- tz - lists.iana.org

[PROPOSED] Fix some unsequenced/reproducible commentary
by Paul Eggert Dec. 21, 2025

Dec. 21, 2025

* private.h (ATTRIBUTE_CONST, ATTRIBUTE_PURE): In commentary do not say that __attribute__((const)) is stricter than [[unsequenced]], as Jens Gustedt, an originator of [[unsequenced]], wrote in <https://lists.gnu.org/r/bug-gnulib/2025-01/msg00190.html> that some const functions are not unsequenced. To be safe, make a similar change for __attribute__((pure)) and [[reproducible]], though I don’t have confirmation from Jens. --- private.h | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/private.h b/private.h index 074da284..50d6a52a 100644 --- a/private.h +++ b/private.h @@ -619,11 +619,11 @@ typedef unsigned long uintmax_t; # define ATTRIBUTE_UNSEQUENCED /* empty */ #endif -/* GCC attributes that are useful in tzcode. - __attribute__((const)) is stricter than [[unsequenced]], - so the latter is an adequate substitute in non-GCC C23 platforms. - __attribute__((pure)) is stricter than [[reproducible]], - so the latter is an adequate substitute in non-GCC C23 platforms. */ +/* 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 */ -- 2.51.0

2 5

TYPE_SIGNED(t) warnings..
by Christos Zoulas Dec. 19, 2025

Dec. 19, 2025

Hi, Gcc with -Wall -Wextra -Wtype-limits produces warnings with: #define TYPE_SIGNED(T) ((T)-1) < 0) error: comparison of unsigned expression in '< 0' is always false [-Werror=type-limits] Perhaps use something that checks the sign bit directly? #define TYPE_SIGNED(T) \ (((T)~(T)0 >> (sizeof(T) * CHAR_BIT - 1)) == (T)-1) Or something from NetBSD's <sys/cdefs.h>? christos

2 2

2025c zic.c calls fflush(3) on read only file pointer.
by Christos Zoulas Dec. 18, 2025

Dec. 18, 2025

Hello, This is undefined behavior. NetBSD returns EOF and sets errno to EBADFD. I am not sure why fflush(3) is called explicitly now, since fclose(3) does it anyway. Best, christos

6 7

[Feature Request] Add validation in zic for invalid DST intervals
by 任春辉 Dec. 17, 2025

Dec. 17, 2025

Hello tzdata maintainers, I would like to propose a feature enhancement for the zic compiler. Currently, the NOS time zone rules are defined as: Rule NOS 1970 2037 - Jan 27 23:43:01 2:0 SDST Rule NOS 1970 2037 - Jan 28 0:51:01 0 S Zone OS/NOS +7:0:0 NOS NO%s Issue: - The first rule enters DST at 23:43:01 with a +2 hour offset, moving local time to 01:43:01. - The second rule is intended to revert to standard time at 00:51:01. - Due to the offset, 00:51:01 local time never occurs, so the second rule will never trigger. - This creates unreachable local times and may cause incorrect results in time calculations. Proposal: - Add validation in the zic command to detect DST rules where the offset causes the subsequent rule's local time to be unreachable. - The validation could issue a warning or error to alert the author of potentially invalid rules. - This would improve the robustness of time zone definitions and prevent subtle bugs in downstream applications. Thank you for your consideration. Best regards

