buf);
	  err = tzheadsize <= nread ? 0 : nread < 0 ? errno : EFTYPE;
	}
	close(fid);
	if (err)
	  return err < 0 ? 0 : err;

	for (stored = 4; stored <= 8; stored *= 2) {
	    char version = up->tzhead.tzh_version[0];
	    bool skip_datablock = stored == 4 && version;
	    int_fast32_t datablock_size;
	    int_fast32_2s
	      ttisstdcnt = detzcode(up->tzhead.tzh_ttisstdcnt),
	      ttisutcnt = detzcode(up->tzhead.tzh_ttisutcnt),
	      leapcnt = detzcode(up->tzhead.tzh_leapcnt),
	      timecnt = detzcode(up->tzhead.tzh_timecnt),
	      typecnt = detzcode(up->tzhead.tzh_typecnt),
	      charcnt = detzcode(up->tzhead.tzh_charcnt);
	    char const *p = up->buf + tzheadsize;
	    /* Although tzfile(5) currently requires typecnt to be nonzero,
	       support future formats that may allow zero typecnt
	       in files that have a TZ string and no transitions.  */
	    if (! (0 <= leapcnt
		   && leapcnt <= (TZ_RUNTIME_LEAPS ? TZ_MAX_LEAPS : 0)
		   && 0 <= typecnt && typecnt <= TZ_MAX_TYPES
		   && 0 <= timecnt && timecnt <= TZ_MAX_TIMES
		   && 0 <= charcnt && charcnt <= TZ_MAX_CHARS
		   && 0 <= ttisstdcnt && ttisstdcnt <= TZ_MAX_TYPES
		   && 0 <= ttisutcnt && ttisutcnt <= TZ_MAX_TYPES))
	      return EFTYPE;
	    datablock_size
		    = (timecnt * stored		/* ats */
		       + timecnt		/* types */
		       + typecnt * 6		/* ttinfos */
		       + charcnt		/* chars */
		       + leapcnt * (stored + 4)	/* lsinfos */
		       + ttisstdcnt		/* ttisstds */
		       + ttisutcnt);		/* ttisuts */
	    if (nread < tzheadsize + datablock_size)
	      return EFTYPE;
	    if (skip_datablock)
		p += datablock_size;
	    else if (! ((ttisstdcnt == typecnt || ttisstdcnt == 0)
			&& (ttisutcnt == typecnt || ttisutcnt == 0)))
	      return EFTYPE;
	    else {
		int_fast64_t prevtr = -1;
		int_fast32_2s prevcorr;
		set_leapcount(sp, leapcnt);
		sp->timecnt = timecnt;
		sp->typecnt = typecnt;
		sp->charcnt = charcnt;

		/* Read transitions, discarding those out of time_t range.
		   But pretend the last transition before TIME_T_MIN
		   occurred at TIME_T_MIN.  */
		timecnt = 0;
		for (i = 0; i < sp->timecnt; ++i) {
			int_fast64_t at
			  = stored == 4 ? detzcode(p) : detzcode64(p);
			sp->types[i] = at <= TIME_T_MAX;
			if (sp->types[i]) {
			  time_t attime
			    = ((TYPE_SIGNED(time_t) ? at < TIME_T_MIN : at < 0)
			       ? TIME_T_MIN : at);
			  if (timecnt && attime <= sp->ats[timecnt - 1]) {
			    if (attime < sp->ats[timecnt - 1])
			      return EFTYPE;
			    sp->types[i - 1] = 0;
			    timecnt--;
			  }
			  sp->ats[timecnt++] = attime;
			}
			p += stored;
		}

		timecnt = 0;
		for (i = 0; i < sp->timecnt; ++i) {
			unsigned char typ = *p++;
			if (sp->typecnt <= typ)
			  return EFTYPE;
			if (sp->types[i])
				sp->types[timecnt++] = typ;
		}
		sp->timecnt = timecnt;
		for (i = 0; i < sp->typecnt; ++i) {
			register struct ttinfo *	ttisp;
			unsigned char isdst, desigidx;
			int_fast32_2s utoff = detzcode(p);

			/* Reject a UT offset equal to -2**31, as it might
			   cause trouble both in this file and in callers.
			   Also, it violates RFC 9636 section 3.2.  */
			if (utoff < -TWO_31_MINUS_1)
			  return EFTYPE;

			ttisp = &sp->ttis[i];
			ttisp->tt_utoff = utoff;
			p += 4;
			isdst = *p++;
			if (! (isdst < 2))
			  return EFTYPE;
			ttisp->tt_isdst = isdst;
			desigidx = *p++;
			if (! (desigidx < sp->charcnt))
			  return EFTYPE;
			ttisp->tt_desigidx = desigidx;
		}
		for (i = 0; i < sp->charcnt; ++i)
			sp->chars[i] = *p++;
		/* Ensure '\0'-terminated, and make it safe to call
		   ttunspecified later.  */
		memset(&sp->chars[i], 0, CHARS_EXTRA);

		/* Read leap seconds, discarding those out of time_t range.  */
		leapcnt = 0;
		for (i = 0; i < leapcount(sp); i++) {
		  int_fast64_t tr = stored == 4 ? detzcode(p) : detzcode64(p);
		  int_fast32_2s corr = detzcode(p + stored);
		  p += stored + 4;

		  /* Leap seconds cannot occur before the Epoch,
		     or out of order.  */
		  if (tr <= prevtr)
		    return EFTYPE;

		  /* To avoid other botches in this code, each leap second's
		     correction must differ from the previous one's by 1
		     second or less, except that the first correction can be
		     any value; these requirements are more generous than
		     RFC 9636, to allow future RFC extensions.  */
		  if (! (i == 0
			 || (prevcorr < corr
			     ? corr == prevcorr + 1
			     : (corr == prevcorr
				|| corr == prevcorr - 1))))
		    return EFTYPE;
		  prevtr = tr;
		  prevcorr = corr;

		  if (tr <= TIME_T_MAX) {
		    struct lsinfo ls;
		    ls.ls_trans = tr;
		    ls.ls_corr = corr;
		    set_lsinfo(sp, leapcnt, ls);
		    leapcnt++;
		  }
		}
		set_leapcount(sp, leapcnt);

		/* Do not bother to validate standard/wall and UT/local
		   indicators, as they are no longer used here.  */
		p += ttisstdcnt + ttisutcnt;
	    }

	    nread -= p - up->buf;
	    memmove(up->buf, p, nread);

	    /* If this is an old file, we're done.  */
	    if (!version)
	      break;
	}
	if ((tzloadflags & TZLOAD_TZSTRING) && nread > 2 &&
		up->buf[0] == '\n' && up->buf[nread - 1] == '\n') {
			struct state	*ts = &lsp->u.st;

			up->buf[nread - 1] = '\0';
			if (!tzparse(&up->buf[1], ts, sp))
			  return EFTYPE;
			else {
			  /* Attempt to reuse existing abbreviations.
			     Without this, America/Anchorage would
			     consume 50 bytes for abbreviations, as
			     sp->charcnt equals 40 (for LMT AST AWT APT AHST
			     AHDT YST AKDT AKST) and ts->charcnt equals 10
			     (for AKST AKDT).  Reusing means sp->charcnt can
			     stay 40 in this example.  */
			  int charcnt = sp->charcnt;
			  for (i = 0; i < ts->typecnt; i++) {
			    char *tsabbr = ts->chars + ts->ttis[i].tt_desigidx;
			    int j;
			    for (j = 0; j < charcnt; j++)
			      if (strcmp(sp->chars + j, tsabbr) == 0) {
				ts->ttis[i].tt_desigidx = j;
				break;
			      }
			    if (! (j < charcnt)) {
			      int tsabbrlen = strnlen(tsabbr, TZ_MAX_CHARS - j);
			      if (TZ_MAX_CHARS <= j + tsabbrlen)
				return EOVERFLOW;
			      else {
				char *cp = sp->chars + j;
				cp = mempcpy(cp, tsabbr, tsabbrlen);
				*cp = '\0';
				charcnt = j + tsabbrlen + 1;
				ts->ttis[i].tt_desigidx = j;
			      }
			    }
			  }

			  if (TZ_MAX_TYPES - sp->typecnt < ts->typecnt)
			    return EOVERFLOW;
			  else {
			    sp->charcnt = charcnt;

			    /* Ignore any trailing, no-op transitions generated
			       by zic as they don't help here and can run afoul
			       of bugs in zic 2016j or earlier.  */
			    while (1 < sp->timecnt
				   && (sp->types[sp->timecnt - 1]
				       == sp->types[sp->timecnt - 2]))
			      sp->timecnt--;

			    sp->goahead = ts->goahead;

			    for (i = 0; i < ts->timecnt; i++) {
			      time_t t = ts->ats[i];
			      if (increment_overflow_time(&t, leapcorr(sp, t))
				  || (0 < sp->timecnt
				      && t <= sp->ats[sp->timecnt - 1]))
				continue;
			      if (TZ_MAX_TIMES <= sp->timecnt)
				return EOVERFLOW;
			      sp->ats[sp->timecnt] = t;
			      sp->types[sp->timecnt] = (sp->typecnt
							+ ts->types[i]);
			      sp->timecnt++;
			    }
			    for (i = 0; i < ts->typecnt; i++)
			      sp->ttis[sp->typecnt++] = ts->ttis[i];
			  }
			}
	}
	if (sp->typecnt == 0)
	  return EFTYPE;

	return 0;
}

/* Load tz data from the file named NAME into *SP.  Respect TZLOADFLAGS.
   Return 0 on success, an errno value on failure.  */
static int
tzload(char const *name, struct state *sp, char tzloadflags)
{
  int r;
  union local_storage *lsp0;
  union local_storage *lsp;
#if ALL_STATE
  lsp = NULL;
#else
  union local_storage ls;
  lsp = &ls;
#endif
  lsp0 = lsp;
  r = tzloadbody(name, sp, tzloadflags, &lsp);
  if (lsp != lsp0)
    free(lsp);
  return r;
}

static const int	mon_lengths[2][MONSPERYEAR] = {
	{ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 },
	{ 31, 29, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 }
};

/* Is C an ASCII digit?  */
static bool
is_digit(char c)
{
  return '0' <= c && c <= '9';
}

/*
** Given a pointer into a timezone string, scan until a character that is not
** a valid character in a time zone abbreviation is found.
** Return a pointer to that character.
*/

ATTRIBUTE_PURE_114833 static const char *
getzname(register const char *strp)
{
	register char	c;

	while ((c = *strp) != '\0' && !is_digit(c) && c != ',' && c != '-' &&
		c != '+')
			++strp;
	return strp;
}

/*
** Given a pointer into an extended timezone string, scan until the ending
** delimiter of the time zone abbreviation is located.
** Return a pointer to the delimiter.
**
** As with getzname above, the legal character set is actually quite
** restricted, with other characters producing undefined results.
** We don't do any checking here; checking is done later in common-case code.
*/

ATTRIBUTE_PURE_114833 static const char *
getqzname(register const char *strp, const int delim)
{
	register int	c;

	while ((c = *strp) != '\0' && c != delim)
		++strp;
	return strp;
}

/*
** Given a pointer into a timezone string, extract a number from that string.
** Check that the number is within a specified range; if it is not, return
** NULL.
** Otherwise, return a pointer to the first character not part of the number.
*/

static const char *
getnum(register const char *strp, int *const nump, const int min, const int max)
{
	register char	c;
	register int	num;

	if (strp == NULL || !is_digit(c = *strp))
		return NULL;
	num = 0;
	do {
		num = num * 10 + (c - '0');
		if (num > max)
			return NULL;	/* illegal value */
		c = *++strp;
	} while (is_digit(c));
	if (num < min)
		return NULL;		/* illegal value */
	*nump = num;
	return strp;
}

/*
** Given a pointer into a timezone string, extract a number of seconds,
** in hh[:mm[:ss]] form, from the string.
** If any error occurs, return NULL.
** Otherwise, return a pointer to the first character not part of the number
** of seconds.
*/

static const char *
getsecs(register const char *strp, int_fast32_t *const secsp)
{
	int	num;
	int_fast32_t secsperhour = SECSPERHOUR;

	/*
	** 'HOURSPERDAY * DAYSPERWEEK - 1' allows quasi-POSIX rules like
	** "M10.4.6/26", which does not conform to POSIX,
	** but which specifies the equivalent of
	** "02:00 on the first Sunday on or after 23 Oct".
	*/
	strp = getnum(strp, &num, 0, HOURSPERDAY * DAYSPERWEEK - 1);
	if (strp == NULL)
		return NULL;
	*secsp = num * secsperhour;
	if (*strp == ':') {
		++strp;
		strp = getnum(strp, &num, 0, MINSPERHOUR - 1);
		if (strp == NULL)
			return NULL;
		*secsp += num * SECSPERMIN;
		if (*strp == ':') {
			++strp;
			/* 'SECSPERMIN' allows for leap seconds.  */
			strp = getnum(strp, &num, 0, SECSPERMIN);
			if (strp == NULL)
				return NULL;
			*secsp += num;
		}
	}
	return strp;
}

/*
** Given a pointer into a timezone string, extract an offset, in
** [+-]hh[:mm[:ss]] form, from the string.
** If any error occurs, return NULL.
** Otherwise, return a pointer to the first character not part of the time.
*/

static const char *
getoffset(register const char *strp, int_fast32_t *const offsetp)
{
	register bool neg = false;

	if (*strp == '-') {
		neg = true;
		++strp;
	} else if (*strp == '+')
		++strp;
	strp = getsecs(strp, offsetp);
	if (strp == NULL)
		return NULL;		/* illegal time */
	if (neg)
		*offsetp = -*offsetp;
	return strp;
}

/*
** Given a pointer into a timezone string, extract a rule in the form
** date[/time]. See POSIX Base Definitions section 8.3 variable TZ
** for the format of "date" and "time".
** If a valid rule is not found, return NULL.
** Otherwise, return a pointer to the first character not part of the rule.
*/

static const char *
getrule(const char *strp, register struct rule *const rulep)
{
	if (*strp == 'J') {
		/*
		** Julian day.
		*/
		rulep->r_type = JULIAN_DAY;
		++strp;
		strp = getnum(strp, &rulep->r_day, 1, DAYSPERNYEAR);
	} else if (*strp == 'M') {
		/*
		** Month, week, day.
		*/
		rulep->r_type = MONTH_NTH_DAY_OF_WEEK;
		++strp;
		strp = getnum(strp, &rulep->r_mon, 1, MONSPERYEAR);
		if (strp == NULL)
			return NULL;
		if (*strp++ != '.')
			return NULL;
		strp = getnum(strp, &rulep->r_week, 1, 5);
		if (strp == NULL)
			return NULL;
		if (*strp++ != '.')
			return NULL;
		strp = getnum(strp, &rulep->r_day, 0, DAYSPERWEEK - 1);
	} else if (is_digit(*strp)) {
		/*
		** Day of year.
		*/
		rulep->r_type = DAY_OF_YEAR;
		strp = getnum(strp, &rulep->r_day, 0, DAYSPERLYEAR - 1);
	} else	return NULL;		/* invalid format */
	if (strp == NULL)
		return NULL;
	if (*strp == '/') {
		/*
		** Time specified.
		*/
		++strp;
		strp = getoffset(strp, &rulep->r_time);
	} else	rulep->r_time = 2 * SECSPERHOUR;	/* default = 2:00:00 */
	return strp;
}

/*
** Given a year, a rule, and the offset from UT at the time that rule takes
** effect, calculate the year-relative time that rule takes effect.
*/

static int_fast32_t
transtime(time_t year, register const struct rule *const rulep,
	  const int_fast32_t offset)
{
	int d;  /* Day of year (zero-origin).  */
	register bool	leapyear;

	leapyear = isleap(year);

	if (rulep->r_type <= DAY_OF_YEAR) {
		/*
		** Jn - Julian day, 1 == January 1, 60 == March 1 even in leap
		** years.
		** In non-leap years, or if the day number is 59 or less, just
		** use the day number - 1.
		**
		** n - day of year.
		*/
		d = rulep->r_day - ((rulep->r_type < DAY_OF_YEAR)
				    & (!leapyear | (rulep->r_day <= 59)));
	} else {
		/*
		** Mm.n.d - nth "dth day" of month m.
		*/

		/*
		** Use Zeller's Congruence to get day-of-week of first day of
		** month.
		*/
		int i;
		bool janfeb = rulep->r_mon <= 2;
		int month = (rulep->r_mon
			     + (janfeb ? MONSPERYEAR : 0)); /* 3..14  */
		int adjustment = (TYPE_SIGNED(time_t) ? 0 : 400) - janfeb;
		int ay_rem = (year + adjustment) % YEARSPERREPEAT;
		int y = ay_rem + (ay_rem < 0 ? YEARSPERREPEAT : 0);
		int dow = (((13 * (month + 1)) / 5
			    + y + y / 4 - y / 100 + y / 400)
			   % DAYSPERWEEK);

		/*
		** "dow" is the day-of-week of the first day of the month. Get
		** the day-of-month (zero-origin) of the first "dow" day of the
		** month.
		*/
		d = rulep->r_day - dow;
		d += (rulep->r_week - (0 <= d)) * DAYSPERWEEK;
		if (mon_lengths[leapyear][rulep->r_mon - 1] <= d)
		  d -= DAYSPERWEEK;

		/*
		** "d" is the day-of-month (zero-origin) of the day we want.
		*/
		for (i = 0; i < rulep->r_mon - 1; ++i)
		  d += mon_lengths[leapyear][i];
	}

	/*
	** d is the origin-0 year-relative day in
	** question. To get the year-relative time of the specified local
	** time on that day, add the transition time and the current offset
	** from UT to d * SECSPERDAY.
	*/
	return d * SECSPERDAY + rulep->r_time + offset;
}

/*
** Given a POSIX.1 proleptic TZ string, fill in the rule tables as
** appropriate.
*/

static bool
tzparse(const char *name, struct state *sp, struct state const *basep)
{
	const char *			stdname;
	const char *			dstname;
	int_fast32_t			stdoffset;
	int_fast32_t			dstoffset;
	register char *			cp;
	ptrdiff_t stdlen, dstlen, charcnt;
	time_t atlo = TIME_T_MIN, leaplo = TIME_T_MIN;

	stdname = name;
	if (*name == '<') {
	  name++;
	  stdname = name;
	  name = getqzname(name, '>');
	  if (*name != '>')
	    return false;
	  stdlen = name - stdname;
	  name++;
	} else {
	  name = getzname(name);
	  stdlen = name - stdname;
	}
	if (! (0 < stdlen && stdlen <= TZNAME_MAXIMUM))
	  return false;
	name = getoffset(name, &stdoffset);
	if (name == NULL)
	  return false;
	charcnt = stdlen + 1;
	if (basep) {
	  if (0 < basep->timecnt)
	    atlo = basep->ats[basep->timecnt - 1];
	  set_leapcount(sp, leapcount(basep));
	  if (0 < leapcount(sp)) {
	    int i;
	    for (i = 0; i < leapcount(sp); i++)
	      set_lsinfo(sp, i, lsinfo(basep, i));
	    leaplo = lsinfo(sp, leapcount(sp) - 1).ls_trans;
	  }
	} else
	  set_leapcount(sp, 0);	/* So, we're off a little.  */
	sp->goahead = false;
	if (*name != '\0') {
		struct rule start, end;
		int timecnt;
		time_t janfirst, repeatbeg, year, yearbeg, yearlim;
		int_fast32_t janoffset = 0;

		if (*name == '<') {
			dstname = ++name;
			name = getqzname(name, '>');
			if (*name != '>')
			  return false;
			dstlen = name - dstname;
			name++;
		} else {
			dstname = name;
			name = getzname(name);
			dstlen = name - dstname; /* length of DST abbr. */
		}
		if (! (0 < dstlen && dstlen <= TZNAME_MAXIMUM))
		  return false;
		charcnt += dstlen + 1;
		if (*name != '\0' && *name != ',' && *name != ';') {
			name = getoffset(name, &dstoffset);
			if (name == NULL)
			  return false;
		} else	dstoffset = stdoffset - SECSPERHOUR;

		if (*name == '\0')
			name = TZDEFRULESTRING;
		if (! (*name == ',' || *name == ';'))
		  return false;

		name = getrule(name + 1, &start);
		if (!name)
		  return false;
		if (*name++ != ',')
		  return false;
		name = getrule(name, &end);
		if (!name || *name)
		  return false;
		sp->typecnt = 2;	/* standard time and DST */

		/* Two transitions per year, from atlo forward, and going on
		   for years_of_observations past max(atlo, leaplo).  */
		init_ttinfo(&sp->ttis[0], -stdoffset, false, 0);
		init_ttinfo(&sp->ttis[1], -dstoffset, true, stdlen + 1);
		timecnt = 0;
		repeatbeg = (atlo / SECSPERREPEAT
			     + (atlo < 0 && 0 < atlo % SECSPERREPEAT));
		yearbeg = repeatbeg * YEARSPERREPEAT + EPOCH_YEAR;
		janfirst = repeatbeg * SECSPERREPEAT;

		do {
		  int_fast32_t yearsecs
		    = year_days(yearbeg - 1) * SECSPERDAY;
		  time_t janfirst1 = janfirst;
		  yearbeg--;
		  if (increment_overflow_time(&janfirst1, -yearsecs)) {
		    janoffset = -yearsecs;
		    break;
		  }
		  janfirst = janfirst1;
		} while (atlo < janfirst);

		while (true) {
		  int_fast32_t yearsecs
		    = year_days(yearbeg) * SECSPERDAY;
		  time_t janfirst1 = janfirst;
		  if (increment_overflow_time(&janfirst1, yearsecs)
		      || atlo <= janfirst1)
		    break;
		  yearbeg++;
		  janfirst = janfirst1;
		}

		yearlim = yearbeg + years_of_observations;
		for (year = yearbeg; year < yearlim; year++) {
		  int_fast32_t
		    starttime = transtime(year, &start, stdoffset),
		    endtime = transtime(year, &end, dstoffset),
		    yearsecs = year_days(year) * SECSPERDAY;
		  bool reversed = endtime < starttime;
		  if (reversed) {
		    int_fast32_t swap = starttime;
		    starttime = endtime;
		    endtime = swap;
		  }
		  if (reversed
		      || (starttime < endtime
			  && endtime - starttime < yearsecs)) {
		    time_t at_added = TIME_T_MAX;
		    time_t at = janfirst;
		    if (! increment_overflow_time(&at, janoffset + starttime)
			&& atlo <= at) {
		      if (TZ_MAX_TIMES <= timecnt)
			return false;
		      sp->ats[timecnt] = at_added = at;
		      sp->types[timecnt++] = !reversed;
		    }
		    at = janfirst;
		    if (! increment_overflow_time(&at, janoffset + endtime)
			&& atlo <= at) {
		      if (TZ_MAX_TIMES <= timecnt)
			return false;
		      sp->ats[timecnt] = at_added = at;
		      sp->types[timecnt++] = reversed;
		    }
		    if (at_added < leaplo)
		      yearlim = year + years_of_observations;
		  }
		  if (increment_overflow_time(&janfirst, janoffset + yearsecs))
		    break;
		  janoffset = 0;
		}
		sp->timecnt = timecnt;
		if (! timecnt) {
		  sp->ttis[0] = sp->ttis[1];
		  sp->typecnt = 1;	/* Perpetual DST.  */
		} else if (years_of_observations <= year - yearbeg)
		  sp->goahead = true;
	} else {
		dstlen = 0;
		sp->typecnt = 1;		/* only standard time */
		sp->timecnt = 0;
		init_ttinfo(&sp->ttis[0], -stdoffset, false, 0);
	}
	sp->charcnt = charcnt;
	cp = sp->chars;
	cp = mempcpy(cp, stdname, stdlen);
	*cp++ = '\0';
	if (dstlen != 0) {
	  cp = mempcpy(cp, dstname, dstlen);
	  *cp = '\0';
	}
	return true;
}

static void
gmtload(struct state *const sp)
{
	if (!TZ_RUNTIME_LEAPS || tzload(etc_utc, sp, TZLOAD_TZSTRING) != 0)
	  tzparse("UTC0", sp, NULL);
}

#if !USE_TIMEX_T || !defined TM_GMTOFF

/* Return true if primary cached time zone data are fresh,
   i.e., if this function is known to have recently returned false.
   A call is recent if it occurred less than tz_change_interval seconds ago.
   NOW should be the current time.  */
static bool
fresh_tzdata(monotime_t now)
{
  /* If nonzero, the time of the last false return.  */
  static monotime_t last_checked;

  if (last_checked && now - last_checked < tz_change_interval)
    return true;
  last_checked = now;
  return false;
}

/* Initialize *SP to a value appropriate for the TZ setting NAME.
   Respect TZLOADFLAGS.
   Return 0 on success, an errno value on failure.  */
static int
zoneinit(struct state *sp, char const *name, char tzloadflags)
{
  if (name && ! name[0]) {
    /*
    ** User wants it fast rather than right.
    */
    set_leapcount(sp, 0);		/* so, we're off a little */
    sp->timecnt = 0;
    sp->typecnt = 0;
    sp->charcnt = 0;
    sp->goahead = false;
    init_ttinfo(&sp->ttis[0], 0, false, 0);
    strcpy(sp->chars, utc);
    return 0;
  } else {
    int err = tzload(name, sp, tzloadflags);
    if (err != 0 && name && name[0] != ':' && !(tzloadflags & TZLOAD_TZDIR_SUB)
	&& tzparse(name, sp, NULL))
      err = 0;
    if (err == 0)
      err = scrub_abbrs(sp);
    return err;
  }
}

/* If THREADED, upgrade a read lock to a write lock.
   Return 0 on success, a positive errno value otherwise.  */
static int
rd2wrlock(ATTRIBUTE_MAYBE_UNUSED bool threaded)
{
# if THREAD_RWLOCK
  if (threaded) {
    dounlock();
    return pthread_rwlock_wrlock(&locallock);
  }
# endif
  return 0;
}

/* Like tzset(), but in a critical section.
   If THREADED && THREAD_RWLOCK the caller has a read lock,
   and this function might upgrade it to a write lock.
   If WALL, act as if TZ is unset; although always false in this file,
   a wrapper .c file's obsolete and ineffective tzsetwall function can use it.
   If tz_change_interval is positive the time is NOW; otherwise ignore NOW.  */
static void
tzset_unlocked(bool threaded, bool wall, monotime_t now)
{
  char const *name;
  struct state *sp;
  char tzloadflags;
  size_t namelen;
  bool writing = false;

  for (;;) {
    name = wall ? NULL : getenv("TZ");
    sp = lclptr;
    tzloadflags = TZLOAD_FROMENV | TZLOAD_TZSTRING;
    namelen = sizeof lcl_TZname + 1; /* placeholder for no name */

    if (name) {
      namelen = strnlen(name, sizeof lcl_TZname);

      /* Abbreviate a string like "/usr/share/zoneinfo/America/Los_Angeles"
	 to its shorter equivalent "America/Los_Angeles".  */
      if (!SUPPRESS_TZDIR && tzdirslashlen < namelen
	  && memcmp(name, tzdirslash, tzdirslashlen) == 0) {
	char const *p = name + tzdirslashlen;
	while (*p == '/')
	  p++;
	if (*p && *p != ':') {
	  name = p;
	  namelen = strnlen(name, sizeof lcl_TZname);
	  tzloadflags |= TZLOAD_TZDIR_SUB;
	}
      }
    }

    if ((tz_change_interval <= 0 ? tz_change_interval < 0 : fresh_tzdata(now))
	&& (name
	    ? 0 < lcl_is_set && strcmp(lcl_TZname, name) == 0
	    : lcl_is_set < 0))
      return;

    if (!THREAD_RWLOCK || writing)
      break;
    if (rd2wrlock(threaded) != 0)
      return;
    writing = true;
  }

# if ALL_STATE
  if (! sp)
    lclptr = sp = malloc(sizeof *lclptr);
# endif
  if (sp) {
    int err = zoneinit(sp, name, tzloadflags);
    if (err != 0) {
      zoneinit(sp, "", 0);
      /* Abbreviate with "-00" if there was an error.
	 Do not treat a missing TZDEFAULT file as an error.  */
      if (name || err != ENOENT)
	strcpy(sp->chars, UNSPEC);
    }
    if (namelen < sizeof lcl_TZname) {
      char *cp = lcl_TZname;
      cp = mempcpy(cp, name, namelen);
      *cp = '\0';
    }
  }
  settzname();
  lcl_is_set = (sizeof lcl_TZname > namelen) - (sizeof lcl_TZname < namelen);
}

#endif

#if !defined TM_GMTOFF || !USE_TIMEX_T

/* If tz_change_interval is positive,
   return the current time as a monotonically nondecreasing value.
   Otherwise the return value does not matter.  */
static monotime_t
get_monotonic_time(void)
{
  struct timespec now;
  now.tv_sec = 0;
  if (0 < tz_change_interval)
    clock_gettime(CLOCK_MONOTONIC_COARSE, &now);
  return now.tv_sec;
}
#endif

#if !USE_TIMEX_T

void
tzset(void)
{
  monotime_t now = get_monotonic_time();
  int err = lock();
  if (0 < err) {
    errno = err;
    return;
  }
  tzset_unlocked(!err, false, now);
  unlock(!err);
}
#endif

static void
gmtcheck1(void)
{
#if ALL_STATE
  gmtptr = malloc(sizeof *gmtptr);
#endif
  if (gmtptr)
    gmtload(gmtptr);
}

static void
gmtcheck(void)
{
  static once_t gmt_once = ONCE_INIT;
  once(&gmt_once, gmtcheck1);
}

#if NETBSD_INSPIRED && !USE_TIMEX_T

timezone_t
tzalloc(char const *name)
{
  timezone_t sp = malloc(sizeof *sp);
  if (sp) {
    int err = zoneinit(sp, name, TZLOAD_TZSTRING);
    if (err != 0) {
      free(sp);
      errno = err;
      return NULL;
    }
  } else if (!HAVE_MALLOC_ERRNO)
    errno = ENOMEM;
  return sp;
}

#ifndef FREE_PRESERVES_ERRNO
# if ((defined _POSIX_VERSION && 202405 <= _POSIX_VERSION) \
      || (defined __GLIBC__ && 2 < __GLIBC__ + (33 <= __GLIBC_MINOR__)) \
      || defined __OpenBSD__ || defined __sun)
#  define FREE_PRESERVES_ERRNO 1
# else
#  define FREE_PRESERVES_ERRNO 0
# endif
#endif

void
tzfree(timezone_t sp)
{
  int err;
  if (!FREE_PRESERVES_ERRNO)
    err = errno;
  free(sp);
  if (!FREE_PRESERVES_ERRNO)
    errno = err;
}

/*
** NetBSD 6.1.4 has ctime_rz, but omit it because C23 deprecates ctime and
** POSIX.1-2024 removes ctime_r.  Both have potential security problems that
** ctime_rz would share.  Callers can instead use localtime_rz + strftime.
**
** NetBSD 6.1.4 has tzgetname, but omit it because it doesn't work
** in zones with three or more time zone abbreviations.
** Callers can instead use localtime_rz + strftime.
*/

#endif

#if !USE_TIMEX_T || !defined TM_GMTOFF

/*
** The easy way to behave "as if no library function calls" localtime
** is to not call it, so we drop its guts into "localsub", which can be
** freely called. (And no, the PANS doesn't require the above behavior,
** but it *is* desirable.)
**
** If successful and SETNAME is nonzero,
** set the applicable parts of tzname, timezone and altzone;
** however, it's OK to omit this step for proleptic TZ strings
** since in that case tzset should have already done this step correctly.
** SETNAME's type is int_fast32_t for compatibility with gmtsub,
** but it is actually a boolean and its value should be 0 or 1.
**
** If LTM_YEAR, store the resulting year-1900 into *LTM_YEAR rather
** than into the default TMP->tm_year; this prevents year overflow.
*/

/*ARGSUSED*/
static struct tm *
localsub(struct state const *sp, time_t const *timep, int_fast32_t setname,
	 struct tm *tmp, time_t *ltm_year)
{
	register const struct ttinfo *	ttisp;
	register int			i;
	register struct tm *		result;
	const time_t			t = *timep;

	if (sp == NULL) {
	  /* Don't bother to set tzname etc.; tzset has already done it.  */
	  return gmtsub(gmtptr, timep, 0, tmp, ltm_year);
	}
	if (sp->goahead && sp->ats[sp->timecnt - 1] < t) {
			/* Avoid integer overflow when time_t is signed, by
			   using secs_div_2 twice; the full value would
			   always be even, so halving does not round.  */
			time_t
			  tlo = sp->ats[sp->timecnt - 1],
			  diffyears = ((t / 2 - tlo / 2
					+ ((t % 2 - tlo % 2 + 2) / 2 - 1))
				       / (SECSPERREPEAT / 2)
				       * YEARSPERREPEAT),
			  years = diffyears + YEARSPERREPEAT,
			  secs_div_2 = (diffyears * (AVGSECSPERYEAR / 2)
					+ SECSPERREPEAT / 2),
			  newt = t - secs_div_2 - secs_div_2,
			  ryear;

			result = localsub(sp, &newt, setname, tmp, &ryear);
			if (result) {
			  if (ltm_year)
			    *ltm_year = ryear + years;
			  else {
# ifdef ckd_add
				if (ckd_add(&result->tm_year, ryear, years))
				  return NULL;
# else
				time_t newy = ryear + years;
				if (! (INT_MIN <= newy && newy <= INT_MAX))
					return NULL;
				result->tm_year = newy;
# endif
			  }
			}
			return result;
	}
	if (sp->timecnt == 0 || t < sp->ats[0]) {
		i = 0;
	} else {
		register int	lo = 1;
		register int	hi = sp->timecnt;

		while (lo < hi) {
			register int	mid = (lo + hi) >> 1;

			if (t < sp->ats[mid])
				hi = mid;
			else	lo = mid + 1;
		}
		i = sp->types[lo - 1];
	}
	ttisp = &sp->ttis[i];
	/*
	** To get (wrong) behavior that's compatible with System V Release 2.0
	** you'd replace the statement below with
	**	t += ttisp->tt_utoff;
	**	timesub(&t, 0, sp, tmp, ltm_year);
	*/
	result = timesub(&t, ttisp->tt_utoff, sp, tmp, ltm_year);
	if (result) {
	  result->tm_isdst = ttisp->tt_isdst;
# ifdef TM_ZONE
	  result->TM_ZONE = UNCONST(&sp->chars[ttisp->tt_desigidx]);
# endif
	  if (setname)
	    update_tzname_etc(sp, ttisp);
	}
	return result;
}
#endif

#if !USE_TIMEX_T

/* Return TMP, or a thread-specific struct tm * selected by WHICH.  */
static struct tm *
tm_multi(struct tm *tmp, ATTRIBUTE_MAYBE_UNUSED enum tm_multi which)
{
# if THREAD_SAFE && THREAD_TM_MULTI
  /* It is OK to check is_threaded() separately here; even if it
     returns a different value in other places in the caller,
     this function's behavior is still valid.  */
  if (is_threaded()) {
    /* Try to get a thread-specific struct tm *.
       Fall back on TMP if this fails.  */
    static pthread_once_t tm_multi_once = PTHREAD_ONCE_INIT;
    pthread_once(&tm_multi_once, tm_multi_key_init);
    if (!tm_multi_key_err) {
      struct tm *p = pthread_getspecific(tm_multi_key);
      if (!p) {
	p = malloc(N_TM_MULTI * sizeof *p);
	if (p && pthread_setspecific(tm_multi_key, p) != 0) {
	  free(p);
	  p = NULL;
	}
      }
      if (p)
	return &p[which];
    }
  }
# endif
  return tmp;
}

# if NETBSD_INSPIRED
struct tm *
localtime_rz(struct state *restrict sp, time_t const *restrict timep,
	     struct tm *restrict tmp)
{
  return localsub(sp, timep, 0, tmp, NULL);
}
# endif

static struct tm *
localtime_tzset(time_t const *timep, struct tm *tmp, bool setname)
{
  monotime_t now = get_monotonic_time();
  int err = lock();
  if (0 < err) {
    errno = err;
    return NULL;
  }
  if (0 <= tz_change_interval || setname || !lcl_is_set)
    tzset_unlocked(!err, false, now);
  tmp = localsub(lclptr, timep, setname, tmp, NULL);
  unlock(!err);
  return tmp;
}

struct tm *
localtime(const time_t *timep)
{
# if !SUPPORT_C89
  static struct tm tm;
# endif
  return localtime_tzset(timep, tm_multi(&tm, LOCALTIME_TM_MULTI), true);
}

struct tm *
localtime_r(const time_t *restrict timep, struct tm *restrict tmp)
{
  return localtime_tzset(timep, tmp, false);
}
#endif

/*
** gmtsub is to gmtime as localsub is to localtime.
*/

static struct tm *
gmtsub(ATTRIBUTE_MAYBE_UNUSED struct state const *sp, time_t const *timep,
       int_fast32_t offset, struct tm *tmp, time_t *ltm_year)
{
	register struct tm *	result;

	result = timesub(timep, offset, gmtptr, tmp, ltm_year);
#ifdef TM_ZONE
	/*
	** Could get fancy here and deliver something such as
	** "+xx" or "-xx" if offset is non-zero,
	** but this is no time for a treasure hunt.
	*/
	tmp->TM_ZONE = UNCONST(offset ? wildabbr
			       : gmtptr ? gmtptr->chars : utc);
#endif /* defined TM_ZONE */
	return result;
}

#if !USE_TIMEX_T

/*
* Re-entrant version of gmtime.
*/

struct tm *
gmtime_r(time_t const *restrict timep, struct tm *restrict tmp)
{
  gmtcheck();
  return gmtsub(gmtptr, timep, 0, tmp, NULL);
}

struct tm *
gmtime(const time_t *timep)
{
# if !SUPPORT_C89
  static struct tm tm;
# endif
  return gmtime_r(timep, tm_multi(&tm, GMTIME_TM_MULTI));
}

# if STD_INSPIRED

/* This function is obsolescent and may disappear in future releases.
   Callers can instead use localtime_rz with a fixed-offset zone.  */

struct tm *
offtime_r(time_t const *restrict timep, long offset, struct tm *restrict tmp)
{
  gmtcheck();
  return gmtsub(gmtptr, timep, offset, tmp, NULL);
}

struct tm *
offtime(time_t const *timep, long offset)
{
#  if !SUPPORT_C89
  static struct tm tm;
#  endif
  return offtime_r(timep, offset, tm_multi(&tm, OFFTIME_TM_MULTI));
}

