- tz - lists.iana.org

[PROPOSED] Fix pre-C23 typos in recent overflow fixes
by Paul Eggert Jan. 21, 2026

Jan. 21, 2026

* localtime.c (time2sub) [!defined ckd_sub]: Fix typos in Friday’s “Fix remaining Naveed8951-reported overflows” changes. This is needed for GCC 13 and earlier, and for other pre-C23 compilers that lack <stdckdint.h>. --- localtime.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/localtime.c b/localtime.c index 698f61bf..40a4d7ee 100644 --- a/localtime.c +++ b/localtime.c @@ -2697,8 +2697,8 @@ time2sub(struct tm *const tmp, 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 <= yourm.TM_GMTOFF - && your.TM_GMTOFF <= TWO_31_MINUS_1); + 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 -- 2.52.0

1 0

[PROPOSED] Support slightly-larger TZif files
by Paul Eggert Jan. 21, 2026

Jan. 21, 2026

* NEWS: Mention this. * localtime.c (tzloadbody): Fix off-by-one errors in tests that mistakenly caused the code to reject TZif files that actually fit. Problem introduced in release 2014g, by my commit 13a277175c603cfb62ae5bdcf26e23faf5619ef1 dated 2014-08-19. --- NEWS | 3 +++ localtime.c | 12 ++++++------ 2 files changed, 9 insertions(+), 6 deletions(-) diff --git a/NEWS b/NEWS index 71bd91e4..995507f0 100644 --- a/NEWS +++ b/NEWS @@ -43,6 +43,9 @@ Unreleased, experimental changes outlandish but conforming UT offsets or leap second corrections, has also been fixed. (Also reported by Naveed8951.) + 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. diff --git a/localtime.c b/localtime.c index aa9b48ca..698f61bf 100644 --- a/localtime.c +++ b/localtime.c @@ -1018,12 +1018,12 @@ tzloadbody(char const *name, struct state *sp, char tzloadflags, /* 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_MAX_LEAPS - && 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)) + if (! (0 <= leapcnt && leapcnt <= TZ_MAX_LEAPS + && 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 EINVAL; datablock_size = (timecnt * stored /* ats */ -- 2.51.0

1 0

[PROPOSED 0/3] Leap second overflow bugs
by Paul Eggert Jan. 20, 2026

Jan. 20, 2026

This patch series fixes some bugs involving integer overflow in leap second calculations at runtime. The main problem is with the optional time2posix and posix2time functions; without the patches the functions can crash or loop indefinitely or return completely-wrong answers. With these patches the functions have new, documented behavior: they return (time_t) -1 and set errno to EOVERFLOW when the result is not representable, instead of behaving badly in tricky situations. Given the problems I've been seeing with the leap second runtime code, and given the worry that similar problems remain, and given that the leap second runtime code is rarely used in practice, I'm thinking it may be a good idea to add a compile-time option to disable localtime.c's support for leap seconds. Using the option would shrink the attack surface on tzcode. Paul Eggert (3): Fix theoretical -2**31 leapcorr bug Improve time2posix man page Fix overflow bugs in time2posix and posix2time NEWS | 13 ++++- localtime.c | 90 +++++++++++++++++++++------------- time2posix.3 | 134 ++++++++++++++++++++++++++++++++++----------------- 3 files changed, 156 insertions(+), 81 deletions(-) -- 2.51.0

1 3

[PROPOSED] Check with -Wzero-as-null-pointer-constant
by Paul Eggert Jan. 19, 2026

Jan. 19, 2026

* Makefile (GCC_DEBUG_FLAGS): Omit -Wstringop-truncation, as it’s the default. Add -Wzero-as-null-pointer-constant. --- Makefile | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 570518c4..021186fc 100644 --- a/Makefile +++ b/Makefile @@ -380,12 +380,13 @@ GCC_DEBUG_FLAGS = -DGCC_LINT -g3 -O3 \ -Wold-style-definition -Woverlength-strings -Wpointer-arith \ -Wshadow -Wshift-overflow=2 -Wstrict-overflow \ -Wstrict-prototypes -Wstringop-overflow=4 \ - -Wstringop-truncation -Wsuggest-attribute=cold \ + -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 -- 2.51.0