3 5

[PROPOSED] Make HTML view better on cell phones
by Paul Eggert Dec. 16, 2025

Dec. 16, 2025

Unfortunately without the meta viewport element described below, most mobile browers default to unreadably tiny text for backward compatibility with how the iPhone misbehaved in 2007. While we’re giving up on vanilla style, improve readability in minor ways for bigger screens too. Here are the changes made: * Add to HTML files the element <meta name="viewport" content="width=device-width, initial-scale=1">. This is the most common setting mentioned in <https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/Elements/meta/n…>. * Style pre elements with "overflow: auto", not "white-space: prewrap". * Decrease pre elements’ margins from 2em to 1.3rem, and change margins/padding for dd and ul to 1.3rem. * In tz-how-to.html, stop trying to package a wideish pre element inside a table, as that’s too painful to do portably and nicely on both narrow and wide screens. Just do a pre followed by a table. --- theory.html | 5 +++- tz-art.html | 4 +++ tz-how-to.html | 72 +++++++++----------------------------------------- tz-link.html | 5 +++- 4 files changed, 25 insertions(+), 61 deletions(-) diff --git a/theory.html b/theory.html index e6e64177..8c802858 100644 --- a/theory.html +++ b/theory.html @@ -3,8 +3,11 @@ <head> <title>Theory and pragmatics of the tz code and data</title> <meta charset="UTF-8"> + <meta name="viewport" content="width=device-width, initial-scale=1"> <style> - pre {margin-left: 2em; white-space: pre-wrap;} + dd {margin-left: 1.3rem;} + pre {margin-left: 1.3rem; overflow: auto;} + ul {padding-left: 1.3rem;} </style> </head> diff --git a/tz-art.html b/tz-art.html index 3039a86f..5f92fa1b 100644 --- a/tz-art.html +++ b/tz-art.html @@ -2,6 +2,10 @@ <html lang="en"> <head> <meta charset="UTF-8"> +<meta name="viewport" content="width=device-width, initial-scale=1"> +<style> +ul {padding-left: 1.3rem;} +</style> <title>Time and the Arts</title> </head> <body> diff --git a/tz-how-to.html b/tz-how-to.html index ccfdc9eb..ddafb2aa 100644 --- a/tz-how-to.html +++ b/tz-how-to.html @@ -3,14 +3,15 @@ <head> <title>How to Read the tz Database</title> <meta charset="UTF-8"> +<meta name="viewport" content="width=device-width, initial-scale=1"> <style> -pre {margin-left: 2em; white-space: pre-wrap;} -pre.td {margin-left: 0;} +pre {margin-left: 1.3rem; overflow: auto;} td {text-align: center;} table {border: 1px outset;} th, td {border: 1px inset;} table.rule {border: none; margin: auto;} td.footnote {text-align: left;} +ul {padding-left: 1.3rem;} </style> </head> <body> @@ -38,16 +39,7 @@ about the zones.</p> for Chicago (from the <code>northamerica</code> file in the <code>data</code> subdirectory):</p> -<table> -<tr> - <th colspan="6">From the Source File</th> -</tr> -<tr> - <td colspan="6"> - <table class="rule"> - <tr><td style="border:none;text-align:left"> -<pre class="td"> -#Rule NAME FROM TO - IN ON AT SAVE LETTER +<pre>#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 @@ -55,8 +47,7 @@ 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 </pre> - </td></tr></table></td> -</tr> +<table> <tr> <th colspan="6">Reformatted a Bit</th> </tr> @@ -160,16 +151,7 @@ time changed in 1955. Got it?</p> <p>OK, now for the somewhat more interesting “US” rules:</p> -<table> -<tr> - <th colspan="6">From the Source File</th> -</tr> -<tr> - <td colspan="6"> - <table class="rule"> - <tr><td style="border:none;text-align:left"> -<pre class="td"> -#Rule NAME FROM TO - IN ON AT SAVE LETTER/S +<pre>#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 @@ -184,8 +166,7 @@ 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 </pre> - </td></tr></table></td> -</tr> +<table> <tr> <th colspan="6">Reformatted a Bit</th> </tr> @@ -328,15 +309,7 @@ rule, so there should be no change.</li> <p>OK, now let’s look at a Zone record:</p> -<table> -<tr> - <th colspan="5">From the Source File</th> -</tr> -<tr> - <td colspan="5"> - <table class="rule"> - <tr><td style="border:none;text-align:left"> -<pre class="td"> +<pre> #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 @@ -347,8 +320,7 @@ Zone America/Chicago -5:50:36 - LMT 1883 Nov 18 12:09:24 -6:00 Chicago C%sT 1967 -6:00 US C%sT </pre> - </td></tr></table></td> -</tr> +<table> <tr> <th colspan="5">Columns Renamed</th> </tr> @@ -576,31 +548,14 @@ the true offset is undefined. <p>As a final example, here’s the complete history for Hawaii:</p> -<table> -<tr> - <th colspan="6">Relevant Excerpts from the US Rules</th> -</tr> -<tr> - <td colspan="6"> - <table class="rule"> - <tr><td style="border:none;text-align:left"> -<pre class="td"> +<pre># Relevant Excerpts from the US Rules</th> #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 -</pre> - </td></tr></table></td> -</tr> -<tr> - <th colspan="6">The Zone Record</th> -</tr> -<tr> - <td colspan="6"> - <table class="rule"> - <tr><td style="border:none;text-align:left"> -<pre class="td"> + +# 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 @@ -608,8 +563,7 @@ Zone Pacific/Honolulu -10:31:26 - LMT 1896 Jan 13 12:00 -10:30 US H%sT 1947 Jun 8 2:00 -10:00 - HST </pre> - </td></tr></table></td> -</tr> +<table> <tr> <th colspan="6">What We Infer</th> </tr> diff --git a/tz-link.html b/tz-link.html index ba10a68b..5e941a8d 100644 --- a/tz-link.html +++ b/tz-link.html @@ -3,8 +3,11 @@ <head> <title>Time zone and daylight saving time data</title> <meta charset="UTF-8"> +<meta name="viewport" content="width=device-width, initial-scale=1"> <style> -pre {margin-left: 2em; white-space: pre-wrap;} +dd {margin-left: 1.3rem;} +pre {margin-left: 1.3rem; overflow: auto;} +ul {padding-left: 1.3rem;} </style> </head> <body> -- 2.51.0

1 0

Re: [PROPOSED] Mention -Dst_ctim=M in CFLAGS
by Paul Eggert Dec. 12, 2025

Dec. 12, 2025

On 2025-12-11 22:08, Robert Elz wrote: > Prior to this (since inception) we (everything BSD derived) had > st_ctimespec ... you might want to give an example of an actual use > of -Dst_ctim=M and that might be a useful one to use, as it was so > common Thanks, good suggestion. Come to think of it, it's probably clearer to give only the st_ctimespec example and let any (probably zero) builders for obscure OS versions deduce the more-general case. I installed the attached further proposed patch.

1 0

[PROPOSED] Mention -Dst_ctim=M in CFLAGS
by Paul Eggert Dec. 11, 2025

Dec. 11, 2025

* Makefile: Add comment about using -Dst_ctim=M in CFLAGS. As far as I know this matters only on POSIXish platforms so old that they’re no longer supported by their suppliers, e.g., FreeBSD 8.4 (2013), old NetBSD, old UnixWare. --- Makefile | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/Makefile b/Makefile index eda8d53d..09f20100 100644 --- a/Makefile +++ b/Makefile @@ -284,8 +284,10 @@ LDLIBS= # -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 member st_ctim -# and code should use st_ctime instead+ +# -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 is called M rather than st_ctim, +# use -Dst_ctim=M 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 <sys/stat.h> does not work* -- 2.51.0

1 0

[tz-announce] 2025c release of tz code and data available
by Paul Eggert via tz-announce Dec. 11, 2025