# endif
#endif

/*
** Return the number of leap years through the end of the given year
** where, to make the math easy, the answer for year zero is defined as zero.
*/

static time_t
leaps_thru_end_of_nonneg(time_t y)
{
  return y / 4 - y / 100 + y / 400;
}

static time_t
leaps_thru_end_of(time_t y)
{
  return (y < 0
	  ? -1 - leaps_thru_end_of_nonneg(-1 - y)
	  : leaps_thru_end_of_nonneg(y));
}

static struct tm *
timesub(const time_t *timep, int_fast32_t offset,
	const struct state *sp, struct tm *tmp, time_t *ltm_year)
{
	register time_t			tdays;
	register const int *		ip;
	int_fast32_2s corr;
	register int			i;
	int_fast32_t idays, rem, dayoff, dayrem;
	time_t y;

	/* If less than SECSPERMIN, the number of seconds since the
	   most recent positive leap second; otherwise, do not add 1
	   to localtime tm_sec because of leap seconds.  */
	time_t secs_since_posleap = SECSPERMIN;

	corr = 0;
	i = sp ? leapcount(sp) : 0;
	while (--i >= 0) {
		struct lsinfo ls = lsinfo(sp, i);
		if (ls.ls_trans <= *timep) {
			corr = ls.ls_corr;
			if ((i == 0 ? 0 : lsinfo(sp, i - 1).ls_corr) < corr)
			  secs_since_posleap = *timep - ls.ls_trans;
			break;
		}
	}

	/* Calculate the year, avoiding integer overflow even if
	   time_t is unsigned.  */
	tdays = *timep / SECSPERDAY;
	rem = *timep % SECSPERDAY;
	rem += offset % SECSPERDAY - corr % SECSPERDAY + 3 * SECSPERDAY;
	dayoff = offset / SECSPERDAY - corr / SECSPERDAY + rem / SECSPERDAY - 3;
	rem %= SECSPERDAY;
	/* y = (EPOCH_YEAR
		+ floor((tdays + dayoff) / DAYSPERREPEAT) * YEARSPERREPEAT),
	   sans overflow.  But calculate against 1570 (EPOCH_YEAR -
	   YEARSPERREPEAT) instead of against 1970 so that things work
	   for localtime values before 1970 when time_t is unsigned.  */
	dayrem = tdays % DAYSPERREPEAT;
	dayrem += dayoff % DAYSPERREPEAT;
	y = (EPOCH_YEAR - YEARSPERREPEAT
	     + ((1 + dayoff / DAYSPERREPEAT + dayrem / DAYSPERREPEAT
		 - ((dayrem % DAYSPERREPEAT) < 0)
		 + tdays / DAYSPERREPEAT)
		* YEARSPERREPEAT));
	/* idays = (tdays + dayoff) mod DAYSPERREPEAT, sans overflow.  */
	idays = tdays % DAYSPERREPEAT;
	idays += dayoff % DAYSPERREPEAT + 2 * DAYSPERREPEAT;
	idays %= DAYSPERREPEAT;
	/* Increase Y and decrease IDAYS until IDAYS is in range for Y.  */
	while (year_days(y) <= idays) {
		int tdelta = idays / DAYSPERLYEAR;
		int_fast32_t ydelta = tdelta + !tdelta;
		time_t newy = y + ydelta;
		register int	leapdays;
		leapdays = leaps_thru_end_of(newy - 1) -
			leaps_thru_end_of(y - 1);
		idays -= ydelta * DAYSPERNYEAR;
		idays -= leapdays;
		y = newy;
	}

	if (ltm_year) {
	  *ltm_year = y - TM_YEAR_BASE;
	} else {
#ifdef ckd_add
	  if (ckd_add(&tmp->tm_year, y, -TM_YEAR_BASE)) {
	    errno = EOVERFLOW;
	    return NULL;
	  }
#else
	  if (!TYPE_SIGNED(time_t) && y < TM_YEAR_BASE) {
	    int signed_y = y;
	    tmp->tm_year = signed_y - TM_YEAR_BASE;
	  } else if ((!TYPE_SIGNED(time_t) || INT_MIN + TM_YEAR_BASE <= y)
		     && y - TM_YEAR_BASE <= INT_MAX)
	    tmp->tm_year = y - TM_YEAR_BASE;
	  else {
	    errno = EOVERFLOW;
	    return NULL;
	  }
#endif
	}
	tmp->tm_yday = idays;
	/*
	** The "extra" mods below avoid overflow problems.
	*/
	tmp->tm_wday = (TM_WDAY_BASE
			+ ((y % DAYSPERWEEK - TM_YEAR_BASE % DAYSPERWEEK)
			   % DAYSPERWEEK
			   * (DAYSPERNYEAR % DAYSPERWEEK))
			+ leaps_thru_end_of(y - 1)
			- leaps_thru_end_of(TM_YEAR_BASE - 1)
			+ idays);
	tmp->tm_wday %= DAYSPERWEEK;
	if (tmp->tm_wday < 0)
		tmp->tm_wday += DAYSPERWEEK;
	tmp->tm_hour = rem / SECSPERHOUR;
	rem %= SECSPERHOUR;
	tmp->tm_min = rem / SECSPERMIN;
	tmp->tm_sec = rem % SECSPERMIN;

	/* Use "... ??:??:60" at the end of the localtime minute containing
	   the second just before the positive leap second.  */
	tmp->tm_sec += secs_since_posleap <= tmp->tm_sec;

	ip = mon_lengths[isleap(y)];
	for (tmp->tm_mon = 0; idays >= ip[tmp->tm_mon]; ++(tmp->tm_mon))
		idays -= ip[tmp->tm_mon];
	tmp->tm_mday = idays + 1;
	tmp->tm_isdst = 0;
#ifdef TM_GMTOFF
	tmp->TM_GMTOFF = offset;
#endif /* defined TM_GMTOFF */
	return tmp;
}

/*
** Adapted from code provided by Robert Elz, who writes:
**	The "best" way to do mktime I think is based on an idea of Bob
**	Kridle's (so its said...) from a long time ago.
**	It does a binary search of the time_t space. Since time_t's are
**	just 32 bits, its a max of 32 iterations (even at 64 bits it
**	would still be very reasonable).
*/

#ifndef WRONG
# define WRONG (-1)
#endif /* !defined WRONG */

/*
** Normalize logic courtesy Paul Eggert.
*/

static bool
increment_overflow_64(int *ip, int_fast64_t j)
{
#ifdef ckd_add
  return ckd_add(ip, *ip, j);
#else
  if (j < 0 ? *ip < INT_MIN - j : INT_MAX - j < *ip)
    return true;
  *ip += j;
  return false;
#endif
}

static bool
increment_overflow_time_iinntt(time_t *tp, iinntt j)
{
#ifdef ckd_add
  return ckd_add(tp, *tp, j);
#else
  if (j < 0
      ? (TYPE_SIGNED(time_t) ? *tp < TIME_T_MIN - j : *tp <= -1 - j)
      : TIME_T_MAX - j < *tp)
    return true;
  *tp += j;
  return false;
#endif
}

static bool
increment_overflow_time_64(time_t *tp, int_fast64_t j)
{
#ifdef ckd_add
  return ckd_add(tp, *tp, j);
#else
  if (j < 0
      ? (TYPE_SIGNED(time_t) ? *tp < TIME_T_MIN - j : *tp <= -1 - j)
      : TIME_T_MAX - j < *tp)
    return true;
  *tp += j;
  return false;
#endif
}

static bool
increment_overflow_time(time_t *tp, int_fast32_2s j)
{
#ifdef ckd_add
	return ckd_add(tp, *tp, j);
#else
	/*
	** This is like
	** 'if (! (TIME_T_MIN <= *tp + j && *tp + j <= TIME_T_MAX)) ...',
	** except that it does the right thing even if *tp + j would overflow.
	*/
	if (! (j < 0
	       ? (TYPE_SIGNED(time_t) ? TIME_T_MIN - j <= *tp : -1 - j < *tp)
	       : *tp <= TIME_T_MAX - j))
		return true;
	*tp += j;
	return false;
#endif
}

/* Return A - B, where both are in the range -2**31 + 1 .. 2**31 - 1.
   The result cannot overflow.  */
static int_fast64_t
utoff_diff (int_fast32_t a, int_fast32_t b)
{
  int_fast64_t aa = a;
  return aa - b;
}

static int
tmcomp(register const struct tm *const atmp,
       register const struct tm *const btmp)
{
	register int	result;

	if (atmp->tm_year != btmp->tm_year)
		return atmp->tm_year < btmp->tm_year ? -1 : 1;
	if ((result = (atmp->tm_mon - btmp->tm_mon)) == 0 &&
		(result = (atmp->tm_mday - btmp->tm_mday)) == 0 &&
		(result = (atmp->tm_hour - btmp->tm_hour)) == 0 &&
		(result = (atmp->tm_min - btmp->tm_min)) == 0)
			result = atmp->tm_sec - btmp->tm_sec;
	return result;
}

/* Copy to *DEST from *SRC.  Copy only the members needed for mktime,
   as other members might not be initialized.  */
static void
mktmcpy(struct tm *dest, struct tm const *src)
{
  dest->tm_sec = src->tm_sec;
  dest->tm_min = src->tm_min;
  dest->tm_hour = src->tm_hour;
  dest->tm_mday = src->tm_mday;
  dest->tm_mon = src->tm_mon;
  dest->tm_year = src->tm_year;
  dest->tm_isdst = src->tm_isdst;
#if defined TM_GMTOFF && ! UNINIT_TRAP
  dest->TM_GMTOFF = src->TM_GMTOFF;
#endif
}

static time_t
time2sub(struct tm *const tmp,
	 struct tm *funcp(struct state const *, time_t const *,
			  int_fast32_t, struct tm *, time_t *),
	 struct state const *sp,
	 const int_fast32_t offset,
	 bool *okayp,
	 bool do_norm_secs)
{
	register int			dir;
	register int			i, j;
	register time_t			lo;
	register time_t			hi;
	iinntt y, mday, hour, min, saved_seconds;
	time_t				newt;
	time_t				t;
	struct tm			yourtm, mytm;

	*okayp = false;
	mktmcpy(&yourtm, tmp);

	min = yourtm.tm_min;
	if (do_norm_secs) {
	  min += yourtm.tm_sec / SECSPERMIN;
	  yourtm.tm_sec %= SECSPERMIN;
	  if (yourtm.tm_sec < 0) {
	    yourtm.tm_sec += SECSPERMIN;
	    min--;
	  }
	}

	hour = yourtm.tm_hour;
	hour += min / MINSPERHOUR;
	yourtm.tm_min = min % MINSPERHOUR;
	if (yourtm.tm_min < 0) {
	  yourtm.tm_min += MINSPERHOUR;
	  hour--;
	}

	mday = yourtm.tm_mday;
	mday += hour / HOURSPERDAY;
	yourtm.tm_hour = hour % HOURSPERDAY;
	if (yourtm.tm_hour < 0) {
	  yourtm.tm_hour += HOURSPERDAY;
	  mday--;
	}

	y = yourtm.tm_year;
	y += yourtm.tm_mon / MONSPERYEAR;
	yourtm.tm_mon %= MONSPERYEAR;
	if (yourtm.tm_mon < 0) {
	  yourtm.tm_mon += MONSPERYEAR;
	  y--;
	}

	/*
	** Turn y into an actual year number for now.
	** It is converted back to an offset from TM_YEAR_BASE later.
	*/
	y += TM_YEAR_BASE;

	while (mday <= 0) {
	  iinntt li = y - (yourtm.tm_mon <= 1);
	  mday += year_days(li);
	  y--;
	}
	while (DAYSPERLYEAR < mday) {
	  iinntt li = y + (1 < yourtm.tm_mon);
	  mday -= year_days(li);
	  y++;
	}
	yourtm.tm_mday = mday;
	for ( ; ; ) {
		i = mon_lengths[isleap(y)][yourtm.tm_mon];
		if (yourtm.tm_mday <= i)
			break;
		yourtm.tm_mday -= i;
		if (++yourtm.tm_mon >= MONSPERYEAR) {
			yourtm.tm_mon = 0;
			y++;
		}
	}
#ifdef ckd_add
	if (ckd_add(&yourtm.tm_year, y, -TM_YEAR_BASE))
	  return WRONG;
#else
	y -= TM_YEAR_BASE;
	if (! (INT_MIN <= y && y <= INT_MAX))
		return WRONG;
	yourtm.tm_year = y;
#endif
	if (yourtm.tm_sec >= 0 && yourtm.tm_sec < SECSPERMIN)
		saved_seconds = 0;
	else if (yourtm.tm_year < EPOCH_YEAR - TM_YEAR_BASE) {
		/*
		** We can't set tm_sec to 0, because that might push the
		** time below the minimum representable time.
		** Set tm_sec to 59 instead.
		** This assumes that the minimum representable time is
		** not in the same minute that a leap second was deleted from,
		** which is a safer assumption than using 58 would be.
		*/
		saved_seconds = yourtm.tm_sec;
		saved_seconds -= SECSPERMIN - 1;
		yourtm.tm_sec = SECSPERMIN - 1;
	} else {
		saved_seconds = yourtm.tm_sec;
		yourtm.tm_sec = 0;
	}
	/*
	** Do a binary search (this works whatever time_t's type is).
	*/
	lo = TIME_T_MIN;
	hi = TIME_T_MAX;
	for ( ; ; ) {
		t = lo / 2 + hi / 2;
		if (t < lo)
			t = lo;
		else if (t > hi)
			t = hi;
		if (! funcp(sp, &t, offset, &mytm, NULL)) {
			/*
			** Assume that t is too extreme to be represented in
			** a struct tm; arrange things so that it is less
			** extreme on the next pass.
			*/
			dir = (t > 0) ? 1 : -1;
		} else	dir = tmcomp(&mytm, &yourtm);
		if (dir != 0) {
			if (t == lo) {
				if (t == TIME_T_MAX)
					return WRONG;
				++t;
				++lo;
			} else if (t == hi) {
				if (t == TIME_T_MIN)
					return WRONG;
				--t;
				--hi;
			}
			if (lo > hi)
				return WRONG;
			if (dir > 0)
				hi = t;
			else	lo = t;
			continue;
		}
#if defined TM_GMTOFF && ! UNINIT_TRAP
		if (mytm.TM_GMTOFF != yourtm.TM_GMTOFF
		    && (yourtm.TM_GMTOFF < 0
			? (-SECSPERDAY <= yourtm.TM_GMTOFF
			   && (mytm.TM_GMTOFF <=
			       (min(INT_FAST32_MAX, LONG_MAX)
				+ yourtm.TM_GMTOFF)))
			: (yourtm.TM_GMTOFF <= SECSPERDAY
			   && ((max(INT_FAST32_MIN, LONG_MIN)
				+ yourtm.TM_GMTOFF)
			       <= mytm.TM_GMTOFF)))) {
		  /* MYTM matches YOURTM except with the wrong UT offset.
		     YOURTM.TM_GMTOFF is plausible, so try it instead.
		     It's OK if YOURTM.TM_GMTOFF contains uninitialized data,
		     since the guess gets checked.  */
		  time_t altt = t;
		  int_fast64_t offdiff;
		  bool v;
# ifdef ckd_sub
		  v = ckd_sub(&offdiff, mytm.TM_GMTOFF, yourtm.TM_GMTOFF);
# else
		  /* A ckd_sub approximation that is good enough here.  */
		  v = !(-TWO_31_MINUS_1 <= yourtm.TM_GMTOFF
			&& yourtm.TM_GMTOFF <= TWO_31_MINUS_1);
		  if (!v)
		    offdiff = utoff_diff(mytm.TM_GMTOFF, yourtm.TM_GMTOFF);
# endif
		  if (!v && !increment_overflow_time_64(&altt, offdiff)) {
		    struct tm alttm;
		    if (funcp(sp, &altt, offset, &alttm, NULL)
			&& alttm.tm_isdst == mytm.tm_isdst
			&& alttm.TM_GMTOFF == yourtm.TM_GMTOFF
			&& tmcomp(&alttm, &yourtm) == 0) {
		      t = altt;
		      mytm = alttm;
		    }
		  }
		}
#endif
		if (yourtm.tm_isdst < 0 || mytm.tm_isdst == yourtm.tm_isdst)
			break;
		/*
		** Right time, wrong type.
		** Hunt for right time, right type.
		** It's okay to guess wrong since the guess
		** gets checked.
		*/
		if (sp == NULL)
			return WRONG;
		for (i = sp->typecnt - 1; i >= 0; --i) {
			if (sp->ttis[i].tt_isdst != yourtm.tm_isdst)
				continue;
			for (j = sp->typecnt - 1; j >= 0; --j) {
				if (sp->ttis[j].tt_isdst == yourtm.tm_isdst)
					continue;
				if (ttunspecified(sp, j))
				  continue;
				newt = t;
				if (increment_overflow_time_64
				    (&newt,
				     utoff_diff(sp->ttis[j].tt_utoff,
						sp->ttis[i].tt_utoff)))
				  continue;
				if (! funcp(sp, &newt, offset, &mytm, NULL))
					continue;
				if (tmcomp(&mytm, &yourtm) != 0)
					continue;
				if (mytm.tm_isdst != yourtm.tm_isdst)
					continue;
				/*
				** We have a match.
				*/
				t = newt;
				goto label;
			}
		}
		return WRONG;
	}
label:
	if (increment_overflow_time_iinntt(&t, saved_seconds))
		return WRONG;
	if (funcp(sp, &t, offset, tmp, NULL))
		*okayp = true;
	return t;
}

static time_t
time2(struct tm * const	tmp,
      struct tm *funcp(struct state const *, time_t const *,
		       int_fast32_t, struct tm *, time_t *),
      struct state const *sp,
      const int_fast32_t offset,
      bool *okayp)
{
	time_t	t;

	/*
	** First try without normalization of seconds
	** (in case tm_sec contains a value associated with a leap second).
	** If that fails, try with normalization of seconds.
	*/
	t = time2sub(tmp, funcp, sp, offset, okayp, false);
	return *okayp ? t : time2sub(tmp, funcp, sp, offset, okayp, true);
}

static time_t
time1(struct tm *const tmp,
      struct tm *funcp(struct state const *, time_t const *,
		       int_fast32_t, struct tm *, time_t *),
      struct state const *sp,
      const int_fast32_t offset)
{
	register time_t			t;
	register int			samei, otheri;
	register int			sameind, otherind;
	register int			i;
	register int			nseen;
	char				seen[TZ_MAX_TYPES];
	unsigned char			types[TZ_MAX_TYPES];
	bool				okay;

	if (tmp == NULL) {
		errno = EINVAL;
		return WRONG;
	}
	if (tmp->tm_isdst > 1)
		tmp->tm_isdst = 1;
	t = time2(tmp, funcp, sp, offset, &okay);
	if (okay)
		return t;
	if (tmp->tm_isdst < 0)
#ifdef PCTS
		/*
		** POSIX Conformance Test Suite code courtesy Grant Sullivan.
		*/
		tmp->tm_isdst = 0;	/* reset to std and try again */
#else
		return t;
#endif /* !defined PCTS */
	/*
	** We're supposed to assume that somebody took a time of one type
	** and did some math on it that yielded a "struct tm" that's bad.
	** We try to divine the type they started from and adjust to the
	** type they need.
	*/
	if (sp == NULL)
		return WRONG;
	for (i = 0; i < sp->typecnt; ++i)
		seen[i] = false;
	nseen = 0;
	for (i = sp->timecnt - 1; i >= 0; --i)
		if (!seen[sp->types[i]] && !ttunspecified(sp, sp->types[i])) {
			seen[sp->types[i]] = true;
			types[nseen++] = sp->types[i];
		}
	for (sameind = 0; sameind < nseen; ++sameind) {
		samei = types[sameind];
		if (sp->ttis[samei].tt_isdst != tmp->tm_isdst)
			continue;
		for (otherind = 0; otherind < nseen; ++otherind) {
			otheri = types[otherind];
			if (sp->ttis[otheri].tt_isdst != tmp->tm_isdst) {
			  int sec = tmp->tm_sec;
			  if (!increment_overflow_64
			      (&tmp->tm_sec,
			       utoff_diff(sp->ttis[otheri].tt_utoff,
					  sp->ttis[samei].tt_utoff))) {
			    tmp->tm_isdst = !tmp->tm_isdst;
			    t = time2(tmp, funcp, sp, offset, &okay);
			    if (okay)
			      return t;
			    tmp->tm_isdst = !tmp->tm_isdst;
			  }
			  tmp->tm_sec = sec;
			}
		}
	}
	return WRONG;
}

#if !defined TM_GMTOFF || !USE_TIMEX_T

static time_t
mktime_tzname(struct state *sp, struct tm *tmp, bool setname)
{
  if (sp)
    return time1(tmp, localsub, sp, setname);
  else {
    gmtcheck();
    return time1(tmp, gmtsub, gmtptr, 0);
  }
}

# if USE_TIMEX_T
static
# endif
time_t
mktime(struct tm *tmp)
{
  monotime_t now = get_monotonic_time();
  time_t t;
  int err = lock();
  if (0 < err) {
    errno = err;
    return -1;
  }
  tzset_unlocked(!err, false, now);
  t = mktime_tzname(lclptr, tmp, true);
  unlock(!err);
  return t;
}

#endif

#if NETBSD_INSPIRED && !USE_TIMEX_T
time_t
mktime_z(struct state *restrict sp, struct tm *restrict tmp)
{
  return mktime_tzname(sp, tmp, false);
}
#endif

#if STD_INSPIRED && !USE_TIMEX_T
/* This function is obsolescent and may disappear in future releases.
   Callers can instead use mktime.  */
time_t
timelocal(struct tm *tmp)
{
	if (tmp != NULL)
		tmp->tm_isdst = -1;	/* in case it wasn't initialized */
	return mktime(tmp);
}
#endif

#if defined TM_GMTOFF || !USE_TIMEX_T

# ifndef EXTERN_TIMEOFF
#  ifndef timeoff
#   define timeoff my_timeoff /* Don't collide with OpenBSD 7.4 .  */
#  endif
#  define EXTERN_TIMEOFF static
# endif

/* This function is obsolescent and may disappear in future releases.
   Callers can instead use mktime_z with a fixed-offset zone.  */
EXTERN_TIMEOFF time_t
timeoff(struct tm *tmp, long offset)
{
  if (tmp)
    tmp->tm_isdst = 0;
  gmtcheck();
  return time1(tmp, gmtsub, gmtptr, offset);
}
#endif

#if !USE_TIMEX_T
time_t
timegm(struct tm *tmp)
{
  time_t t;
  struct tm tmcpy;
  mktmcpy(&tmcpy, tmp);
  tmcpy.tm_wday = -1;
  t = timeoff(&tmcpy, 0);
  if (0 <= tmcpy.tm_wday)
    *tmp = tmcpy;
  return t;
}
#endif

static int_fast32_2s
leapcorr(struct state const *sp, time_t t)
{
	register int			i;

	i = leapcount(sp);
	while (--i >= 0) {
	  struct lsinfo ls = lsinfo(sp, i);
	  if (ls.ls_trans <= t)
	    return ls.ls_corr;
	}
	return 0;
}

/*
** XXX--is the below the right way to conditionalize??
*/

#if !USE_TIMEX_T
# if STD_INSPIRED

static bool
decrement_overflow_time(time_t *tp, int_fast32_2s j)
{
#ifdef ckd_sub
  return ckd_sub(tp, *tp, j);
#else
  if (! (j < 0
	 ? *tp <= TIME_T_MAX + j
	 : (TYPE_SIGNED(time_t) ? TIME_T_MIN + j <= *tp : j <= *tp)))
    return true;
  *tp -= j;
  return false;
#endif
}

/* NETBSD_INSPIRED_EXTERN functions are exported to callers if
   NETBSD_INSPIRED is defined, and are private otherwise.  */
#  if NETBSD_INSPIRED
#   define NETBSD_INSPIRED_EXTERN
#  else
#   define NETBSD_INSPIRED_EXTERN static
#  endif

/*
** IEEE Std 1003.1 (POSIX) says that 536457599
** shall correspond to "Wed Dec 31 23:59:59 UTC 1986", which
** is not the case if we are accounting for leap seconds.
** So, we provide the following conversion routines for use
** when exchanging timestamps with POSIX conforming systems.
*/

NETBSD_INSPIRED_EXTERN time_t
time2posix_z(struct state *sp, time_t t)
{
  if (decrement_overflow_time(&t, leapcorr(sp, t))) {
    /* Overflow near maximum time_t value with negative correction.
       This can happen with unrealistic-but-valid TZif files.  */
    errno = EOVERFLOW;
    return -1;
  }
  return t;
}

time_t
time2posix(time_t t)
{
  monotime_t now = get_monotonic_time();
  int err = lock();
  if (0 < err) {
    errno = err;
    return -1;
  }
  if (0 <= tz_change_interval || !lcl_is_set)
    tzset_unlocked(!err, false, now);
  if (lclptr)
    t = time2posix_z(lclptr, t);
  unlock(!err);
  return t;
}

NETBSD_INSPIRED_EXTERN time_t
posix2time_z(struct state *sp, time_t t)
{
  int i;
  for (i = leapcount(sp); 0 <= --i; ) {
    struct lsinfo ls = lsinfo(sp, i);
    time_t t_corr = t;

    if (increment_overflow_time(&t_corr, ls.ls_corr)) {
      if (0 <= ls.ls_corr) {
	/* Overflow near maximum time_t value with positive correction.
	   This can happen with ordinary TZif files with leap seconds.  */
	errno = EOVERFLOW;
	return -1;
      } else {
	/* A negative correction overflowed, so keep going.
	   This can happen with unrealistic-but-valid TZif files.  */
      }
    } else if (ls.ls_trans <= t_corr)
      return (t_corr
	      - (ls.ls_trans == t_corr
		 && (i == 0 ? 0 : lsinfo(sp, i - 1).ls_corr) < ls.ls_corr));
  }
  return t;
}

time_t
posix2time(time_t t)
{
  monotime_t now = get_monotonic_time();
  int err = lock();
  if (0 < err) {
    errno = err;
    return -1;
  }
  if (0 <= tz_change_interval || !lcl_is_set)
    tzset_unlocked(!err, false, now);
  if (lclptr)
    t = posix2time_z(lclptr, t);
  unlock(!err);
  return t;
}

# endif /* STD_INSPIRED */

# if TZ_TIME_T

#  if !USG_COMPAT
#   define timezone 0
#  endif

/* Convert from the underlying system's time_t to the ersatz time_tz,
   which is called 'time_t' in this file.  Typically, this merely
   converts the time's integer width.  On some platforms, the system
   time is local time not UT, or uses some epoch other than the POSIX
   epoch.

   Although this code appears to define a function named 'time' that
   returns time_t, the macros in private.h cause this code to actually
   define a function named 'tz_time' that returns tz_time_t.  The call
   to sys_time invokes the underlying system's 'time' function.  */

time_t
time(time_t *p)
{
  time_t r = sys_time(NULL);
  if (r != (time_t) -1) {
    iinntt offset = EPOCH_LOCAL ? timezone : 0;
    if (offset < IINNTT_MIN + EPOCH_OFFSET
	|| increment_overflow_time_iinntt(&r, offset - EPOCH_OFFSET)) {
      errno = EOVERFLOW;
      r = -1;
    }
  }
  if (p)
    *p = r;
  return r;
}

# endif
#endif
asctime.c0000644000000000000000000001074615251675224007616 0ustar00/* asctime a la ISO C.  */

/*
** This file is in the public domain, so clarified as of
** 1996-06-05 by Arthur David Olson.
*/

/*
** Avoid the temptation to punt entirely to strftime;
** strftime can behave badly when tm components are out of range, and
** the output of strftime is supposed to be locale specific
** whereas the output of asctime is supposed to be constant.
*/

/*LINTLIBRARY*/

#include "private.h"
#include 

enum { STD_ASCTIME_BUF_SIZE = 26 };
/*
** Big enough for something such as
** ??? ???-2147483648 -2147483648:-2147483648:-2147483648     -2147483648\n
** (two three-character abbreviations, five strings denoting integers,
** seven explicit spaces, two explicit colons, a newline,
** and a trailing NUL byte).
** The values above are for systems where an int is 32 bits and are provided
** as an example; the size expression below is a bound for the system at
** hand.
*/
static char buf_asctime[2*3 + 5*INT_STRLEN_MAXIMUM(int) + 7 + 2 + 1 + 1];

/* On pre-C99 platforms, a snprintf substitute good enough for us.  */
#if !HAVE_SNPRINTF
# include 
ATTRIBUTE_FORMAT((printf, 3, 4)) static int
my_snprintf(char *s, size_t size, char const *format, ...)
{
  int n;
  va_list args;
  char stackbuf[sizeof buf_asctime];
  va_start(args, format);
  n = vsprintf(stackbuf, format, args);
  va_end (args);
  if (0 <= n && n < size)
    memcpy (s, stackbuf, n + 1);
  return n;
}
# undef snprintf
# define snprintf my_snprintf
#endif

/* Publish asctime_r and ctime_r only when supporting older POSIX.  */
#if SUPPORT_POSIX2008
# define asctime_static
#else
# define asctime_static static
# undef asctime_r
# undef ctime_r
# define asctime_r static_asctime_r
# define ctime_r static_ctime_r
#endif

asctime_static
char *
asctime_r(struct tm const *restrict timeptr, char *restrict buf)
{
	static const char	wday_name[][4] = {
		"Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat"
	};
	static const char	mon_name[][4] = {
		"Jan", "Feb", "Mar", "Apr", "May", "Jun",
		"Jul", "Aug", "Sep", "Oct", "Nov", "Dec"
	};
	register const char *	wn;
	register const char *	mn;
	int year, mday, hour, min, sec;
	long long_TM_YEAR_BASE = TM_YEAR_BASE;
	int bufsize = (buf == buf_asctime
		       ? sizeof buf_asctime : STD_ASCTIME_BUF_SIZE);

	if (timeptr == NULL) {
		strcpy(buf, "??? ??? ?? ??:??:?? ????\n");
		/* Set errno now, since strcpy might change it in
		   POSIX.1-2017 and earlier.  */
		errno = EINVAL;
		return buf;
	}
	if (timeptr->tm_wday < 0 || timeptr->tm_wday >= DAYSPERWEEK)
		wn = "???";
	else	wn = wday_name[timeptr->tm_wday];
	if (timeptr->tm_mon < 0 || timeptr->tm_mon >= MONSPERYEAR)
		mn = "???";
	else	mn = mon_name[timeptr->tm_mon];

	year = timeptr->tm_year;
	mday = timeptr->tm_mday;
	hour = timeptr->tm_hour;
	min = timeptr->tm_min;
	sec = timeptr->tm_sec;

	/* Vintage programs are coded for years that are always four bytes long
	   and may assume that the newline always lands in the same place.
	   For years that are less than four bytes, pad the output with
	   leading zeroes to get the newline in the traditional place.
	   For years longer than four bytes, put extra spaces before the year
	   so that vintage code trying to overwrite the newline
	   won't overwrite a digit within a year and truncate the year,
	   using the principle that no output is better than wrong output.
	   This conforms to ISO C and POSIX standards, which say behavior
	   is undefined when the year is less than 1000 or greater than 9999.

	   Also, avoid overflow when formatting tm_year + TM_YEAR_BASE.  */

	if ((year <= LONG_MAX - TM_YEAR_BASE
	     ? snprintf (buf, bufsize,
			 ((-999 - TM_YEAR_BASE <= year
			   && year <= 9999 - TM_YEAR_BASE)
			  ? "%s %s%3d %.2d:%.2d:%.2d %04ld\n"
			  : "%s %s%3d %.2d:%.2d:%.2d     %ld\n"),
			 wn, mn, mday, hour, min, sec,
			 year + long_TM_YEAR_BASE)
	     : snprintf (buf, bufsize,
			 "%s %s%3d %.2d:%.2d:%.2d     %d%d\n",
			 wn, mn, mday, hour, min, sec,
			 year / 10 + TM_YEAR_BASE / 10,
			 year % 10))
	    < bufsize)
		return buf;
	else {
		errno = EOVERFLOW;
		return NULL;
	}
}

char *
asctime(register const struct tm *timeptr)
{
	return asctime_r(timeptr, buf_asctime);
}

asctime_static
char *
ctime_r(const time_t *timep, char *buf)
{
  struct tm mytm;
  struct tm *tmp = localtime_r(timep, &mytm);
  return tmp ? asctime_r(tmp, buf) : NULL;
}

char *
ctime(const time_t *timep)
{
  /* Do not call localtime_r, as C23 requires ctime to initialize the
     static storage that localtime updates.  */
  struct tm *tmp = localtime(timep);
  return tmp ? asctime(tmp) : NULL;
}
difftime.c0000644000000000000000000000263014031301343007730 0ustar00/* Return the difference between two timestamps.  */

/*
** This file is in the public domain, so clarified as of
** 1996-06-05 by Arthur David Olson.
*/

/*LINTLIBRARY*/

#include "private.h"	/* for time_t and TYPE_SIGNED */

/* Return -X as a double.  Using this avoids casting to 'double'.  */
static double
dminus(double x)
{
  return -x;
}

double
difftime(time_t time1, time_t time0)
{
	/*
	** If double is large enough, simply convert and subtract
	** (assuming that the larger type has more precision).
	*/
	if (sizeof(time_t) < sizeof(double)) {
	  double t1 = time1, t0 = time0;
	  return t1 - t0;
	}

	/*
	** The difference of two unsigned values can't overflow
	** if the minuend is greater than or equal to the subtrahend.
	*/
	if (!TYPE_SIGNED(time_t))
	  return time0 <= time1 ? time1 - time0 : dminus(time0 - time1);

	/* Use uintmax_t if wide enough.  */
	if (sizeof(time_t) <= sizeof(uintmax_t)) {
	  uintmax_t t1 = time1, t0 = time0;
	  return time0 <= time1 ? t1 - t0 : dminus(t0 - t1);
	}

	/*
	** Handle cases where both time1 and time0 have the same sign
	** (meaning that their difference cannot overflow).
	*/
	if ((time1 < 0) == (time0 < 0))
	  return time1 - time0;

	/*
	** The values have opposite signs and uintmax_t is too narrow.
	** This suffers from double rounding; attempt to lessen that
	** by using long double temporaries.
	*/
	{
	  long double t1 = time1, t0 = time0;
	  return t1 - t0;
	}
}
strftime.c0000644000000000000000000004556015244464512010026 0ustar00/* Convert a broken-down timestamp to a string.  */

/* Copyright 1989 The Regents of the University of California.
   All rights reserved.

   Redistribution and use in source and binary forms, with or without
   modification, are permitted provided that the following conditions
   are met:
   1. Redistributions of source code must retain the above copyright
      notice, this list of conditions and the following disclaimer.
   2. Redistributions in binary form must reproduce the above copyright
      notice, this list of conditions and the following disclaimer in the
      documentation and/or other materials provided with the distribution.
   3. Neither the name of the University nor the names of its contributors
      may be used to endorse or promote products derived from this software
      without specific prior written permission.

   THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS "AS IS" AND
   ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
   IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
   ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
   FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
   DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
   OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
   HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
   LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
   OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
   SUCH DAMAGE.  */

/*
** Based on the UCB version with the copyright notice appearing above.
**
** This is ANSIish only when "multibyte character == plain character".
*/

#include "private.h"

#include 
#include 
#include 

/* If true, the value returned by an idealized unlimited-range mktime
   always fits into an integer type with bounds MIN and MAX.
   If false, the value might not fit.
   This macro is usable in #if if its arguments are.
   Add or subtract 2**31 - 1 for the maximum UT offset allowed in a TZif file,
   divide by the maximum number of non-leap seconds in a year,
   divide again by two just to be safe,
   and account for the tm_year origin (1900) and time_t origin (1970).  */
#define MKTIME_FITS_IN(min, max) \
  ((min) < 0 \
   && (((min) + TWO_31_MINUS_1) / 366 / 24 / 60 / 60 / 2 + 1970 - 1900 \
       < INT_MIN)						\
   && INT_MAX < ((max) - TWO_31_MINUS_1) / 366 / 24 / 60 / 60 / 2 + 1970 - 1900)

/* MKTIME_MIGHT_OVERFLOW is true if mktime can fail due to time_t overflow
   or if it is not known whether mktime can fail,
   and is false if mktime definitely cannot fail.
   This macro is usable in #if, and so does not use TIME_T_MAX or sizeof.
   If the builder has not configured this macro, guess based on what
   known platforms do.  When in doubt, guess true.  */
#ifndef MKTIME_MIGHT_OVERFLOW
# if defined __FreeBSD__ || defined __NetBSD__ || defined __OpenBSD__
#  include 
# endif
# if ((/* The following heuristics assume native time_t.  */ \
       defined_time_tz) \
      || ((/* Traditional time_t is 'long', so if 'long' is not wide enough \
	      assume overflow unless we're on a known-safe host.  */ \
	   !MKTIME_FITS_IN(LONG_MIN, LONG_MAX)) \
	  && (/* GNU C Library 2.29 (2019-02-01) and later has 64-bit time_t \
		 if __TIMESIZE is 64.  */ \
	      !defined __TIMESIZE || __TIMESIZE < 64) \
	  && (/* FreeBSD 12 r320347 (__FreeBSD_version 1200036; 2017-06-26), \
		 and later has 64-bit time_t on all platforms but i386 which \
		 is currently scheduled for end-of-life on 2028-11-30.  */ \
	      !defined __FreeBSD_version || __FreeBSD_version < 1200036 \
	      || defined __i386) \
	  && (/* NetBSD 6.0 (2012-10-17) and later has 64-bit time_t.  */ \
	      !defined __NetBSD_Version__ || __NetBSD_Version__ < 600000000) \
	  && (/* OpenBSD 5.5 (2014-05-01) and later has 64-bit time_t.  */ \
	      !defined OpenBSD || OpenBSD < 201405)))
#  define MKTIME_MIGHT_OVERFLOW 1
# else
#  define MKTIME_MIGHT_OVERFLOW 0
# endif
#endif
/* Check that MKTIME_MIGHT_OVERFLOW is consistent with time_t's range.  */
static_assert(MKTIME_MIGHT_OVERFLOW
	      || MKTIME_FITS_IN(TIME_T_MIN, TIME_T_MAX));

#if MKTIME_MIGHT_OVERFLOW
/* Do this after system includes as it redefines time_t, mktime, timeoff.  */
# define USE_TIMEX_T true
# include "localtime.c"
#endif

#ifndef DEPRECATE_TWO_DIGIT_YEARS
# define DEPRECATE_TWO_DIGIT_YEARS 0
#endif

struct lc_time_T {
	const char *	mon[MONSPERYEAR];
	const char *	month[MONSPERYEAR];
	const char *	wday[DAYSPERWEEK];
	const char *	weekday[DAYSPERWEEK];
	const char *	X_fmt;
	const char *	x_fmt;
	const char *	c_fmt;
	const char *	am;
	const char *	pm;
	const char *	date_fmt;
};

static const struct lc_time_T	C_time_locale = {
	{
		"Jan", "Feb", "Mar", "Apr", "May", "Jun",
		"Jul", "Aug", "Sep", "Oct", "Nov", "Dec"
	}, {
		"January", "February", "March", "April", "May", "June",
		"July", "August", "September", "October", "November", "December"
	}, {
		"Sun", "Mon", "Tue", "Wed",
		"Thu", "Fri", "Sat"
	}, {
		"Sunday", "Monday", "Tuesday", "Wednesday",
		"Thursday", "Friday", "Saturday"
	},

	/* X_fmt */
	"%H:%M:%S",

	/*
	** x_fmt
	** C99 and later require this format.
	** Using just numbers (as here) makes Quakers happier;
	** it's also compatible with SVR4.
	*/
	"%m/%d/%y",

	/*
	** c_fmt
	** C99 and later require this format.
	** Previously this code used "%D %X", but we now conform to C99.
	** Note that
	**	"%a %b %d %H:%M:%S %Y"
	** is used by Solaris 2.3.
	*/
	"%a %b %e %T %Y",

	/* am */
	"AM",

	/* pm */
	"PM",

	/* date_fmt */
	"%a %b %e %H:%M:%S %Z %Y"
};

enum warn { IN_NONE, IN_SOME, IN_THIS, IN_ALL };

static char *	_add(const char *, char *, const char *);
static char *	_conv(int, const char *, char *, const char *);
static char *	_fmt(const char *, const struct tm *, char *, const char *,
		     enum warn *);
static char *	_yconv(int, int, bool, bool, char *, char const *);

#ifndef YEAR_2000_NAME
# define YEAR_2000_NAME "CHECK_STRFTIME_FORMATS_FOR_TWO_DIGIT_YEARS"
#endif /* !defined YEAR_2000_NAME */

#if HAVE_STRFTIME_L
size_t
strftime_l(char *restrict s, size_t maxsize, char const *restrict format,
	   struct tm const *restrict t,
	   ATTRIBUTE_MAYBE_UNUSED locale_t locale)
{
  /* Just call strftime, as only the C locale is supported.  */
  return strftime(s, maxsize, format, t);
}
#endif

size_t
strftime(char *restrict s, size_t maxsize, char const *restrict format,
	 struct tm const *restrict t)
{
	char *	p;
	int saved_errno = errno;
	enum warn warn = IN_NONE;

	tzset();
	p = _fmt(format, t, s, s + maxsize, &warn);
	if (DEPRECATE_TWO_DIGIT_YEARS
	    && warn != IN_NONE && getenv(YEAR_2000_NAME)) {
		fprintf(stderr, "\n");
		fprintf(stderr, "strftime format \"%s\" ", format);
		fprintf(stderr, "yields only two digits of years in ");
		if (warn == IN_SOME)
			fprintf(stderr, "some locales");
		else if (warn == IN_THIS)
			fprintf(stderr, "the current locale");
		else	fprintf(stderr, "all locales");
		fprintf(stderr, "\n");
	}
	if (p == s + maxsize) {
		errno = ERANGE;
		return 0;
	}
	*p = '\0';
	errno = saved_errno;
	return p - s;
}

static char *
_fmt(const char *format, const struct tm *t, char *pt,
     const char *ptlim, enum warn *warnp)
{
	struct lc_time_T const *Locale = &C_time_locale;

	for ( ; *format; ++format) {
		if (*format == '%') {
label:
			switch (*++format) {
			default:
				/* Output unknown conversion specifiers as-is,
				   to aid debugging.  This includes '%' at
				   format end.  This conforms to C23 section
				   7.29.3.5 paragraph 6, which says behavior
				   is undefined here.  */
				--format;
				break;
			case 'A':
				pt = _add((t->tm_wday < 0 ||
					t->tm_wday >= DAYSPERWEEK) ?
					"?" : Locale->weekday[t->tm_wday],
					pt, ptlim);
				continue;
			case 'a':
				pt = _add((t->tm_wday < 0 ||
					t->tm_wday >= DAYSPERWEEK) ?
					"?" : Locale->wday[t->tm_wday],
					pt, ptlim);
				continue;
			case 'B':
				pt = _add((t->tm_mon < 0 ||
					t->tm_mon >= MONSPERYEAR) ?
					"?" : Locale->month[t->tm_mon],
					pt, ptlim);
				continue;
			case 'b':
			case 'h':
				pt = _add((t->tm_mon < 0 ||
					t->tm_mon >= MONSPERYEAR) ?
					"?" : Locale->mon[t->tm_mon],
					pt, ptlim);
				continue;
			case 'C':
				/*
				** %C used to do a...
				**	_fmt("%a %b %e %X %Y", t);
				** ...whereas now POSIX 1003.2 calls for
				** something completely different.
				** (ado, 1993-05-24)
				*/
				pt = _yconv(t->tm_year, TM_YEAR_BASE,
					    true, false, pt, ptlim);
				continue;
			case 'c':
				{
				enum warn warn2 = IN_SOME;

				pt = _fmt(Locale->c_fmt, t, pt, ptlim, &warn2);
				if (warn2 == IN_ALL)
					warn2 = IN_THIS;
				if (warn2 > *warnp)
					*warnp = warn2;
				}
				continue;
			case 'D':
				pt = _fmt("%m/%d/%y", t, pt, ptlim, warnp);
				continue;
			case 'd':
				pt = _conv(t->tm_mday, "%02d", pt, ptlim);
				continue;
			case 'E':
			case 'O':
				/*
				** Locale modifiers of C99 and later.
				** The sequences
				**	%Ec %EC %Ex %EX %Ey %EY
				**	%Od %oe %OH %OI %Om %OM
				**	%OS %Ou %OU %OV %Ow %OW %Oy
				** are supposed to provide alternative
				** representations.
				*/
				goto label;
			case 'e':
				pt = _conv(t->tm_mday, "%2d", pt, ptlim);
				continue;
			case 'F':
				pt = _fmt("%Y-%m-%d", t, pt, ptlim, warnp);
				continue;
			case 'H':
				pt = _conv(t->tm_hour, "%02d", pt, ptlim);
				continue;
			case 'I':
				pt = _conv((t->tm_hour % 12) ?
					(t->tm_hour % 12) : 12,
					"%02d", pt, ptlim);
				continue;
			case 'j':
				pt = _conv(t->tm_yday + 1, "%03d", pt, ptlim);
				continue;
			case 'k':
				/*
				** This used to be...
				**	_conv(t->tm_hour % 12 ?
				**		t->tm_hour % 12 : 12, 2, ' ');
				** ...and has been changed to the below to
				** match SunOS 4.1.1 and Arnold Robbins'
				** strftime version 3.0. That is, "%k" and
				** "%l" have been swapped.
				** (ado, 1993-05-24)
				*/
				pt = _conv(t->tm_hour, "%2d", pt, ptlim);
				continue;
#ifdef KITCHEN_SINK
			case 'K':
				/*
				** After all this time, still unclaimed!
				*/
				pt = _add("kitchen sink", pt, ptlim);
				continue;
#endif /* defined KITCHEN_SINK */
			case 'l':
				/*
				** This used to be...
				**	_conv(t->tm_hour, 2, ' ');
				** ...and has been changed to the below to
				** match SunOS 4.1.1 and Arnold Robbin's
				** strftime version 3.0. That is, "%k" and
				** "%l" have been swapped.
				** (ado, 1993-05-24)
				*/
				pt = _conv((t->tm_hour % 12) ?
					(t->tm_hour % 12) : 12,
					"%2d", pt, ptlim);
				continue;
			case 'M':
				pt = _conv(t->tm_min, "%02d", pt, ptlim);
				continue;
			case 'm':
				pt = _conv(t->tm_mon + 1, "%02d", pt, ptlim);
				continue;
			case 'n':
				pt = _add("\n", pt, ptlim);
				continue;
			case 'p':
				pt = _add((t->tm_hour >= (HOURSPERDAY / 2)) ?
					Locale->pm :
					Locale->am,
					pt, ptlim);
				continue;
			case 'R':
				pt = _fmt("%H:%M", t, pt, ptlim, warnp);
				continue;
			case 'r':
				pt = _fmt("%I:%M:%S %p", t, pt, ptlim, warnp);
				continue;
			case 'S':
				pt = _conv(t->tm_sec, "%02d", pt, ptlim);
				continue;
			case 's':
				{
					struct tm	tm;
					char		buf[INT_STRLEN_MAXIMUM(
								time_t) + 1];
					time_t		mkt;

					tm.tm_sec = t->tm_sec;
					tm.tm_min = t->tm_min;
					tm.tm_hour = t->tm_hour;
					tm.tm_mday = t->tm_mday;
					tm.tm_mon = t->tm_mon;
					tm.tm_year = t->tm_year;

					/* Get the time_t value for TM.
					   Native time_t, or its redefinition
					   by localtime.c above, is wide enough
					   so that this cannot overflow.  */
#ifdef TM_GMTOFF
					mkt = timeoff(&tm, t->TM_GMTOFF);
#else
					tm.tm_isdst = t->tm_isdst;
					mkt = mktime(&tm);
#endif
					if (TYPE_SIGNED(time_t)) {
					  intmax_t n = mkt;
					  sprintf(buf, "%"PRIdMAX, n);
					} else {
					  uintmax_t n = mkt;
					  sprintf(buf, "%"PRIuMAX, n);
					}
					pt = _add(buf, pt, ptlim);
				}
				continue;
			case 'T':
				pt = _fmt("%H:%M:%S", t, pt, ptlim, warnp);
				continue;
			case 't':
				pt = _add("\t", pt, ptlim);
				continue;
			case 'U':
				pt = _conv((t->tm_yday + DAYSPERWEEK -
					t->tm_wday) / DAYSPERWEEK,
					"%02d", pt, ptlim);
				continue;
			case 'u':
				/*
				** From Arnold Robbins' strftime version 3.0:
				** "ISO 8601: Weekday as a decimal number
				** [1 (Monday) - 7]"
				** (ado, 1993-05-24)
				*/
				pt = _conv((t->tm_wday == 0) ?
					DAYSPERWEEK : t->tm_wday,
					"%d", pt, ptlim);
				continue;
			case 'V':	/* ISO 8601 week number */
			case 'G':	/* ISO 8601 year (four digits) */
			case 'g':	/* ISO 8601 year (two digits) */
/*
** From Arnold Robbins' strftime version 3.0: "the week number of the
** year (the first Monday as the first day of week 1) as a decimal number
** (01-53)."
** (ado, 1993-05-24)
**
** From  by Markus Kuhn:
** "Week 01 of a year is per definition the first week which has the
** Thursday in this year, which is equivalent to the week which contains
** the fourth day of January. In other words, the first week of a new year
** is the week which has the majority of its days in the new year. Week 01
** might also contain days from the previous year and the week before week
** 01 of a year is the last week (52 or 53) of the previous year even if
** it contains days from the new year. A week starts with Monday (day 1)
** and ends with Sunday (day 7). For example, the first week of the year
** 1997 lasts from 1996-12-30 to 1997-01-05..."
** (ado, 1996-01-02)
*/
				{
					int	year;
					int	base;
					int	yday;
					int	wday;
					int	w;

					year = t->tm_year;
					base = TM_YEAR_BASE;
					yday = t->tm_yday;
					wday = t->tm_wday;
					for ( ; ; ) {
						int	len;
						int	bot;
						int	top;

						len = year_sum_days(year, base);
						/*
						** What yday (-3 ... 3) does
						** the ISO year begin on?
						*/
						bot = ((yday + 11 - wday) %
							DAYSPERWEEK) - 3;
						/*
						** What yday does the NEXT
						** ISO year begin on?
						*/
						top = bot -
							(len % DAYSPERWEEK);
						if (top < -3)
							top += DAYSPERWEEK;
						top += len;
						if (yday >= top) {
							++base;
							w = 1;
							break;
						}
						if (yday >= bot) {
							w = 1 + ((yday - bot) /
								DAYSPERWEEK);
							break;
						}
						--base;
						yday += year_sum_days(year,
								      base);
					}
#ifdef XPG4_1994_04_09
					if ((w == 52 &&
						t->tm_mon == TM_JANUARY) ||
						(w == 1 &&
						t->tm_mon == TM_DECEMBER))
							w = 53;
#endif /* defined XPG4_1994_04_09 */
					if (*format == 'V')
						pt = _conv(w, "%02d",
							pt, ptlim);
					else if (*format == 'g') {
						*warnp = IN_ALL;
						pt = _yconv(year, base,
							false, true,
							pt, ptlim);
					} else	pt = _yconv(year, base,
							true, true,
							pt, ptlim);
				}
				continue;
			case 'v':
				/*
				** From Arnold Robbins' strftime version 3.0:
				** "date as dd-bbb-YYYY"
				** (ado, 1993-05-24)
				*/
				pt = _fmt("%e-%b-%Y", t, pt, ptlim, warnp);
				continue;
			case 'W':
				pt = _conv((t->tm_yday + DAYSPERWEEK -
					(t->tm_wday ?
					(t->tm_wday - 1) :
					(DAYSPERWEEK - 1))) / DAYSPERWEEK,
					"%02d", pt, ptlim);
				continue;
			case 'w':
				pt = _conv(t->tm_wday, "%d", pt, ptlim);
				continue;
			case 'X':
				pt = _fmt(Locale->X_fmt, t, pt, ptlim, warnp);
				continue;
			case 'x':
				{
				enum warn warn2 = IN_SOME;

				pt = _fmt(Locale->x_fmt, t, pt, ptlim, &warn2);
				if (warn2 == IN_ALL)
					warn2 = IN_THIS;
				if (warn2 > *warnp)
					*warnp = warn2;
				}
				continue;
			case 'y':
				*warnp = IN_ALL;
				pt = _yconv(t->tm_year, TM_YEAR_BASE,
					false, true,
					pt, ptlim);
				continue;
			case 'Y':
				pt = _yconv(t->tm_year, TM_YEAR_BASE,
					true, true,
					pt, ptlim);
				continue;
			case 'Z':
#ifdef TM_ZONE
				pt = _add(t->TM_ZONE, pt, ptlim);
#elif HAVE_TZNAME
				if (t->tm_isdst >= 0)
					pt = _add(tzname[t->tm_isdst != 0],
						pt, ptlim);
#endif
				/*
				** C99 and later say that %Z must be
				** replaced by the empty string if the
				** time zone abbreviation is not
				** determinable.
				*/
				continue;
			case 'z':
#if defined TM_GMTOFF || USG_COMPAT || ALTZONE
				{
				long		diff;
				char const *	sign;
				bool negative;

# ifdef TM_GMTOFF
				diff = t->TM_GMTOFF;
# else
				/*
				** C99 and later say that the UT offset must
				** be computed by looking only at
				** tm_isdst. This requirement is
				** incorrect, since it means the code
				** must rely on magic (in this case
				** altzone and timezone), and the
				** magic might not have the correct
				** offset. Doing things correctly is
				** tricky and requires disobeying the standard;
				** see GNU C strftime for details.
				** For now, punt and conform to the
				** standard, even though it's incorrect.
				**
				** C99 and later say that %z must be replaced by
				** the empty string if the time zone is not
				** determinable, so output nothing if the
				** appropriate variables are not available.
				*/
				if (t->tm_isdst < 0)
					continue;
				if (t->tm_isdst == 0)
#  if USG_COMPAT
					diff = -timezone;
#  else
					continue;
#  endif
				else
#  if ALTZONE
					diff = -altzone;
#  else
					continue;
#  endif
# endif
				negative = diff < 0;
				if (diff == 0) {
# ifdef TM_ZONE
				  negative = t->TM_ZONE[0] == '-';
# else
				  negative = t->tm_isdst < 0;
#  if HAVE_TZNAME
				  if (tzname[t->tm_isdst != 0][0] == '-')
				    negative = true;
#  endif
# endif
				}
				if (negative) {
					sign = "-";
					diff = -diff;
				} else	sign = "+";
				pt = _add(sign, pt, ptlim);
				diff /= SECSPERMIN;
				diff = (diff / MINSPERHOUR) * 100 +
					(diff % MINSPERHOUR);
				pt = _conv(diff, "%04d", pt, ptlim);
				}
#endif
				continue;
			case '+':
				pt = _fmt(Locale->date_fmt, t, pt, ptlim,
					warnp);
				continue;
			case '%':
				break;
			}
		}
		if (pt == ptlim)
			break;
		*pt++ = *format;
	}
	return pt;
}

static char *
_conv(int n, const char *format, char *pt, const char *ptlim)
{
	char	buf[INT_STRLEN_MAXIMUM(int) + 1];

	sprintf(buf, format, n);
	return _add(buf, pt, ptlim);
}

static char *
_add(const char *str, char *pt, const char *ptlim)
{
	while (pt < ptlim && (*pt = *str++) != '\0')
		++pt;
	return pt;
}

/*
** POSIX and the C Standard are unclear or inconsistent about
** what %C and %y do if the year is negative or exceeds 9999.
** Use the convention that %C concatenated with %y yields the
** same output as %Y, and that %Y contains at least 4 bytes,
** with more only if necessary.
*/

static char *
_yconv(int a, int b, bool convert_top, bool convert_yy,
       char *pt, const char *ptlim)
{
	register int	lead;
	register int	trail;

	int DIVISOR = 100;
	trail = a % DIVISOR + b % DIVISOR;
	lead = a / DIVISOR + b / DIVISOR + trail / DIVISOR;
	trail %= DIVISOR;
	if (trail < 0 && lead > 0) {
		trail += DIVISOR;
		--lead;
	} else if (lead < 0 && trail > 0) {
		trail -= DIVISOR;
		++lead;
	}
	if (convert_top) {
		if (lead == 0 && trail < 0)
			pt = _add("-0", pt, ptlim);
		else	pt = _conv(lead, "%02d", pt, ptlim);
	}
	if (convert_yy)
		pt = _conv(((trail < 0) ? -trail : trail), "%02d", pt, ptlim);
	return pt;
}
zic.c0000644000000000000000000034211715246336075006760 0ustar00/* Compile .zi time zone data into TZif binary files.  */

/*
** This file is in the public domain, so clarified as of
** 2006-07-17 by Arthur David Olson.
*/

/* Use the system 'time' function, instead of any private replacement.
   This avoids creating an unnecessary dependency on localtime.c.  */
#undef EPOCH_LOCAL
#undef EPOCH_OFFSET
#undef RESERVE_STD_EXT_IDS
#undef time_tz

#include "version.h"
#include "private.h"
#include "tzdir.h"
#include "tzfile.h"

#include 
#ifndef O_BINARY
# define O_BINARY 0 /* MS-Windows */
#endif

#include 
#include 
#include 
#include 

typedef int_fast64_t	zic_t;
static zic_t const
  ZIC_MIN = INT_FAST64_MIN,
  ZIC_MAX = INT_FAST64_MAX,
  ZIC32_MIN = -1 - (zic_t) TWO_31_MINUS_1,
  ZIC32_MAX = TWO_31_MINUS_1;
#define SCNdZIC SCNdFAST64

#ifndef ZIC_MAX_ABBR_LEN_WO_WARN
# define ZIC_MAX_ABBR_LEN_WO_WARN 6
#endif /* !defined ZIC_MAX_ABBR_LEN_WO_WARN */

/* Minimum and maximum years, assuming signed 32-bit time_t.  */
enum { YEAR_32BIT_MIN = 1901, YEAR_32BIT_MAX = 2038 };

/* An upper bound on how much a format might grow due to concatenation.  */
enum { FORMAT_LEN_GROWTH_BOUND = 5 };

#ifdef HAVE_DIRECT_H
# include 
# include 
# define mkdir(name, mode) _mkdir(name)
typedef unsigned short gid_t, mode_t, uid_t;
#endif

#ifndef HAVE_GETRANDOM
# ifdef __has_include
#  if __has_include()
#   include 
#  endif
# elif 2 < __GLIBC__ + (25 <= __GLIBC_MINOR__)
#  include 
# endif
# define HAVE_GETRANDOM GRND_RANDOM
#elif HAVE_GETRANDOM
# include 
#endif


#if HAVE_SYS_STAT_H
# include 
#endif

#ifndef S_IRWXU
# define S_IRUSR 0400
# define S_IWUSR 0200
# define S_IXUSR 0100
# define S_IRGRP 0040
# define S_IWGRP 0020
# define S_IXGRP 0010
# define S_IROTH 0004
# define S_IWOTH 0002
# define S_IXOTH 0001
# define S_IRWXU (S_IRUSR | S_IWUSR | S_IXUSR)
# define S_IRWXG (S_IRGRP | S_IWGRP | S_IXGRP)
# define S_IRWXO (S_IROTH | S_IWOTH | S_IXOTH)
#endif

/* All file permission bits.  */
#define ALL_PERMS (S_IRWXU | S_IRWXG | S_IRWXO)

/* Troublesome file permission bits.  */
#define TROUBLE_PERMS (S_IWGRP | S_IWOTH)

/* File permission bits for making directories.
   The umask modifies these bits.  */
#define MKDIR_PERMS (ALL_PERMS & ~TROUBLE_PERMS)

/* File permission bits for making regular files.
   The umask modifies these bits.  */
#define CREAT_PERMS (MKDIR_PERMS & ~(S_IXUSR | S_IXGRP | S_IXOTH))
static mode_t creat_perms = CREAT_PERMS;

#ifndef HAVE_PWD_H
# ifdef __has_include
#  if __has_include() && __has_include()
#   define HAVE_PWD_H 1
#  else
#   define HAVE_PWD_H 0
#  endif
# endif
#endif
#ifndef HAVE_PWD_H
# define HAVE_PWD_H 1
#endif
#if HAVE_PWD_H
# include 
# include 
#else
struct group { gid_t gr_gid; };
struct passwd { uid_t pw_uid; };
# define getgrnam(arg) NULL
# define getpwnam(arg) NULL
# define fchown(fd, owner, group) ((fd) < 0 ? -1 : 0)
#endif
static gid_t const no_gid = -1;
static uid_t const no_uid = -1;
static gid_t output_group = -1;
static uid_t output_owner = -1;
#ifndef GID_T_MAX
# define GID_T_MAX_NO_PADDING MAXVAL(gid_t, TYPE_BIT(gid_t))
# if HAVE__GENERIC
#  define GID_T_MAX \
    (TYPE_SIGNED(gid_t) \
     ? _Generic((gid_t) 0, \
		signed char: SCHAR_MAX, short: SHRT_MAX, \
		int: INT_MAX, long: LONG_MAX, long long: LLONG_MAX, \
		default: GID_T_MAX_NO_PADDING) \
     : (gid_t) -1)
# else
#  define GID_T_MAX GID_T_MAX_NO_PADDING
# endif
#endif
#ifndef UID_T_MAX
# define UID_T_MAX_NO_PADDING MAXVAL(uid_t, TYPE_BIT(uid_t))
# if HAVE__GENERIC
#  define UID_T_MAX \
    (TYPE_SIGNED(uid_t) \
     ? _Generic((uid_t) 0, \
		signed char: SCHAR_MAX, short: SHRT_MAX, \
		int: INT_MAX, long: LONG_MAX, long long: LLONG_MAX, \
		default: UID_T_MAX_NO_PADDING) \
     : (uid_t) -1)
# else
#  define UID_T_MAX UID_T_MAX_NO_PADDING
# endif
#endif

/* The minimum alignment of a type, for pre-C23 platforms.
   The __SUNPRO_C test is because Oracle Developer Studio 12.6 lacks
    even though __STDC_VERSION__ == 201112.  */
#if __STDC_VERSION__ < 201112 || defined __SUNPRO_C
# define alignof(type) offsetof(struct { char a; type b; }, b)
#elif __STDC_VERSION__ < 202311
# include 
#endif

/* The name used for the file implementing the obsolete -p option.  */
#ifndef TZDEFRULES
# define TZDEFRULES "posixrules"
#endif

/* The maximum length of a text line, including the trailing newline.  */
#ifndef _POSIX2_LINE_MAX
# define _POSIX2_LINE_MAX 2048
#endif

/* The type for line numbers.  Use PRIdMAX to format them; formerly
   there was also "#define PRIdLINENO PRIdMAX" and formats used
   PRIdLINENO, but xgettext cannot grok that.  */
typedef intmax_t lineno;

struct rule {
	int		r_filenum;
	lineno		r_linenum;
	const char *	r_name;

	zic_t		r_loyear;	/* for example, 1986 */
	zic_t		r_hiyear;	/* for example, 1986 */
	bool		r_hiwasnum;

	int		r_month;	/* 0..11 */

	int		r_dycode;	/* see below */
	int		r_dayofmonth;
	int		r_wday;

	zic_t		r_tod;		/* time from midnight */
	bool		r_todisstd;	/* is r_tod standard time? */
	bool		r_todisut;	/* is r_tod UT? */
	bool		r_isdst;	/* is this daylight saving time? */
	zic_t		r_save;		/* offset from standard time */
	const char *	r_abbrvar;	/* variable part of abbreviation */

	bool		r_todo;		/* a rule to do (used in outzone) */
	zic_t		r_temp;		/* used in outzone */
};

/*
** r_dycode	r_dayofmonth	r_wday
*/
enum {
  DC_DOM,	/* 1..31 */	/* unused */
  DC_DOWGEQ,	/* 1..31 */	/* 0..6 (Sun..Sat) */
  DC_DOWLEQ	/* 1..31 */	/* 0..6 (Sun..Sat) */
};

struct zone {
	int		z_filenum;
	lineno		z_linenum;

	const char *	z_name;
	zic_t		z_stdoff;
	char *		z_rule;
	const char *	z_format;
	char		z_format_specifier;

	bool		z_isdst;
	zic_t		z_save;

	struct rule *	z_rules;
	ptrdiff_t	z_nrules;

	struct rule	z_untilrule;
	zic_t		z_untiltime;
};

#if ! HAVE_SYMLINK
static ssize_t
readlink(char const *restrict file, char *restrict buf, size_t size)
{
  errno = ENOTSUP;
  return -1;
}
static int
symlink(char const *target, char const *linkname)
{
  errno = ENOTSUP;
  return -1;
}
#endif
#ifndef AT_SYMLINK_FOLLOW
# define linkat(targetdir, target, linknamedir, linkname, flag) \
   (errno = ENOTSUP, -1)
#endif

static int	addabbr(char[TZ_MAX_CHARS], int *, char const *);
static void	addtt(zic_t starttime, int type);
static int	addtype(zic_t, char const *, bool, bool, bool);
static void	adjleap(void);
static void	associate(void);
static void	checkabbr(char const *);
static void	check_for_signal(void);
static void	dolink(const char *, const char *, bool);
static int	getfields(char *, char **, int);
static zic_t	getsave(char *, bool *);
static void	inexpires(char **, int);
static void	infile(int, char const *);
static void	inleap(char ** fields, int nfields);
static void	inlink(char ** fields, int nfields);
static void	inrule(char ** fields, int nfields);
static bool	inzcont(char ** fields, int nfields);
static bool	inzone(char ** fields, int nfields);
static bool	inzsub(char **, int, bool);
static bool	is_alpha(char a);
static int	itssymlink(char const *, int *);
static void	leapadd(zic_t, int, int);
static char	lowerit(char);
static void	mkdirs(char const *, bool);
static zic_t	oadd(zic_t t1, zic_t t2);
static zic_t	omul(zic_t, zic_t);
static void	outzone(const struct zone * zp, ptrdiff_t ntzones);
static void	remove_temp(char const *);
static zic_t	rpytime(const struct rule * rp, zic_t wantedy);
static bool	rulesub(struct rule * rp,
			const char * loyearp, const char * hiyearp,
			const char * typep, const char * monthp,
			const char * dayp, const char * timep);
static zic_t	tadd(zic_t t1, zic_t t2);

/* Is C an ASCII digit?  */
static bool
is_digit(char c)
{
  return '0' <= c && c <= '9';
}

/* Bound on length of what %z can expand to.  */
enum { PERCENT_Z_LEN_BOUND = sizeof "+995959" - 1 };

static int		charcnt;
static bool		errors;
static bool		warnings;
static int		filenum;
static ptrdiff_t	leapcnt;
static ptrdiff_t	leap_alloc;
static bool		leapseen;
static zic_t		leapminyear;
static zic_t		leapmaxyear;
static lineno		linenum;
static int		max_abbrvar_len = PERCENT_Z_LEN_BOUND;
static int		max_format_len;
static zic_t		max_year;
static zic_t		min_year;
static bool		noise;
static bool		skip_mkdir;
static int		rfilenum;
static lineno		rlinenum;
static const char *	progname;
static char const *	leapsec;
static char *const *	main_argv;
static ptrdiff_t	timecnt;
static ptrdiff_t	timecnt_alloc;
static int		typecnt;
static int		unspecifiedtype;

/*
** Line codes.
*/

enum {
  LC_RULE,
  LC_ZONE,
  LC_LINK,
  LC_LEAP,
  LC_EXPIRES
};

/*
** Which fields are which on a Zone line.
*/

enum {
  ZF_NAME = 1,
  ZF_STDOFF,
  ZF_RULE,
  ZF_FORMAT,
  ZF_TILYEAR,
  ZF_TILMONTH,
  ZF_TILDAY,
  ZF_TILTIME,
  ZONE_MAXFIELDS,
  ZONE_MINFIELDS = ZF_TILYEAR
};

/*
** Which fields are which on a Zone continuation line.
*/

enum {
  ZFC_STDOFF,
  ZFC_RULE,
  ZFC_FORMAT,
  ZFC_TILYEAR,
  ZFC_TILMONTH,
  ZFC_TILDAY,
  ZFC_TILTIME,
  ZONEC_MAXFIELDS,
  ZONEC_MINFIELDS = ZFC_TILYEAR
};

/*
** Which files are which on a Rule line.
*/

enum {
  RF_NAME = 1,
  RF_LOYEAR,
  RF_HIYEAR,
  RF_COMMAND,
  RF_MONTH,
  RF_DAY,
  RF_TOD,
  RF_SAVE,
  RF_ABBRVAR,
  RULE_FIELDS
};

/*
** Which fields are which on a Link line.
*/

enum {
  LF_TARGET = 1,
  LF_LINKNAME,
  LINK_FIELDS
};

/*
** Which fields are which on a Leap line.
*/

enum {
  LP_YEAR = 1,
  LP_MONTH,
  LP_DAY,
  LP_TIME,
  LP_CORR,
  LP_ROLL,
  LEAP_FIELDS,

  /* Expires lines are like Leap lines, except without CORR and ROLL fields.  */
  EXPIRES_FIELDS = LP_TIME + 1
};

/* The maximum number of fields on any of the above lines.
   (This uses INT_PROMOTE to pacify gcc -Wenum-compare.)  */
enum {
  MAX_FIELDS = max(max(INT_PROMOTE(RULE_FIELDS), INT_PROMOTE(LINK_FIELDS)),
		   max(INT_PROMOTE(LEAP_FIELDS), INT_PROMOTE(EXPIRES_FIELDS)))
};

/*
** Year synonyms.
*/

enum {
  YR_MINIMUM, /* "minimum" is for backward compatibility only */
  YR_MAXIMUM,
  YR_ONLY
};

static struct rule *	rules;
static ptrdiff_t	nrules;	/* number of rules */
static ptrdiff_t	nrules_alloc;

static struct zone *	zones;
static ptrdiff_t	nzones;	/* number of zones */
static ptrdiff_t	nzones_alloc;

struct link {
	int		l_filenum;
	lineno		l_linenum;
	const char *	l_target;
	const char *	l_linkname;
};

static struct link *	links;
static ptrdiff_t	nlinks;
static ptrdiff_t	nlinks_alloc;

struct lookup {
	const char *	l_word;
	const int	l_value;
};

static struct lookup const *	byword(const char * string,
					const struct lookup * lp);

static struct lookup const zi_line_codes[] = {
	{ "Rule",	LC_RULE },
	{ "Zone",	LC_ZONE },
	{ "Link",	LC_LINK },
	{ NULL,		0 }
};
static struct lookup const leap_line_codes[] = {
	{ "Leap",	LC_LEAP },
	{ "Expires",	LC_EXPIRES },
	{ NULL,		0}
};

static struct lookup const	mon_names[] = {
	{ "January",	TM_JANUARY },
	{ "February",	TM_FEBRUARY },
	{ "March",	TM_MARCH },
	{ "April",	TM_APRIL },
	{ "May",	TM_MAY },
	{ "June",	TM_JUNE },
	{ "July",	TM_JULY },
	{ "August",	TM_AUGUST },
	{ "September",	TM_SEPTEMBER },
	{ "October",	TM_OCTOBER },
	{ "November",	TM_NOVEMBER },
	{ "December",	TM_DECEMBER },
	{ NULL,		0 }
};

static struct lookup const	wday_names[] = {
	{ "Sunday",	TM_SUNDAY },
	{ "Monday",	TM_MONDAY },
	{ "Tuesday",	TM_TUESDAY },
	{ "Wednesday",	TM_WEDNESDAY },
	{ "Thursday",	TM_THURSDAY },
	{ "Friday",	TM_FRIDAY },
	{ "Saturday",	TM_SATURDAY },
	{ NULL,		0 }
};

static struct lookup const	lasts[] = {
	{ "last-Sunday",	TM_SUNDAY },
	{ "last-Monday",	TM_MONDAY },
	{ "last-Tuesday",	TM_TUESDAY },
	{ "last-Wednesday",	TM_WEDNESDAY },
	{ "last-Thursday",	TM_THURSDAY },
	{ "last-Friday",	TM_FRIDAY },
	{ "last-Saturday",	TM_SATURDAY },
	{ NULL,			0 }
};

static struct lookup const	begin_years[] = {
	{ "minimum",	YR_MINIMUM },
	{ NULL,		0 }
};

static struct lookup const	end_years[] = {
	{ "maximum",	YR_MAXIMUM },
	{ "only",	YR_ONLY },
	{ NULL,		0 }
};

static struct lookup const	leap_types[] = {
	{ "Rolling",	true },
	{ "Stationary",	false },
	{ NULL,		0 }
};

static const int	len_months[2][MONSPERYEAR] = {
	{ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 },
	{ 31, 29, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 }
};

static struct attype {
	zic_t		at;
	bool		dontmerge;
	unsigned char	type;
} *			attypes;
static zic_t		utoffs[TZ_MAX_TYPES];
static char		isdsts[TZ_MAX_TYPES];
static unsigned char	desigidx[TZ_MAX_TYPES];
static bool		ttisstds[TZ_MAX_TYPES];
static bool		ttisuts[TZ_MAX_TYPES];
static char		chars[TZ_MAX_CHARS];
static struct {
  zic_t trans;
  zic_t corr;
  char roll;
} *leap;

/*
** Memory allocation.
*/

ATTRIBUTE_NORETURN static void
memory_exhausted(const char *msg)
{
	fprintf(stderr, _("%s: Memory exhausted: %s\n"), progname, msg);
	exit(EXIT_FAILURE);
}

ATTRIBUTE_NORETURN static void
size_overflow(void)
{
  memory_exhausted(_("size overflow"));
}

ATTRIBUTE_PURE_114833_HACK
static ptrdiff_t
size_sum(size_t a, size_t b)
{
#ifdef ckd_add
  ptrdiff_t sum;
  if (!ckd_add(&sum, a, b) && sum <= INDEX_MAX)
    return sum;
#else
  if (a <= INDEX_MAX && b <= INDEX_MAX - a)
    return a + b;
#endif
  size_overflow();
}

ATTRIBUTE_PURE_114833_HACK
static ptrdiff_t
size_product(ptrdiff_t nitems, ptrdiff_t itemsize)
{
#ifdef ckd_mul
  ptrdiff_t product;
  if (!ckd_mul(&product, nitems, itemsize) && product <= INDEX_MAX)
    return product;
#else
  ptrdiff_t nitems_max = INDEX_MAX / itemsize;
  if (nitems <= nitems_max)
    return nitems * itemsize;
#endif
  size_overflow();
}

ATTRIBUTE_PURE_114833_HACK
static ptrdiff_t
align_to(ptrdiff_t size, ptrdiff_t alignment)
{
  ptrdiff_t lo_bits = alignment - 1, sum = size_sum(size, lo_bits);
  return sum & ~lo_bits;
}

#if !HAVE_STRDUP
static char *
strdup(char const *str)
{
  char *result = malloc(strlen(str) + 1);
  return result ? strcpy(result, str) : result;
}
#endif

static void *
memcheck(void *ptr)
{
	if (ptr == NULL)
	  memory_exhausted(strerror(HAVE_MALLOC_ERRNO ? errno : ENOMEM));
	return ptr;
}

static void *
xmalloc(size_t size)
{
  return memcheck(malloc(size));
}

static void *
xrealloc(void *ptr, size_t size)
{
  return memcheck(realloc(ptr, size));
}

static char *
xstrdup(char const *str)
{
  return memcheck(strdup(str));
}

static ptrdiff_t
grow_nitems_alloc(ptrdiff_t *nitems_alloc, ptrdiff_t itemsize)
{
  ptrdiff_t addend = (*nitems_alloc >> 1) + 1;
#if defined ckd_add && defined ckd_mul
  ptrdiff_t product;
  if (!ckd_add(nitems_alloc, *nitems_alloc, addend)
      && !ckd_mul(&product, *nitems_alloc, itemsize) && product <= INDEX_MAX)
    return product;
#else
  if (*nitems_alloc <= ((INDEX_MAX - 1) / 3 * 2) / itemsize) {
    *nitems_alloc += addend;
    return *nitems_alloc * itemsize;
  }
#endif
  memory_exhausted(_("integer overflow"));
}

static void *
growalloc(void *ptr, ptrdiff_t itemsize, ptrdiff_t nitems,
	  ptrdiff_t *nitems_alloc)
{
  return (nitems < *nitems_alloc
	  ? ptr
	  : xrealloc(ptr, grow_nitems_alloc(nitems_alloc, itemsize)));
}

/*
** Error handling.
*/

/* In most of the code, an input file name is represented by its index
   into the main argument vector, except that LEAPSEC_FILENUM stands
   for leapsec and COMMAND_LINE_FILENUM stands for the command line.  */
enum { LEAPSEC_FILENUM = -2, COMMAND_LINE_FILENUM = -1 };

/* Return the name of the Ith input file, for diagnostics.  */
static char const *
filename(int i)
{
  if (i == COMMAND_LINE_FILENUM)
    return _("command line");
  else {
    char const *fname = i == LEAPSEC_FILENUM ? leapsec : main_argv[i];
    return strcmp(fname, "-") == 0 ? _("standard input") : fname;
  }
}

static void
eats(int fnum, lineno num, int rfnum, lineno rnum)
{
	filenum = fnum;
	linenum = num;
	rfilenum = rfnum;
	rlinenum = rnum;
}

static void
eat(int fnum, lineno num)
{
	eats(fnum, num, 0, -1);
}

ATTRIBUTE_FORMAT((printf, 1, 0)) static void
verror(char const *msgid, va_list args)
{
	check_for_signal();
	/*
	** Match the format of "cc" to allow sh users to
	**	zic ... 2>&1 | error -t "*" -v
	** on BSD systems.
	*/
	if (filenum)
	  fprintf(stderr, _("\"%s\", line %"PRIdMAX": "),
		  filename(filenum), linenum);
	vfprintf(stderr, _(msgid), args);
	if (rfilenum)
		fprintf(stderr, _(" (rule from \"%s\", line %"PRIdMAX")"),
			filename(rfilenum), rlinenum);
	fprintf(stderr, "\n");
}