1 0

[PROPOSED] Simplify int_fast32_t definition on C89 platforms
by Paul Eggert Jan. 17, 2026

Jan. 17, 2026

* private.h (int_fast32_t, INT_FAST32_MAX, INT_FAST32_MIN): When PORT_TO_C89 is defined, simplify these fallback definitions by using long rather than attempting to use int on platforms where int suffices. The attempt’s test was too strict: its ‘INT_MAX >> 31 == 0’ should have been ‘INT_MAX >> 30 == 0’. This led to a possible performance bug on C89 platforms where int is 32 bits, as int_fast32_t was defined to be long when int would suffice. C89 platforms are now museum pieces and this code is currently scheduled to be removed in 2029 or so anyway, so it is not worth trying to measure and debug any performance issues. Simplify the definitions instead. That mirrors what is already being done for uint_fast32_t. --- private.h | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-) diff --git a/private.h b/private.h index 42a1c892..e8cf9c22 100644 --- a/private.h +++ b/private.h @@ -410,15 +410,9 @@ typedef long long int_fast64_t; # endif # ifndef INT_FAST32_MAX -# if INT_MAX >> 31 == 0 typedef long int_fast32_t; -# define INT_FAST32_MAX LONG_MAX -# define INT_FAST32_MIN LONG_MIN -# else -typedef int int_fast32_t; -# define INT_FAST32_MAX INT_MAX -# define INT_FAST32_MIN INT_MIN -# endif +# define INT_FAST32_MAX LONG_MAX +# define INT_FAST32_MIN LONG_MIN # endif # ifndef INT_LEAST32_MAX -- 2.51.0

1 0

Re: tzload: validate tt_utoff to avoid signed overflow
by Paul Eggert Jan. 17, 2026

Jan. 17, 2026

In <https://github.com/eggert/tz/pull/37> (2026-01-10) the GitHub user Naveed8951 writes: > The TZif specification guarantees that tt_utoff is never INT32_MIN > so that 32-bit clients can safely negate it. However, tzloadbody() > previously accepted unvalidated values from untrusted TZif input. > A malformed or malicious file could therefore trigger signed integer > overflow during later offset arithmetic (e.g., negation for timezone / > altzone or offset-difference calculations), which is undefined behavior > in C. Naveed8951 submitted a patch that would cause localtime.c to reject TZif files containing UT offsets outside the POSIX range -89999..93599 (-24:59:50 .. 25:59:59), the idea being that small offsets like that would not cause overflow. Unfortunately, I think such a change would still allow undefined behavior in some cases, as overflow could still happen when adding a small offset to a large tm_sec or time_t value. Also, Internet RFC 9636 section 3.2 allows (though it discourages) UT offsets outside the POSIX range. So instead, I fixed localtime.c to handle the outlandish offsets correctly. As a special case, localtime.c (and zic.c) now reject a UT offset exactly equal to -2**31, as RFC 9636 does not allow that and it can easily cause trouble in callers that negate a UT offset stored in a 32-bit integer.

1 0

[PROPOSED] Pacify gcc -Wunused-function with 32-bit time_t
by Paul Eggert Jan. 17, 2026