Dec. 11, 2025

The 2025c release of the tz code and data is available. This release mostly changes code and commentary. The only changed data are leap second table expiration and pre-1976 time in Baja California. This release contains the following changes: 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 <sys/auxv.h> 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. 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 -TTHREAD_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 unportable 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”. Here are links to the release files: https://www.iana.org/time-zones/repository/releases/tzcode2025c.tar.gz https://www.iana.org/time-zones/repository/releases/tzdata2025c.tar.gz https://www.iana.org/time-zones/repository/releases/tzdb-2025c.tar.lz The following convenience links are also available, although they may point to the previous release until the relevant caches are refreshed: https://www.iana.org/time-zones/repository/tzcode-latest.tar.gz https://www.iana.org/time-zones/repository/tzdata-latest.tar.gz https://www.iana.org/time-zones/repository/tzdb-latest.tar.lz Links are also available via plain HTTP, and via FTP from ftp://ftp.iana.org/tz/releases with the same basenames as above. Each release file has a GPG signature, which can be retrieved by appending ".asc" to the above URLs. Copies of these signatures are appended to this message. This release corresponds to commit 7e1145bfdb9630c127841dc8ce808a937a300938 dated 2025-03-22 13:40:46 -0700 and tagged '2025c' in the development GitHub repository at <https://github.com/eggert/tz>. Here are the SHA-512 checksums for the release files: d14bc0680b9cb33d6526a2c0e483f2abafc2593cfe5fbfe7dcab940aafdc1baaec065dbaa1bb58b03586a8631ffdf43ebb22ff5a70503298b6ea2cf061d82c16 tzcode2025c.tar.gz 1e33f7212fd0ae2ad3c16e68f0c1fc7a6ad26a126b8406c379a5768d79604c6a816054bd0fe3a63228d70cd6a1fc2b1bae2a9f8014e102d3727eb9d21affa1f1 tzdata2025c.tar.gz 6fa233d6a884acd24b52c44b2eecf683f9403f44b39dcbe85b1ec92fc11f4daf6ee5143f05332faf9258b8bd8f7ca7d4882e19b1e3d892b4761be43fd4f39d51 tzdb-2025c.tar.lz Here are GPG digital signatures for the release files: -----BEGIN PGP SIGNATURE----- iQIzBAABCgAdFiEEfjeSqdis99YzvBWI7ZfpDmKqfjQFAmk5+AQACgkQ7ZfpDmKq fjS7FQ/+OcQWOkAheWXB17w3dzqEaBTzzOFPoVWJkYDYYqRBFYcZHt6SL8NsJ6Vr 4PRv8UXvGZZhyamQY6CA3Te6/SRBTn4vKAPPYrwRxRENBHJtwf00H4khIqzL4iuu Lni7d3MyEurnQ99YKdCdsDM7V2GqGov8SFOIQWohAazL29n/396Me68Excr7WYh2 FY1qmbqQhBuh9WrQ9+YznWkKd5FcPj0a8+XBReABUR3yV58qZ4rvDhDpx+476U09 5PlOHgC3FdEII+I8ihKI+Y0FW20dDfyTc7Eechef2z3izlzaxSFbAXXQX1gs1Vj+ yhjp+MiJC9UiFP0VmafGdpuhNWafrRSF8uyhHEAPC2WRRSlz3hUBFchOWRBpgAkT yIhN7dUWPPT80Pu2vWn3Y647I8397b6mS5o8tLpgcU9S82phwY3zVtsXG4WjVVC/ RcNCyG0J2ybP+IA8r/jPdINLNm+lsdJKoU6Y19R+BJyAV1uycltUEx4yzxs8Mg/8 lxvg+JcKe94UhUHIQXo3j9lXg7KcFXCVjGBBBpfBn4FfcRmdA+IkgONF3MxWz86J 8jBJdSVi2/CjlteWVQ2pBKEEesjXjT0tX8dbVA2x4KAcWYT5jMMQxDEP+Q2uuKfj NhE4D99CiZwdYgi4JP4+86cTDArj9l17tWR9bmVXgyviFU4o3Go= =R12Y -----END PGP SIGNATURE----- -----BEGIN PGP SIGNATURE----- iQIzBAABCgAdFiEEfjeSqdis99YzvBWI7ZfpDmKqfjQFAmk5+AUACgkQ7ZfpDmKq fjTtEg//VloILc1FvXSuFccnpY8G7rmCkL48q5628VbogfqLc3MSQ8FqlfBw3OkI LqiDAF22/nFcMhtQ9j61cFJ5GrAUF28MIKDl1ol5YEswQgiLxDxFbK2c5uWdYJed 4I5bqU8rs6RTA/5ZZ/MMDe7JIGEYClZhHzT9Hs9vLrg2aOW7vSzXjjuAy7FqUwpX k1m6KQtL6UPl4tkbDY9wLSifDtg6GqfZT9bYtPxxLKc89Mm0tFJP7aXuooELb0/k 1YsCFZDU7OK8ttgAlPbtAebfrFJbk1jKuBQqoqyH59QtA4OegHFjRVPFcHYevJsl FMhA10XEXS1C3T2KC9De9Xypt1aJIVIhtEHz7GjlubZRfgLGb+xAvTsJj2h20TIm mTAaJ9tBdr7TiVRlLdUgWdsLRwrVSOLbFCjL5Fa2+XpTNoowPmDp7YHmFsk+itDH 40fRFvBc842huCIsFbQ1EInjVpuUQqaMb1CmZlNMhJz6DoXEicOoo9iwqjRxZyVt 7NB4AQaqBwAKgvTw7iV4Iw0naWaKhwN9pP2AbnC5cMd3SRMneJ6Sxt4MtA9dqpFB mKIMF9Tlb+twuyOZXezaD/RqJs19gusdFmVlpJbNzez//pYMJ/9LcjB+xz61c5Zf mIPWE0mkzSRvhqca4L3inlC73aTRDp2OEt9+n+AJPnk27BRe5Xs= =1kyi -----END PGP SIGNATURE----- -----BEGIN PGP SIGNATURE----- iQIzBAABCgAdFiEEfjeSqdis99YzvBWI7ZfpDmKqfjQFAmk5+AYACgkQ7ZfpDmKq fjRXzA/9GHRwVUZRXXmlwgpPZDVU5BtB2xrcJVQk3pFLs0J6cz+nUVRtsm3Gz400 qaksn6T+azBX5Lv6pSPYmDtj+cSbCA7vcaLxHh414gKu6SKkNTm566xGM3RIKome odhacZtLvPYpfONIGt0YRgwl6oO+Q1dXuu4aoi2H4jdYrHEjpTFY1Pg1TfRrpM/p JxarI/VUOvuXZHAsJRBJshEc5GcOnaUAzYKr/yAelbdowjW2DLetRhWU0YZ9zgZk w6xaVNvcU0i9fX2Ri0MMiTpo92xzRdZd1WZ8wfU3hqUMfolVSMJuk8Qs/NyYm8v9 DMtiQb0Ewi10Y0YyGUnQi3r6wpKQ2K1D2JVZj5eGLsbTfBpi+3bilJHtxBc05NaK HQyZhhgXSTwLkR4LUcsyTtWdiS9uUB2D18EtfxODWoeJsIcoU2fRPX3zVIn0EjTq YLSrGvbk69xAiyhDQlcYFEUoL6GQb0WbDOz1b1IwRJ6de7YcNkkCAlWTt4qIat65 6OfZURSGURRWd9BBNqF2m1OgaGfv4fn/HJL4TYYaEXATpTUxVoPujxl73In5Mu3O A8zVP2QouGD25RYmoQLdAYMyap9XJIS1HWRNBPAyRC8z5YZE1ptm+L0KzBD9+vwe rTejW79L/QMsu9Orz6Mnptraz6C83q7C1hi7STqwNPd4rE8grV8= =MFmx -----END PGP SIGNATURE-----