ATTRIBUTE_FORMAT((printf, 1, 2)) static void
error(char const *msgid, ...)
{
	va_list args;
	va_start(args, msgid);
	verror(msgid, args);
	va_end(args);
	errors = true;
}

ATTRIBUTE_FORMAT((printf, 1, 2)) static void
warning(char const *msgid, ...)
{
	va_list args;
	fprintf(stderr, _("warning: "));
	va_start(args, msgid);
	verror(msgid, args);
	va_end(args);
	warnings = true;
}

/* Convert ARG, a string in base BASE, to an unsigned long value no
   greater than MAXVAL.  On failure, diagnose with MSGID and exit.  */
static unsigned long
arg2num(char const *arg, int base, unsigned long maxval, char const *msgid)
{
  unsigned long n;
  char *ep;
  errno = 0;
  n = strtoul(arg, &ep, base);
  if (ep == arg || *ep || maxval < n || errno) {
    fprintf(stderr, _(msgid), progname, arg);
    exit(EXIT_FAILURE);
  }
  return n;
}

#ifndef MODE_T_MAX
# define MODE_T_MAX_NO_PADDING MAXVAL(mode_t, TYPE_BIT(mode_t))
# if HAVE__GENERIC
#  define MODE_T_MAX \
    (TYPE_SIGNED(mode_t) \
     ? _Generic((mode_t) 0, \
		signed char: SCHAR_MAX, short: SHRT_MAX, \
		int: INT_MAX, long: LONG_MAX, long long: LLONG_MAX, \
		default: MODE_T_MAX_NO_PADDING) \
     : (mode_t) -1)
# else
#  define MODE_T_MAX MODE_T_MAX_NO_PADDING
# endif
#endif

#ifndef HAVE_FCHMOD
# define HAVE_FCHMOD 1
#endif
#if !HAVE_FCHMOD
# define fchmod(fd, mode) 0
#endif

#ifndef HAVE_SETMODE
# if (defined __FreeBSD__ || defined __NetBSD__ || defined __OpenBSD__ \
      || defined __DragonFly__ \
      || (defined __APPLE__ && defined __MACH__))
#  define HAVE_SETMODE 1
# else
#  define HAVE_SETMODE 0
# endif
#endif

static mode_t const no_mode = -1;
static mode_t output_mode = -1;

static mode_t
mode_option(char const *arg)
{
#if HAVE_SETMODE
  void *set = setmode(arg);
  if (set) {
    mode_t mode = getmode(set, CREAT_PERMS);
    free(set);
    return mode;
  }
#endif
  return arg2num(arg, 8, min(MODE_T_MAX, ULONG_MAX),
		 N_("%s: -m '%s': invalid mode\n"));
}

static int
chmetadata(FILE *stream)
{
  if (output_owner != no_uid || output_group != no_gid) {
    int r = fchown(fileno(stream), output_owner, output_group);
    if (r < 0)
      return r;
  }
  return output_mode == no_mode ? 0 : fchmod(fileno(stream), output_mode);
}

/* Close STREAM.
   If it had an I/O error, report it against DIR/NAME,
   remove TEMPNAME if nonnull, and then exit.
   If TEMPNAME is nonnull, and if requested,
   change the stream's metadata before closing.  */
static void
close_file(FILE *stream, char const *dir, char const *name,
	   char const *tempname)
{
  char const *e = (ferror(stream) ? _("I/O error")
		   : ((tempname
		       && (fflush(stream) < 0 || chmetadata(stream) < 0))
		      || fclose(stream) < 0)
		   ? strerror(errno) : NULL);
  if (e) {
    if (name && *name == '/')
      dir = NULL;
    fprintf(stderr, "%s: %s%s%s%s%s\n", progname,
	    dir ? dir : "", dir ? "/" : "",
	    name ? name : "", name ? ": " : "",
	    e);
    if (tempname)
      remove_temp(tempname);
    exit(EXIT_FAILURE);
  }
}

ATTRIBUTE_NORETURN static void
duplicate_options(char const *opt)
{
  fprintf(stderr, _("%s: More than one %s option specified\n"), progname, opt);
  exit(EXIT_FAILURE);
}

ATTRIBUTE_NORETURN static void
usage(FILE *stream, int status)
{
  fprintf(stream,
	  _("%s: usage is %s [ --version ] [ --help ] [ -v ] \\\n"
	    "\t[ -b {slim|fat} ] [ -d directory ] [ -D ] \\\n"
	    "\t[ -l localtime ] [ -L leapseconds ] [ -m mode ] \\\n"
	    "\t[ -p posixrules ] [ -r '[@lo][/@hi]' ] [ -R @hi ] \\\n"
	    "\t[ -t localtime-link ] [ -u 'owner[:group]' ] \\\n"
	    "\t[ filename ... ]\n\n"
	    "Report bugs to %s.\n"),
	  progname, progname, REPORT_BUGS_TO);
  if (status == EXIT_SUCCESS)
    close_file(stream, NULL, NULL, NULL);
  exit(status);
}

static void
group_option(char const *arg)
{
  if (*arg) {
    if (output_group != no_gid) {
      fprintf(stderr, _("multiple groups specified"));
      exit(EXIT_FAILURE);
    } else {
      struct group *gr = getgrnam(arg);
      output_group = (gr ? gr->gr_gid
		      : arg2num(arg, 10, min(GID_T_MAX, ULONG_MAX),
				N_("%s: invalid group: %s\n")));
    }
  }
}

static void
owner_option(char const *arg)
{
  if (*arg) {
    if (output_owner != no_uid) {
      fprintf(stderr, _("multiple owners specified"));
      exit(EXIT_FAILURE);
    } else {
      struct passwd *pw = getpwnam(arg);
      output_owner = (pw ? pw->pw_uid
		      : arg2num(arg, 10, min(UID_T_MAX, ULONG_MAX),
				N_("%s: invalid owner: %s\n")));
    }
  }
}

/* If setting owner or group, use temp file permissions that avoid
   security races before the fchmod at the end.  */
static void
use_safe_temp_permissions(void)
{
  if (output_owner != no_uid || output_group != no_gid) {

    /* The mode when done with the file.  */
    mode_t omode;
    if (output_mode == no_mode) {
      mode_t cmask = umask(0);
      umask(cmask);
      omode = CREAT_PERMS & ~cmask;
    } else
      omode = output_mode;

    /* The mode passed to open+O_CREAT.  Do not bother with executable
       permissions, as they should not be used and this mode is merely
       a nicety (even a mode of 0 still work).  */
    creat_perms = ((((omode & (S_IRUSR | S_IRGRP | S_IROTH))
		     == (S_IRUSR | S_IRGRP | S_IROTH))
		    ? S_IRUSR | S_IRGRP | S_IROTH : 0)
		   | (((omode & (S_IWUSR | S_IWGRP | S_IWOTH))
		       == (S_IWUSR | S_IWGRP | S_IWOTH))
		      ? S_IWUSR | S_IWGRP | S_IWOTH : 0));

    /* If creat_perms is not the final mode, arrange to run
       fchmod later, even if -m was not used.  */
    if (creat_perms != omode)
      output_mode = omode;
  }
}

/* Change the working directory to DIR, possibly creating DIR and its
   ancestors.  After this is done, all files are accessed with names
   relative to DIR.  */
static void
change_directory(char const *dir)
{
  if (chdir(dir) != 0) {
    int chdir_errno = errno;
    if (chdir_errno == ENOENT) {
      mkdirs(dir, false);
      chdir_errno = chdir(dir) == 0 ? 0 : errno;
    }
    if (chdir_errno != 0) {
      fprintf(stderr, _("%s: Can't chdir to %s: %s\n"),
	      progname, dir, strerror(chdir_errno));
      exit(EXIT_FAILURE);
    }
  }
}

/* Compare the two links A and B, for a stable sort by link name.  */
static int
qsort_linkcmp(void const *a, void const *b)
{
  struct link const *l = a;
  struct link const *m = b;
  int cmp = strcmp(l->l_linkname, m->l_linkname);
  if (cmp)
    return cmp;

  /* The link names are the same.  Make the sort stable by comparing
     file numbers (where subtraction cannot overflow) and possibly
     line numbers (where it can).  */
  cmp = l->l_filenum - m->l_filenum;
  if (cmp)
    return cmp;
  return (l->l_linenum > m->l_linenum) - (l->l_linenum < m->l_linenum);
}

/* Compare the string KEY to the link B, for bsearch.  */
static int
bsearch_linkcmp(void const *key, void const *b)
{
  struct link const *m = b;
  return strcmp(key, m->l_linkname);
}

/* Make the links specified by the Link lines.  */
static void
make_links(void)
{
  ptrdiff_t i, j, nalinks, pass_size;
  if (1 < nlinks)
    qsort(links, nlinks, sizeof *links, qsort_linkcmp);

  /* Ignore each link superseded by a later link with the same name.  */
  j = 0;
  for (i = 0; i < nlinks; i++) {
    while (i + 1 < nlinks
	   && strcmp(links[i].l_linkname, links[i + 1].l_linkname) == 0)
      i++;
    links[j++] = links[i];
  }
  nlinks = pass_size = j;

  /* Walk through the link array making links.  However,
     if a link's target has not been made yet, append a copy to the
     end of the array.  The end of the array will gradually fill
     up with a small sorted subsequence of not-yet-made links.
     nalinks counts all the links in the array, including copies.
     When we reach the copied subsequence, it may still contain
     a link to a not-yet-made link, so the process repeats.
     At any given point in time, the link array consists of the
     following subregions, where 0 <= i <= j <= nalinks and
     0 <= nlinks <= nalinks:

       0 .. (i - 1):
	 links that either have been made, or have been copied to a
	 later point point in the array (this later point can be in
	 any of the three subregions)
       i .. (j - 1):
	 not-yet-made links for this pass
       j .. (nalinks - 1):
	 not-yet-made links that this pass has skipped because
	 they were links to not-yet-made links

     The first subregion might not be sorted if nlinks < i;
     the other two subregions are sorted.  This algorithm does
     not alter entries 0 .. (nlinks - 1), which remain sorted.

     If there are L links, this algorithm is O(C*L*log(L)) where
     C is the length of the longest link chain.  Usually C is
     short (e.g., 3) though its worst-case value is L.  */

  j = nalinks = nlinks;

  for (i = 0; i < nalinks; i++) {
    struct link *l;

    eat(links[i].l_filenum, links[i].l_linenum);

    /* If this pass examined all its links, start the next pass.  */
    if (i == j) {
      if (nalinks - i == pass_size) {
	error(N_("\"Link %s %s\" is part of a link cycle"),
	      links[i].l_target, links[i].l_linkname);
	break;
      }
      j = nalinks;
      pass_size = nalinks - i;
    }

    /* Diagnose self links, which the cycle detection algorithm would not
       otherwise catch.  */
    if (strcmp(links[i].l_target, links[i].l_linkname) == 0) {
      error(N_("link %s targets itself"), links[i].l_target);
      continue;
    }

    /* Make this link unless its target has not been made yet.  */
    l = bsearch(links[i].l_target, &links[i + 1], j - (i + 1),
		sizeof *links, bsearch_linkcmp);
    if (!l)
      l = bsearch(links[i].l_target, &links[j], nalinks - j,
		  sizeof *links, bsearch_linkcmp);
    if (!l)
      dolink(links[i].l_target, links[i].l_linkname, false);
    else {
      /* The link target has not been made yet; copy the link to the end.  */
      links = growalloc(links, sizeof *links, nalinks, &nlinks_alloc);
      links[nalinks++] = links[i];
    }

    if (noise && i < nlinks) {
      if (l)
	warning(N_("link %s targeting link %s mishandled by pre-2023 zic"),
		links[i].l_linkname, links[i].l_target);
      else if (bsearch(links[i].l_target, links, nlinks, sizeof *links,
		       bsearch_linkcmp))
	warning(N_("link %s targeting link %s"),
		links[i].l_linkname, links[i].l_target);
    }
    check_for_signal();
  }
}

/* Simple signal handling: just set a flag that is checked
   periodically outside critical sections.  To set up the handler,
   prefer sigaction if available to close a signal race.  */

static sig_atomic_t got_signal;

static void
signal_handler(int sig)
{
#ifndef SA_SIGINFO
  signal(sig, signal_handler);
#endif
  got_signal = sig;
}

/* Arrange for SIGINT etc. to be caught by the handler.  */
static void
catch_signals(void)
{
  static int const signals[] = {
#ifdef SIGHUP
    SIGHUP,
#endif
    SIGINT,
#ifdef SIGPIPE
    SIGPIPE,
#endif
    SIGTERM
  };
  int i;
  for (i = 0; i < sizeof signals / sizeof signals[0]; i++) {
#ifdef SA_SIGINFO
    struct sigaction act0, act;
    act.sa_handler = signal_handler;
    sigemptyset(&act.sa_mask);
    act.sa_flags = 0;
    if (sigaction(signals[i], &act, &act0) == 0
	&& ! (act0.sa_flags & SA_SIGINFO) && act0.sa_handler == SIG_IGN) {
      sigaction(signals[i], &act0, NULL);
      got_signal = 0;
    }
#else
    if (signal(signals[i], signal_handler) == SIG_IGN) {
      signal(signals[i], SIG_IGN);
      got_signal = 0;
    }
#endif
  }
}

/* If a signal has arrived, terminate zic with appropriate status.  */
static void
check_for_signal(void)
{
  int sig = got_signal;
  if (sig) {
    signal(sig, SIG_DFL);
    raise(sig);
    abort(); /* A bug in 'raise'.  */
  }
}

enum { TIME_T_BITS_IN_FILE = 64 };

/* The minimum and maximum values representable in a TZif file.  */
static zic_t const min_time = MINVAL(zic_t, TIME_T_BITS_IN_FILE);
static zic_t const max_time = MAXVAL(zic_t, TIME_T_BITS_IN_FILE);

/* The minimum, and one less than the maximum, values specified by
   the -r option.  These default to MIN_TIME and MAX_TIME.  */
static zic_t lo_time = MINVAL(zic_t, TIME_T_BITS_IN_FILE);
static zic_t hi_time = MAXVAL(zic_t, TIME_T_BITS_IN_FILE);

/* The time specified by the -R option, defaulting to MIN_TIME;
   or lo_time, whichever is greater.  */
static zic_t redundant_time = MINVAL(zic_t, TIME_T_BITS_IN_FILE);

/* The time specified by an Expires line, or negative if no such line.  */
static zic_t leapexpires = -1;

/* Set the time range of the output to TIMERANGE.
   Return true if successful.  */
static bool
timerange_option(char *timerange)
{
  intmax_t lo = min_time, hi = max_time;
  char *lo_end = timerange, *hi_end;
  if (*timerange == '@') {
    errno = 0;
    lo = strtoimax(timerange + 1, &lo_end, 10);
    if (lo_end == timerange + 1 || (lo == INTMAX_MAX && errno == ERANGE))
      return false;
  }
  hi_end = lo_end;
  if (lo_end[0] == '/' && lo_end[1] == '@') {
    errno = 0;
    hi = strtoimax(lo_end + 2, &hi_end, 10);
    if (hi_end == lo_end + 2 || hi == INTMAX_MIN)
      return false;
    hi -= ! (hi == INTMAX_MAX && errno == ERANGE);
  }
  if (*hi_end || hi < lo || max_time < lo || hi < min_time)
    return false;
  lo_time = max(lo, min_time);
  hi_time = min(hi, max_time);
  return true;
}

/* Generate redundant time stamps up to OPT.  Return true if successful.  */
static bool
redundant_time_option(char *opt)
{
  if (*opt == '@') {
    intmax_t redundant;
    char *opt_end;
    redundant = strtoimax(opt + 1, &opt_end, 10);
    if (opt_end != opt + 1 && !*opt_end) {
      redundant_time = max(redundant_time, redundant);
      return true;
    }
  }
  return false;
}

static const char *	psxrules;
static const char *	lcltime;
static const char *	directory;
static const char *	tzdefault;

/* True if DIRECTORY ends in '/'.  */
static bool directory_ends_in_slash;

/* -1 if the TZif output file should be slim, 0 if default, 1 if the
   output should be fat for backward compatibility.  ZIC_BLOAT_DEFAULT
   determines the default.  */
static int bloat;

static bool
want_bloat(void)
{
  return 0 <= bloat;
}

#ifndef ZIC_BLOAT_DEFAULT
# define ZIC_BLOAT_DEFAULT "slim"
#endif

int
main(int argc, char **argv)
{
	register int c, k;
	register ptrdiff_t i, j;
	bool timerange_given = false;

#if HAVE_GETTEXT
	setlocale(LC_ALL, "");
# ifdef TZ_DOMAINDIR
	bindtextdomain(TZ_DOMAIN, TZ_DOMAINDIR);
# endif /* defined TEXTDOMAINDIR */
	textdomain(TZ_DOMAIN);
#endif /* HAVE_GETTEXT */
	main_argv = argv;
	progname = argv[0] ? argv[0] : "zic";
	if (TYPE_BIT(zic_t) < 64) {
		fprintf(stderr, "%s: %s\n", progname,
			_("wild compilation-time specification of zic_t"));
		return EXIT_FAILURE;
	}
	for (k = 1; k < argc; k++)
		if (strcmp(argv[k], "--version") == 0) {
			printf("zic %s%s\n", PKGVERSION, TZVERSION);
			close_file(stdout, NULL, NULL, NULL);
			return EXIT_SUCCESS;
		} else if (strcmp(argv[k], "--help") == 0) {
			usage(stdout, EXIT_SUCCESS);
		}
	while ((c = getopt(argc, argv, "b:d:Dg:l:L:m:p:r:R:st:u:vy:")) != -1)
		switch (c) {
			default:
				usage(stderr, EXIT_FAILURE);
			case 'b':
				if (strcmp(optarg, "slim") == 0) {
				  if (0 < bloat)
				    error(N_("incompatible -b options"));
				  bloat = -1;
				} else if (strcmp(optarg, "fat") == 0) {
				  if (bloat < 0)
				    error(N_("incompatible -b options"));
				  bloat = 1;
				} else
				  error(N_("invalid option: -b '%s'"), optarg);
				break;
			case 'd':
				if (directory)
				  duplicate_options("-d");
				directory = optarg;
				break;
			case 'D':
				skip_mkdir = true;
				break;
			case 'g':
				/* This undocumented option is present for
				   compatibility with FreeBSD 14.  */
				group_option(optarg);
				break;
			case 'l':
				if (lcltime)
				  duplicate_options("-l");
				lcltime = optarg;
				break;
			case 'm':
				if (output_mode != no_mode)
				  duplicate_options("-m");
				output_mode = mode_option(optarg);
				break;
			case 'p':
				if (psxrules)
				  duplicate_options("-p");
				if (strcmp(optarg, "-") != 0)
				  warning(N_("-p is obsolete"
					     " and likely ineffective"));
				psxrules = optarg;
				break;
			case 't':
				if (tzdefault)
				  duplicate_options("-t");
				tzdefault = optarg;
				break;
			case 'u':
				{
				  char *colon = strchr(optarg, ':');
				  if (colon)
				    *colon = '\0';
				  owner_option(optarg);
				  if (colon)
				    group_option(colon + 1);
				}
				break;
			case 'y':
				warning(N_("-y ignored"));
				break;
			case 'L':
				if (leapsec)
				  duplicate_options("-L");
				leapsec = optarg;
				break;
			case 'v':
				noise = true;
				break;
			case 'r':
				if (timerange_given)
				  duplicate_options("-r");
				if (! timerange_option(optarg)) {
				  fprintf(stderr,
					  _("%s: invalid time range: %s\n"),
					  progname, optarg);
				  return EXIT_FAILURE;
				}
				timerange_given = true;
				break;
			case 'R':
				if (! redundant_time_option(optarg)) {
				  fprintf(stderr, _("%s: invalid time: %s\n"),
					  progname, optarg);
				  return EXIT_FAILURE;
				}
				break;
			case 's':
				warning(N_("-s ignored"));
				break;
		}
	if (optind == argc - 1 && strcmp(argv[optind], "=") == 0)
		usage(stderr, EXIT_FAILURE);	/* usage message by request */
	if (hi_time + (hi_time < ZIC_MAX) < redundant_time) {
	  fprintf(stderr, _("%s: -R time exceeds -r cutoff\n"), progname);
	  return EXIT_FAILURE;
	}
	if (redundant_time < lo_time)
	  redundant_time = lo_time;
	if (bloat == 0) {
	  static char const bloat_default[] = ZIC_BLOAT_DEFAULT;
	  if (strcmp(bloat_default, "slim") == 0)
	    bloat = -1;
	  else if (strcmp(bloat_default, "fat") == 0)
	    bloat = 1;
	  else
	    abort(); /* Configuration error.  */
	}
	if (directory == NULL)
		directory = TZDIR;
	if (tzdefault == NULL)
		tzdefault = TZDEFAULT;

	if (optind < argc && leapsec != NULL) {
		infile(LEAPSEC_FILENUM, leapsec);
		adjleap();
	}

	for (k = optind; k < argc; k++)
	  infile(k, argv[k]);
	if (errors)
		return EXIT_FAILURE;
	associate();
	use_safe_temp_permissions();
	change_directory(directory);
	directory_ends_in_slash = directory[strlen(directory) - 1] == '/';
	catch_signals();
	for (i = 0; i < nzones; i = j) {
		/*
		** Find the next non-continuation zone entry.
		*/
		for (j = i + 1; j < nzones && zones[j].z_name == NULL; ++j)
			continue;
		outzone(&zones[i], j - i);
		check_for_signal();
	}
	make_links();
	if (lcltime != NULL) {
		eat(COMMAND_LINE_FILENUM, 1);
		dolink(lcltime, tzdefault, true);
	}
	if (psxrules != NULL) {
		eat(COMMAND_LINE_FILENUM, 1);
		dolink(psxrules, TZDEFRULES, true);
	}
	if (warnings && (ferror(stderr) || fclose(stderr) != 0))
	  return EXIT_FAILURE;
	return errors ? EXIT_FAILURE : EXIT_SUCCESS;
}

static bool
componentcheck(char const *name, char const *component,
	       char const *component_end)
{
	enum { component_len_max = 14 };
	ptrdiff_t component_len = component_end - component;
	if (component_len == 0) {
	  if (!*name)
	    error(N_("empty file name"));
	  else
	    error((component == name
		   ? N_("file name '%s' begins with '/'")
		   : *component_end
		   ? N_("file name '%s' contains '//'")
		   : N_("file name '%s' ends with '/'")),
		  name);
	  return false;
	}
	if (0 < component_len && component_len <= 2
	    && component[0] == '.' && component_end[-1] == '.') {
	  int len = component_len;
	  error(N_("file name '%s' contains '%.*s' component"),
		name, len, component);
	  return false;
	}
	if (noise) {
	  if (0 < component_len && component[0] == '-')
	    warning(N_("file name '%s' component contains leading '-'"),
		    name);
	  if (component_len_max < component_len)
	    warning(N_("file name '%s' contains overlength component"
		       " '%.*s...'"),
		    name, component_len_max, component);
	}
	return true;
}

static bool
namecheck(const char *name)
{
	register char const *cp;

	/* Benign characters in a portable file name.  */
	static char const benign[] =
	  "-/_"
	  "abcdefghijklmnopqrstuvwxyz"
	  "ABCDEFGHIJKLMNOPQRSTUVWXYZ";

	/* Non-control chars in the POSIX portable character set,
	   excluding the benign characters.  */
	static char const printable_and_not_benign[] =
	  " !\"#$%&'()*+,.0123456789:;<=>?@[\\]^`{|}~";

	register char const *component = name;
	for (cp = name; *cp; cp++) {
		unsigned char c = *cp;
#if defined _WIN32 || defined __CYGWIN__
		if (c == ':' || c == '\\') {
		  error(N_("file name '%s' contains '%c'"), name, c);
		  return false;
		}
#endif
		if (noise && !strchr(benign, c)) {
			warning((strchr(printable_and_not_benign, c)
				 ? N_("file name '%s' contains byte '%c'")
				 : N_("file name '%s' contains byte '\\%o'")),
				name, c);
		}
		if (c == '/') {
			if (!componentcheck(name, component, cp))
			  return false;
			component = cp + 1;
		}
	}
	return componentcheck(name, component, cp);
}

/* Return a random uint_fast64_t.  */
static uint_fast64_t
get_rand_u64(void)
{
#if HAVE_GETRANDOM
  static uint_fast64_t entropy_buffer[max(1, 256 / sizeof(uint_fast64_t))];
  static int nwords;
  if (!nwords) {
    ssize_t s;
    for (;; check_for_signal()) {
      s = getrandom(entropy_buffer, sizeof entropy_buffer, 0);
      if (! (s < 0 && errno == EINTR))
	break;
    }

    nwords = s < 0 ? -1 : s / sizeof *entropy_buffer;
  }
  if (0 < nwords)
    return entropy_buffer[--nwords];
#endif

  /* getrandom didn't work, so fall back on portable code that is
     not the best because the seed isn't cryptographically random and
     'rand' might not be cryptographically secure.  */
  {
    static bool initialized;
    if (!initialized) {
      srand(time(NULL));
      initialized = true;
    }
  }

  /* Return a random number if rand() yields a random number and in
     the typical case where RAND_MAX is one less than a power of two.
     In other cases this code yields a sort-of-random number.  */
  {
    uint_fast64_t rand_max = RAND_MAX,
      nrand = rand_max < UINT_FAST64_MAX ? rand_max + 1 : 0,
      rmod = INT_MAX < UINT_FAST64_MAX ? 0 : UINT_FAST64_MAX / nrand + 1,
      r = 0, rmax = 0;

    for (;; check_for_signal()) {
      uint_fast64_t rmax1 = rmax;
      if (rmod) {
	/* Avoid signed integer overflow on theoretical platforms
	   where uint_fast64_t promotes to int.  */
	rmax1 %= rmod;
	r %= rmod;
      }
      rmax1 = nrand * rmax1 + rand_max;
      r = nrand * r + rand();
      rmax = rmax < rmax1 ? rmax1 : UINT_FAST64_MAX;
      if (UINT_FAST64_MAX <= rmax)
	break;
    }

    return r;
  }
}

/* Generate a randomish name in the same directory as *NAME.  If
   *NAMEALLOC, put the name into *NAMEALLOC which is assumed to be
   that returned by a previous call and is thus already almost set up
   and equal to *NAME; otherwise, allocate a new name and put its
   address into both *NAMEALLOC and *NAME.  */
static void
random_dirent(char const **name, char **namealloc)
{
  char const *src = *name;
  char *dst = *namealloc;
  static char const prefix[] = ".zic";
  static char const alphabet[] =
    "abcdefghijklmnopqrstuvwxyz"
    "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
    "0123456789";
  enum { prefixlen = sizeof prefix - 1, alphabetlen = sizeof alphabet - 1 };
  int suffixlen = 6;
  char const *lastslash = strrchr(src, '/');
  ptrdiff_t dirlen = lastslash ? lastslash + 1 - src : 0;
  int i;
  uint_fast64_t r;
  uint_fast64_t base = alphabetlen;

  /* BASE**6 */
  uint_fast64_t base__6 = base * base * base * base * base * base;

  /* The largest uintmax_t that is a multiple of BASE**6.  Any random
     uintmax_t value that is this value or greater, yields a biased
     remainder when divided by BASE**6.  UNFAIR_MIN equals the
     mathematical value of ((UINTMAX_MAX + 1) - (UINTMAX_MAX + 1) % BASE**6)
     computed without overflow.  */
  uint_fast64_t unfair_min = - ((UINTMAX_MAX % base__6 + 1) % base__6);

  if (!dst) {
    char *cp = dst = xmalloc(size_sum(dirlen, prefixlen + suffixlen + 1));
    cp = mempcpy(cp, src, dirlen);
    cp = mempcpy(cp, prefix, prefixlen);
    cp[suffixlen] = '\0';
    *name = *namealloc = dst;
  }

  for (;; check_for_signal()) {
    r = get_rand_u64();
    if (r < unfair_min)
      break;
  }

  for (i = 0; i < suffixlen; i++) {
    dst[dirlen + prefixlen + i] = alphabet[r % alphabetlen];
    r /= alphabetlen;
  }
}

/* For diagnostics the directory, and file name relative to that
   directory, respectively.  A diagnostic routine can name FILENAME by
   outputting diagdir(FILENAME), then diagslash(FILENAME), then FILENAME.  */
static char const *
diagdir(char const *filename)
{
  return *filename == '/' ? "" : directory;
}
static char const *
diagslash(char const *filename)
{
  return &"/"[*filename == '/' || directory_ends_in_slash];
}

/* Prepare to write to the file *OUTNAME, using *TEMPNAME to store the
   name of the temporary file that will eventually be renamed to
   *OUTNAME.  Assign the temporary file's name to both *OUTNAME and
   *TEMPNAME.  If *TEMPNAME is null, allocate the name of any such
   temporary file; otherwise, reuse *TEMPNAME's storage, which is
   already set up and only needs its trailing suffix updated.  */
static FILE *
open_outfile(char const **outname, char **tempname)
{
  bool dirs_made = false;
  if (!*tempname)
    random_dirent(outname, tempname);

  for (;; check_for_signal()) {
    int oflags = O_WRONLY | O_BINARY | O_CREAT | O_EXCL;
    int fd = open(*outname, oflags, creat_perms);
    int err;
    if (fd < 0)
      err = errno;
    else {
      FILE *fp = fdopen(fd, "wb");
      if (fp)
	return fp;
      err = errno;
      close(fd);
    }
    if (err == ENOENT && !dirs_made) {
      mkdirs(*outname, true);
      dirs_made = true;
    } else if (err == EEXIST)
      random_dirent(outname, tempname);
    else {
      fprintf(stderr, _("%s: Can't create %s%s%s: %s\n"),
	      progname, diagdir(*outname), diagslash(*outname), *outname,
	      strerror(err));
      exit(EXIT_FAILURE);
    }
  }
}

/* If TEMPNAME, the result is in the temporary file TEMPNAME even
   though the user wanted it in NAME, so rename TEMPNAME to NAME.
   Report an error and exit if there is trouble.  Also, free TEMPNAME.  */
static void
rename_dest(char *tempname, char const *name)
{
  if (tempname) {
    if (rename(tempname, name) != 0) {
      int rename_errno = errno;
      fprintf(stderr, _("%s: rename to %s%s%s: %s\n"),
	      progname, diagdir(name), diagslash(name), name,
	      strerror(rename_errno));
      remove_temp(tempname);
      exit(EXIT_FAILURE);
    }
    free(tempname);
  }
}

/* Remove the temporary file TEMP, diagnosing any failure.  */
static void
remove_temp(char const *temp)
{
  if (remove(temp) < 0)
    fprintf(stderr, _("%s: Can't remove temporary file %s%s%s: %s\n"),
	    progname, diagdir(temp), diagslash(temp), temp, strerror(errno));
}

/* Create symlink contents suitable for symlinking TARGET to LINKNAME, as a
   freshly allocated string.  TARGET should be a relative file name, and
   is relative to the global variable DIRECTORY.  LINKNAME can be either
   relative or absolute.  Return a null pointer if the symlink contents
   was not computed because LINKNAME is absolute but DIRECTORY is not.  */
static char *
relname(char const *target, char const *linkname)
{
  size_t i, taillen, dir_len = 0, dotdots = 0;
  ptrdiff_t dotdotetcsize, linksize = INDEX_MAX;
  char const *f = target;
  char *result = NULL;
  if (*linkname == '/') {
    /* Make F absolute too.  */
    size_t len = strlen(directory);
    bool needs_slash = len && directory[len - 1] != '/';
    size_t lenslash = len + needs_slash;
    size_t targetsize = strlen(target) + 1;
    char *cp;
    if (*directory != '/')
      return NULL;
    linksize = size_sum(lenslash, targetsize);
    f = cp = result = xmalloc(linksize);
    cp = mempcpy(cp, directory, len);
    *cp = '/';
    memcpy(cp + needs_slash, target, targetsize);
  }
  for (i = 0; f[i] && f[i] == linkname[i]; i++)
    if (f[i] == '/')
      dir_len = i + 1;
  for (; linkname[i]; i++)
    dotdots += linkname[i] == '/' && linkname[i - 1] != '/';
  taillen = strlen(f + dir_len);
  dotdotetcsize = size_sum(size_product(dotdots, 3), taillen + 1);
  if (dotdotetcsize <= linksize) {
    char *cp;
    if (!result)
      result = xmalloc(dotdotetcsize);
    cp = result;
    for (i = 0; i < dotdots; i++)
      cp = mempcpy(cp, "../", 3);
    memmove(cp, f + dir_len, taillen + 1);
  }
  return result;
}

/* Return true if A and B must have the same parent dir if A and B exist.
   Return false if this is not necessarily true (though it might be true).
   Keep it simple, and do not inspect the file system.  */
ATTRIBUTE_PURE_114833
static bool
same_parent_dirs(char const *a, char const *b)
{
  for (; *a == *b; a++, b++)
    if (!*a)
      return true;
  return ! (strchr(a, '/') || strchr(b, '/'));
}

static void
dolink(char const *target, char const *linkname, bool staysymlink)
{
	bool linkdirs_made = false;
	int link_errno;
	char *tempname = NULL;
	char const *outname = linkname;
	int targetissym = -2, linknameissym = -2;

	if (strcmp(target, "-") == 0) {
	  if (remove(linkname) == 0 || errno == ENOENT || errno == ENOTDIR)
	    return;
	  else {
	    char const *e = strerror(errno);
	    fprintf(stderr, _("%s: Can't remove %s%s%s: %s\n"),
		    progname, diagdir(linkname), diagslash(linkname), linkname,
		    e);
	    exit(EXIT_FAILURE);
	  }
	}

	for (;; check_for_signal()) {
	  if (linkat(AT_FDCWD, target, AT_FDCWD, outname, AT_SYMLINK_FOLLOW)
	      == 0) {
	    link_errno = 0;
	    break;
	  }
	  link_errno = errno;
	  /* Linux 2.6.16 and 2.6.17 mishandle AT_SYMLINK_FOLLOW.  */
	  if (link_errno == EINVAL)
	    link_errno = ENOTSUP;
#if HAVE_LINK
	  /* If linkat is not supported, fall back on link(A, B).
	     However, skip this if A is a relative symlink
	     and A and B might not have the same parent directory.
	     On some platforms link(A, B) does not follow a symlink A,
	     and if A is relative it might misbehave elsewhere.  */
	  if (link_errno == ENOTSUP
	      && (same_parent_dirs(target, outname)
		  || 0 <= itssymlink(target, &targetissym))) {
	    if (link(target, outname) == 0) {
	      link_errno = 0;
	      break;
	    }
	    link_errno = errno;
	    /* When hard links are not supported, some MS-Windows file system
	       drivers fail with EINVAL, contrary to the intent of
	       MS-FSA 42.0 (2025) section 2.1.5.15.7.  */
	    if (link_errno == EINVAL)
	      link_errno = ENOTSUP;
	  }
#endif
	  /* On platforms like AIX, the Linux kernel, macOS, and Solaris,
	     link/linkat can fail with ENOSYS or EPERM if the file system
	     does not support hard links, or if other problems occur.
	     It is too much trouble to suss out the other problems.  */
	  if (link_errno == ENOSYS || link_errno == EPERM)
	    link_errno = ENOTSUP;

	  if (link_errno == EXDEV || link_errno == ENOTSUP)
	    break;

	  if (link_errno == EEXIST) {
	    staysymlink &= !tempname;
	    random_dirent(&outname, &tempname);
	    if (staysymlink && itssymlink(linkname, &linknameissym))
	      break;
	  } else if (link_errno == ENOENT && !linkdirs_made) {
	    mkdirs(linkname, true);
	    linkdirs_made = true;
	  } else {
	    fprintf(stderr, _("%s: Can't link %s%s%s to %s%s%s: %s\n"),
		    progname, diagdir(target), diagslash(target), target,
		    diagdir(outname), diagslash(outname), outname,
		    strerror(link_errno));
	    exit(EXIT_FAILURE);
	  }
	}
	if (link_errno != 0) {
	  bool absolute = *target == '/';
	  char *linkalloc = absolute ? NULL : relname(target, linkname);
	  char const *contents = absolute ? target : linkalloc;
	  int symlink_errno = -1;

	  if (contents) {
	    for (;; check_for_signal()) {
	      if (symlink(contents, outname) == 0) {
		symlink_errno = 0;
		break;
	      }
	      symlink_errno = errno;
	      if (symlink_errno == EEXIST)
		random_dirent(&outname, &tempname);
	      else if (symlink_errno == ENOENT && !linkdirs_made) {
		mkdirs(linkname, true);
		linkdirs_made = true;
	      } else
		break;
	    }
	  }
	  free(linkalloc);
	  if (symlink_errno == 0) {
	    if (link_errno != ENOTSUP && link_errno != EEXIST)
	      warning(N_("symbolic link used because hard link failed: %s"),
		      strerror(link_errno));
	  } else {
	    FILE *fp, *tp;
	    int c;
	    fp = fopen(target, "rb");
	    if (!fp) {
	      char const *e = strerror(errno);
	      fprintf(stderr, _("%s: Can't read %s%s%s: %s\n"),
		      progname, diagdir(target), diagslash(target), target, e);
	      exit(EXIT_FAILURE);
	    }
	    tp = open_outfile(&outname, &tempname);
	    for (; (c = getc(fp)) != EOF; check_for_signal())
	      putc(c, tp);
	    close_file(tp, directory, linkname, tempname);
	    close_file(fp, directory, target, NULL);
	    if (link_errno != ENOTSUP)
	      warning(N_("copy used because hard link failed: %s"),
		      strerror(link_errno));
	    else if (symlink_errno < 0)
	      warning(N_("copy used because symbolic link not obvious"));
	    else if (symlink_errno != ENOSYS && symlink_errno != ENOTSUP
		     && symlink_errno != EPERM)
	      warning(N_("copy used because symbolic link failed: %s"),
		      strerror(symlink_errno));
	  }
	}
	rename_dest(tempname, linkname);
}

/* Return 1 if NAME is an absolute symbolic link, -1 if it is relative,
   0 if it is not a symbolic link.  If *CACHE is not -2, it is the
   cached result of a previous call to this function with the same NAME.  */
static int
itssymlink(char const *name, int *cache)
{
  if (*cache == -2) {
    char c = '\0';
    *cache = readlink(name, &c, 1) < 0 ? 0 : c == '/' ? 1 : -1;
  }
  return *cache;
}

/*
** Associate sets of rules with zones.
*/

/*
** Sort by rule name.
*/

static int
rcomp(const void *cp1, const void *cp2)
{
  struct rule const *r1 = cp1, *r2 = cp2;
  return strcmp(r1->r_name, r2->r_name);
}

static void
associate(void)
{
	register struct zone *	zp;
	register struct rule *	rp;
	register ptrdiff_t i, j, base, out;

	if (1 < nrules) {
		qsort(rules, nrules, sizeof *rules, rcomp);
		for (i = 0; i < nrules - 1; ++i) {
			if (strcmp(rules[i].r_name,
				rules[i + 1].r_name) != 0)
					continue;
			if (rules[i].r_filenum == rules[i + 1].r_filenum)
					continue;
			eat(rules[i].r_filenum, rules[i].r_linenum);
			warning(N_("same rule name in multiple files"));
			eat(rules[i + 1].r_filenum, rules[i + 1].r_linenum);
			warning(N_("same rule name in multiple files"));
			for (j = i + 2; j < nrules; ++j) {
				if (strcmp(rules[i].r_name,
					rules[j].r_name) != 0)
						break;
				if (rules[i].r_filenum == rules[j].r_filenum)
						continue;
				if (rules[i + 1].r_filenum
				    == rules[j].r_filenum)
						continue;
				break;
			}
			i = j - 1;
		}
	}
	for (i = 0; i < nzones; ++i) {
		zp = &zones[i];
		zp->z_rules = NULL;
		zp->z_nrules = 0;
	}
	for (base = 0; base < nrules; base = out) {
		rp = &rules[base];
		for (out = base + 1; out < nrules; ++out)
			if (strcmp(rp->r_name, rules[out].r_name) != 0)
				break;
		for (i = 0; i < nzones; ++i) {
			zp = &zones[i];
			if (strcmp(zp->z_rule, rp->r_name) != 0)
				continue;
			zp->z_rules = rp;
			zp->z_nrules = out - base;
		}
	}
	for (i = 0; i < nzones; ++i) {
		zp = &zones[i];
		if (zp->z_nrules == 0) {
			/*
			** Maybe we have a local standard time offset.
			*/
			eat(zp->z_filenum, zp->z_linenum);
			zp->z_save = getsave(zp->z_rule, &zp->z_isdst);
			/*
			** Note, though, that if there's no rule,
			** a '%s' in the format is a bad thing.
			*/
			if (zp->z_format_specifier == 's')
			  error(N_("%%s in ruleless zone"));
		}
	}
	if (errors)
		exit(EXIT_FAILURE);
}

/* Read a text line from FP into BUF, which is of size BUFSIZE.
   Terminate it with a NUL byte instead of a newline.
   Return true if successful, false if EOF.
   On error, report the error and exit.  */
static bool
inputline(FILE *fp, char *buf, ptrdiff_t bufsize)
{
  ptrdiff_t linelen = 0, ch;
  for (; (ch = getc(fp)) != '\n'; check_for_signal()) {
    if (ch < 0) {
      if (ferror(fp)) {
	error(N_("input error"));
	exit(EXIT_FAILURE);
      }
      if (linelen == 0)
	return false;
      error(N_("unterminated line"));
      exit(EXIT_FAILURE);
    }
    if (!ch) {
      error(N_("NUL input byte"));
      exit(EXIT_FAILURE);
    }
    buf[linelen++] = ch;
    if (linelen == bufsize) {
      error(N_("line too long"));
      exit(EXIT_FAILURE);
    }
  }
  buf[linelen] = '\0';
  return true;
}

static void
infile(int fnum, char const *name)
{
	register FILE *			fp;
	register const struct lookup *	lp;
	register bool			wantcont;
	register lineno			num;

	if (strcmp(name, "-") == 0) {
		fp = stdin;
	} else if ((fp = fopen(name, "r")) == NULL) {
		const char *e = strerror(errno);

		fprintf(stderr, _("%s: Can't open %s: %s\n"),
			progname, name, e);
		exit(EXIT_FAILURE);
	}
	wantcont = false;
	for (num = 1; ; ++num) {
		enum { bufsize_bound
		  = (min(INT_MAX, INDEX_MAX) / FORMAT_LEN_GROWTH_BOUND) };
		char buf[min(_POSIX2_LINE_MAX, bufsize_bound)];
		int nfields;
		char *fields[MAX_FIELDS];
		eat(fnum, num);
		if (!inputline(fp, buf, sizeof buf))
		  break;
		nfields = getfields(buf, fields,
				    sizeof fields / sizeof *fields);
		if (nfields == 0) {
			/* nothing to do */
		} else if (wantcont) {
			wantcont = inzcont(fields, nfields);
		} else {
			struct lookup const *line_codes
			  = fnum < 0 ? leap_line_codes : zi_line_codes;
			lp = byword(fields[0], line_codes);
			if (lp == NULL)
				error(N_("input line of unknown type"));
			else switch (lp->l_value) {
				case LC_RULE:
					inrule(fields, nfields);
					wantcont = false;
					break;
				case LC_ZONE:
					wantcont = inzone(fields, nfields);
					break;
				case LC_LINK:
					inlink(fields, nfields);
					wantcont = false;
					break;
				case LC_LEAP:
					inleap(fields, nfields);
					wantcont = false;
					break;
				case LC_EXPIRES:
					inexpires(fields, nfields);
					wantcont = false;
					break;
				default: unreachable();
			}
		}
		check_for_signal();
	}
	close_file(fp, NULL, filename(fnum), NULL);
	if (wantcont)
		error(N_("expected continuation line not found"));
}

/*
** Convert a string of one of the forms
**	h	-h	hh:mm	-hh:mm	hh:mm:ss	-hh:mm:ss
** into a number of seconds.
** A null string maps to zero.
** Call error with msgid and return zero on errors.
*/

static zic_t
gethms(char const *string, char const *msgid)
{
	zic_t hh, r;
	int sign, mm = 0, ss = 0;
	char hhx, mmx, ssx, xr = '0', xs;
	int tenths = 0;
	bool ok = true;

	if (string == NULL || *string == '\0')
		return 0;
	if (*string == '-') {
		sign = -1;
		++string;
	} else	sign = 1;
	switch (sscanf(string,
		       "%"SCNdZIC"%c%d%c%d%c%1d%*[0]%c%*[0123456789]%c",
		       &hh, &hhx, &mm, &mmx, &ss, &ssx, &tenths, &xr, &xs)) {
	  default: ok = false; break;
	  case 8:
	    ok = is_digit(xr);
	    ATTRIBUTE_FALLTHROUGH;
	  case 7:
	    ok &= ssx == '.';
	    if (ok && noise)
	      warning(N_("fractional seconds rejected by"
			 " pre-2018 versions of zic"));
	    ATTRIBUTE_FALLTHROUGH;
	  case 5: ok &= mmx == ':'; ATTRIBUTE_FALLTHROUGH;
	  case 3: ok &= hhx == ':'; ATTRIBUTE_FALLTHROUGH;
	  case 1: break;
	}
	if (!ok || hh < 0 ||
		mm < 0 || mm >= MINSPERHOUR ||
		ss < 0 || ss > SECSPERMIN) {
			error(N_("%s: %s"), string, _(msgid));
			return 0;
	}
	ss += 5 + ((ss ^ 1) & (xr == '0')) <= tenths; /* Round to even.  */
	if (noise && (hh > HOURSPERDAY ||
		(hh == HOURSPERDAY && (mm != 0 || ss != 0))))
	  warning(N_("values over 24 hours"
		     " not handled by pre-2007 versions of zic"));
	r = oadd(omul(hh, sign * SECSPERHOUR), sign * (mm * SECSPERMIN + ss));

	/* Check that 4 * R fits in zic_t.  This is more than enough
	   for times of day and for UT offsets in TZif files, and it
	   avoids later problems during arithmetic on these numbers,
	   as stringzone can compute up to 3 times one of these values
	   when computing fake timezones, and stringrule can do a bit more.
	   Do not limit R strictly to 32 bits here, as it is OK to go
	   over the TZif limit temporarily so long as the result fits.  */
	omul(r, 4);
	/* Discard omul's return value, as it is not needed here.  */

	return r;
}

static zic_t
getsave(char *field, bool *isdst)
{
  int dst = -1;
  zic_t save;
  ptrdiff_t fieldlen = strlen(field);
  if (fieldlen != 0) {
    char *ep = field + fieldlen - 1;
    switch (*ep) {
      case 'd': dst = 1; *ep = '\0'; break;
      case 's': dst = 0; *ep = '\0'; break;
    }
  }
  save = gethms(field, N_("invalid saved time"));
  *isdst = dst < 0 ? save != 0 : dst;
  return save;
}

static void
inrule(char **fields, int nfields)
{
	struct rule r;

	if (nfields != RULE_FIELDS) {
		error(N_("wrong number of fields on Rule line"));
		return;
	}
	switch (*fields[RF_NAME]) {
	  case '\0':
	  case ' ': case '\f': case '\n': case '\r': case '\t': case '\v':
	  case '+': case '-':
	  case '0': case '1': case '2': case '3': case '4':
	  case '5': case '6': case '7': case '8': case '9':
		error(N_("Invalid rule name \"%s\""), fields[RF_NAME]);
		return;
	}
	r.r_filenum = filenum;
	r.r_linenum = linenum;
	r.r_save = getsave(fields[RF_SAVE], &r.r_isdst);
	if (!rulesub(&r, fields[RF_LOYEAR], fields[RF_HIYEAR],
		     fields[RF_COMMAND], fields[RF_MONTH], fields[RF_DAY],
		     fields[RF_TOD]))
	  return;
	r.r_name = xstrdup(fields[RF_NAME]);
	r.r_abbrvar = xstrdup(fields[RF_ABBRVAR]);
	if (max_abbrvar_len < strlen(r.r_abbrvar))
		max_abbrvar_len = strlen(r.r_abbrvar);
	rules = growalloc(rules, sizeof *rules, nrules, &nrules_alloc);
	rules[nrules++] = r;
}

static bool
inzone(char **fields, int nfields)
{
	register ptrdiff_t i;

	if (nfields < ZONE_MINFIELDS || nfields > ZONE_MAXFIELDS) {
		error(N_("wrong number of fields on Zone line"));
		return false;
	}
	if (lcltime != NULL && strcmp(fields[ZF_NAME], tzdefault) == 0) {
	  error(N_("\"Zone %s\" line and -l option are mutually exclusive"),
		tzdefault);
	  return false;
	}
	if (strcmp(fields[ZF_NAME], TZDEFRULES) == 0 && psxrules != NULL) {
	  error(N_("\"Zone %s\" line and -p option are mutually exclusive"),
		TZDEFRULES);
	  return false;
	}
	for (i = 0; i < nzones; ++i)
		if (zones[i].z_name != NULL &&
			strcmp(zones[i].z_name, fields[ZF_NAME]) == 0) {
				error(N_("duplicate zone name %s"
					" (file \"%s\", line %"PRIdMAX")"),
				      fields[ZF_NAME],
				      filename(zones[i].z_filenum),
				      zones[i].z_linenum);
				return false;
		}
	return inzsub(fields, nfields, false);
}

static bool
inzcont(char **fields, int nfields)
{
	if (nfields < ZONEC_MINFIELDS || nfields > ZONEC_MAXFIELDS) {
		error(N_("wrong number of fields on Zone continuation line"));
		return false;
	}
	return inzsub(fields, nfields, true);
}

static bool
inzsub(char **fields, int nfields, bool iscont)
{
	register char *		cp;
	char *			cp1;
	struct zone z;
	int format_len;
	register int		i_stdoff, i_rule, i_format;
	register int		i_untilyear, i_untilmonth;
	register int		i_untilday, i_untiltime;
	register bool		hasuntil;

	if (iscont) {
		i_stdoff = ZFC_STDOFF;
		i_rule = ZFC_RULE;
		i_format = ZFC_FORMAT;
		i_untilyear = ZFC_TILYEAR;
		i_untilmonth = ZFC_TILMONTH;
		i_untilday = ZFC_TILDAY;
		i_untiltime = ZFC_TILTIME;
	} else if (!namecheck(fields[ZF_NAME]))
		return false;
	else {
		i_stdoff = ZF_STDOFF;
		i_rule = ZF_RULE;
		i_format = ZF_FORMAT;
		i_untilyear = ZF_TILYEAR;
		i_untilmonth = ZF_TILMONTH;
		i_untilday = ZF_TILDAY;
		i_untiltime = ZF_TILTIME;
	}
	z.z_filenum = filenum;
	z.z_linenum = linenum;
	z.z_stdoff = gethms(fields[i_stdoff], N_("invalid UT offset"));
	cp = strchr(fields[i_format], '%');
	if (cp) {
		if ((*++cp != 's' && *cp != 'z') || strchr(cp, '%')
		    || strchr(fields[i_format], '/')) {
			error(N_("invalid abbreviation format"));
			return false;
		}
	}
	z.z_format_specifier = cp ? *cp : '\0';
	format_len = strlen(fields[i_format]);
	if (max_format_len < format_len)
	  max_format_len = format_len;
	hasuntil = nfields > i_untilyear;
	if (hasuntil) {
		z.z_untilrule.r_filenum = filenum;
		z.z_untilrule.r_linenum = linenum;
		if (!rulesub(
			&z.z_untilrule,
			fields[i_untilyear],
			"only",
			"",
			(nfields > i_untilmonth) ?
			fields[i_untilmonth] : "Jan",
			(nfields > i_untilday) ? fields[i_untilday] : "1",
			(nfields > i_untiltime) ? fields[i_untiltime] : "0"))
		  return false;
		z.z_untiltime = rpytime(&z.z_untilrule,
			z.z_untilrule.r_loyear);
		if (iscont && nzones > 0 &&
			zones[nzones - 1].z_untiltime >= z.z_untiltime) {
		  error(N_("Zone continuation line end time is"
			   " not after end time of previous line"));
		  return false;
		}
	}
	z.z_name = iscont ? NULL : xstrdup(fields[ZF_NAME]);
	z.z_rule = xstrdup(fields[i_rule]);
	z.z_format = cp1 = xstrdup(fields[i_format]);
	if (z.z_format_specifier == 'z') {
	  cp1[cp - fields[i_format]] = 's';
	  if (noise)
	    warning(N_("format '%s' not handled by pre-2015 versions of zic"),
		    fields[i_format]);
	}
	zones = growalloc(zones, sizeof *zones, nzones, &nzones_alloc);
	zones[nzones++] = z;
	/*
	** If there was an UNTIL field on this line,
	** there's more information about the zone on the next line.
	*/
	return hasuntil;
}

static zic_t
getleapdatetime(char **fields, bool expire_line)
{
	register const struct lookup *	lp;
	zic_t				year;
	zic_t				t;
	char xs;
	struct rule r;

	if (sscanf(fields[LP_YEAR], "%"SCNdZIC"%c", &year, &xs) != 1) {
		/*
		** Leapin' Lizards!
		*/
		error(N_("invalid leaping year"));
		return -1;
	}
	if (!expire_line) {
	    if (!leapseen || leapmaxyear < year)
		leapmaxyear = year;
	    if (!leapseen || leapminyear > year)
		leapminyear = year;
	    leapseen = true;
	}
	if ((lp = byword(fields[LP_MONTH], mon_names)) == NULL) {
		error(N_("invalid month name"));
		return -1;
	}
	r.r_month = lp->l_value;
	if (! (sscanf(fields[LP_DAY], "%d%c", &r.r_dayofmonth, &xs) == 1
	       && 0 < r.r_dayofmonth
	       && r.r_dayofmonth <= len_months[isleap(year)][r.r_month])) {
			error(N_("invalid day of month"));
			return -1;
	}
	r.r_dycode = DC_DOM;
	r.r_tod = gethms(fields[LP_TIME], N_("invalid time of day"));
	t = rpytime(&r, year);
	if (t < 0)
	  error(N_("leap second precedes Epoch"));
	return t;
}

static void
inleap(char **fields, int nfields)
{
  if (nfields != LEAP_FIELDS)
    error(N_("wrong number of fields on Leap line"));
  else {
    zic_t t = getleapdatetime(fields, false);
    if (0 <= t) {
      struct lookup const *lp = byword(fields[LP_ROLL], leap_types);
      if (!lp)
	error(N_("invalid Rolling/Stationary field on Leap line"));
      else {
	int correction = 0;
	if (!fields[LP_CORR][0]) /* infile() turns "-" into "".  */
	  correction = -1;
	else if (strcmp(fields[LP_CORR], "+") == 0)
	  correction = 1;
	else
	  error(N_("invalid CORRECTION field on Leap line"));
	if (correction)
	  leapadd(t, correction, lp->l_value);
      }
    }
  }
}

static void
inexpires(char **fields, int nfields)
{
  if (nfields != EXPIRES_FIELDS)
    error(N_("wrong number of fields on Expires line"));
  else if (0 <= leapexpires)
    error(N_("multiple Expires lines"));
  else
    leapexpires = getleapdatetime(fields, true);
}

static void
inlink(char **fields, int nfields)
{
	struct link	l;

	if (nfields != LINK_FIELDS) {
		error(N_("wrong number of fields on Link line"));
		return;
	}
	if (*fields[LF_TARGET] == '\0') {
		error(N_("blank TARGET field on Link line"));
		return;
	}
	if (! namecheck(fields[LF_TARGET]))
	  return;
	if (! namecheck(fields[LF_LINKNAME]))
	  return;
	l.l_filenum = filenum;
	l.l_linenum = linenum;
	l.l_target = xstrdup(fields[LF_TARGET]);
	l.l_linkname = xstrdup(fields[LF_LINKNAME]);
	links = growalloc(links, sizeof *links, nlinks, &nlinks_alloc);
	links[nlinks++] = l;
}

static bool
rulesub(struct rule *rp, const char *loyearp, const char *hiyearp,
	const char *typep, const char *monthp, const char *dayp,
	const char *timep)
{
	register const struct lookup *	lp;
	register const char *		cp;
	register char *			dp;
	register char *			ep;
	char xs;

	if ((lp = byword(monthp, mon_names)) == NULL) {
		error(N_("invalid month name"));
		return false;
	}
	rp->r_month = lp->l_value;
	rp->r_todisstd = false;
	rp->r_todisut = false;
	dp = xstrdup(timep);
	if (*dp != '\0') {
		ep = dp + strlen(dp) - 1;
		switch (lowerit(*ep)) {
			case 's':	/* Standard */
				rp->r_todisstd = true;
				rp->r_todisut = false;
				*ep = '\0';
				break;
			case 'w':	/* Wall */
				rp->r_todisstd = false;
				rp->r_todisut = false;
				*ep = '\0';
				break;
			case 'g':	/* Greenwich */
			case 'u':	/* Universal */
			case 'z':	/* Zulu */
				rp->r_todisstd = true;
				rp->r_todisut = true;
				*ep = '\0';
				break;
		}
	}
	rp->r_tod = gethms(dp, N_("invalid time of day"));
	free(dp);
	/*
	** Year work.
	*/
	cp = loyearp;
	lp = byword(cp, begin_years);
	if (lp) switch (lp->l_value) {
		case YR_MINIMUM:
			warning(N_("FROM year \"%s\" is obsolete;"
				   " treated as %d"),
				cp, YEAR_32BIT_MIN - 1);
			rp->r_loyear = YEAR_32BIT_MIN - 1;
			break;
		default: unreachable();
	} else if (sscanf(cp, "%"SCNdZIC"%c", &rp->r_loyear, &xs) != 1) {
		error(N_("invalid starting year"));
		return false;
	}
	cp = hiyearp;
	lp = byword(cp, end_years);
	rp->r_hiwasnum = lp == NULL;
	if (!rp->r_hiwasnum) switch (lp->l_value) {
		case YR_MAXIMUM:
			rp->r_hiyear = ZIC_MAX;
			break;
		case YR_ONLY:
			rp->r_hiyear = rp->r_loyear;
			break;
		default: unreachable();
	} else if (sscanf(cp, "%"SCNdZIC"%c", &rp->r_hiyear, &xs) != 1) {
		error(N_("invalid ending year"));
		return false;
	}
	if (rp->r_loyear > rp->r_hiyear) {
		error(N_("starting year greater than ending year"));
		return false;
	}
	if (*typep != '\0') {
		error(N_("year type \"%s\" is unsupported; use \"-\" instead"),
		      typep);
		return false;
	}
	/*
	** Day work.
	** Accept things such as:
	**	1
	**	lastSunday
	**	last-Sunday (undocumented; warn about this)
	**	Sun<=20
	**	Sun>=7
	*/
	dp = xstrdup(dayp);
	if ((lp = byword(dp, lasts)) != NULL) {
		rp->r_dycode = DC_DOWLEQ;
		rp->r_wday = lp->l_value;
		rp->r_dayofmonth = len_months[1][rp->r_month];
	} else {
		ep = strchr(dp, '<');
		if (ep)
		    rp->r_dycode = DC_DOWLEQ;
		else {
		    ep = strchr(dp, '>');
		    if (ep)
			rp->r_dycode = DC_DOWGEQ;
		    else {
			ep = dp;
			rp->r_dycode = DC_DOM;
		    }
		}
		if (rp->r_dycode != DC_DOM) {
			*ep++ = 0;
			if (*ep++ != '=') {
				error(N_("invalid day of month"));
				free(dp);
				return false;
			}
			if ((lp = byword(dp, wday_names)) == NULL) {
				error(N_("invalid weekday name"));
				free(dp);
				return false;
			}
			rp->r_wday = lp->l_value;
		}
		if (sscanf(ep, "%d%c", &rp->r_dayofmonth, &xs) != 1 ||
			rp->r_dayofmonth <= 0 ||
			(rp->r_dayofmonth > len_months[1][rp->r_month])) {
				error(N_("invalid day of month"));
				free(dp);
				return false;
		}
	}
	free(dp);
	return true;
}

static void
convert(uint_fast32_t val, char *buf)
{
	register int	i;
	register int	shift;
	unsigned char *const b = (unsigned char *) buf;

	for (i = 0, shift = 24; i < 4; ++i, shift -= 8)
	  b[i] = (val >> shift) & 0xff;
}

static void
convert64(uint_fast64_t val, char *buf)
{
	register int	i;
	register int	shift;
	unsigned char *const b = (unsigned char *) buf;

	for (i = 0, shift = 56; i < 8; ++i, shift -= 8)
	  b[i] = (val >> shift) & 0xff;
}

static void
puttzcode(zic_t val, FILE *fp)
{
	char	buf[4];

	convert(val, buf);
	fwrite(buf, sizeof buf, 1, fp);
}

static void
puttzcodepass(zic_t val, FILE *fp, int pass)
{
  if (pass == 1)
    puttzcode(val, fp);
  else {
	char	buf[8];

	convert64(val, buf);
	fwrite(buf, sizeof buf, 1, fp);
  }
}

static int
atcomp(const void *avp, const void *bvp)
{
  struct attype const *ap = avp, *bp = bvp;
  zic_t a = ap->at, b = bp->at;
  return a < b ? -1 : a > b;
}

struct timerange {
  int defaulttype;
  ptrdiff_t base, count;
  ptrdiff_t leapbase, leapcount;
  bool leapexpiry;
};

static struct timerange
limitrange(struct timerange r, zic_t lo, zic_t hi,
	   zic_t const *ats, unsigned char const *types)
{
  /* Omit ordinary transitions < LO.  */
  while (0 < r.count && ats[r.base] < lo) {
    r.defaulttype = types[r.base];
    r.count--;
    r.base++;
  }

  /* Omit as many initial leap seconds as possible, such that the
     first leap second in the truncated list is <= LO, and is a
     positive leap second if and only if it has a positive correction.
     This supports common TZif readers that assume that the first leap
     second is positive if and only if its correction is positive.  */
  while (1 < r.leapcount && leap[r.leapbase + 1].trans <= lo) {
    r.leapcount--;
    r.leapbase++;
  }
  while (0 < r.leapbase
	 && ((leap[r.leapbase - 1].corr < leap[r.leapbase].corr)
	     != (0 < leap[r.leapbase].corr))) {
    r.leapcount++;
    r.leapbase--;
  }


  /* Omit ordinary and leap second transitions greater than HI + 1.  */
  if (hi < max_time) {
    while (0 < r.count && hi + 1 < ats[r.base + r.count - 1])
      r.count--;
    while (0 < r.leapcount && hi + 1 < leap[r.leapbase + r.leapcount - 1].trans)
      r.leapcount--;
  }

  /* Determine whether to append an expiration to the leap second table.  */
  r.leapexpiry = 0 <= leapexpires && leapexpires - 1 <= hi;

  return r;
}

static void
writezone(const char *const name, const char *const string, char version,
	  int defaulttype)
{
	register FILE *			fp;
	register ptrdiff_t		i, j;
	register int			pass;
	char *tempname = NULL;
	char const *outname = name;

	/* Allocate the ATS and TYPES arrays via a single malloc,
	   as this is a bit faster.  Do not malloc(0) if !timecnt,
	   as that might return NULL even on success.  */
	zic_t *ats = xmalloc(align_to(size_product(timecnt + !timecnt,
						   sizeof *ats + 1),
				      alignof(zic_t)));
	void *typesptr = ats + timecnt;
	unsigned char *types = typesptr;
	struct timerange rangeall = {0}, range32, range64;

	/*
	** Sort.
	*/
	if (timecnt > 1)
		qsort(attypes, timecnt, sizeof *attypes, atcomp);
	/*
	** Optimize and skip unwanted transitions.
	*/
	{
		ptrdiff_t fromi, toi;

		toi = 0;
		fromi = 0;
		for ( ; fromi < timecnt; ++fromi) {
			if (toi != 0) {
			    /* Skip the previous transition if it is unwanted
			       because its local time is not earlier.
			       The UT offset additions can't overflow because
			       of how the times were calculated.  */
			    unsigned char type_2 =
			      toi == 1 ? 0 : attypes[toi - 2].type;
			    if ((attypes[fromi].at
				 + utoffs[attypes[toi - 1].type])
				<= attypes[toi - 1].at + utoffs[type_2]) {
				    if (attypes[fromi].type == type_2)
					toi--;
				    else
					attypes[toi - 1].type =
						attypes[fromi].type;
				    continue;
			    }
			}

			/* Use a transition if it is the first one,
			   or if it cannot be merged for other reasons,
			   or if it transitions to different timekeeping.  */
			if (toi == 0
			    || attypes[fromi].dontmerge
			    || (utoffs[attypes[toi - 1].type]
				!= utoffs[attypes[fromi].type])
			    || (isdsts[attypes[toi - 1].type]
				!= isdsts[attypes[fromi].type])
			    || (desigidx[attypes[toi - 1].type]
				!= desigidx[attypes[fromi].type]))
					attypes[toi++] = attypes[fromi];
		}
		timecnt = toi;
	}

	if (noise) {
	  if (1200 < timecnt) {
	    if (TZ_MAX_TIMES < timecnt)
		warning(N_("reference clients mishandle"
			   " more than %d transition times"),
			TZ_MAX_TIMES);
	    else
		warning(N_("pre-2014 clients may mishandle"
			   " more than 1200 transition times"));
	  }
	  if (TZ_MAX_LEAPS < leapcnt)
	    warning(N_("reference clients mishandle more than %d leap seconds"),
		    TZ_MAX_LEAPS);
	}
	/*
	** Transfer.
	*/
	for (i = 0; i < timecnt; ++i) {
		ats[i] = attypes[i].at;
		types[i] = attypes[i].type;
	}

	/*
	** Correct for leap seconds.
	*/
	for (i = 0; i < timecnt; ++i) {
		j = leapcnt;
		while (--j >= 0)
			if (leap[j].trans - leap[j].corr < ats[i]) {
				ats[i] = tadd(ats[i], leap[j].corr);
				break;
			}
	}

	rangeall.defaulttype = defaulttype;
	rangeall.count = timecnt;
	rangeall.leapcount = leapcnt;
	range64 = limitrange(rangeall, lo_time,
			     max(hi_time,
				 redundant_time - (ZIC_MIN < redundant_time)),
			     ats, types);
	range32 = limitrange(range64, ZIC32_MIN, ZIC32_MAX, ats, types);

	/* TZif version 4 is needed if a no-op transition is appended to
	   indicate the expiration of the leap second table, or if the first
	   leap second transition is not to a +1 or -1 correction.  */
	for (pass = 1; pass <= 2; pass++) {
	  struct timerange const *r = pass == 1 ? &range32 : &range64;
	  if (pass == 1 && !want_bloat())
	    continue;
	  if (r->leapexpiry) {
	    if (noise)
	      warning(N_("%s: pre-2021b clients may mishandle"
			 " leap second expiry"),
		      name);
	    version = '4';
	  }
	  if (0 < r->leapcount
	      && leap[r->leapbase].corr != 1 && leap[r->leapbase].corr != -1) {
	    if (noise)
	      warning(N_("%s: pre-2021b clients may mishandle"
			 " leap second table truncation"),
		      name);
	    version = '4';
	  }
	  if (version == '4')
	    break;
	}

	fp = open_outfile(&outname, &tempname);

	for (pass = 1; pass <= 2; ++pass) {
		register ptrdiff_t thistimei, thistimecnt, thistimelim;
		register ptrdiff_t thisleapi, thisleapcnt, thisleaplim;
		struct tzhead tzh;
		int pretranstype = -1, thisdefaulttype;
		bool locut, hicut, thisleapexpiry;
		zic_t lo, thismin, thismax;
		int old0;
		char		omittype[TZ_MAX_TYPES];
		int		typemap[TZ_MAX_TYPES];
		int		thistypecnt, stdcnt, utcnt;
		char		thischars[TZ_MAX_CHARS];
		int		thischarcnt;
		bool		toomanytimes;
		int		indmap[TZ_MAX_CHARS];

		if (pass == 1) {
			thisdefaulttype = range32.defaulttype;
			thistimei = range32.base;
			thistimecnt = range32.count;
			toomanytimes = thistimecnt >> 31 >> 1 != 0;
			thisleapi = range32.leapbase;
			thisleapcnt = range32.leapcount;
			thisleapexpiry = range32.leapexpiry;
			thismin = ZIC32_MIN;
			thismax = ZIC32_MAX;
		} else {
			thisdefaulttype = range64.defaulttype;
			thistimei = range64.base;
			thistimecnt = range64.count;
			toomanytimes = thistimecnt >> 31 >> 31 >> 2 != 0;
			thisleapi = range64.leapbase;
			thisleapcnt = range64.leapcount;
			thisleapexpiry = range64.leapexpiry;
			thismin = min_time;
			thismax = max_time;
		}
		if (toomanytimes)
		  error(N_("too many transition times"));

		locut = thismin < lo_time && lo_time <= thismax;
		hicut = thismin <= hi_time && hi_time < thismax;
		thistimelim = thistimei + thistimecnt;
		memset(omittype, true, typecnt);

		/* Determine whether to output a transition before the first
		   transition in range.  This is needed when the output is
		   truncated at the start, and is also useful when catering to
		   buggy 32-bit clients that do not use time type 0 for
		   timestamps before the first transition.  */
		if ((locut || (pass == 1 && thistimei))
		    && ! (thistimecnt && ats[thistimei] == lo_time)) {
		  pretranstype = thisdefaulttype;
		  omittype[pretranstype] = false;
		}

		/* Arguably the default time type in the 32-bit data
		   should be range32.defaulttype, which is suited for
		   timestamps just before ZIC32_MIN.  However, zic
		   traditionally used the time type of the indefinite
		   past instead.  Internet RFC 8532 says readers should
		   ignore 32-bit data, so this discrepancy matters only
		   to obsolete readers where the traditional type might
		   be more appropriate even if it's "wrong".  So, use
		   the historical zic value, unless -r specifies a low
		   cutoff that excludes some 32-bit timestamps.  */
		if (pass == 1 && lo_time <= thismin)
		  thisdefaulttype = range64.defaulttype;

		if (locut)
		  thisdefaulttype = unspecifiedtype;
		omittype[thisdefaulttype] = false;
		for (i = thistimei; i < thistimelim; i++)
		  omittype[types[i]] = false;
		if (hicut)
		  omittype[unspecifiedtype] = false;

		/* Reorder types to make THISDEFAULTTYPE type 0.
		   Use TYPEMAP to swap OLD0 and THISDEFAULTTYPE so that
		   THISDEFAULTTYPE appears as type 0 in the output instead
		   of OLD0.  TYPEMAP also omits unused types.  */
		old0 = strlen(omittype);

#ifndef LEAVE_SOME_PRE_2011_SYSTEMS_IN_THE_LURCH
		/*
		** For some pre-2011 systems: if the last-to-be-written
		** standard (or daylight) type has an offset different from the
		** most recently used offset,
		** append an (unused) copy of the most recently used type
		** (to help get global "altzone" and "timezone" variables
		** set correctly).
		*/
		if (want_bloat()) {
			register int	mrudst, mrustd, hidst, histd, type;

			hidst = histd = mrudst = mrustd = -1;
			if (0 <= pretranstype) {
			  if (isdsts[pretranstype])
			    mrudst = pretranstype;
			  else
			    mrustd = pretranstype;
			}
			for (i = thistimei; i < thistimelim; i++)
				if (isdsts[types[i]])
					mrudst = types[i];
				else	mrustd = types[i];
			for (i = old0; i < typecnt; i++) {
			  int h = (i == old0 ? thisdefaulttype
				   : i == thisdefaulttype ? old0 : i);
			  if (!omittype[h]) {
			    if (isdsts[h])
			      hidst = i;
			    else
			      histd = i;
			  }
			}
			if (hidst >= 0 && mrudst >= 0 && hidst != mrudst &&
				utoffs[hidst] != utoffs[mrudst]) {
					isdsts[mrudst] = -1;
					type = addtype(utoffs[mrudst],
						&chars[desigidx[mrudst]],
						true,
						ttisstds[mrudst],
						ttisuts[mrudst]);
					isdsts[mrudst] = 1;
					omittype[type] = false;
			}
			if (histd >= 0 && mrustd >= 0 && histd != mrustd &&
				utoffs[histd] != utoffs[mrustd]) {
					isdsts[mrustd] = -1;
					type = addtype(utoffs[mrustd],
						&chars[desigidx[mrustd]],
						false,
						ttisstds[mrustd],
						ttisuts[mrustd]);
					isdsts[mrustd] = 0;
					omittype[type] = false;
			}
		}
#endif /* !defined LEAVE_SOME_PRE_2011_SYSTEMS_IN_THE_LURCH */
		thistypecnt = 0;
		for (i = old0; i < typecnt; i++)
		  if (!omittype[i])
		    typemap[i == old0 ? thisdefaulttype
			    : i == thisdefaulttype ? old0 : i]
		      = thistypecnt++;

		thischarcnt = stdcnt = utcnt = 0;
		for (i = old0; i < typecnt; i++) {
			if (omittype[i])
				continue;
			if (ttisstds[i])
			  stdcnt = thistypecnt;
			if (ttisuts[i])
			  utcnt = thistypecnt;
			addabbr(thischars, &thischarcnt, &chars[desigidx[i]]);
		}

		/* Now that all abbrevs have been added to THISCHARS,
		   it is safe to set INDMAP without worrying about
		   whether the abbrevs might move later.  */
		for (i = 0; i < TZ_MAX_CHARS; i++)
		  indmap[i] = -1;
		for (i = old0; i < typecnt; i++)
		  if (!omittype[i] && indmap[desigidx[i]] < 0)
		    indmap[desigidx[i]] = addabbr(thischars, &thischarcnt,
						  &chars[desigidx[i]]);

		if (pass == 1 && !want_bloat()) {
		  hicut = thisleapexpiry = false;
		  pretranstype = -1;
		  thistimecnt = thisleapcnt = 0;
		  thistypecnt = thischarcnt = 1;
		}
#define DO(field)	fwrite(tzh.field, sizeof tzh.field, 1, fp)
		memset(&tzh, 0, sizeof tzh);
		memcpy(tzh.tzh_magic, TZ_MAGIC, sizeof tzh.tzh_magic);
		tzh.tzh_version[0] = version;
		convert(utcnt, tzh.tzh_ttisutcnt);
		convert(stdcnt, tzh.tzh_ttisstdcnt);
		convert(thisleapcnt + thisleapexpiry, tzh.tzh_leapcnt);
		convert((0 <= pretranstype) + thistimecnt + hicut,
			tzh.tzh_timecnt);
		convert(thistypecnt, tzh.tzh_typecnt);
		convert(thischarcnt, tzh.tzh_charcnt);
		DO(tzh_magic);
		DO(tzh_version);
		DO(tzh_reserved);
		DO(tzh_ttisutcnt);
		DO(tzh_ttisstdcnt);
		DO(tzh_leapcnt);
		DO(tzh_timecnt);
		DO(tzh_typecnt);
		DO(tzh_charcnt);
#undef DO
		if (pass == 1 && !want_bloat()) {
		  /* Output a minimal data block with just one time type.  */
		  puttzcode(0, fp);	/* utoff */
		  putc(0, fp);		/* dst */
		  putc(0, fp);		/* index of abbreviation */
		  putc(0, fp);		/* empty-string abbreviation */
		  continue;
		}

		if (pass == 2 && noise && 50 < thischarcnt)
		  warning(N_("%s: pre-2026 reference clients mishandle"
			     " more than 50 bytes of abbreviations"),
			  name);