1 1

Question about leap-seconds.list expiry
by Marcus van Dam Dec. 11, 2025

Dec. 11, 2025

Hi, I hope this is the appropriate list for questions about the leap-seconds file. It seems the `leap-seconds.list` file hosted on the IANA server [1] is about to expire on 28 December 2025, and its latest update NTP timestamp (01 July 2025) appears to be older than the timestamp on the upstream bulletin [2] (07 July 2025). The obspm.fr servers also appear to host a leap-seconds.list file [3] with an expiry date of 28 June 2026. With the current expiry date approaching, I was wondering if the IANA-hosted file will be updated to have an expiry date in the future. As a side note, the leap-seconds.list file shipped in tzdata on Debian [4] already seems to use the 2026 expiry date. [1] https://data.iana.org/time-zones/tzdb/leap-seconds.list [2] https://hpiers.obspm.fr/iers/bul/bulc/bulletinc.dat [3] https://hpiers.obspm.fr/iers/bul/bulc/ntp/leap-seconds.list [4] https://sources.debian.org/src/tzdata/2025b-0%2Bdeb12u2/leap-seconds.list -- Kind regards Marcus van Dam

3 2

[PROPOSED] Say more clearly that the tz mailing list is public
by Paul Eggert Dec. 10, 2025

Dec. 10, 2025

* CONTRIBUTING, africa, asia, australasia, backzone, europe, * northamerica, southamerica, theory.html, tz-art.html * tz-link.html: Say that the tz mailing list and its archive are public, and that people should not contribute confidential information. --- CONTRIBUTING | 6 +++++- africa | 6 +++--- asia | 6 +++--- australasia | 6 +++--- backzone | 8 ++++---- europe | 8 ++++---- northamerica | 6 +++--- southamerica | 6 +++--- theory.html | 11 ++++++++--- tz-art.html | 7 +++++-- tz-link.html | 12 +++++++++--- 11 files changed, 50 insertions(+), 32 deletions(-) diff --git a/CONTRIBUTING b/CONTRIBUTING index 1f064fa0..c5fa803f 100644 --- a/CONTRIBUTING +++ b/CONTRIBUTING @@ -10,7 +10,8 @@ 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 <tz(a)iana.org> for use in the future. In your email, please give -reliable sources that reviewers can check. +reliable sources that reviewers can check. The mailing list and its +archives are public, so please do not send confidential information. ## Contributing technical changes @@ -84,6 +85,9 @@ If you use Git the following workflow may be helpful: 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 <tz(a)iana.org> for others to review. diff --git a/africa b/africa index 11738a1e..ec067bee 100644 --- a/africa +++ b/africa @@ -4,9 +4,9 @@ # 2009-05-17 by Arthur David Olson. # This file is by no means authoritative; if you think you know better, -# go ahead and edit the file (and please send any changes to -# tz(a)iana.org for general use in the future). For more, please see -# the file CONTRIBUTING in the tz distribution. +# go ahead and edit the file, and please send any changes to +# the public mailing list tz(a)iana.org for general use in the future. +# For more, please see the file CONTRIBUTING in the tz distribution. # From Paul Eggert (2018-05-27): # diff --git a/asia b/asia index d3d1059a..dd76bdfc 100644 --- a/asia +++ b/asia @@ -4,9 +4,9 @@ # 2009-05-17 by Arthur David Olson. # This file is by no means authoritative; if you think you know better, -# go ahead and edit the file (and please send any changes to -# tz(a)iana.org for general use in the future). For more, please see -# the file CONTRIBUTING in the tz distribution. +# go ahead and edit the file, and please send any changes to +# the public mailing list tz(a)iana.org for general use in the future. +# For more, please see the file CONTRIBUTING in the tz distribution. # From Paul Eggert (2019-07-11): # diff --git a/australasia b/australasia index 5cf8f117..e52aab10 100644 --- a/australasia +++ b/australasia @@ -937,9 +937,9 @@ Zone Pacific/Efate 11:13:16 - LMT 1912 Jan 13 # Vila # NOTES # This file is by no means authoritative; if you think you know better, -# go ahead and edit the file (and please send any changes to -# tz(a)iana.org for general use in the future). For more, please see -# the file CONTRIBUTING in the tz distribution. +# go ahead and edit the file, and please send any changes to +# the public mailing list tz(a)iana.org for general use in the future. +# For more, please see the file CONTRIBUTING in the tz distribution. # From Paul Eggert (2018-11-18): # diff --git a/backzone b/backzone index dd7a8675..3c7e720a 100644 --- a/backzone +++ b/backzone @@ -2,10 +2,10 @@ # This file is in the public domain. -# This file is by no means authoritative; if you think you know -# better, go ahead and edit it (and please send any changes to -# tz(a)iana.org for general use in the future). For more, please see -# the file CONTRIBUTING in the tz distribution. +# This file is by no means authoritative; if you think you know better, +# go ahead and edit the file, and please send any changes to +# the public mailing list tz(a)iana.org for general use in the future. +# For more, please see the file CONTRIBUTING in the tz distribution. # When proposing changes to this file, please use 'git format-patch' # format, either by attaching the resulting .patch file to your email, diff --git a/europe b/europe index 4d7e1f3b..af521710 100644 --- a/europe +++ b/europe @@ -4,9 +4,9 @@ # 2009-05-17 by Arthur David Olson. # This file is by no means authoritative; if you think you know better, -# go ahead and edit the file (and please send any changes to -# tz(a)iana.org for general use in the future). For more, please see -# the file CONTRIBUTING in the tz distribution. +# go ahead and edit the file, and please send any changes to +# the public mailing list tz(a)iana.org for general use in the future. +# For more, please see the file CONTRIBUTING in the tz distribution. # From Paul Eggert (2017-02-10): # @@ -42,7 +42,7 @@ # <https://www.jstor.org/stable/1774359>. He writes: # "It is requested that corrections and additions to these tables # may be sent to Mr. John Milne, Royal Geographical Society, -# Savile Row, London." Nowadays please email them to tz(a)iana.org. +# Savile Row, London." Nowadays please see the file CONTRIBUTING. # # Byalokoz EL. New Counting of Time in Russia since July 1, 1919. # This Russian-language source was consulted by Vladimir Karpinsky; see diff --git a/northamerica b/northamerica index bb3d1b43..16b79af2 100644 --- a/northamerica +++ b/northamerica @@ -6,9 +6,9 @@ # also includes Central America and the Caribbean # This file is by no means authoritative; if you think you know better, -# go ahead and edit the file (and please send any changes to -# tz(a)iana.org for general use in the future). For more, please see -# the file CONTRIBUTING in the tz distribution. +# go ahead and edit the file, and please send any changes to +# the public mailing list tz(a)iana.org for general use in the future. +# For more, please see the file CONTRIBUTING in the tz distribution. # From Paul Eggert (1999-03-22): # A reliable and entertaining source about time zones is diff --git a/southamerica b/southamerica index 06fee598..bcd4abdf 100644 --- a/southamerica +++ b/southamerica @@ -4,9 +4,9 @@ # 2009-05-17 by Arthur David Olson. # This file is by no means authoritative; if you think you know better, -# go ahead and edit the file (and please send any changes to -# tz(a)iana.org for general use in the future). For more, please see -# the file CONTRIBUTING in the tz distribution. +# go ahead and edit the file, and please send any changes to +# the public mailing list tz(a)iana.org for general use in the future. +# For more, please see the file CONTRIBUTING in the tz distribution. # From Paul Eggert (2016-12-05): # diff --git a/theory.html b/theory.html index cdc6093a..6e52a929 100644 --- a/theory.html +++ b/theory.html @@ -1584,9 +1584,14 @@ Sources for time on other planets: </section> <footer> - <hr> - This file is in the public domain, so clarified as of 2009-05-17 by - Arthur David Olson. +<hr> +This web page is in the public domain, so clarified as of +2009-05-17 by Arthur David Olson. +<br> +Please send corrections to this web page to the +<a href="mailto:tz@iana.org">time zone mailing list</a>. +The mailing list and its archives are public, +so please do not send confidential information. </footer> </body> </html> diff --git a/tz-art.html b/tz-art.html index 01a00297..3039a86f 100644 --- a/tz-art.html +++ b/tz-art.html @@ -611,13 +611,16 @@ but maybe in your time zone I haven’t finished yet. So stay tuned!” <li><a href="tz-link.html">Time Zone and Daylight Saving Time Data</a></li> </ul> + +<footer> <hr> -<address> This web page is in the public domain, so clarified as of 2009-05-17 by Arthur David Olson. <br> Please send corrections to this web page to the <a href="mailto:tz@iana.org">time zone mailing list</a>. -</address> +The mailing list and its archives are public, +so please do not send confidential information. +</footer> </body> </html> diff --git a/tz-link.html b/tz-link.html index 669a3cb8..9267fb6f 100644 --- a/tz-link.html +++ b/tz-link.html @@ -211,7 +211,10 @@ location.</p> The <code><abbr>tz</abbr></code> code and data are by no means authoritative. If you find errors, please email changes to <a href="mailto:tz@iana.org"><code>tz(a)iana.org</code></a>, -the time zone mailing list. See +the time zone mailing list. +The mailing list and its archives are public, +so please do not send confidential information. +See <a href="https://lists.iana.org/postorius/lists/tz.iana.org/">the mailing list’s main page</a> to subscribe or to browse its archive of old messages. <a href="https://tzdata-meta.timtimeonline.com">Metadata for mailing list @@ -283,8 +286,9 @@ collaboration among governments, the time zone database volunteer community, and data distributors downstream. <p> If your government plans to change its time zone boundaries or -daylight saving rules, please send email to <a -href="mailto:tz@iana.org"><code>tz(a)iana.org</code></a> well in advance, +daylight saving rules, please send email as described in +"<a href="#changes">Changes to the <code><abbr>tz</abbr></code> database</a>". +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 @@ -1250,6 +1254,8 @@ This web page is in the public domain, so clarified as of <br> Please send corrections to this web page to the <a href="mailto:tz@iana.org">time zone mailing list</a>. +The mailing list and its archives are public, +so please do not send confidential information. </footer> </body> </html> -- 2.51.0

1 0

An odd juxtaposition...
by Arthur Olson Dec. 10, 2025

Dec. 10, 2025

...of sex appeal and time zones at the Hyattsville, Maryland Mall at Prince George's: a watch kiosk called "Time Zone 2" by a Victoria's Secret store. An image is part of the social media posting linked to below. @dashdashado https://bsky.app/profile/dashdashado.bsky.social/post/3m7ld53qtj22s

3 2

Greenland parliament to vote February on time zone change
by Paul Eggert Dec. 8, 2025

Dec. 8, 2025

Hanna Hviid reported in Sermitsiaq last month[1] (in Danish) that Inatsisartut, Greenland's parliament, plans to debate in February whether to revert the 2023 change to Greenland's clocks. It seems like there's a good chance something will pass. Although the details are unclear, presumably this would affect clocks starting in fall 2026. I found out about this by running across last month's "Iceland’s Old Timezone Controversy Echoes in Greenland"[2], by the Iceland Review's Michael Chapman. [1]: https://www.sermitsiaq.ag/samfund/pele-beskylder-udvalg-for-at-forhale-over… [2]: https://www.icelandreview.com/news/icelands-old-timezone-controversy-echoes…

2 3

Request for Clarification or Alternative Label for “Indian/Mauritius” Time Zone
by James Lai Dec. 4, 2025

Dec. 4, 2025

Hello, I hope you are well. I am writing regarding the naming of the IANA time zone “Indian/Mauritius.” Many users in Mauritius have reported confusion because the word “Indian” is often interpreted as referring to India rather than the Indian Ocean region. This creates misunderstandings when scheduling international meetings, especially on platforms like Zoom, Google Calendar, and others that rely directly on IANA time zone identifiers. I understand that renaming time zones is avoided due to compatibility concerns. However, I would like to kindly ask whether: 1. An alternative label such as “IndianOcean/Mauritius” could be considered, or 2. A documented alias or clarification could be added to help reduce confusion for end users. Even a non-breaking solution (e.g., an alias) would greatly help clarify the intended meaning without disrupting existing systems. Thank you for your time and for maintaining this important global standard. Kind regards, Kind regards, [cid:asset3_0cde460c-bb22-4017-a4dd-b19716304dc9.png] James Lai IT | Head of Operations Intercontinental Trust Limited Level 3, Alexander House, 35 Cybercity, Ebene 72201, Mauritius E: jlai(a)intercontinentaltrust.com<mailto:jlai@intercontinentaltrust.com> W: www.intercontinentaltrust.com<http://www.intercontinentaltrust.com/> T: (230) 403 0800 DID: (230) 403 0869 M: (230) 5258 5889 F: (230) 403 0801 BRN: C07023546 [cid:itl_emailbanner_03.10.25v2_d5878fa8-0af2-4ccb-b63e-55a10f1a1c87.jpg] Intercontinental Trust Ltd is regulated by the Financial Services Commission in Mauritius. If you are not the intended recipient, please advise the sender immediately and delete this message. [cid:facebooklogo_9db4cba5-e3da-4106-9849-2d5d2e6d37db.gif]<https://www.facebook.com/IntercontinentalTrust?fref=ts> [cid:LinkedInlogo_86f1ad59-c943-4f8d-a70a-699b2322ca7e.gif] <http://www.linkedin.com/company/2502998?trk=pro_other_cmpy> [cid:Twitterlogo_3b4b7685-ec4d-4906-8971-931729a1de68.gif] <https://twitter.com/ITLMauritius> Disclaimer<http://intercontinentaltrust.com/disclaimer> Please consider the environment before you print

8 10

backward file for zonenow.tab?
by kirill.gagarski＠gmail.com Nov. 17, 2025

Nov. 17, 2025

Hi. Our software tries to expose a list of supported timezones to an end user for setting up some periodic scheduled task in the UI. Till now we've tried to populate the list of Java supported timezones, filtering out aliases (from backward file). I noticed that zonedata now ships with zonenow.tab file which is exactly what we need for our UI and meets our needs more precisely. Unfortunately it seems quite impossible to migrate old data to this new notation automatically. For example, hte fact that 'Europe/Berlin' is covered by 'Europe/Paris' in zonenow.tab is not mentioned anywhere. Is there a way to determine this fact from tzdata without deep analysis of the rules done by myself? Other simple question: is there a way to know what is "now" in zonenow? Is it defined by the release date of a new version?

4 6

Calendrical Issues Russia section
by Kirill Gagarski Nov. 14, 2025

Nov. 14, 2025

I fully understand that I am trying to engage into a 30 y.o discussion now, but the Russia section in calendars file repeats quite widespread myths about Eternal Calendar with 30th February. I used to believe in that myself, given quite a little information on the internet about the issue. Even Wikipedia spread this myth back in the day. Even the Russian version of Wikipedia spreaded the myth: https://ru.wikipedia.org/w/index.php?title=%D0%9D%D0%B5%D0%BF%D1%80%D0%B5%D… Nowadays, Wikipedia is more clean about the issue, providing some source links: https://ru.wikipedia.org/wiki/%D0%9D%D0%B5%D0%BF%D1%80%D0%B5%D1%80%D1%8B%D0… https://en.wikipedia.org/wiki/Soviet_calendar The truth is: - Gregorian calendar was continuously used during the mentioned time - 30th February in USSR is an absolute myth (to be more clear, this is a declined proposal of a new calendar) - Mentioned 5-day weeks and 6-day weeks were in fact an attempt to relabel weekdays (Monday-Sunday) - Relabeling was done in a way, not conflicting with traditional weekdays (they used words like "third day of six-week day" or colors). Traditional names were used in parallel I could not say, the calendar file does not say the truth (after all, it raises doubts in a second quoted message), and all this information is indeed very interesting from a historical point of view, but does not seem relevant to Calendrical Issues. In fact 100 years later no one (unless someone is specifically interested in these weird weekdays) cares if 12th of December of 1937 was "sixth day of six-day week", because it was Sunday at the same time. I'd suggest removing the whole Russia section from this file (you can keep the small note about switching from Julia to Gregorian calendar in 1918), since for the purpose of it it raises nothing but confusion.