		/* Output a LO_TIME transition if needed; see limitrange.
		   But do not go below the minimum representable value
		   for this pass.  */
		lo = pass == 1 && lo_time < ZIC32_MIN ? ZIC32_MIN : lo_time;

		if (0 <= pretranstype)
		  puttzcodepass(lo, fp, pass);
		for (i = thistimei; i < thistimelim; ++i) {
		  puttzcodepass(ats[i], fp, pass);
		}
		if (hicut)
		  puttzcodepass(hi_time + 1, fp, pass);
		if (0 <= pretranstype)
		  putc(typemap[pretranstype], fp);
		for (i = thistimei; i < thistimelim; i++)
		  putc(typemap[types[i]], fp);
		if (hicut)
		  putc(typemap[unspecifiedtype], fp);

		for (i = old0; i < typecnt; i++) {
		  int h = (i == old0 ? thisdefaulttype
			   : i == thisdefaulttype ? old0 : i);
		  if (!omittype[h]) {
		    puttzcode(utoffs[h], fp);
		    putc(isdsts[h], fp);
		    putc(indmap[desigidx[h]], fp);
		  }
		}
		if (thischarcnt != 0)
			fwrite(thischars, sizeof thischars[0],
				      thischarcnt, fp);
		thisleaplim = thisleapi + thisleapcnt;
		for (i = thisleapi; i < thisleaplim; ++i) {
			register zic_t	todo;

			if (leap[i].roll) {
				if (timecnt == 0 || leap[i].trans < ats[0]) {
					j = 0;
					while (isdsts[j])
						if (++j >= typecnt) {
							j = 0;
							break;
						}
				} else {
					j = 1;
					while (j < timecnt &&
						ats[j] <= leap[i].trans)
							++j;
					j = types[j - 1];
				}
				todo = tadd(leap[i].trans, -utoffs[j]);
			} else	todo = leap[i].trans;
			puttzcodepass(todo, fp, pass);
			puttzcode(leap[i].corr, fp);
		}
		if (thisleapexpiry) {
		  /* Append a no-op leap correction indicating when the leap
		     second table expires.  Although this does not conform to
		     Internet RFC 9636, most clients seem to accept this and
		     the plan is to amend the RFC to allow this in version 4
		     TZif files.  */
		  puttzcodepass(leapexpires, fp, pass);
		  puttzcode(thisleaplim ? leap[thisleaplim - 1].corr : 0, fp);
		}
		if (stdcnt != 0)
		  for (i = old0; i < typecnt; i++)
			if (!omittype[i])
				putc(ttisstds[i], fp);
		if (utcnt != 0)
		  for (i = old0; i < typecnt; i++)
			if (!omittype[i])
				putc(ttisuts[i], fp);
	}
	fprintf(fp, "\n%s\n", string);
	close_file(fp, directory, name, tempname);
	rename_dest(tempname, name);
	free(ats);
}

static char const *
abbroffset(char *buf, zic_t offset)
{
  zic_t offset_lim = 100L * SECSPERHOUR;
  if (! (-offset_lim < offset && offset < offset_lim)) {
    error(N_("%%z UT offset magnitude exceeds 99:59:59"));
    return "%z";
  } else {
    char sign = offset < 0 ? '-' : '+';
    int_fast32_t abs_offset = offset < 0 ? -offset : offset;
    int seconds = abs_offset % SECSPERMIN;
    int abs_minutes_offset = abs_offset / SECSPERMIN;
    int minutes = abs_minutes_offset % MINSPERHOUR;
    int hours = abs_minutes_offset / MINSPERHOUR;
    char *p = buf;
    *p++ = sign;
    *p++ = '0' + hours / 10;
    *p++ = '0' + hours % 10;
    if (minutes | seconds) {
      *p++ = '0' + minutes / 10;
      *p++ = '0' + minutes % 10;
      if (seconds) {
	*p++ = '0' + seconds / 10;
	*p++ = '0' + seconds % 10;
      }
    }
    *p = '\0';
    return buf;
  }
}

static char const disable_percent_s[] = "";

static ptrdiff_t
doabbr(char *abbr, struct zone const *zp, char const *letters,
       bool isdst, zic_t save, bool doquotes)
{
	register char *	cp;
	ptrdiff_t len;
	char const *format = zp->z_format;
	char const *slashp = strchr(format, '/');

	if (slashp == NULL) {
	  char letterbuf[PERCENT_Z_LEN_BOUND + 1];
	  if (zp->z_format_specifier == 'z')
	    letters = abbroffset(letterbuf, zp->z_stdoff + save);
	  else if (!letters)
	    letters = "%s";
	  else if (letters == disable_percent_s)
	    return 0;
	  sprintf(abbr, format, letters);
	} else if (isdst)
	  strcpy(abbr, slashp + 1);
	else {
	  char *abbrend = mempcpy(abbr, format, slashp - format);
	  *abbrend = '\0';
	}
	len = strlen(abbr);
	if (!doquotes)
		return len;
	for (cp = abbr; is_alpha(*cp); cp++)
		continue;
	if (len > 0 && *cp == '\0')
		return len;
	abbr[len + 2] = '\0';
	abbr[len + 1] = '>';
	memmove(abbr + 1, abbr, len);
	abbr[0] = '<';
	return len + 2;
}

static void
updateminmax(const zic_t x)
{
	if (min_year > x)
		min_year = x;
	if (max_year < x)
		max_year = x;
}

static int
stringoffset(char *result, zic_t offset)
{
	register int	hours;
	register int	minutes;
	register int	seconds;
	bool negative = offset < 0;
	int len = negative;

	if (negative) {
		offset = -offset;
		result[0] = '-';
	}
	seconds = offset % SECSPERMIN;
	offset /= SECSPERMIN;
	minutes = offset % MINSPERHOUR;
	offset /= MINSPERHOUR;
	if (offset >= HOURSPERDAY * DAYSPERWEEK) {
		result[0] = '\0';
		return 0;
	}
	hours = offset;
	len += sprintf(result + len, "%d", hours);
	if (minutes != 0 || seconds != 0) {
		len += sprintf(result + len, ":%02d", minutes);
		if (seconds != 0)
			len += sprintf(result + len, ":%02d", seconds);
	}
	return len;
}

static int
stringrule(char *result, struct rule *const rp, zic_t save, zic_t stdoff)
{
	register zic_t	tod = rp->r_tod;
	register int	compat = 0;

	if (rp->r_dycode == DC_DOM) {
		register int	month, total;

		if (rp->r_dayofmonth == 29 && rp->r_month == TM_FEBRUARY)
			return -1;
		total = 0;
		for (month = 0; month < rp->r_month; ++month)
			total += len_months[0][month];
		/* Omit the "J" in Jan and Feb, as that's shorter.  */
		if (rp->r_month <= 1)
		  result += sprintf(result, "%d", total + rp->r_dayofmonth - 1);
		else
		  result += sprintf(result, "J%d", total + rp->r_dayofmonth);
	} else {
		register int	week;
		register int	wday = rp->r_wday;
		register int	wdayoff;

		if (rp->r_dycode == DC_DOWGEQ) {
			wdayoff = (rp->r_dayofmonth - 1) % DAYSPERWEEK;
			if (wdayoff)
				compat = 2013;
			wday -= wdayoff;
			tod += wdayoff * SECSPERDAY;
			week = 1 + (rp->r_dayofmonth - 1) / DAYSPERWEEK;
		} else if (rp->r_dycode == DC_DOWLEQ) {
			if (rp->r_dayofmonth == len_months[1][rp->r_month])
				week = 5;
			else {
				wdayoff = rp->r_dayofmonth % DAYSPERWEEK;
				if (wdayoff)
					compat = 2013;
				wday -= wdayoff;
				tod += wdayoff * SECSPERDAY;
				week = rp->r_dayofmonth / DAYSPERWEEK;
			}
		} else	return -1;	/* "cannot happen" */
		if (wday < 0)
			wday += DAYSPERWEEK;
		result += sprintf(result, "M%d.%d.%d",
				  rp->r_month + 1, week, wday);
	}
	if (rp->r_todisut)
	  tod += stdoff;
	if (rp->r_todisstd && !rp->r_isdst)
	  tod += save;
	if (tod != 2 * SECSPERMIN * MINSPERHOUR) {
		*result++ = '/';
		if (! stringoffset(result, tod))
			return -1;
		if (tod < 0) {
			if (compat < 2013)
				compat = 2013;
		} else if (SECSPERDAY <= tod) {
			if (compat < 1994)
				compat = 1994;
		}
	}
	return compat;
}

static int
rule_cmp(struct rule const *a, struct rule const *b)
{
	if (!a)
		return -!!b;
	if (!b)
		return 1;
	if (a->r_hiyear != b->r_hiyear)
		return a->r_hiyear < b->r_hiyear ? -1 : 1;
	if (a->r_hiyear == ZIC_MAX)
		return 0;
	if (a->r_month - b->r_month != 0)
		return a->r_month - b->r_month;
	return a->r_dayofmonth - b->r_dayofmonth;
}

/* Store into RESULT a proleptic TZ string that represent the future
   predictions for the zone ZPFIRST with ZONECOUNT entries.  Return a
   compatibility indicator (a TZDB release year) if successful, a
   negative integer if no such TZ string exists.  */
static int
stringzone(char *result, struct zone const *zpfirst, ptrdiff_t zonecount)
{
	register const struct zone *	zp;
	register struct rule *		rp;
	register struct rule *		stdrp;
	register struct rule *		dstrp;
	register ptrdiff_t		i;
	register int			compat = 0;
	register int			c;
	int				offsetlen;
	struct rule			stdr, dstr;
	ptrdiff_t len;
	int dstcmp;
	struct rule *lastrp[2] = { NULL, NULL };
	struct zone zstr[2];
	struct zone const *stdzp;
	struct zone const *dstzp;

	result[0] = '\0';

	/* Internet RFC 9636 section 6.1 says to use an empty TZ string if
	   future timestamps are truncated.  */
	if (hi_time < max_time)
	  return -1;

	zp = zpfirst + zonecount - 1;
	for (i = 0; i < zp->z_nrules; ++i) {
		struct rule **last;
		int cmp;
		rp = &zp->z_rules[i];
		last = &lastrp[rp->r_isdst];
		cmp = rule_cmp(*last, rp);
		if (cmp < 0)
		  *last = rp;
		else if (cmp == 0)
		  return -1;
	}
	stdrp = lastrp[false];
	dstrp = lastrp[true];
	dstcmp = zp->z_nrules ? rule_cmp(dstrp, stdrp) : zp->z_isdst ? 1 : -1;
	stdzp = dstzp = zp;

	if (dstcmp < 0) {
	  /* Standard time all year.  */
	  dstrp = NULL;
	} else if (0 < dstcmp) {
	  /* DST all year.  Use an abbreviation like
	     "XXX3EDT4,0/0,J365/23" for EDT (-04) all year.  */
	  zic_t save = dstrp ? dstrp->r_save : zp->z_save;
	  if (0 <= save)
	    {
	      /* Positive DST, the typical case for all-year DST.
		 Fake a timezone with negative DST.  */
	      stdzp = &zstr[0];
	      dstzp = &zstr[1];
	      zstr[0].z_stdoff = zp->z_stdoff + 2 * save;
	      zstr[0].z_format = "XXX";  /* Any 3 letters will do.  */
	      zstr[0].z_format_specifier = 0;
	      zstr[1].z_stdoff = zstr[0].z_stdoff;
	      zstr[1].z_format = zp->z_format;
	      zstr[1].z_format_specifier = zp->z_format_specifier;
	    }
	  dstr.r_month = TM_JANUARY;
	  dstr.r_dycode = DC_DOM;
	  dstr.r_dayofmonth = 1;
	  dstr.r_tod = 0;
	  dstr.r_todisstd = dstr.r_todisut = false;
	  dstr.r_isdst = true;
	  dstr.r_save = save < 0 ? save : -save;
	  dstr.r_abbrvar = dstrp ? dstrp->r_abbrvar : NULL;
	  stdr.r_month = TM_DECEMBER;
	  stdr.r_dycode = DC_DOM;
	  stdr.r_dayofmonth = 31;
	  stdr.r_tod = SECSPERDAY + dstr.r_save;
	  stdr.r_todisstd = stdr.r_todisut = false;
	  stdr.r_isdst = false;
	  stdr.r_save = 0;
	  stdr.r_abbrvar = save < 0 && stdrp ? stdrp->r_abbrvar : NULL;
	  dstrp = &dstr;
	  stdrp = &stdr;
	}
	len = doabbr(result, stdzp, stdrp ? stdrp->r_abbrvar : NULL,
		     false, 0, true);
	offsetlen = stringoffset(result + len, - stdzp->z_stdoff);
	if (! offsetlen) {
		result[0] = '\0';
		return -1;
	}
	len += offsetlen;
	if (dstrp == NULL)
		return compat;
	len += doabbr(result + len, dstzp, dstrp->r_abbrvar,
		      dstrp->r_isdst, dstrp->r_save, true);
	if (dstrp->r_save != SECSPERMIN * MINSPERHOUR) {
	  offsetlen = stringoffset(result + len,
				   - (dstzp->z_stdoff + dstrp->r_save));
	  if (! offsetlen) {
	    result[0] = '\0';
	    return -1;
	  }
	  len += offsetlen;
	}
	result[len++] = ',';
	c = stringrule(result + len, dstrp, dstrp->r_save, stdzp->z_stdoff);
	if (c < 0) {
		result[0] = '\0';
		return -1;
	}
	if (compat < c)
		compat = c;
	len += strlen(result + len);
	result[len++] = ',';
	c = stringrule(result + len, stdrp, dstrp->r_save, stdzp->z_stdoff);
	if (c < 0) {
		result[0] = '\0';
		return -1;
	}
	if (compat < c)
		compat = c;
	return compat;
}

static void
outzone(const struct zone *zpfirst, ptrdiff_t zonecount)
{
	register ptrdiff_t		i, j;
	register zic_t			starttime, untiltime;
	register bool			startttisstd;
	register bool			startttisut;
	register char *			startbuf;
	register char *			ab;
	register char *			envvar;
	register int			max_abbr_len;
	register int			max_envvar_len;
	register int			compat;
	register bool			do_extend;
	register char			version;
	zic_t nonTZlimtime = ZIC_MIN;
	int nonTZlimtype = -1;
	zic_t max_year0;
	int defaulttype = -1;
	int max_stringoffset_len = sizeof "-167:59:59" - 1;
	int max_comma_stringrule_len = (sizeof ",M12.5.6/" - 1
					+ max_stringoffset_len);

	check_for_signal();

	/* This cannot overflow; see FORMAT_LEN_GROWTH_BOUND.  */
	max_abbr_len = 2 + max_format_len + max_abbrvar_len;
	max_envvar_len = 2 * (max_abbr_len + max_stringoffset_len
			      + max_comma_stringrule_len);

	startbuf = xmalloc(max_abbr_len + 1);
	ab = xmalloc(max_abbr_len + 1);
	envvar = xmalloc(max_envvar_len + 1);
	INITIALIZE(untiltime);
	INITIALIZE(starttime);
	/*
	** Now. . .finally. . .generate some useful data!
	*/
	timecnt = 0;
	typecnt = 0;
	charcnt = 0;
	/*
	** Thanks to Earl Chew
	** for noting the need to unconditionally initialize startttisstd.
	*/
	startttisstd = false;
	startttisut = false;
	min_year = max_year = EPOCH_YEAR;
	if (leapseen) {
		updateminmax(leapminyear);
		updateminmax(leapmaxyear + (leapmaxyear < ZIC_MAX));
	}
	for (i = 0; i < zonecount; ++i) {
		struct zone const *zp = &zpfirst[i];
		if (i < zonecount - 1)
			updateminmax(zp->z_untilrule.r_loyear);
		for (j = 0; j < zp->z_nrules; ++j) {
			struct rule *rp = &zp->z_rules[j];
			updateminmax(rp->r_loyear);
			if (rp->r_hiwasnum)
				updateminmax(rp->r_hiyear);
		}
	}
	/*
	** Generate lots of data if a rule can't cover all future times.
	*/
	compat = stringzone(envvar, zpfirst, zonecount);
	version = compat < 2013 ? '2' : '3';
	do_extend = compat < 0;
	if (noise) {
		if (!*envvar)
			warning(N_("no proleptic TZ string for zone %s"),
				zpfirst->z_name);
		else if (compat != 0) {
			/* Circa-COMPAT clients, and earlier clients, might
			   not work for this zone when given dates before
			   1970 or after 2038.  */
			warning(N_("%s: pre-%d clients may mishandle"
				   " distant timestamps"),
				zpfirst->z_name, compat);
		}
	}
	if (do_extend) {
		if (min_year >= ZIC_MIN + years_of_observations)
			min_year -= years_of_observations;
		else	min_year = ZIC_MIN;
		if (max_year <= ZIC_MAX - years_of_observations)
			max_year += years_of_observations;
		else	max_year = ZIC_MAX;
	}
	max_year = max(max_year, (redundant_time / (SECSPERDAY * DAYSPERNYEAR)
				  + EPOCH_YEAR + 1));
	max_year0 = max_year;
	if (want_bloat()) {
	  /* For the benefit of older systems,
	     generate data from 1900 through 2038.  */
	  if (min_year > YEAR_32BIT_MIN - 1)
		min_year = YEAR_32BIT_MIN - 1;
	  if (max_year < YEAR_32BIT_MAX)
		max_year = YEAR_32BIT_MAX;
	}

	if (min_time < lo_time || hi_time < max_time)
	  unspecifiedtype = addtype(0, "-00", false, false, false);

	for (i = 0; i < zonecount; ++i) {
		/*
		** A guess that may well be corrected later.
		*/
		zic_t save = 0;
		struct zone const *zp = &zpfirst[i];
		bool usestart = i > 0 && (zp - 1)->z_untiltime > min_time;
		bool useuntil = i < (zonecount - 1);
		zic_t stdoff = zp->z_stdoff;
		zic_t startoff = stdoff;
		if (useuntil && zp->z_untiltime <= min_time)
			continue;
		eat(zp->z_filenum, zp->z_linenum);
		*startbuf = '\0';
		if (zp->z_nrules == 0) {
			int type;
			save = zp->z_save;
			doabbr(startbuf, zp, NULL, zp->z_isdst, save, false);
			type = addtype(zp->z_stdoff + save,
				startbuf, zp->z_isdst, startttisstd,
				startttisut);
			if (usestart) {
				addtt(starttime, type);
				if (nonTZlimtime < starttime) {
				  nonTZlimtime = starttime;
				  nonTZlimtype = type;
				}
				usestart = false;
			} else
				defaulttype = type;
		} else {
		  zic_t year, next_year;
		  bool try_next_year = true;
		  for (year = min_year; try_next_year; year = next_year) {
			if (useuntil && year > zp->z_untilrule.r_hiyear)
				break;
			try_next_year = false;
			next_year = max_year;
			/*
			** Mark which rules to do in the current year.
			** For those to do, calculate rpytime(rp, year);
			** The former TYPE field was also considered here.
			*/
			for (j = 0; j < zp->z_nrules; ++j) {
				zic_t one = 1;
				zic_t y2038_boundary = one << 31;
				struct rule *rp = &zp->z_rules[j];
				eats(zp->z_filenum, zp->z_linenum,
				     rp->r_filenum, rp->r_linenum);

				/* Keep track of the earliest year after YEAR
				   in which some rule <= J applies.  */
				if (year < rp->r_hiyear) {
				  zic_t next = max(year + 1, rp->r_loyear);
				  if (next <= next_year) {
				    next_year = next;
				    try_next_year = true;
				  }
				}

				rp->r_todo = year >= rp->r_loyear &&
						year <= rp->r_hiyear;
				if (rp->r_todo) {
					rp->r_temp = rpytime(rp, year);
					rp->r_todo
					  = (rp->r_temp < y2038_boundary
					     || year <= max_year0);
				}
			}
			for ( ; ; ) {
				register ptrdiff_t k;
				register zic_t	jtime, ktime;
				register zic_t	offset;
				struct rule *rp;
				int type;

				INITIALIZE(ktime);
				if (useuntil) {
					/*
					** Turn untiltime into UT
					** assuming the current stdoff and
					** save values.
					*/
					untiltime = zp->z_untiltime;
					if (!zp->z_untilrule.r_todisut)
						untiltime = tadd(untiltime,
								 -stdoff);
					if (!zp->z_untilrule.r_todisstd)
						untiltime = tadd(untiltime,
								 -save);
				}
				/*
				** Find the rule (of those to do, if any)
				** that takes effect earliest in the year.
				*/
				k = -1;
				for (j = 0; j < zp->z_nrules; ++j) {
					struct rule *r = &zp->z_rules[j];
					if (!r->r_todo)
						continue;
					eats(zp->z_filenum, zp->z_linenum,
					     r->r_filenum, r->r_linenum);
					offset = r->r_todisut ? 0 : stdoff;
					if (!r->r_todisstd)
						offset += save;
					jtime = r->r_temp;
					jtime = tadd(jtime, -offset);
					if (k < 0 || jtime < ktime) {
						k = j;
						ktime = jtime;
					} else if (jtime == ktime) {
					  static char const dup_rules_msgid[] =
					    N_("two rules for same instant");
					  eats(zp->z_filenum, zp->z_linenum,
					       r->r_filenum, r->r_linenum);
					  warning(dup_rules_msgid);
					  r = &zp->z_rules[k];
					  eats(zp->z_filenum, zp->z_linenum,
					       r->r_filenum, r->r_linenum);
					  error(dup_rules_msgid);
					}
				}
				if (k < 0)
					break;	/* go on to next year */
				rp = &zp->z_rules[k];
				rp->r_todo = false;
				if (useuntil && ktime >= untiltime) {
					if (!*startbuf
					    && (zp->z_stdoff + rp->r_save
						== startoff))
					  doabbr(startbuf, zp, rp->r_abbrvar,
						 rp->r_isdst, rp->r_save,
						 false);
					break;
				}
				save = rp->r_save;
				if (usestart && ktime == starttime)
					usestart = false;
				if (usestart) {
					if (ktime < starttime) {
						startoff = zp->z_stdoff + save;
						doabbr(startbuf, zp,
							rp->r_abbrvar,
							rp->r_isdst,
							rp->r_save,
							false);
						continue;
					}
					if (*startbuf == '\0'
					    && startoff == (zp->z_stdoff
							    + save)) {
							doabbr(startbuf,
								zp,
								rp->r_abbrvar,
								rp->r_isdst,
								rp->r_save,
								false);
					}
				}
				eats(zp->z_filenum, zp->z_linenum,
				     rp->r_filenum, rp->r_linenum);
				doabbr(ab, zp, rp->r_abbrvar,
				       rp->r_isdst, rp->r_save, false);
				offset = zp->z_stdoff + rp->r_save;
				type = addtype(offset, ab, rp->r_isdst,
					rp->r_todisstd, rp->r_todisut);
				if (defaulttype < 0 && !rp->r_isdst)
				  defaulttype = type;
				addtt(ktime, type);
				if (nonTZlimtime < ktime
				    && (useuntil || rp->r_hiyear != ZIC_MAX)) {
				  nonTZlimtime = ktime;
				  nonTZlimtype = type;
				}
			}
		  }
		}
		if (usestart) {
			bool isdst = startoff != zp->z_stdoff;
			if (*startbuf == '\0' && zp->z_format)
			  doabbr(startbuf, zp, disable_percent_s,
				 isdst, save, false);
			eat(zp->z_filenum, zp->z_linenum);
			if (*startbuf == '\0')
			  error(N_("can't determine time zone abbreviation"
				  " to use just after until time"));
			else {
			  int type = addtype(startoff, startbuf, isdst,
					     startttisstd, startttisut);
			  if (defaulttype < 0 && !isdst)
			    defaulttype = type;
			  addtt(starttime, type);
			}
		}
		/*
		** Now we may get to set starttime for the next zone line.
		*/
		if (useuntil) {
			startttisstd = zp->z_untilrule.r_todisstd;
			startttisut = zp->z_untilrule.r_todisut;
			starttime = zp->z_untiltime;
			if (!startttisstd)
			  starttime = tadd(starttime, -save);
			if (!startttisut)
			  starttime = tadd(starttime, -stdoff);
		}
	}
	if (defaulttype < 0)
	  defaulttype = 0;
	if (!do_extend && !want_bloat()) {
	  /* Keep trailing transitions that are no greater than this.  */
	  zic_t keep_at_max;

	  /* The earliest transition into a time governed by the TZ string.  */
	  zic_t TZstarttime = ZIC_MAX;
	  for (i = 0; i < timecnt; i++) {
	    zic_t at = attypes[i].at;
	    if (nonTZlimtime < at && at < TZstarttime)
	      TZstarttime = at;
	  }
	  if (TZstarttime == ZIC_MAX)
	    TZstarttime = nonTZlimtime;

	  /* Omit trailing transitions deducible from the TZ string,
	     and not needed for -r or -R.  */
	  keep_at_max = max(TZstarttime, redundant_time);
	  for (i = j = 0; i < timecnt; i++)
	    if (attypes[i].at <= keep_at_max) {
	      attypes[j].at = attypes[i].at;
	      attypes[j].dontmerge = (attypes[i].at == TZstarttime
				      && (nonTZlimtype != attypes[i].type
					  || strchr(envvar, ',')));
	      attypes[j].type = attypes[i].type;
	      j++;
	    }
	  timecnt = j;
	}
	if (do_extend) {
		/*
		** If we're extending the explicitly listed observations for
		** 400 years because we can't fill the proleptic TZ field,
		** check whether we actually ended up explicitly listing
		** observations through that period.  If there aren't any
		** near the end of the 400-year period, add a redundant
		** one at the end of the final year, to make it clear
		** that we are claiming to have definite knowledge of
		** the lack of transitions up to that point.
		*/
		struct rule xr;
		struct attype *lastat;
		xr.r_month = TM_JANUARY;
		xr.r_dycode = DC_DOM;
		xr.r_dayofmonth = 1;
		xr.r_tod = 0;
		for (lastat = attypes, i = 1; i < timecnt; i++)
			if (attypes[i].at > lastat->at)
				lastat = &attypes[i];
		if (!lastat || lastat->at < rpytime(&xr, max_year - 1)) {
			addtt(rpytime(&xr, max_year + 1),
			      lastat ? lastat->type : defaulttype);
			attypes[timecnt - 1].dontmerge = true;
		}
	}
	writezone(zpfirst->z_name, envvar, version, defaulttype);
	free(startbuf);
	free(ab);
	free(envvar);
}

static void
addtt(zic_t starttime, int type)
{
	attypes = growalloc(attypes, sizeof *attypes, timecnt, &timecnt_alloc);
	attypes[timecnt].at = starttime;
	attypes[timecnt].dontmerge = false;
	attypes[timecnt].type = type;
	++timecnt;
}

static int
addtype(zic_t utoff, char const *abbr, bool isdst, bool ttisstd, bool ttisut)
{
	register int	i, j;
	int charcnt0;

	/* RFC 9636 section 3.2 specifies this range for utoff.  */
	if (! (-TWO_31_MINUS_1 <= utoff && utoff <= TWO_31_MINUS_1)) {
		error(N_("UT offset out of range"));
		exit(EXIT_FAILURE);
	}
	if (!want_bloat())
	  ttisstd = ttisut = false;

	checkabbr(abbr);

	charcnt0 = charcnt;
	j = addabbr(chars, &charcnt, abbr);
	if (charcnt0 < charcnt) {
	  /* If an abbreviation was inserted, increment indexes no
	     earlier than the insert by the size of the insertion,
	     so that they continue to point to the same contents.  */
	  for (i = 0; i < typecnt; i++)
	    if (j <= desigidx[i])
	      desigidx[i] += charcnt - charcnt0;
	} else {
	  /* If there's already an entry, return its index.  */
	  for (i = 0; i < typecnt; i++)
	    if (utoff == utoffs[i] && isdst == isdsts[i] && j == desigidx[i]
		&& ttisstd == ttisstds[i] && ttisut == ttisuts[i])
	      return i;
	}
	/*
	** There isn't one; add a new one, unless there are already too
	** many.
	*/
	if (typecnt >= TZ_MAX_TYPES) {
		error(N_("too many local time types"));
		exit(EXIT_FAILURE);
	}
	i = typecnt++;
	utoffs[i] = utoff;
	isdsts[i] = isdst;
	ttisstds[i] = ttisstd;
	ttisuts[i] = ttisut;
	desigidx[i] = j;
	return i;
}

static void
leapadd(zic_t t, int correction, int rolling)
{
	register ptrdiff_t i;

	if (rolling && (lo_time != min_time || hi_time != max_time)) {
	  error(N_("Rolling leap seconds not supported with -r"));
	  exit(EXIT_FAILURE);
	}
	leap = growalloc(leap, sizeof *leap, leapcnt, &leap_alloc);
	for (i = 0; i < leapcnt; ++i)
		if (t <= leap[i].trans)
			break;
	memmove(&leap[i + 1], &leap[i], (leapcnt - i) * sizeof *leap);
	leap[i].trans = t;
	leap[i].corr = correction;
	leap[i].roll = rolling;
	++leapcnt;
}

static void
adjleap(void)
{
	register ptrdiff_t i;
	register zic_t	last = 0;
	register zic_t	prevtrans = 0;

	/*
	** propagate leap seconds forward
	*/
	for (i = 0; i < leapcnt; ++i) {
		if (leap[i].trans - prevtrans < 28 * SECSPERDAY) {
		  error(N_("Leap seconds too close together"));
		  exit(EXIT_FAILURE);
		}
		prevtrans = leap[i].trans;
		leap[i].trans = tadd(prevtrans, last);
		last = leap[i].corr += last;
	}

	if (0 <= leapexpires) {
	  leapexpires = oadd(leapexpires, last);
	  if (! (leapcnt == 0 || (leap[leapcnt - 1].trans < leapexpires))) {
	    error(N_("last Leap time does not precede Expires time"));
	    exit(EXIT_FAILURE);
	  }
	}
}

/* Is A a space character in the C locale?  */
static bool
is_space(char a)
{
	switch (a) {
	  default:
		return false;
	  case ' ': case '\f': case '\n': case '\r': case '\t': case '\v':
		return true;
	}
}

/* Is A an alphabetic character in the C locale?  */
static bool
is_alpha(char a)
{
	switch (a) {
	  default:
		return false;
	  case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
	  case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
	  case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
	  case 'V': case 'W': case 'X': case 'Y': case 'Z':
	  case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
	  case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
	  case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
	  case 'v': case 'w': case 'x': case 'y': case 'z':
		return true;
	}
}

/* If A is an uppercase character in the C locale, return its lowercase
   counterpart.  Otherwise, return A.  */
static char
lowerit(char a)
{
	switch (a) {
	  default: return a;
	  case 'A': return 'a'; case 'B': return 'b'; case 'C': return 'c';
	  case 'D': return 'd'; case 'E': return 'e'; case 'F': return 'f';
	  case 'G': return 'g'; case 'H': return 'h'; case 'I': return 'i';
	  case 'J': return 'j'; case 'K': return 'k'; case 'L': return 'l';
	  case 'M': return 'm'; case 'N': return 'n'; case 'O': return 'o';
	  case 'P': return 'p'; case 'Q': return 'q'; case 'R': return 'r';
	  case 'S': return 's'; case 'T': return 't'; case 'U': return 'u';
	  case 'V': return 'v'; case 'W': return 'w'; case 'X': return 'x';
	  case 'Y': return 'y'; case 'Z': return 'z';
	}
}

/* case-insensitive equality */
ATTRIBUTE_PURE_114833
static bool
ciequal(register const char *ap, register const char *bp)
{
	while (lowerit(*ap) == lowerit(*bp++))
		if (*ap++ == '\0')
			return true;
	return false;
}

ATTRIBUTE_PURE_114833
static bool
itsabbr(register const char *abbr, register const char *word)
{
	if (lowerit(*abbr) != lowerit(*word))
		return false;
	++word;
	while (*++abbr != '\0')
		do {
			if (*word == '\0')
				return false;
		} while (lowerit(*word++) != lowerit(*abbr));
	return true;
}

/* Return true if ABBR is an initial prefix of WORD, ignoring ASCII case.  */

ATTRIBUTE_PURE_114833
static bool
ciprefix(char const *abbr, char const *word)
{
  do
    if (!*abbr)
      return true;
  while (lowerit(*abbr++) == lowerit(*word++));

  return false;
}

static const struct lookup *
byword(const char *word, const struct lookup *table)
{
	register const struct lookup *	foundlp;
	register const struct lookup *	lp;

	if (word == NULL || table == NULL)
		return NULL;

	/* If TABLE is LASTS and the word starts with "last" followed
	   by a non-'-', skip the "last" and look in WDAY_NAMES instead.
	   Warn about any usage of the undocumented prefix "last-".  */
	if (table == lasts && ciprefix("last", word) && word[4]) {
	  if (word[4] == '-')
	    warning(N_("\"%s\" is undocumented; use \"last%s\" instead"),
		    word, word + 5);
	  else {
	    word += 4;
	    table = wday_names;
	  }
	}

	/*
	** Look for exact match.
	*/
	for (lp = table; lp->l_word != NULL; ++lp)
		if (ciequal(word, lp->l_word))
			return lp;
	/*
	** Look for inexact match.
	*/
	foundlp = NULL;
	for (lp = table; lp->l_word != NULL; ++lp)
		if (ciprefix(word, lp->l_word)) {
			if (foundlp == NULL)
				foundlp = lp;
			else	return NULL;	/* multiple inexact matches */
		}

	if (foundlp && noise) {
	  /* Warn about any backward-compatibility issue with pre-2017c zic.  */
	  bool pre_2017c_match = false;
	  for (lp = table; lp->l_word; lp++)
	    if (itsabbr(word, lp->l_word)) {
	      if (pre_2017c_match) {
		warning(N_("\"%s\" is ambiguous in pre-2017c zic"), word);
		break;
	      }
	      pre_2017c_match = true;
	    }
	}

	return foundlp;
}

static int
getfields(char *cp, char **array, int arrayelts)
{
	register char *		dp;
	register int		nsubs;

	nsubs = 0;
	for ( ; ; ) {
		char *dstart;
		while (is_space(*cp))
				++cp;
		if (*cp == '\0' || *cp == '#')
			break;
		dstart = dp = cp;
		do {
			if ((*dp = *cp++) != '"')
				++dp;
			else while ((*dp = *cp++) != '"')
				if (*dp != '\0')
					++dp;
				else {
				  error(N_("Odd number of quotation marks"));
				  exit(EXIT_FAILURE);
				}
		} while (*cp && *cp != '#' && !is_space(*cp));
		if (is_space(*cp))
			++cp;
		*dp = '\0';
		if (nsubs == arrayelts) {
		  error(N_("Too many input fields"));
		  exit(EXIT_FAILURE);
		}
		array[nsubs++] = dstart + (*dstart == '-' && dp == dstart + 1);
	}
	return nsubs;
}

ATTRIBUTE_NORETURN static void
time_overflow(void)
{
  error(N_("time overflow"));
  exit(EXIT_FAILURE);
}

/* Return T1 + T2, but diagnose any overflow and exit.  */
ATTRIBUTE_PURE_114833_HACK
static zic_t
oadd(zic_t t1, zic_t t2)
{
#ifdef ckd_add
  zic_t sum;
  if (!ckd_add(&sum, t1, t2))
    return sum;
#else
  if (t1 < 0 ? ZIC_MIN - t1 <= t2 : t2 <= ZIC_MAX - t1)
    return t1 + t2;
#endif
  time_overflow();
}

/* Return T1 + T2, but diagnose any overflow and exit.
   This is like oadd, except the result must fit in min_time..max_time range,
   which on oddball machines can be a smaller range than ZIC_MIN..ZIC_MAX.  */
ATTRIBUTE_PURE_114833_HACK
static zic_t
tadd(zic_t t1, zic_t t2)
{
  zic_t sum = oadd(t1, t2);
  if (min_time <= sum && sum <= max_time)
    return sum;
  time_overflow();
}

/* Return T1 * T2, but diagnose any overflow and exit.  */
ATTRIBUTE_PURE_114833_HACK
static zic_t
omul(zic_t t1, zic_t t2)
{
#ifdef ckd_mul
  zic_t product;
  if (!ckd_mul(&product, t1, t2))
    return product;
#else
  if (t2 < 0
      ? ZIC_MAX / t2 <= t1 && (t2 == -1 || t1 <= ZIC_MIN / t2)
      : t2 == 0 || (ZIC_MIN / t2 <= t1 && t1 <= ZIC_MAX / t2))
    return t1 * t2;
#endif
  time_overflow();
}

/*
** Given a rule, and a year, compute the date (in seconds since January 1,
** 1970, 00:00 LOCAL time) in that year that the rule refers to.
** Do not count leap seconds.  On error, diagnose and exit.
*/

static zic_t
rpytime(const struct rule *rp, zic_t wantedy)
{
	register int	m, i;
	register zic_t	dayoff;			/* with a nod to Margaret O. */
	register zic_t	t, y;
	int yrem;

	m = TM_JANUARY;
	y = EPOCH_YEAR;

	/* dayoff = floor((wantedy - y) / YEARSPERREPEAT) * DAYSPERREPEAT,
	   checking for overflow.  */
	yrem = wantedy % YEARSPERREPEAT - y % YEARSPERREPEAT;
	dayoff = omul ((wantedy / YEARSPERREPEAT - y / YEARSPERREPEAT
			+ yrem / YEARSPERREPEAT - (yrem % YEARSPERREPEAT < 0)),
		       DAYSPERREPEAT);
	/* wantedy = y + ((wantedy - y) mod YEARSPERREPEAT), sans overflow.  */
	wantedy = y + (yrem + 2 * YEARSPERREPEAT) % YEARSPERREPEAT;

	while (wantedy != y) {
		i = year_days(y);
		dayoff = oadd(dayoff, i);
		y++;
	}
	while (m != rp->r_month) {
		i = len_months[isleap(y)][m];
		dayoff = oadd(dayoff, i);
		++m;
	}
	i = rp->r_dayofmonth;
	if (m == TM_FEBRUARY && i == 29 && !isleap(y)) {
		if (rp->r_dycode == DC_DOWLEQ)
			--i;
		else {
			error(N_("use of 2/29 in non leap-year"));
			exit(EXIT_FAILURE);
		}
	}
	--i;
	dayoff = oadd(dayoff, i);
	if (rp->r_dycode == DC_DOWGEQ || rp->r_dycode == DC_DOWLEQ) {
		/*
		** Don't trust mod of negative numbers.
		*/
		zic_t wday = ((EPOCH_WDAY + dayoff % DAYSPERWEEK + DAYSPERWEEK)
			      % DAYSPERWEEK);
		while (wday != rp->r_wday)
			if (rp->r_dycode == DC_DOWGEQ) {
				dayoff = oadd(dayoff, 1);
				if (++wday >= DAYSPERWEEK)
					wday = 0;
				++i;
			} else {
				dayoff = oadd(dayoff, -1);
				if (--wday < 0)
					wday = DAYSPERWEEK - 1;
				--i;
			}
		if (i < 0 || i >= len_months[isleap(y)][m]) {
			if (noise)
			  warning(N_("rule goes past start/end of month;"
				     " will not work with pre-2004"
				     " versions of zic"));
		}
	}
	t = omul(dayoff, SECSPERDAY);
	return tadd(t, rp->r_tod);
}

static void
checkabbr(char const *string)
{
	if (strcmp(string, GRANDPARENTED) != 0) {
		register const char *	cp;
		const char *		mp;

		cp = string;
		mp = NULL;
		while (is_alpha(*cp) || is_digit(*cp)
		       || *cp == '-' || *cp == '+')
				++cp;
		if (noise && cp - string < 3)
		  mp = N_("time zone abbreviation has fewer than 3 characters");
		if (cp - string > ZIC_MAX_ABBR_LEN_WO_WARN)
		  mp = N_("time zone abbreviation has too many characters");
		if (*cp != '\0')
		  mp = N_("time zone abbreviation differs from POSIX standard");
		if (mp != NULL)
		  warning(N_("%s (%s)"), _(mp), string);
	}
}

/* Put into CHS, which currently contains *PNCHS bytes containing
   NUL-terminated abbreviations none of which are suffixes of another,
   the abbreviation ABBR including its trailing NUL.
   If ABBR does not already appear in CHS,
   possibly as a suffix of an existing abbreviation,
   add ABBR to CHS, remove from CHS any abbreviation
   that is a suffix of ABBR, and increment *PNCHS accordingly.
   Return the index of ABBR after any modifications to CHS are made.

   If all abbreviations have already been added, this function
   lets the caller look up the index of an existing abbreviation.  */
static int
addabbr(char chs[TZ_MAX_CHARS], int *pnchs, char const *abbr)
{
  int nchs = *pnchs;
  int alen = strlen(abbr), nchs_incr = alen + 1;
  int i;
  for (i = 0; i < nchs; ) {
    int clen = strlen(&chs[i]);
    if (alen <= clen) {
      /* If ABBR is a suffix of an abbreviation in CHS,
	 return the index of ABBR in CHS.  */
      int isuff = i + (clen - alen);
      if (memcmp(&chs[isuff], abbr, alen) == 0)
	return isuff;
    } else if (memcmp(&chs[i], &abbr[alen - clen], clen) == 0) {
      /* An abbreviation in CHS is a substring of ABBR.
	 Replace it with ABBR, instead of the more-common
	 actions of appending ABBR or doing nothing.  */
      nchs_incr = alen - clen;
      break;
    }
    i += clen + 1;
  }
  if (TZ_MAX_CHARS < nchs + nchs_incr) {
    error(N_("too many, or too long, time zone abbreviations"));
    exit(EXIT_FAILURE);
  }
  memmove(&chs[i + nchs_incr], &chs[i], nchs - i);
  memcpy(&chs[i], abbr, nchs_incr);
  *pnchs = nchs + nchs_incr;
  return i;
}

/* Ensure that the directories of ARGNAME exist, by making any missing
   ones.  If ANCESTORS, do this only for ARGNAME's ancestors; otherwise,
   do it for ARGNAME too.  Exit with failure if there is trouble.
   Do not consider an existing file to be trouble.  */
static void
mkdirs(char const *argname, bool ancestors)
{
    /* If -D was specified, do not create directories.
       If a file operation's parent directory is missing,
       the operation will fail and be diagnosed.  */
    if (!skip_mkdir) {

	char *name = xstrdup(argname);
	char *cp = name;

	/* On MS-Windows systems, do not worry about drive letters or
	   backslashes, as this should suffice in practice.  Time zone
	   names do not use drive letters and backslashes.  If the -d
	   option of zic does not name an already-existing directory,
	   it can use slashes to separate the already-existing
	   ancestor prefix from the to-be-created subdirectories.  */

	/* Do not mkdir a root directory, as it must exist.  */
	while (*cp == '/')
	  cp++;

	while (cp && ((cp = strchr(cp, '/')) || !ancestors)) {
		if (cp)
		  *cp = '\0';
		/*
		** Try to create it.  It's OK if creation fails because
		** the directory already exists, perhaps because some
		** other process just created it.  For simplicity do
		** not check first whether it already exists, as that
		** is checked anyway if the mkdir fails.
		*/
		if (mkdir(name, MKDIR_PERMS) < 0) {
			/* Do not report an error if err == EEXIST, because
			   some other process might have made the directory
			   in the meantime.  Likewise for ENOSYS, because
			   Solaris 10 mkdir fails with ENOSYS if the
			   directory is an automounted mount point.
			   Likewise for EACCES, since mkdir can fail
			   with EACCES merely because the parent directory
			   is unwritable.  Likewise for most other error
			   numbers.  */
			int err = errno;
			if (err == ELOOP || err == ENAMETOOLONG
			    || err == ENOENT || err == ENOTDIR) {
				error(N_("%s: Can't create directory %s: %s"),
				      progname, name, strerror(err));
				exit(EXIT_FAILURE);
			}
		}
		if (cp)
		  *cp++ = '/';
	}
	free(name);
    }
}
zdump.c0000644000000000000000000010013115244464512007312 0ustar00/* Dump time zone data in a textual format.  */

/*
** This file is in the public domain, so clarified as of
** 2009-05-17 by Arthur David Olson.
*/

#include "version.h"

#ifndef NETBSD_INSPIRED
# define NETBSD_INSPIRED 1
#endif

#include "private.h"
#include 

#ifndef HAVE_LOCALTIME_R
# define HAVE_LOCALTIME_R 1
#endif

#ifndef HAVE_LOCALTIME_RZ
# ifdef TM_ZONE
#  define HAVE_LOCALTIME_RZ (NETBSD_INSPIRED && USE_LTZ)
# else
#  define HAVE_LOCALTIME_RZ 0
# endif
#endif

#ifndef HAVE_TZSET
# define HAVE_TZSET 1
#endif

#ifndef ZDUMP_LO_YEAR
# define ZDUMP_LO_YEAR (-500)
#endif /* !defined ZDUMP_LO_YEAR */

#ifndef ZDUMP_HI_YEAR
# define ZDUMP_HI_YEAR 2500
#endif /* !defined ZDUMP_HI_YEAR */

#define SECSPERNYEAR	(SECSPERDAY * DAYSPERNYEAR)
#define SECSPERLYEAR	(SECSPERNYEAR + SECSPERDAY)
#define SECSPER400YEARS	(SECSPERNYEAR * (intmax_t) (300 + 3)	\
			 + SECSPERLYEAR * (intmax_t) (100 - 3))

/*
** True if SECSPER400YEARS is known to be representable as an
** intmax_t.  It's OK that SECSPER400YEARS_FITS can in theory be false
** even if SECSPER400YEARS is representable, because when that happens
** the code merely runs a bit more slowly, and this slowness doesn't
** occur on any practical platform.
*/
enum { SECSPER400YEARS_FITS = SECSPERLYEAR <= INTMAX_MAX / 400 };

#if HAVE_GETTEXT
# include  /* for setlocale */
#endif /* HAVE_GETTEXT */

#if ! HAVE_LOCALTIME_RZ
# undef  timezone_t
# define timezone_t char **
#endif

/* The minimum and maximum finite time values.  */
enum { atime_shift = CHAR_BIT * sizeof(time_t) - 2 };
static time_t const absolute_min_time =
  ((time_t) -1 < 0
   ? (- ((time_t) ~ (time_t) 0 < 0)
      - (((time_t) 1 << atime_shift) - 1 + ((time_t) 1 << atime_shift)))
   : 0);
static time_t const absolute_max_time =
  ((time_t) -1 < 0
   ? (((time_t) 1 << atime_shift) - 1 + ((time_t) 1 << atime_shift))
   : -1);
static int	longest;
static char const *progname;
static bool	warned;
static bool	errout;

static char const *abbr(struct tm const *);
static intmax_t delta(struct tm *, struct tm *);
static void dumptime(struct tm const *);
static time_t hunt(timezone_t, time_t, time_t, bool);
static void show(timezone_t, char *, time_t, bool);
static void showextrema(timezone_t, char *, time_t, struct tm *, time_t);
static void showtrans(char const *, struct tm const *, time_t, char const *,
		      char const *);
static const char *tformat(void);
ATTRIBUTE_PURE_114833 static time_t yeartot(intmax_t);

/* Is C an ASCII digit?  */
static bool
is_digit(char c)
{
  return '0' <= c && c <= '9';
}

/* Is A an alphabetic character in the C locale?  */
static bool
is_alpha(char a)
{
	switch (a) {
	  default:
		return false;
	  case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
	  case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
	  case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
	  case 'V': case 'W': case 'X': case 'Y': case 'Z':
	  case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
	  case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
	  case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
	  case 'v': case 'w': case 'x': case 'y': case 'z':
		return true;
	}
}

ATTRIBUTE_NORETURN static void
size_overflow(void)
{
  fprintf(stderr, _("%s: size overflow\n"), progname);
  exit(EXIT_FAILURE);
}

/* Return A + B, exiting if the result would overflow either ptrdiff_t
   or size_t.  A and B are both nonnegative.  */
ATTRIBUTE_PURE_114833_HACK
static ptrdiff_t
sumsize(ptrdiff_t a, ptrdiff_t b)
{
#ifdef ckd_add
  ptrdiff_t sum;
  if (!ckd_add(&sum, a, b) && sum <= INDEX_MAX)
    return sum;
#else
  if (a <= INDEX_MAX && b <= INDEX_MAX - a)
    return a + b;
#endif
  size_overflow();
}


/* Return a pointer to a newly allocated buffer of size SIZE, exiting
   on failure.  SIZE should be positive.  */
static void *
xmalloc(ptrdiff_t size)
{
  void *p = malloc(size);
  if (!p) {
    fprintf(stderr, _("%s: Memory exhausted\n"), progname);
    exit(EXIT_FAILURE);
  }
  return p;
}

#if ! HAVE_TZSET
# undef tzset
# define tzset zdump_tzset
static void tzset(void) { }
#endif

/* Assume gmtime_r works if localtime_r does.
   A replacement localtime_r is defined below if needed.  */
#if ! HAVE_LOCALTIME_R

# undef gmtime_r
# define gmtime_r zdump_gmtime_r

static struct tm *
gmtime_r(time_t *tp, struct tm *tmp)
{
  struct tm *r = gmtime(tp);
  if (r) {
    *tmp = *r;
    r = tmp;
  }
  return r;
}

#endif

/* Platforms with TM_ZONE don't need tzname, so they can use the
   faster localtime_rz or localtime_r if available.  */

#if defined TM_ZONE && HAVE_LOCALTIME_RZ
# define USE_LOCALTIME_RZ true
#else
# define USE_LOCALTIME_RZ false
#endif

#if ! USE_LOCALTIME_RZ

# if !defined TM_ZONE || ! HAVE_LOCALTIME_R || ! HAVE_TZSET
#  undef localtime_r
#  define localtime_r zdump_localtime_r
static struct tm *
localtime_r(time_t *tp, struct tm *tmp)
{
  struct tm *r = localtime(tp);
  if (r) {
    *tmp = *r;
    r = tmp;
  }
  return r;
}
# endif

# undef localtime_rz
# define localtime_rz zdump_localtime_rz
static struct tm *
localtime_rz(ATTRIBUTE_MAYBE_UNUSED timezone_t rz, time_t *tp, struct tm *tmp)
{
  return localtime_r(tp, tmp);
}

# undef tzalloc
# undef tzfree
# define tzalloc zdump_tzalloc
# define tzfree zdump_tzfree

static timezone_t
tzalloc(char const *val)
{
# if HAVE_SETENV
  if (setenv("TZ", val, 1) != 0) {
    char const *e = strerror(errno);
    fprintf(stderr, _("%s: setenv: %s\n"), progname, e);
    exit(EXIT_FAILURE);
  }
  tzset();
  return &optarg;  /* Any valid non-null char ** will do.  */
# else
  enum { TZeqlen = 3 };
  static char const TZeq[TZeqlen] = "TZ=";
  static char **fakeenv;
  static ptrdiff_t fakeenv0size;
  void *freeable = NULL;
  char **env = fakeenv, **initial_environ;
  ptrdiff_t valsize = strlen(val) + 1;
  if (fakeenv0size < valsize) {
    char **e = environ, **to;
    ptrdiff_t initial_nenvptrs = 1;  /* Counting the trailing NULL pointer.  */

    while (*e++) {
#  ifdef ckd_add
      if (ckd_add(&initial_nenvptrs, initial_nenvptrs, 1)
	  || INDEX_MAX < initial_nenvptrs)
	size_overflow();
#  else
      if (initial_nenvptrs == INDEX_MAX / sizeof *environ)
	size_overflow();
      initial_nenvptrs++;
#  endif
    }
    fakeenv0size = sumsize(valsize, valsize);
    fakeenv0size = max(fakeenv0size, 64);
    freeable = env;
    fakeenv = env =
      xmalloc(sumsize(sumsize(sizeof *environ,
			      initial_nenvptrs * sizeof *environ),
		      sumsize(TZeqlen, fakeenv0size)));
    to = env + 1;
    for (e = environ; (*to = *e); e++)
      to += strncmp(*e, TZeq, TZeqlen) != 0;
    env[0] = memcpy(to + 1, TZeq, TZeqlen);
  }
  memcpy(env[0] + TZeqlen, val, valsize);
  initial_environ = environ;
  environ = env;
  tzset();
  free(freeable);
  return initial_environ;
# endif
}

static void
tzfree(ATTRIBUTE_MAYBE_UNUSED timezone_t initial_environ)
{
# if !HAVE_SETENV
  environ = initial_environ;
  tzset();
# endif
}
#endif /* ! USE_LOCALTIME_RZ */

/* A UT time zone, and its initializer.  */
static timezone_t gmtz;
static void
gmtzinit(void)
{
  if (USE_LOCALTIME_RZ) {
    /* Try "GMT" first to find out whether this is one of the rare
       platforms where time_t counts leap seconds; this works due to
       the "Zone GMT 0 - GMT" line in the "etcetera" file.  If "GMT"
       fails, fall back on "GMT0" which might be similar due to the
       "Link GMT GMT0" line in the "backward" file, and which
       should work on all POSIX platforms.  The rest of zdump does not
       use the "GMT" abbreviation that comes from this setting, so it
       is OK to use "GMT" here rather than the modern "UTC" which
       would not work on platforms that omit the "backward" file.  */
    gmtz = tzalloc("GMT");
    if (!gmtz) {
      static char const gmt0[] = "GMT0";
      gmtz = tzalloc(gmt0);
      if (!gmtz) {
	char const *e = strerror(errno);
	fprintf(stderr, _("%s: unknown timezone '%s': %s\n"),
		progname, gmt0, e);
	exit(EXIT_FAILURE);
      }
    }
  }
}

/* Convert *TP to UT, storing the broken-down time into *TMP.
   Return TMP if successful, NULL otherwise.  This is like gmtime_r(TP, TMP),
   except typically faster if USE_LOCALTIME_RZ.  */
static struct tm *
my_gmtime_r(time_t *tp, struct tm *tmp)
{
  return USE_LOCALTIME_RZ ? localtime_rz(gmtz, tp, tmp) : gmtime_r(tp, tmp);
}

static void
abbrok(const char *const abbrp, const char *const zone)
{
	register const char *	cp;
	register const char *	wp;

	if (warned)
		return;
	cp = abbrp;
	while (is_alpha(*cp) || is_digit(*cp) || *cp == '-' || *cp == '+')
		++cp;
	if (*cp)
	  wp = _("has characters other than ASCII alphanumerics, '-' or '+'");
	else if (cp - abbrp < 3)
	  wp = _("has fewer than 3 characters");
	else if (cp - abbrp > 6)
	  wp = _("has more than 6 characters");
	else
	  return;
	fflush(stdout);
	fprintf(stderr,
		_("%s: warning: zone \"%s\" abbreviation \"%s\" %s\n"),
		progname, zone, abbrp, wp);
	warned = errout = true;
}

/* Return a time zone abbreviation.  If the abbreviation needs to be
   saved, use *BUF (of size *BUFALLOC) to save it, and return the
   abbreviation in the possibly reallocated *BUF.  Otherwise, just
   return the abbreviation.  Get the abbreviation from TMP.
   Exit on memory allocation failure.  */
static char const *
saveabbr(char **buf, ptrdiff_t *bufalloc, struct tm const *tmp)
{
  char const *ab = abbr(tmp);
  if (HAVE_LOCALTIME_RZ)
    return ab;
  else {
    ptrdiff_t absize = strlen(ab) + 1;
    if (*bufalloc < absize) {
      free(*buf);

      /* Make the new buffer at least twice as long as the old,
	 to avoid O(N**2) behavior on repeated calls.  */
      *bufalloc = sumsize(*bufalloc, absize);

      *buf = xmalloc(*bufalloc);
    }
    return strcpy(*buf, ab);
  }
}

static void
close_file(FILE *stream)
{
  char const *e = (ferror(stream) ? _("I/O error")
		   : fclose(stream) != 0 ? strerror(errno) : NULL);
  if (e) {
    fprintf(stderr, "%s: %s\n", progname, e);
    exit(EXIT_FAILURE);
  }
}

static void
usage(FILE * const stream, const int status)
{
	fprintf(stream,
_("%s: usage: %s OPTIONS TIMEZONE ...\n"
  "Options include:\n"
  "  -c [L,]U   Start at year L (default -500), end before year U (default 2500)\n"
  "  -t [L,]U   Start at time L, end before time U (in seconds since 1970)\n"
  "  -i         List transitions briefly (format is experimental)\n" \
  "  -v         List transitions verbosely\n"
  "  -V         List transitions a bit less verbosely\n"
  "  --help     Output this help\n"
  "  --version  Output version info\n"
  "\n"
  "Report bugs to %s.\n"),
		progname, progname, REPORT_BUGS_TO);
	if (status == EXIT_SUCCESS)
	  close_file(stream);
	exit(status);
}

int
main(int argc, char *argv[])
{
	/* These are static so that they're initially zero.  */
	static char *		abbrev;
	static ptrdiff_t	abbrevsize;

	register int		i;
	register bool		vflag;
	register bool		Vflag;
	register char *		cutarg;
	register char *		cuttimes;
	register time_t		cutlotime;
	register time_t		cuthitime;
	time_t			now;
	bool iflag = false;
	size_t arglenmax = 0;

	cutlotime = absolute_min_time;
	cuthitime = absolute_max_time;
#if HAVE_GETTEXT
	setlocale(LC_ALL, "");
# ifdef TZ_DOMAINDIR
	bindtextdomain(TZ_DOMAIN, TZ_DOMAINDIR);
# endif /* defined TEXTDOMAINDIR */
	textdomain(TZ_DOMAIN);
#endif /* HAVE_GETTEXT */
	progname = argv[0] ? argv[0] : "zdump";
	for (i = 1; i < argc; ++i)
		if (strcmp(argv[i], "--version") == 0) {
			printf("zdump %s%s\n", PKGVERSION, TZVERSION);
			return EXIT_SUCCESS;
		} else if (strcmp(argv[i], "--help") == 0) {
			usage(stdout, EXIT_SUCCESS);
		}
	vflag = Vflag = false;
	cutarg = cuttimes = NULL;
	for (;;)
	  switch (getopt(argc, argv, "c:it:vV")) {
	  case 'c': cutarg = optarg; break;
	  case 't': cuttimes = optarg; break;
	  case 'i': iflag = true; break;
	  case 'v': vflag = true; break;
	  case 'V': Vflag = true; break;
	  case -1:
	    if (! (optind == argc - 1 && strcmp(argv[optind], "=") == 0))
	      goto arg_processing_done;
	    ATTRIBUTE_FALLTHROUGH;
	  default:
	    usage(stderr, EXIT_FAILURE);
	  }
 arg_processing_done:;

	if (iflag | vflag | Vflag) {
		intmax_t	lo;
		intmax_t	hi;
		char *loend, *hiend;
		register intmax_t cutloyear = ZDUMP_LO_YEAR;
		register intmax_t cuthiyear = ZDUMP_HI_YEAR;
		if (cutarg != NULL) {
			lo = strtoimax(cutarg, &loend, 10);
			if (cutarg != loend && !*loend) {
				hi = lo;
				cuthiyear = hi;
			} else if (cutarg != loend && *loend == ','
				   && (hi = strtoimax(loend + 1, &hiend, 10),
				       loend + 1 != hiend && !*hiend)) {
				cutloyear = lo;
				cuthiyear = hi;
			} else {
				fprintf(stderr, _("%s: wild -c argument %s\n"),
					progname, cutarg);
				return EXIT_FAILURE;
			}
		}
		if (cutarg != NULL || cuttimes == NULL) {
			cutlotime = yeartot(cutloyear);
			cuthitime = yeartot(cuthiyear);
		}
		if (cuttimes != NULL) {
			lo = strtoimax(cuttimes, &loend, 10);
			if (cuttimes != loend && !*loend) {
				hi = lo;
				if (hi < cuthitime) {
					if (hi < absolute_min_time + 1)
					  hi = absolute_min_time + 1;
					cuthitime = hi;
				}
			} else if (cuttimes != loend && *loend == ','
				   && (hi = strtoimax(loend + 1, &hiend, 10),
				       loend + 1 != hiend && !*hiend)) {
				if (cutlotime < lo) {
					if (absolute_max_time < lo)
						lo = absolute_max_time;
					cutlotime = lo;
				}
				if (hi < cuthitime) {
					if (hi < absolute_min_time + 1)
					  hi = absolute_min_time + 1;
					cuthitime = hi;
				}
			} else {
				fprintf(stderr,
					_("%s: wild -t argument %s\n"),
					progname, cuttimes);
				return EXIT_FAILURE;
			}
		}
	}
	gmtzinit();
	if (iflag | vflag | Vflag)
	  now = 0;
	else {
	  now = time(NULL);
	  now |= !now;
	}
	for (i = optind; i < argc; i++) {
	  size_t arglen = strlen(argv[i]);
	  if (arglenmax < arglen)
	    arglenmax = arglen;
	}
	if (!HAVE_SETENV && INDEX_MAX <= arglenmax)
	  size_overflow();
	longest = min(arglenmax, INT_MAX - 2);

	for (i = optind; i < argc; ++i) {
		/* Treat "-" as standard input on platforms with /dev/stdin.
		   It's not worth the bother of supporting "-" on other
		   platforms, as that would need temp files.  */
		timezone_t tz = tzalloc(strcmp(argv[i], "-") == 0
					? "/dev/stdin" : argv[i]);
		char const *ab;
		time_t t;
		struct tm tm, newtm;
		bool tm_ok;
		if (!tz) {
		  char const *e = strerror(errno);
		  fprintf(stderr, _("%s: unknown timezone '%s': %s\n"),
			  progname, argv[i], e);
		  return EXIT_FAILURE;
		}
		if (now) {
			show(tz, argv[i], now, false);
			tzfree(tz);
			continue;
		}
		warned = false;
		t = absolute_min_time;
		if (! (iflag | Vflag)) {
			show(tz, argv[i], t, true);
			if (localtime_rz(tz, &t, &tm) == NULL
			    && t < cutlotime) {
				time_t newt = cutlotime;
				if (localtime_rz(tz, &newt, &newtm) != NULL)
				  showextrema(tz, argv[i], t, NULL, newt);
			}
		}
		if (t + 1 < cutlotime)
		  t = cutlotime - 1;
		tm_ok = localtime_rz(tz, &t, &tm) != NULL;
		if (tm_ok) {
		  ab = saveabbr(&abbrev, &abbrevsize, &tm);
		  if (iflag) {
		    showtrans("\nTZ=%f", &tm, t, ab, argv[i]);
		    showtrans("-\t-\t%Q", &tm, t, ab, argv[i]);
		  }
		} else
		  ab = NULL;
		while (t < cuthitime - 1) {
		  time_t newt = ((t < absolute_max_time - SECSPERDAY / 2
				  && t + SECSPERDAY / 2 < cuthitime - 1)
				 ? t + SECSPERDAY / 2
				 : cuthitime - 1);
		  struct tm *newtmp = localtime_rz(tz, &newt, &newtm);
		  bool newtm_ok = newtmp != NULL;
		  if (tm_ok != newtm_ok
		      || (ab && (delta(&newtm, &tm) != newt - t
				 || newtm.tm_isdst != tm.tm_isdst
				 || strcmp(abbr(&newtm), ab) != 0))) {
		    newt = hunt(tz, t, newt, false);
		    newtmp = localtime_rz(tz, &newt, &newtm);
		    newtm_ok = newtmp != NULL;
		    if (iflag)
		      showtrans("%Y-%m-%d\t%L\t%Q", newtmp, newt,
				newtm_ok ? abbr(&newtm) : NULL, argv[i]);
		    else {
		      show(tz, argv[i], newt - 1, true);
		      show(tz, argv[i], newt, true);
		    }
		  }
		  t = newt;
		  tm_ok = newtm_ok;
		  if (newtm_ok) {
		    ab = saveabbr(&abbrev, &abbrevsize, &newtm);
		    tm = newtm;
		  }
		}
		if (! (iflag | Vflag)) {
			time_t newt = absolute_max_time;
			t = cuthitime;
			if (t < newt) {
			  struct tm *tmp = localtime_rz(tz, &t, &tm);
			  if (tmp != NULL
			      && localtime_rz(tz, &newt, &newtm) == NULL)
			    showextrema(tz, argv[i], t, tmp, newt);
			}
			show(tz, argv[i], absolute_max_time, true);
		}
		tzfree(tz);
	}
	close_file(stdout);
	if (errout && (ferror(stderr) || fclose(stderr) != 0))
	  return EXIT_FAILURE;
	return EXIT_SUCCESS;
}

static time_t
yeartot(intmax_t y)
{
	register intmax_t	myy, seconds, years;
	register time_t		t;

	myy = EPOCH_YEAR;
	t = 0;
	while (myy < y) {
		if (SECSPER400YEARS_FITS && 400 <= y - myy) {
			intmax_t diff400 = (y - myy) / 400;
			if (INTMAX_MAX / SECSPER400YEARS < diff400)
				return absolute_max_time;
			seconds = diff400 * SECSPER400YEARS;
			years = diff400 * 400;
		} else {
			seconds = isleap(myy) ? SECSPERLYEAR : SECSPERNYEAR;
			years = 1;
		}
		myy += years;
		if (t > absolute_max_time - seconds)
			return absolute_max_time;
		t += seconds;
	}
	while (y < myy) {
		if (SECSPER400YEARS_FITS && y + 400 <= myy && myy < 0) {
			intmax_t diff400 = (myy - y) / 400;
			if (INTMAX_MAX / SECSPER400YEARS < diff400)
				return absolute_min_time;
			seconds = diff400 * SECSPER400YEARS;
			years = diff400 * 400;
		} else {
			seconds = isleap(myy - 1) ? SECSPERLYEAR : SECSPERNYEAR;
			years = 1;
		}
		myy -= years;
		if (t < absolute_min_time + seconds)
			return absolute_min_time;
		t -= seconds;
	}
	return t;
}

/* Search for a discontinuity in timezone TZ, in the
   timestamps ranging from LOT through HIT.  LOT and HIT disagree
   about some aspect of timezone.  If ONLY_OK, search only for
   definedness changes, i.e., localtime succeeds on one side of the
   transition but fails on the other side.  Return the timestamp just
   before the transition from LOT's settings.  */

static time_t
hunt(timezone_t tz, time_t lot, time_t hit, bool only_ok)
{
	static char *		loab;
	static ptrdiff_t	loabsize;
	struct tm		lotm;
	struct tm		tm;

	/* Convert LOT into a broken-down time here, even though our
	   caller already did that.  On platforms without TM_ZONE,
	   tzname may have been altered since our caller broke down
	   LOT, and tzname needs to be changed back.  */
	bool lotm_ok = localtime_rz(tz, &lot, &lotm) != NULL;
	bool tm_ok;
	char const *ab = lotm_ok ? saveabbr(&loab, &loabsize, &lotm) : NULL;

	for ( ; ; ) {
		/* T = average of LOT and HIT, rounding down.
		   Avoid overflow.  */
		int rem_sum = lot % 2 + hit % 2;
		time_t t = (rem_sum == 2) - (rem_sum < 0) + lot / 2 + hit / 2;
		if (t == lot)
			break;
		tm_ok = localtime_rz(tz, &t, &tm) != NULL;
		if (lotm_ok == tm_ok
		    && (only_ok
			|| (ab && tm.tm_isdst == lotm.tm_isdst
			    && delta(&tm, &lotm) == t - lot
			    && strcmp(abbr(&tm), ab) == 0))) {
		  lot = t;
		  if (tm_ok)
		    lotm = tm;
		} else	hit = t;
	}
	return hit;
}

/*
** Thanks to Paul Eggert for logic used in delta_nonneg.
*/

static intmax_t
delta_nonneg(struct tm *newp, struct tm *oldp)
{
	intmax_t oldy = oldp->tm_year;
	int cycles = (newp->tm_year - oldy) / YEARSPERREPEAT;
	intmax_t sec = SECSPERREPEAT, result = cycles * sec;
	int tmy = oldp->tm_year + cycles * YEARSPERREPEAT;
	for ( ; tmy < newp->tm_year; ++tmy)
	  result += year_sum_days(tmy, TM_YEAR_BASE);
	result += newp->tm_yday - oldp->tm_yday;
	result *= HOURSPERDAY;
	result += newp->tm_hour - oldp->tm_hour;
	result *= MINSPERHOUR;
	result += newp->tm_min - oldp->tm_min;
	result *= SECSPERMIN;
	result += newp->tm_sec - oldp->tm_sec;
	return result;
}

static intmax_t
delta(struct tm *newp, struct tm *oldp)
{
  return (newp->tm_year < oldp->tm_year
	  ? -delta_nonneg(oldp, newp)
	  : delta_nonneg(newp, oldp));
}

#ifndef TM_GMTOFF
/* Return A->tm_yday, adjusted to compare it fairly to B->tm_yday.
   Assume A and B differ by at most one year.  */
static int
adjusted_yday(struct tm const *a, struct tm const *b)
{
  int yday = a->tm_yday;
  if (b->tm_year < a->tm_year)
    yday += year_sum_days(b->tm_year, TM_YEAR_BASE);
  return yday;
}
#endif

/* If A is the broken-down local time and B the broken-down UT for
   the same instant, return A's UT offset in seconds, where positive
   offsets are east of Greenwich.  On failure, return LONG_MIN.

   If T is nonnull, *T is the timestamp that corresponds to A; call
   my_gmtime_r and use its result instead of B.  Otherwise, B is the
   possibly nonnull result of an earlier call to my_gmtime_r.  */
static long
gmtoff(struct tm const *a, ATTRIBUTE_MAYBE_UNUSED time_t *t,
       ATTRIBUTE_MAYBE_UNUSED struct tm const *b)
{
#ifdef TM_GMTOFF
  return a->TM_GMTOFF;
#else
  struct tm tm;
  if (t)
    b = my_gmtime_r(t, &tm);
  if (! b)
    return LONG_MIN;
  else {
    int ayday = adjusted_yday(a, b);
    int byday = adjusted_yday(b, a);
    int days = ayday - byday;
    long hours = a->tm_hour - b->tm_hour + 24 * days;
    long minutes = a->tm_min - b->tm_min + 60 * hours;
    long seconds = a->tm_sec - b->tm_sec + 60 * minutes;
    return seconds;
  }
#endif
}

static void
show(timezone_t tz, char *zone, time_t t, bool v)
{
	register struct tm *	tmp;
	register struct tm *	gmtmp;
	struct tm tm, gmtm;

	printf("%-*s  ", longest, zone);
	if (v) {
		gmtmp = my_gmtime_r(&t, &gmtm);
		if (gmtmp == NULL) {
			printf(tformat(), t);
			printf(_(" (gmtime failed)"));
		} else {
			dumptime(gmtmp);
			printf(" UT");
		}
		printf(" = ");
	}
	tmp = localtime_rz(tz, &t, &tm);
	if (tmp == NULL) {
		printf(tformat(), t);
		printf(_(" (localtime failed)"));
	} else {
		dumptime(tmp);
		if (*abbr(tmp) != '\0')
			printf(" %s", abbr(tmp));
		if (v) {
			long off = gmtoff(tmp, NULL, gmtmp);
			printf(" isdst=%d", tmp->tm_isdst);
			if (off != LONG_MIN)
			  printf(" gmtoff=%ld", off);
		}
	}
	printf("\n");
	if (tmp != NULL && *abbr(tmp) != '\0')
		abbrok(abbr(tmp), zone);
}

/* Show timestamps just before and just after a transition between
   defined and undefined (or vice versa) in either localtime or
   gmtime.  These transitions are for timezone TZ with name ZONE, in
   the range from LO (with broken-down time LOTMP if that is nonnull)
   through HI.  LO and HI disagree on definedness.  */

static void
showextrema(timezone_t tz, char *zone, time_t lo, struct tm *lotmp, time_t hi)
{
  struct tm localtm[2], gmtm[2];
  time_t t, boundary = hunt(tz, lo, hi, true);
  bool old = false;
  hi = (SECSPERDAY < hi - boundary
	? boundary + SECSPERDAY
	: hi + (hi < TIME_T_MAX));
  if (SECSPERDAY < boundary - lo) {
    lo = boundary - SECSPERDAY;
    lotmp = localtime_rz(tz, &lo, &localtm[old]);
  }
  if (lotmp)
    localtm[old] = *lotmp;
  else
    localtm[old].tm_sec = -1;
  if (! my_gmtime_r(&lo, &gmtm[old]))
    gmtm[old].tm_sec = -1;

  /* Search sequentially for definedness transitions.  Although this
     could be sped up by refining 'hunt' to search for either
     localtime or gmtime definedness transitions, it hardly seems
     worth the trouble.  */
  for (t = lo + 1; t < hi; t++) {
    bool new = !old;
    if (! localtime_rz(tz, &t, &localtm[new]))
      localtm[new].tm_sec = -1;
    if (! my_gmtime_r(&t, &gmtm[new]))
      gmtm[new].tm_sec = -1;
    if (((localtm[old].tm_sec < 0) != (localtm[new].tm_sec < 0))
	| ((gmtm[old].tm_sec < 0) != (gmtm[new].tm_sec < 0))) {
      show(tz, zone, t - 1, true);
      show(tz, zone, t, true);
    }
    old = new;
  }
}

/* On pre-C99 platforms, a snprintf substitute good enough for us.  */
#if !HAVE_SNPRINTF
# include 
ATTRIBUTE_FORMAT((printf, 3, 4)) static int
my_snprintf(char *s, size_t size, char const *format, ...)
{
  int n;
  va_list args;
  char const *arg;
  char *cp;
  size_t arglen, slen;
  char buf[1024];
  va_start(args, format);
  if (strcmp(format, "%s") == 0) {
    arg = va_arg(args, char const *);
    arglen = strlen(arg);
  } else {
    n = vsprintf(buf, format, args);
    if (n < 0) {
      va_end(args);
      return n;
    }
    arg = buf;
    arglen = n;
  }
  slen = arglen < size ? arglen : size - 1;
  cp = s;
  cp = mempcpy(cp, arg, slen);
  *cp = '\0';
  n = arglen <= INT_MAX ? arglen : -1;
  va_end(args);
  return n;
}
# define snprintf my_snprintf
#endif

/* Store into BUF, of size SIZE, a formatted local time taken from *TM.
   Use ISO 8601 format +HH:MM:SS.  Omit :SS if SS is zero, and omit
   :MM too if MM is also zero.

   Return the length of the resulting string.  If the string does not
   fit, return the length that the string would have been if it had
   fit; do not overrun the output buffer.  */
static int
format_local_time(char *buf, ptrdiff_t size, struct tm const *tm)
{
  int ss = tm->tm_sec, mm = tm->tm_min, hh = tm->tm_hour;
  return (ss
	  ? snprintf(buf, size, "%02d:%02d:%02d", hh, mm, ss)
	  : mm
	  ? snprintf(buf, size, "%02d:%02d", hh, mm)
	  : snprintf(buf, size, "%02d", hh));
}

/* Store into BUF, of size SIZE, a formatted UT offset for the
   localtime *TM corresponding to time T.  Use ISO 8601 format
   +HHMMSS, or -HHMMSS for timestamps west of Greenwich; use the
   format -00 for unknown UT offsets.  If the hour needs more than
   two digits to represent, extend the length of HH as needed.
   Otherwise, omit SS if SS is zero, and omit MM too if MM is also
   zero.