2 2

[PROPOSED] Fix unlikely backslash bug in scripts
by Paul Eggert Nov. 12, 2025

Nov. 12, 2025

* tzselect.ksh, workman.sh: Use printf, not echo, if the argument might in theory contain a backslash. --- tzselect.ksh | 9 +++++---- workman.sh | 2 +- 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/tzselect.ksh b/tzselect.ksh index 7cd8679d..85a8c670 100644 --- a/tzselect.ksh +++ b/tzselect.ksh @@ -145,9 +145,11 @@ do t*) # Undocumented option, used for developer testing. zonetabtype=$OPTARG;; -help) - exec echo "$usage";; + say "$usage" + exit;; -version) - exec echo "tzselect $PKGVERSION$TZVERSION";; + say "tzselect $PKGVERSION$TZVERSION" + exit;; -*) say >&2 "$0: -$opt$OPTARG: unknown option; try '$0 --help'"; exit 1;; *) @@ -516,8 +518,7 @@ while ' ="$distance_table" ) echo >&2 'Please select one of the following timezones,' - echo >&2 'listed roughly in increasing order' \ - "of distance from $coord". + say >&2 "listed roughly in increasing order of distance from $coord." doselect $regions region=$select_result tz=$( diff --git a/workman.sh b/workman.sh index 91621b10..5da55654 100644 --- a/workman.sh +++ b/workman.sh @@ -51,6 +51,6 @@ elif (type nroff && type perl) >/dev/null 2>&1; then } ' else - echo >&2 "$0: please install groff, or mandoc and col" + printf >&2 '%s\n' "$0: please install groff, or mandoc and col" exit 1 fi -- 2.51.0

1 0

Melbourne's Shrine of Remembrance
by Tim Parenti Nov. 11, 2025

Nov. 11, 2025

Earlier this year, Julian O'Shea released a short video detailing a peculiarity about the Shrine of Remembrance in Melbourne, "How Daylight Savings Broke this $24 Million Building": https://www.youtube.com/watch?v=mfzsBMUiGGQ The shrine, originally dedicated in 1934 to honor those who served in World War I, was designed with a small aperture in the roof. Each Remembrance Day (11 November) at 11:00am local time, the position of the sun would be perfectly aligned to shine through the aperture, allowing the light of its rays to fall upon an inscribed stone inside, thus commemorating the hour of the Armistice. The problem? Several decades later, in 1971, Victoria adopted DST, but of course the aperture had been tuned to the position of the sun at 11:00am AEST. So what do you do to preserve this meaningful element of an annual ceremony with important dignitaries, when it's effectively been moved an hour earlier to 11:00am AEDT instead? The answer involves a clever system of mirrors. (And the general public still gets to witness the light following its true path an hour later!) Since this is quite a grand physical representation of how we interact with our timekeeping, I've added this to tz-art.html, as well as a few fixups to unclosed HTML tags nearby. -- Tim Parenti

4 3

Cape Town
by Justin Singh Nov. 11, 2025

Nov. 11, 2025

Hello The Timezone for Cape Town is UTC+01:00 The western Northern Cape and Western Cape differ, however. Everywhere on land west of 22°30′ E effectively experiences year-round daylight saving time because of its location in true UTC+01:00 but still being in South African Standard Time. Sunrise and sunset are thus relatively late in Cape Town, compared to the rest of the country.

6 7

Nunavik to stop changing its clocks after March 2026?
by Paul Eggert Nov. 8, 2025

Nov. 8, 2025

Yesterday’s CBC News report “Nunavimmiut have wound back their clocks for the last time”[1] is so light on details that for now I am merely attaching a draft patch to this email for review, rather than installing the patch into the development repository. I found no announcement of the change at www.makivvik.ca. Although I asked for more details by using the form at <https://www.makivvik.ca/contact/>, these sorts of contact forms often do not work, so if anyone has access to more-authoritative info than the CBC report, please let us know. Among other things, details in the news report make me wonder how likely Makivvik would reconsider its decision before the change would take effect about a year from now. [1]: https://www.cbc.ca/news/canada/north/nunavik-daylight-savings-9.6969650

1 0

Phoenix AZ timezone issue
by akbar＠marketing-qa.com Nov. 5, 2025

Nov. 5, 2025

🧭 Problem Summary Arizona remains one of the few U.S. states that does not observe Daylight Saving Time (DST) since 1968. However, it identifies its time zone as Mountain Standard Time (MST) — which does change to MDT when DST applies elsewhere. Because of this, every spring and fall, Arizona’s effective time alignment shifts: In summer, Arizona matches Pacific Daylight Time (PDT). In winter, it matches Mountain Standard Time (MST). This causes systemic issues in digital calendars, automated scheduling, flight bookings, and cross-state business coordination, requiring frequent manual adjustments. 💡 Recommended Solution Create and declare a dedicated time zone: “Phoenix Standard Time.” This would be a fixed, non-DST time zone officially recognized by IANA (the Internet Assigned Numbers Authority) and major platforms (Microsoft, Apple, Google). All digital devices could use it as a unique time zone identifier that never shifts with DST — eliminating confusion and rescheduling errors. It would function similarly to “UTC+7” year-round but would display as “Phoenix Standard Time (PST)” in software systems. Thank you Akbar Jaffer Akbar Jaffer Office: (415) 335-6950 Mobile: (650) 430-0232 Web: Marketing-QA.com <http://www.marketing-qa.com/> Meeting: Book time W/Akbar <https://calendly.com/ajaffer/> LinkedIn <https://www.linkedin.com/in/akbarjaffer/> | POI Blog <https://pointofintersection.org/>

8 9

How to select zones for a TimeZone picker
by Guillermo Rodriguez Garcia Nov. 3, 2025