   Return the length of the resulting string, or -1 if the result is
   not representable as a string.  If the string does not fit, return
   the length that the string would have been if it had fit; do not
   overrun the output buffer.  */
static int
format_utc_offset(char *buf, ptrdiff_t size, struct tm const *tm, time_t t)
{
  long off = gmtoff(tm, &t, NULL);
  char sign = ((off < 0
		|| (off == 0
		    && (*abbr(tm) == '-' || strcmp(abbr(tm), "zzz") == 0)))
	       ? '-' : '+');
  long hh;
  int mm, ss;
  if (off < 0)
    {
      if (off == LONG_MIN)
	return -1;
      off = -off;
    }
  ss = off % 60;
  mm = off / 60 % 60;
  hh = off / 60 / 60;
  return (ss || 100 <= hh
	  ? snprintf(buf, size, "%c%02ld%02d%02d", sign, hh, mm, ss)
	  : mm
	  ? snprintf(buf, size, "%c%02ld%02d", sign, hh, mm)
	  : snprintf(buf, size, "%c%02ld", sign, hh));
}

/* Store into BUF (of size SIZE) a quoted string representation of P.
   If the representation's length is less than SIZE, return the
   length; the representation is not null terminated.  Otherwise
   return SIZE, to indicate that BUF is too small.  */
static ptrdiff_t
format_quoted_string(char *buf, ptrdiff_t size, char const *p)
{
  char *b = buf;
  ptrdiff_t s = size;
  if (!s)
    return size;
  *b++ = '"', s--;
  for (;;) {
    char c = *p++;
    if (s <= 1)
      return size;
    switch (c) {
    default: *b++ = c, s--; continue;
    case '\0': *b++ = '"', s--; return size - s;
    case '"': case '\\': break;
    case ' ': c = 's'; break;
    case '\f': c = 'f'; break;
    case '\n': c = 'n'; break;
    case '\r': c = 'r'; break;
    case '\t': c = 't'; break;
    case '\v': c = 'v'; break;
    }
    *b++ = '\\', *b++ = c, s -= 2;
  }
}

/* Store into BUF (of size SIZE) a timestamp formatted by TIME_FMT.
   TM is the broken-down time, T the seconds count, AB the time zone
   abbreviation, and ZONE_NAME the zone name.  Return true if
   successful, false if the output would require more than SIZE bytes.
   TIME_FMT uses the same format that strftime uses, with these
   additions:

   %f zone name
   %L local time as per format_local_time
   %Q like "U\t%Z\tD" where U is the UT offset as for format_utc_offset
      and D is the isdst flag; except omit D if it is zero, omit %Z if
      it equals U, quote and escape %Z if it contains nonalphabetics,
      and omit any trailing tabs.  */

static bool
istrftime(char *buf, ptrdiff_t size, char const *time_fmt,
	  struct tm const *tm, time_t t, char const *ab, char const *zone_name)
{
  char *b = buf;
  ptrdiff_t s = size;
  char const *f = time_fmt, *p;

  for (p = f; ; p++)
    if (*p == '%' && p[1] == '%')
      p++;
    else if (!*p
	     || (*p == '%'
		 && (p[1] == 'f' || p[1] == 'L' || p[1] == 'Q'))) {
      ptrdiff_t formatted_len;
      ptrdiff_t f_prefix_len = p - f;
      ptrdiff_t f_prefix_copy_size = sumsize(f_prefix_len, 2);
      char fbuf[100];
      bool oversized = sizeof fbuf <= f_prefix_copy_size;
      char *f_prefix_copy = oversized ? xmalloc(f_prefix_copy_size) : fbuf;
      char *cp = f_prefix_copy;
      cp = mempcpy(cp, f, f_prefix_len);
      strcpy(cp, "X");
      formatted_len = strftime(b, s, f_prefix_copy, tm);
      if (oversized)
	free(f_prefix_copy);
      if (formatted_len == 0)
	return false;
      formatted_len--;
      b += formatted_len, s -= formatted_len;
      if (!*p++)
	break;
      switch (*p) {
      case 'f':
	formatted_len = format_quoted_string(b, s, zone_name);
	break;
      case 'L':
	formatted_len = format_local_time(b, s, tm);
	break;
      case 'Q':
	{
	  bool show_abbr;
	  int offlen = format_utc_offset(b, s, tm, t);
	  if (! (0 <= offlen && offlen < s))
	    return false;
	  show_abbr = strcmp(b, ab) != 0;
	  b += offlen, s -= offlen;
	  if (show_abbr) {
	    char const *abp;
	    ptrdiff_t len;
	    if (s <= 1)
	      return false;
	    *b++ = '\t', s--;
	    for (abp = ab; is_alpha(*abp); abp++)
	      continue;
	    len = (!*abp && *ab
		   ? snprintf(b, s, "%s", ab)
		   : format_quoted_string(b, s, ab));
	    if (s <= len)
	      return false;
	    b += len, s -= len;
	  }
	  formatted_len
	    = (tm->tm_isdst
	       ? snprintf(b, s, &"\t\t%d"[show_abbr], tm->tm_isdst)
	       : 0);
	}
	break;
      }
      if (s <= formatted_len)
	return false;
      b += formatted_len, s -= formatted_len;
      f = p + 1;
    }
  *b = '\0';
  return true;
}

/* Show a time transition.  */
static void
showtrans(char const *time_fmt, struct tm const *tm, time_t t, char const *ab,
	  char const *zone_name)
{
  if (!tm) {
    printf(tformat(), t);
    putchar('\n');
  } else {
    char stackbuf[1000];
    ptrdiff_t size = sizeof stackbuf;
    char *buf = stackbuf;
    char *bufalloc = NULL;
    while (! istrftime(buf, size, time_fmt, tm, t, ab, zone_name)) {
      size = sumsize(size, size);
      free(bufalloc);
      buf = bufalloc = xmalloc(size);
    }
    puts(buf);
    free(bufalloc);
  }
}

static char const *
abbr(struct tm const *tmp)
{
#ifdef TM_ZONE
	return tmp->TM_ZONE;
#else
# if HAVE_TZNAME
	if (0 <= tmp->tm_isdst && tzname[0 < tmp->tm_isdst])
	  return tzname[0 < tmp->tm_isdst];
# endif
	return "";
#endif
}

/*
** The code below can fail on certain theoretical systems;
** it works on all known real-world systems as of 2022-01-25.
*/

static const char *
tformat(void)
{
#if HAVE__GENERIC
	/* C11-style _Generic is more likely to return the correct
	   format when distinct types have the same size.  */
	char const *fmt =
	  _Generic(+ (time_t) 0,
		   int: "%d", long: "%ld", long long: "%lld",
		   unsigned: "%u", unsigned long: "%lu",
		   unsigned long long: "%llu",
		   default: NULL);
	if (fmt)
	  return fmt;
	fmt = _Generic((time_t) 0,
		       intmax_t: "%"PRIdMAX, uintmax_t: "%"PRIuMAX,
		       default: NULL);
	if (fmt)
	  return fmt;
#endif
	if (0 > (time_t) -1) {		/* signed */
		if (sizeof(time_t) == sizeof(intmax_t))
			return "%"PRIdMAX;
		if (sizeof(time_t) > sizeof(long))
			return "%lld";
		if (sizeof(time_t) > sizeof(int))
			return "%ld";
		return "%d";
	}
#ifdef PRIuMAX
	if (sizeof(time_t) == sizeof(uintmax_t))
		return "%"PRIuMAX;
#endif
	if (sizeof(time_t) > sizeof(unsigned long))
		return "%llu";
	if (sizeof(time_t) > sizeof(unsigned int))
		return "%lu";
	return "%u";
}

static void
dumptime(register const struct tm *timeptr)
{
	static const char	wday_name[][4] = {
		"Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat"
	};
	static const char	mon_name[][4] = {
		"Jan", "Feb", "Mar", "Apr", "May", "Jun",
		"Jul", "Aug", "Sep", "Oct", "Nov", "Dec"
	};
	register int		lead;
	register int		trail;
	int DIVISOR = 10;

	/*
	** The packaged localtime_rz and gmtime_r never put out-of-range
	** values in tm_wday or tm_mon, but since this code might be compiled
	** with other (perhaps experimental) versions, paranoia is in order.
	*/
	printf("%s %s%3d %.2d:%.2d:%.2d ",
		((0 <= timeptr->tm_wday
		  && timeptr->tm_wday < sizeof wday_name / sizeof wday_name[0])
		 ? wday_name[timeptr->tm_wday] : "???"),
		((0 <= timeptr->tm_mon
		  && timeptr->tm_mon < sizeof mon_name / sizeof mon_name[0])
		 ? mon_name[timeptr->tm_mon] : "???"),
		timeptr->tm_mday, timeptr->tm_hour,
		timeptr->tm_min, timeptr->tm_sec);
	trail = timeptr->tm_year % DIVISOR + TM_YEAR_BASE % DIVISOR;
	lead = timeptr->tm_year / DIVISOR + TM_YEAR_BASE / DIVISOR +
		trail / DIVISOR;
	trail %= DIVISOR;
	if (trail < 0 && lead > 0) {
		trail += DIVISOR;
		--lead;
	} else if (lead < 0 && trail > 0) {
		trail -= DIVISOR;
		++lead;
	}
	if (lead == 0)
		printf("%d", trail);
	else	printf("%d%d", lead, ((trail < 0) ? -trail : trail));
}
date.c0000644000000000000000000001244615076626027007107 0ustar00/* Display or set the current time and date.  */

/* Copyright 1985, 1987, 1988 The Regents of the University of California.
   All rights reserved.

   Redistribution and use in source and binary forms, with or without
   modification, are permitted provided that the following conditions
   are met:
   1. Redistributions of source code must retain the above copyright
      notice, this list of conditions and the following disclaimer.
   2. Redistributions in binary form must reproduce the above copyright
      notice, this list of conditions and the following disclaimer in the
      documentation and/or other materials provided with the distribution.
   3. Neither the name of the University nor the names of its contributors
      may be used to endorse or promote products derived from this software
      without specific prior written permission.

   THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS "AS IS" AND
   ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
   IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
   ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
   FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
   DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
   OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
   HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
   LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
   OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
   SUCH DAMAGE.  */

#include "private.h"
#include 
#include 

static int		retval = EXIT_SUCCESS;

static void		display(const char *, time_t);
static void		dogmt(void);
static void		errensure(void);
static void		timeout(FILE *, const char *, const struct tm *);
ATTRIBUTE_NORETURN static void usage(void);

int
main(const int argc, char *argv[])
{
	register const char *	format = "+%+";
	register int		ch;
	register bool		rflag = false;
	time_t			t;
	intmax_t		secs;
	char *			endarg;

#ifdef LC_ALL
	setlocale(LC_ALL, "");
#endif /* defined(LC_ALL) */
#if HAVE_GETTEXT
# ifdef TZ_DOMAINDIR
	bindtextdomain(TZ_DOMAIN, TZ_DOMAINDIR);
# endif /* defined(TEXTDOMAINDIR) */
	textdomain(TZ_DOMAIN);
#endif /* HAVE_GETTEXT */
	t = time(NULL);
	while ((ch = getopt(argc, argv, "ucr:")) != -1) {
		switch (ch) {
		default:
			usage();
		case 'u':		/* do it in UT */
		case 'c':
			dogmt();
			break;
		case 'r':		/* seconds since 1970 */
			if (rflag) {
				fprintf(stderr,
					_("date: error: multiple -r's used"));
				usage();
			}
			rflag = true;
			errno = 0;
			secs = strtoimax(optarg, &endarg, 0);
			if (*endarg || optarg == endarg)
				errno = EINVAL;
			else if (! (TIME_T_MIN <= secs && secs <= TIME_T_MAX))
				errno = ERANGE;
			if (errno) {
				char const *e = strerror(errno);
				fprintf(stderr, _("date: %s: %s\n"),
					optarg, e);
				errensure();
				exit(retval);
			}
			t = secs;
			break;
		}
	}
	if (optind < argc) {
	  if (argc - optind != 1) {
	    fprintf(stderr,
		    _("date: error: multiple operands in command line\n"));
	    usage();
	  }
	  format = argv[optind];
	  if (*format != '+') {
	    fprintf(stderr, _("date: unknown operand: %s\n"), format);
	    usage();
	  }
	}

	display(format, t);
	return retval;
}

static void
dogmt(void)
{
	static char **	fakeenv;

	if (fakeenv == NULL) {
		static char	tzeutc0[] = "TZ=UTC0";
		ptrdiff_t from, to, n;

		for (n = 0;  environ[n] != NULL;  ++n)
			continue;
#if defined ckd_add && defined ckd_mul
		if (!ckd_add(&n, n, 2) && !ckd_mul(&n, n, sizeof *fakeenv)
		    && n <= INDEX_MAX)
		  fakeenv = malloc(n);
#else
		if (n <= INDEX_MAX / sizeof *fakeenv - 2)
		  fakeenv = malloc((n + 2) * sizeof *fakeenv);
#endif
		if (fakeenv == NULL) {
			fprintf(stderr, _("date: Memory exhausted\n"));
			errensure();
			exit(retval);
		}
		to = 0;
		fakeenv[to++] = tzeutc0;
		for (from = 1; environ[from] != NULL; ++from)
			if (strncmp(environ[from], "TZ=", 3) != 0)
				fakeenv[to++] = environ[from];
		fakeenv[to] = NULL;
		environ = fakeenv;
	}
}

static void
errensure(void)
{
	if (retval == EXIT_SUCCESS)
		retval = EXIT_FAILURE;
}

static void
usage(void)
{
	fprintf(stderr,
		       _("date: usage: date [-u] [-c] [-r seconds]"
			 " [+format]\n"));
	errensure();
	exit(retval);
}

static void
display(char const *format, time_t now)
{
	struct tm *tmp;

	tmp = localtime(&now);
	if (!tmp) {
		fprintf(stderr,
			_("date: error: time out of range\n"));
		errensure();
		return;
	}
	timeout(stdout, format, tmp);
	putchar('\n');
	fflush(stdout);
	fflush(stderr);
	if (ferror(stdout) || ferror(stderr)) {
		fprintf(stderr,
			_("date: error: couldn't write results\n"));
		errensure();
	}
}

static void
timeout(FILE *fp, char const *format, struct tm const *tmp)
{
	char *cp = NULL;
	ptrdiff_t result;
	ptrdiff_t size = 1024 / 2;

	for ( ; ; ) {
#ifdef ckd_mul
		bool bigger = !ckd_mul(&size, size, 2) && size <= INDEX_MAX;
#else
		bool bigger = size <= INDEX_MAX / 2 && (size *= 2, true);
#endif
		char *newcp = bigger ? realloc(cp, size) : NULL;
		if (!newcp) {
			fprintf(stderr,
				_("date: error: can't get memory\n"));
			errensure();
			exit(retval);
		}
		cp = newcp;
		result = strftime(cp, size, format, tmp);
		if (result != 0)
			break;
	}
	fwrite(cp + 1, 1, result - 1, fp);
	free(cp);
}
tzselect.ksh0000644000000000000000000005247315105202464010362 0ustar00#!/bin/bash
# Ask the user about the time zone, and output the resulting TZ value to stdout.
# Interact with the user via stderr and stdin.

PKGVERSION='(tzcode) '
TZVERSION=see_Makefile
REPORT_BUGS_TO=tz@iana.org

# Contributed by Paul Eggert.  This file is in the public domain.

# Porting notes:
#
# This script requires a POSIX-like shell and prefers the extension of a
# 'select' statement.  The 'select' statement was introduced in the
# Korn shell and is available in Bash and other shell implementations.
# If your host lacks both Bash and the Korn shell, you can get their
# source from one of these locations:
#
#	Bash 
#	Korn Shell 
#	MirBSD Korn Shell 
#
# This script also uses several features of POSIX awk.
# If your host lacks awk, or has an old awk that does not conform to POSIX,
# you can use any of the following free programs instead:
#
#	Gawk (GNU awk) 
#	mawk 
#	nawk 
#
# Because 'awk "VAR=VALUE" ...' and 'awk -v "VAR=VALUE" ...' are not portable
# if VALUE contains \, ", or newline, awk scripts in this file use:
#   awk 'BEGIN { VAR = substr(ARGV[1], 2); ARGV[1] = "" } ...' ="VALUE"
# The substr avoids problems when VALUE is of the form X=Y and would be
# misinterpreted as an assignment.

# This script does not want path expansion.
set -f

# Specify default values for environment variables if they are unset.
: ${AWK=awk}
: ${TZDIR=$PWD}

# Output one argument as-is to standard output, with trailing newline.
# Safer than 'echo', which can mishandle '\' or leading '-'.
say() {
  printf '%s\n' "$1"
}

coord=
location_limit=10
zonetabtype=zone1970

usage="Usage: tzselect [--version] [--help] [-c COORD] [-n LIMIT]
Select a timezone interactively.

Options:

  -c COORD
    Instead of asking for continent and then country and then city,
    ask for selection from time zones whose largest cities
    are closest to the location with geographical coordinates COORD.
    COORD should use ISO 6709 notation, for example, '-c +4852+00220'
    for Paris (in degrees and minutes, North and East), or
    '-c -35-058' for Buenos Aires (in degrees, South and West).

  -n LIMIT
    Display at most LIMIT locations when -c is used (default $location_limit).

  --version
    Output version information.

  --help
    Output this help.

Report bugs to $REPORT_BUGS_TO."

# Ask the user to select from the function's arguments,
# and assign the selected argument to the variable 'select_result'.
# Exit on EOF or I/O error.  Use the shell's nicer 'select' builtin if
# available, falling back on a portable substitute otherwise.
if
  case $BASH_VERSION in
  ?*) :;;
  '')
    # '; exit' should be redundant, but Dash doesn't properly fail without it.
    (eval 'set --; select x; do break; done; exit') <>/dev/null 2>&0
  esac
then
  # Do this inside 'eval', as otherwise the shell might exit when parsing it
  # even though it is never executed.
  eval '
    doselect() {
      select select_result
      do
	case $select_result in
	"") echo >&2 "Please enter a number in range.";;
	?*) break
	esac
      done || exit
    }
  '
else
  doselect() {
    # Field width of the prompt numbers.
    select_width=${##}

    select_i=

    while :
    do
      case $select_i in
      '')
	select_i=0
	for select_word
	do
	  select_i=$(($select_i + 1))
	  printf >&2 "%${select_width}d) %s\\n" $select_i "$select_word"
	done;;
      *[!0-9]*)
	echo >&2 'Please enter a number in range.';;
      *)
	if test 1 -le $select_i && test $select_i -le $#; then
	  shift $(($select_i - 1))
	  select_result=$1
	  break
	fi
	echo >&2 'Please enter a number in range.'
      esac

      # Prompt and read input.
      printf >&2 %s "${PS3-#? }"
      read select_i || exit
    done
  }
fi

while getopts c:n:t:-: opt
do
  case $opt$OPTARG in
  c*)
    coord=$OPTARG;;
  n*)
    location_limit=$OPTARG;;
  t*) # Undocumented option, used for developer testing.
    zonetabtype=$OPTARG;;
  -help)
    say "$usage"
    exit;;
  -version)
    say "tzselect $PKGVERSION$TZVERSION"
    exit;;
  -*)
    say >&2 "$0: -$opt$OPTARG: unknown option; try '$0 --help'"; exit 1;;
  *)
    say >&2 "$0: try '$0 --help'"; exit 1
  esac
done

shift $(($OPTIND - 1))
case $# in
0) ;;
*) say >&2 "$0: $1: unknown argument"; exit 1
esac

# translit=: to try transliteration.
# This is false if U+12345 CUNEIFORM SIGN URU TIMES KI has length 1
# which means the shell and (presumably) awk do not need transliteration.
# It is ':' if the byte string has some other length in characters, or
# if this is a POSIX.1-2017 or earlier shell that does not support $'...'.
CUNEIFORM_SIGN_URU_TIMES_KI=$'\360\222\215\205'
if test ${#CUNEIFORM_SIGN_URU_TIMES_KI} = 1
then translit=false
else translit=:
fi

# Read into shell variable $1 the contents of file $2.
# Convert to the current locale's encoding if possible,
# as the shell aligns columns better that way.
# If GNU iconv's //TRANSLIT does not work, fall back on POSIXish iconv;
# if that does not work, fall back on 'cat'.
read_file() {
  { $translit && {
    eval "$1=\$( (iconv -f UTF-8 -t //TRANSLIT) 2>/dev/null <\"\$2\")" ||
    eval "$1=\$( (iconv -f UTF-8) 2>/dev/null <\"\$2\")"
  }; } ||
  eval "$1=\$(cat <\"\$2\")" || {
    say >&2 "$0: time zone files are not set up correctly"
    exit 1
  }
}
read_file TZ_COUNTRY_TABLE "$TZDIR/iso3166.tab"
read_file TZ_ZONETABTYPE_TABLE "$TZDIR/$zonetabtype.tab"
TZ_ZONENOW_TABLE=

newline='
'
IFS=$newline

# Awk script to output a country list.
output_country_list='
  BEGIN {
    continent_re = substr(ARGV[1], 2)
    TZ_COUNTRY_TABLE = substr(ARGV[2], 2)
    TZ_ZONE_TABLE = substr(ARGV[3], 2)
    ARGV[1] = ARGV[2] = ARGV[3] = ""
    FS = "\t"
    nlines = split(TZ_ZONE_TABLE, line, /\n/)
    for (iline = 1; iline <= nlines; iline++) {
      $0 = line[iline]
      commentary = $0 ~ /^#@/
      if (commentary) {
	if ($0 !~ /^#@/)
	  continue
	col1ccs = substr($1, 3)
	conts = $2
      } else {
	col1ccs = $1
	conts = $3
      }
      ncc = split(col1ccs, cc, /,/)
      ncont = split(conts, cont, /,/)
      for (i = 1; i <= ncc; i++) {
	elsewhere = commentary
	for (ci = 1; ci <= ncont; ci++) {
	  if (cont[ci] ~ continent_re) {
	    if (!cc_seen[cc[i]]++)
	      cc_list[++ccs] = cc[i]
	    elsewhere = 0
	  }
	}
	if (elsewhere)
	  for (i = 1; i <= ncc; i++)
	    cc_elsewhere[cc[i]] = 1
      }
    }
    nlines = split(TZ_COUNTRY_TABLE, line, /\n/)
    for (i = 1; i <= nlines; i++) {
      $0 = line[i]
      if ($0 !~ /^#/)
	cc_name[$1] = $2
    }
    for (i = 1; i <= ccs; i++) {
      country = cc_list[i]
      if (cc_elsewhere[country])
	continue
      if (cc_name[country])
	country = cc_name[country]
      print country
    }
  }
'

# Awk script to process a time zone table and output the same table,
# with each row preceded by its distance from 'here'.
# If output_times is set, each row is instead preceded by its local time
# and any apostrophes are escaped for the shell.
output_distances_or_times='
  BEGIN {
    coord = substr(ARGV[1], 2)
    TZ_COUNTRY_TABLE = substr(ARGV[2], 2)
    TZ_ZONE_TABLE = substr(ARGV[3], 2)
    ARGV[1] = ARGV[2] = ARGV[3] = ""
    FS = "\t"
    if (!output_times) {
      nlines = split(TZ_COUNTRY_TABLE, line, /\n/)
      for (i = 1; i <= nlines; i++) {
	$0 = line[i]
	if ($0 ~ /^#/)
	  continue
	country[$1] = $2
      }
      country["US"] = "US" # Otherwise the strings get too long.
    }
  }
  function abs(x) {
    return x < 0 ? -x : x;
  }
  function min(x, y) {
    return x < y ? x : y;
  }
  function convert_coord(coord, deg, minute, ilen, sign, sec) {
    if (coord ~ /^[-+]?[0-9]?[0-9][0-9][0-9][0-9][0-9][0-9]([^0-9]|$)/) {
      degminsec = coord
      intdeg = degminsec < 0 ? -int(-degminsec / 10000) : int(degminsec / 10000)
      minsec = degminsec - intdeg * 10000
      intmin = minsec < 0 ? -int(-minsec / 100) : int(minsec / 100)
      sec = minsec - intmin * 100
      deg = (intdeg * 3600 + intmin * 60 + sec) / 3600
    } else if (coord ~ /^[-+]?[0-9]?[0-9][0-9][0-9][0-9]([^0-9]|$)/) {
      degmin = coord
      intdeg = degmin < 0 ? -int(-degmin / 100) : int(degmin / 100)
      minute = degmin - intdeg * 100
      deg = (intdeg * 60 + minute) / 60
    } else
      deg = coord
    return deg * 0.017453292519943296
  }
  function convert_latitude(coord) {
    match(coord, /..*[-+]/)
    return convert_coord(substr(coord, 1, RLENGTH - 1))
  }
  function convert_longitude(coord) {
    match(coord, /..*[-+]/)
    return convert_coord(substr(coord, RLENGTH))
  }
  # Great-circle distance between points with given latitude and longitude.
  # Inputs and output are in radians.  This uses the great-circle special
  # case of the Vicenty formula for distances on ellipsoids.
  function gcdist(lat1, long1, lat2, long2, dlong, x, y, num, denom) {
    dlong = long2 - long1
    x = cos(lat2) * sin(dlong)
    y = cos(lat1) * sin(lat2) - sin(lat1) * cos(lat2) * cos(dlong)
    num = sqrt(x * x + y * y)
    denom = sin(lat1) * sin(lat2) + cos(lat1) * cos(lat2) * cos(dlong)
    return atan2(num, denom)
  }
  # Parallel distance between points with given latitude and longitude.
  # This is the product of the longitude difference and the cosine
  # of the latitude of the point that is further from the equator.
  # I.e., it considers longitudes to be further apart if they are
  # nearer the equator.
  function pardist(lat1, long1, lat2, long2) {
    return abs(long1 - long2) * min(cos(lat1), cos(lat2))
  }
  # The distance function is the sum of the great-circle distance and
  # the parallel distance.  It could be weighted.
  function dist(lat1, long1, lat2, long2) {
    return gcdist(lat1, long1, lat2, long2) + pardist(lat1, long1, lat2, long2)
  }
  BEGIN {
    coord_lat = convert_latitude(coord)
    coord_long = convert_longitude(coord)
    nlines = split(TZ_ZONE_TABLE, line, /\n/)
    for (h = 1; h <= nlines; h++) {
      $0 = line[h]
      if ($0 ~ /^#/)
	continue
      inline[inlines++] = $0
      ncc = split($1, cc, /,/)
      for (i = 1; i <= ncc; i++)
	cc_used[cc[i]]++
    }
    for (h = 0; h < inlines; h++) {
      $0 = inline[h]
      outline = $1 "\t" $2 "\t" $3
      sep = "\t"
      ncc = split($1, cc, /,/)
      split("", item_seen)
      item_seen[""] = 1
      for (i = 1; i <= ncc; i++) {
	item = cc_used[cc[i]] <= 1 ? country[cc[i]] : $4
	if (item_seen[item]++)
	  continue
	outline = outline sep item
	sep = "; "
      }
      if (output_times) {
	fmt = "TZ='\''%s'\'' date +'\''%d %%Y %%m %%d %%H:%%M %%a %%b\t%s'\''\n"
	gsub(/'\''/, "&\\\\&&", outline)
	printf fmt, $3, h, outline
      } else {
	here_lat = convert_latitude($2)
	here_long = convert_longitude($2)
	printf "%g\t%s\n", dist(coord_lat, coord_long, here_lat, here_long), \
	  outline
      }
    }
  }
'

# Begin the main loop.  We come back here if the user wants to retry.
while

  echo >&2 'Please identify a location' \
    'so that time zone rules can be set correctly.'

  continent=
  country=
  country_result=
  region=
  time=
  TZ_ZONE_TABLE=$TZ_ZONETABTYPE_TABLE

  case $coord in
  ?*)
    continent=coord;;
  '')

    # Ask the user for continent or ocean.

    echo >&2 \
      'Please select a continent, ocean, "coord", "TZ", "time", or "now".'

    quoted_continents=$(
      $AWK '
	function handle_entry(entry) {
	  entry = substr(entry, 1, index(entry, "/") - 1)
	  if (entry == "America")
	    entry = entry "s"
	  if (entry ~ /^(Arctic|Atlantic|Indian|Pacific)$/)
	    entry = entry " Ocean"
	  printf "'\''%s'\''\n", entry
	}
	BEGIN {
	  TZ_ZONETABTYPE_TABLE = substr(ARGV[1], 2)
	  ARGV[1] = ""
	  FS = "\t"
	  nlines = split(TZ_ZONETABTYPE_TABLE, line, /\n/)
	  for (i = 1; i <= nlines; i++) {
	    $0 = line[i]
	    if ($0 ~ /^[^#]/)
	      handle_entry($3)
	    else if ($0 ~ /^#@/) {
	      ncont = split($2, cont, /,/)
	      for (ci = 1; ci <= ncont; ci++)
		handle_entry(cont[ci])
	    }
	  }
	}
      ' ="$TZ_ZONETABTYPE_TABLE" |
      sort -u |
      tr '\n' ' '
      echo ''
    )

    eval '
      doselect '"$quoted_continents"' \
	"coord - I want to use geographical coordinates." \
	"TZ - I want to specify the timezone using a proleptic TZ string." \
	"time - I know local time already." \
	"now - Like \"time\", but configure only for timestamps from now on."
      continent=$select_result
      case $continent in
      Americas) continent=America;;
      *)
	# Get the first word of $continent.  Path expansion is disabled
	# so this works even with "*", which should not happen.
	IFS=" "
	for continent in $continent ""; do break; done
	IFS=$newline;;
      esac
      case $zonetabtype,$continent in
      zonenow,*) ;;
      *,now)
	${TZ_ZONENOW_TABLE:+:} read_file TZ_ZONENOW_TABLE "$TZDIR/zonenow.tab"
	TZ_ZONE_TABLE=$TZ_ZONENOW_TABLE
      esac
    '
  esac

  case $continent in
  TZ)
    # Ask the user for a proleptic TZ string.  Check that it conforms.
    check_POSIX_TZ_string='
      BEGIN {
	tz = substr(ARGV[1], 2)
	ARGV[1] = ""
	tzname = ("(<[[:alnum:]+-][[:alnum:]+-][[:alnum:]+-]+>" \
		  "|[[:alpha:]][[:alpha:]][[:alpha:]]+)")
	sign = "[-+]?"
	hhmm = "(:[0-5][0-9](:[0-5][0-9])?)?"
	offset = sign "(2[0-4]|[0-1]?[0-9])" hhmm
	time = sign "(16[0-7]|(1[0-5]|[0-9]?)[0-9])" hhmm
	mdate = "M([1-9]|1[0-2])\\.[1-5]\\.[0-6]"
	jdate = ("((J[1-9]|[0-9]|J?[1-9][0-9]" \
		 "|J?[1-2][0-9][0-9])|J?3[0-5][0-9]|J?36[0-5])")
	datetime = ",(" mdate "|" jdate ")(/" time ")?"
	tzpattern = ("^(:.*|" tzname offset "(" tzname \
		     "(" offset ")?(" datetime datetime ")?)?)$")
	exit tz ~ tzpattern
      }
    '

    while
      echo >&2 'Please enter the desired value' \
	'of the TZ environment variable.'
      echo >&2 'For example, AEST-10 is abbreviated' \
	'AEST and is 10 hours'
      echo >&2 'ahead (east) of Greenwich,' \
	'with no daylight saving time.'
      read tz
      $AWK "$check_POSIX_TZ_string" ="$tz"
    do
      say >&2 "'$tz' is not a conforming POSIX proleptic TZ string."
    done
    TZ_for_date=$tz;;
  *)
    case $continent in
    coord)
      case $coord in
      '')
	echo >&2 'Please enter coordinates' \
	  'in ISO 6709 notation.'
	echo >&2 'For example, +4042-07403 stands for'
	echo >&2 '40 degrees 42 minutes north,' \
	  '74 degrees 3 minutes west.'
	read coord
      esac
      distance_table=$(
	$AWK \
	  "$output_distances_or_times" \
	  ="$coord" ="$TZ_COUNTRY_TABLE" ="$TZ_ZONE_TABLE" |
	sort -n |
	$AWK "{print} NR == $location_limit { exit }"
      )
      regions=$(
	$AWK '
	  BEGIN {
	    distance_table = substr(ARGV[1], 2)
	    ARGV[1] = ""
	    nlines = split(distance_table, line, /\n/)
	    for (nr = 1; nr <= nlines; nr++) {
	      nf = split(line[nr], f, /\t/)
	      print f[nf]
	    }
	  }
	' ="$distance_table"
      )
      echo >&2 'Please select one of the following timezones,'
      say >&2 "listed roughly in increasing order of distance from $coord."
      doselect $regions
      region=$select_result
      tz=$(
	$AWK '
	  BEGIN {
	    distance_table = substr(ARGV[1], 2)
	    region = substr(ARGV[2], 2)
	    ARGV[1] = ARGV[2] = ""
	    nlines = split(distance_table, line, /\n/)
	    for (nr = 1; nr <= nlines; nr++) {
	      nf = split(line[nr], f, /\t/)
	      if (f[nf] == region)
		print f[4]
	    }
	  }
	' ="$distance_table" ="$region"
      );;
    *)
      case $continent in
      now|time)
	minute_format='%a %b %d %H:%M'
	old_minute=$(TZ=UTC0 date +"$minute_format")
	for i in 1 2 3
	do
	  time_table_command=$(
	    $AWK \
	      -v output_times=1 \
	      "$output_distances_or_times" \
	      = = ="$TZ_ZONE_TABLE"
	  )
	  time_table=$(eval "$time_table_command")
	  new_minute=$(TZ=UTC0 date +"$minute_format")
	  case $old_minute in
	  "$new_minute") break
	  esac
	  old_minute=$new_minute
	done
	echo >&2 "The system says Universal Time is $new_minute."
	echo >&2 "Assuming that's correct, what is the local time?"
	sorted_table=$(say "$time_table" | sort -k2n -k2,5 -k1n) || {
	  say >&2 "$0: cannot sort time table"
	  exit 1
	}
	eval doselect $(
	  $AWK '
	    BEGIN {
	      sorted_table = substr(ARGV[1], 2)
	      ARGV[1] = ""
	      nlines = split(sorted_table, line, /\n/)
	      for (i = 1; i <= nlines; i++) {
		$0 = line[i]
		outline = $6 " " $7 " " $4 " " $5
		if (outline == oldline)
		  continue
		oldline = outline
		gsub(/'\''/, "&\\\\&&", outline)
		printf "'\''%s'\''\n", outline
	      }
	    }
	  ' ="$sorted_table"
	)
	time=$select_result
	continent_re='^'
	zone_table=$(
	  $AWK '
	    BEGIN {
	      time = substr(ARGV[1], 2)
	      time_table = substr(ARGV[2], 2)
	      ARGV[1] = ARGV[2] = ""
	      nlines = split(time_table, line, /\n/)
	      for (i = 1; i <= nlines; i++) {
		$0 = line[i]
		if ($6 " " $7 " " $4 " " $5 == time) {
		  sub(/[^\t]*\t/, "")
		  print
		}
	      }
	    }
	  ' ="$time" ="$time_table"
	)
	countries=$(
	  $AWK \
	    "$output_country_list" \
	    ="$continent_re" ="$TZ_COUNTRY_TABLE" ="$zone_table" |
	  sort -f
	)
	;;
      *)
	continent_re="^$continent/"
	zone_table=$TZ_ZONE_TABLE
      esac

      # Get list of names of countries in the continent or ocean.
      countries=$(
	$AWK \
	  "$output_country_list" \
	  ="$continent_re" ="$TZ_COUNTRY_TABLE" ="$zone_table" |
	sort -f
      )
      # If all zone table entries have comments, and there are
      # at most 22 entries, asked based on those comments.
      # This fits the prompt onto old-fashioned 24-line screens.
      regions=$(
	$AWK '
	  BEGIN {
	    TZ_ZONE_TABLE = substr(ARGV[1], 2)
	    ARGV[1] = ""
	    FS = "\t"
	    nlines = split(TZ_ZONE_TABLE, line, /\n/)
	    for (i = 1; i <= nlines; i++) {
	      $0 = line[i]
	      if ($0 ~ /^[^#]/ && !missing_comment) {
		if ($4)
		  comment[++inlines] = $4
		else
		  missing_comment = 1
	      }
	    }
	    if (!missing_comment && inlines <= 22)
	      for (i = 1; i <= inlines; i++)
		print comment[i]
	  }
	' ="$zone_table"
      )

      # If there's more than one country, ask the user which one.
      case $countries in
      *"$newline"*)
	echo >&2 'Please select a country' \
	  'whose clocks agree with yours.'
	doselect $countries
	country_result=$select_result
	country=$select_result;;
      *)
	country=$countries
      esac


      # Get list of timezones in the country.
      regions=$(
	$AWK '
	  BEGIN {
	    country = substr(ARGV[1], 2)
	    TZ_COUNTRY_TABLE = substr(ARGV[2], 2)
	    TZ_ZONE_TABLE = substr(ARGV[3], 2)
	    ARGV[1] = ARGV[2] = ARGV[3] = ""
	    FS = "\t"
	    cc = country
	    nlines = split(TZ_COUNTRY_TABLE, line, /\n/)
	    for (i = 1; i <= nlines; i++) {
	      $0 = line[i]
	      if ($0 !~ /^#/  &&  country == $2) {
		cc = $1
		break
	      }
	    }
	    nlines = split(TZ_ZONE_TABLE, line, /\n/)
	    for (i = 1; i <= nlines; i++) {
	      $0 = line[i]
	      if ($0 ~ /^#/)
		continue
	      if ($1 ~ cc)
		print $4
	    }
	  }
	' ="$country" ="$TZ_COUNTRY_TABLE" ="$zone_table"
      )

      # If there's more than one region, ask the user which one.
      case $regions in
      *"$newline"*)
	echo >&2 'Please select one of the following timezones.'
	doselect $regions
	region=$select_result
      esac

      # Determine tz from country and region.
      tz=$(
	$AWK '
	  BEGIN {
	    country = substr(ARGV[1], 2)
	    region = substr(ARGV[2], 2)
	    TZ_COUNTRY_TABLE = substr(ARGV[3], 2)
	    TZ_ZONE_TABLE = substr(ARGV[4], 2)
	    ARGV[1] = ARGV[2] = ARGV[3] = ARGV[4] = ""
	    FS = "\t"
	    cc = country
	    nlines = split(TZ_COUNTRY_TABLE, line, /\n/)
	    for (i = 1; i <= nlines; i++) {
	      $0 = line[i]
	      if ($0 !~ /^#/  &&  country == $2) {
		cc = $1
		break
	      }
	    }
	    nlines = split(TZ_ZONE_TABLE, line, /\n/)
	    for (i = 1; i <= nlines; i++) {
	      $0 = line[i]
	      if ($0 ~ /^#/)
		continue
	      if ($1 ~ cc && ($4 == region || !region))
		print $3
	    }
	  }
	' ="$country" ="$region" ="$TZ_COUNTRY_TABLE" ="$zone_table"
      )
    esac

    # Make sure the corresponding zoneinfo file exists.
    TZ_for_date=$TZDIR/$tz
    <"$TZ_for_date" || {
      say >&2 "$0: time zone files are not set up correctly"
      exit 1
    }
  esac


  # Use the proposed TZ to output the current date relative to UTC.
  # Loop until they agree in seconds.
  # Give up after 8 unsuccessful tries.

  extra_info=
  for i in 1 2 3 4 5 6 7 8
  do
    TZdate=$(LANG=C TZ="$TZ_for_date" date)
    UTdate=$(LANG=C TZ=UTC0 date)
    TZsecsetc=${TZdate##*[0-5][0-9]:}
    UTsecsetc=${UTdate##*[0-5][0-9]:}
    if test "${TZsecsetc%%[!0-9]*}" = "${UTsecsetc%%[!0-9]*}"
    then
      extra_info="
Selected time is now:	$TZdate.
Universal Time is now:	$UTdate."
      break
    fi
  done


  # Output TZ info and ask the user to confirm.

  echo >&2 ""
  echo >&2 "Based on the following information:"
  echo >&2 ""
  case $time%$country_result%$region%$coord in
  ?*%?*%?*%)
    say >&2 "	$time$newline	$country_result$newline	$region";;
  ?*%?*%%|?*%%?*%) say >&2 "	$time$newline	$country_result$region";;
  ?*%%%)	say >&2 "	$time";;
  %?*%?*%)	say >&2 "	$country_result$newline	$region";;
  %?*%%)	say >&2 "	$country_result";;
  %%?*%?*)	say >&2 "	coord $coord$newline	$region";;
  %%%?*)	say >&2 "	coord $coord";;
  *)		say >&2 "	TZ='$tz'"
  esac
  say >&2 ""
  say >&2 "TZ='$tz' will be used.$extra_info"
  say >&2 "Is the above information OK?"

  doselect Yes No
  ok=$select_result
  case $ok in
  Yes) break
  esac
do coord=
done

case $SHELL in
*csh) file=.login line="setenv TZ '$tz'";;
*)    file=.profile line="export TZ='$tz'"
esac

test -t 1 && say >&2 "
You can make this change permanent for yourself by appending the line
	$line
to the file '$file' in your home directory; then log out and log in again.

Here is that TZ value again, this time on standard output so that you
can use the $0 command in shell scripts:"

say "$tz"
workman.sh0000644000000000000000000000224615105202464010021 0ustar00#! /bin/sh
# Convert manual page troff stdin to formatted .txt stdout.

# This file is in the public domain, so clarified as of
# 2009-05-17 by Arthur David Olson.

manflags=
while
  case $1 in
    -*) :;;
    *) false;;
  esac
do
  manflags="$manflags $1"
  shift
done

groff="groff -dAD=l -rHY=0 $manflags -mtty-char -man -ww -P-bcou"
if ($groff) /dev/null 2>&1; then
  $groff "$@"
elif (type mandoc && type col) >/dev/null 2>&1; then
  mandoc $manflags -man "$@" | col -bx
elif (type nroff && type perl) >/dev/null 2>&1; then
  printf '%s\n' '.
.\" Left-adjust and do not hyphenate.
.am TH
.na
.hy 0
..
.\" Omit internal page headers and footers.
.\" Unfortunately this also omits the starting header and ending footer,
.\" but that is the best old nroff can easily do.
.rm }H
.rm }F
.' | nroff -man - "$@" | perl -ne '
	binmode STDIN, '\'':encoding(utf8)'\'';
	binmode STDOUT, '\'':encoding(utf8)'\'';
	chomp;
	s/.\010//g;
	s/\s*$//;
	if (/^$/) {
		$sawblank = 1;
		next;
	} else {
		if ($sawblank && $didprint) {
			print "\n";
			$sawblank = 0;
		}
		print "$_\n";
		$didprint = 1;
	}
  '
else
  printf >&2 '%s\n' "$0: please install groff, or mandoc and col"
  exit 1
fi