Nov. 3, 2025

Hi all, I'm building a TimeZone picker with the following constraints: 1. At least one distinct timezone per country (so that e.g. there is an entry for Atlantic/Reykjavik instead of mapping IS to Africa/Abidjan) 2. Within a country, do not show timezones that currently have identical rules to another timezone in the same country (i.e. historical differences don't matter). Because of (1) I started from zone.tab rather than zone1970.tab. However, I'm not sure how to handle (2). For example, Spain has three zones in zone.tab: Europe/Madrid, Africa/Ceuta, and Atlantic/Canary. But as far as most users are concerned, as of today there are only two timezones in Spain: Madrid and Canary Islands. Ceuta follows Madrid rules, so there should be no need to show this as an option to the user. Another example: the U.S. lists 29 zones, but only 8 are distinct today (ignoring historical changes). So, the question: Any suggestions on how to filter timezones within each country to drop those that as of today are not relevant except for historical differences? Thank you, Guillermo Rodriguez Garcia guille.rodriguez(a)gmail.com

8 17

[PROPOSED] tzfree now preserves errno
by Paul Eggert Nov. 3, 2025

Nov. 3, 2025

* Makefile, NEWS, newtzset.3: Mention this. * localtime.c (FREE_PRESERVES_ERRNO): Default to 1 if 'free' is known to preserve errno, 0 otherwise. (tzfree): Preserve errno if 'free' does not. --- Makefile | 2 ++ NEWS | 2 ++ localtime.c | 15 +++++++++++++++ newtzset.3 | 4 ++++ 4 files changed, 23 insertions(+) diff --git a/Makefile b/Makefile index 0fe8a9f8..0d33fa17 100644 --- a/Makefile +++ b/Makefile @@ -239,6 +239,8 @@ LDLIBS= # -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 <time.h> does not declare asctime_r # on POSIX platforms predating POSIX.1-2024 # -DHAVE_DECL_ENVIRON if <unistd.h> declares 'environ' diff --git a/NEWS b/NEWS index 1ed9b2eb..a9b2d321 100644 --- a/NEWS +++ b/NEWS @@ -93,6 +93,8 @@ Unreleased, experimental changes 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. diff --git a/localtime.c b/localtime.c index b85941f2..46acf3bd 100644 --- a/localtime.c +++ b/localtime.c @@ -2061,10 +2061,25 @@ tzalloc(char const *name) 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; } /* diff --git a/newtzset.3 b/newtzset.3 index 3801ab1b..028cfd25 100644 --- a/newtzset.3 +++ b/newtzset.3 @@ -379,6 +379,10 @@ If successful, the 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 -- 2.51.0

1 0

[PROPOSED] Improve TZ="/..." security and speed (NetBSD only)
by Paul Eggert Oct. 28, 2025

Oct. 28, 2025

* NEWS: Mention security. * localtime.c (O_REGULAR): Default to 0. (tzloadbody): Open TZif files with O_REGULAR. Suppress unnecessary call to 'stat' if O_REGULAR is present. --- NEWS | 3 +++ localtime.c | 7 +++++-- 2 files changed, 8 insertions(+), 2 deletions(-) diff --git a/NEWS b/NEWS index 64d32bf3..1ed9b2eb 100644 --- a/NEWS +++ b/NEWS @@ -40,6 +40,9 @@ Unreleased, experimental changes 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. diff --git a/localtime.c b/localtime.c index 0d2d2a93..b85941f2 100644 --- a/localtime.c +++ b/localtime.c @@ -355,6 +355,9 @@ static int openat(int dd, char const *path, int oflag) { unreachable (); } #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 @@ -900,7 +903,7 @@ tzloadbody(char const *name, struct state *sp, char tzloadflags, 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_IGNORE_CTTY | O_NOCTTY | O_REGULAR); int err; struct stat st; st.st_ctime = 0; @@ -927,7 +930,7 @@ tzloadbody(char const *name, struct state *sp, char tzloadflags, continue; else if (issetugid()) return ENOTCAPABLE; - else { + else if (!O_REGULAR) { /* Check for devices, as their mere opening could have unwanted side effects. Though racy, there is no portable way to fix the races. This check is needed -- 2.48.1

1 0

Iana build failures on Windows
by Manuela Friedrich Oct. 28, 2025

Oct. 28, 2025

Hello Paul, I have hit a compile error for zic.c on Windows. Git commit 9e0bf0b42d68b3d9ad0bc7ff70fc92d87f8b40d2 zic.c zic.c(152): error C2081: 'mode_t': name in formal parameter list illegal zic.c(2437): warning C4116: unnamed type definition in parentheses NMAKE : U1077: '"C:\Program Files (x86)\Microsoft Visual Studio\2017\Professional\VC\Tools\MSVC\14.12.25827\bin\HostX64\x64\cl.EXE"' : return code '0x2' Stop. Can you please advise? Please do not release a new version with this issue present. Thank you! Regards Manuela Friedrich

4 17

[PROPOSED] Be more like OpenBSD wrt ".."
by Paul Eggert Oct. 25, 2025

Oct. 25, 2025

* localtime.c (tzloadbody): Do not worry about settings like TZ=".." (i.e., trailing ".." component) as the later read will fail anyway. This simplifies the code slightly, and agrees with the security-conscious OpenBSD libc. --- localtime.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/localtime.c b/localtime.c index f1a82370..0d2d2a93 100644 --- a/localtime.c +++ b/localtime.c @@ -941,13 +941,13 @@ tzloadbody(char const *name, struct state *sp, char tzloadflags, if (relname[0] != '/') { if (!OPENAT_TZDIR || !O_RESOLVE_BENEATH) { - /* Fail if a relative name contains a ".." component, - as such a name could read a file outside TZDIR + /* 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[2]) + && component[2] == '/' && (component == relname || component[-1] == '/')) return ENOTCAPABLE; } -- 2.48.1

1 0