Jan. 17, 2026

* localtime.c (USE_TIMEX_T): Move definition up. (is_threaded, locallock, dolock, dounlock, lock, unlock): Do not define if (defined TM_GMTOFF && USE_TIMEX_T). (rd2wrlock, tm_multi): Move definition down to be closer to use, so that it is nested inside #if and thus defined only when used. --- localtime.c | 117 ++++++++++++++++++++++++++++------------------------ 1 file changed, 62 insertions(+), 55 deletions(-) diff --git a/localtime.c b/localtime.c index a2f92b2a..9b7bbaa9 100644 --- a/localtime.c +++ b/localtime.c @@ -52,6 +52,10 @@ struct stat { char st_ctime, st_dev, st_ino; }; # define THREAD_TM_MULTI 0 #endif +#ifndef USE_TIMEX_T +# define USE_TIMEX_T false +#endif + #if THREAD_SAFE # include <pthread.h> @@ -88,6 +92,10 @@ extern int __isthreaded; # 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. @@ -98,24 +106,25 @@ extern int __isthreaded; static bool is_threaded(void) { -# if THREAD_PREFER_SINGLE && HAVE___ISTHREADED +# if THREAD_PREFER_SINGLE && HAVE___ISTHREADED return !!__isthreaded; -# elif THREAD_PREFER_SINGLE && HAVE_SYS_SINGLE_THREADED_H +# elif THREAD_PREFER_SINGLE && HAVE_SYS_SINGLE_THREADED_H return !__libc_single_threaded; -# else +# else return true; -# endif +# endif } -# if THREAD_RWLOCK +# 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 +# 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 +# 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 @@ -131,25 +140,12 @@ unlock(bool threaded) if (threaded) dounlock(); } -#else +# else static int lock(void) { return -1; } static void unlock(ATTRIBUTE_MAYBE_UNUSED bool threaded) { } +# endif #endif -/* 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; -} - #if THREAD_SAFE typedef pthread_once_t once_t; # define ONCE_INIT PTHREAD_ONCE_INIT @@ -187,36 +183,6 @@ tm_multi_key_init(void) #endif -/* 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; -} - /* 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. */ @@ -293,9 +259,6 @@ typedef time_t monotime_t; 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. */ -#ifndef USE_TIMEX_T -# define USE_TIMEX_T false -#endif #if USE_TIMEX_T # undef TIME_T_MIN # undef TIME_T_MAX @@ -1843,6 +1806,20 @@ zoneinit(struct state *sp, char const *name, char tzloadflags) } } +/* 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. @@ -2134,6 +2111,36 @@ localsub(struct state const *sp, time_t const *timep, int_fast32_t setname, #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, -- 2.52.0

1 0

[PROPOSED 0/6] Fix integer overflows and drop posixrules
by Paul Eggert Jan. 17, 2026

Jan. 17, 2026

GitHub user Naveed8951 reported a potential (but unlikely) security flaw due to undefined behavior in localtime.c after integer overflow. The patches in this series should fix the reported problems. Some of the problem occurred in a part of the localtime.c code that was implementing a feature declared obsolete in TZDB release 2019b. This feature let the sysadmin specify default DST rules for DST-but-ruleless old-style POSIX settings like TZ="AST4ADT". Rather than try to fix the feature this patch series removes it; the feature is hardly ever used and would stop working in 2038 anyway. As a result, tzcode by default uses current US DST for these settings, with no opportunity for sysadmin override (though the default can be overriden when tzcode is compiled). This behavior is already common elsewhere. Paul Eggert (6): Add comment re UT offsets equal to -2**31 Document zic -p better zic now warns about -p Remove POSIXRULES installation option Remove TZDEFRULES ("posixrules") from localtime.c Fix remaining Naveed8951-reported overflows Makefile | 28 +--- NEWS | 19 +++ localtime.c | 388 ++++++++++++++++++++++++---------------------------- newctime.3 | 4 +- newtzset.3 | 29 +--- tzfile.5 | 7 +- tzfile.h | 12 +- zic.8 | 17 ++- zic.c | 8 ++ 9 files changed, 228 insertions(+), 284 deletions(-) -- 2.52.0

1 6

Unable to find Doha in the database
by Paul Menzel Jan. 15, 2026

Jan. 15, 2026

Dear Time Zone Database folks, Visiting Doha, I wanted to change to its time zone, and knowing `Europe/Berlin`, I looked for *Doha*, but didn’t find it in GNOME [1]. Looking through the Time Zone Database archive, there is in `asia` [3]: # Bahrain # Qatar # Zone NAME STDOFF RULES FORMAT [UNTIL] Zone Asia/Qatar 3:26:08 - LMT 1920 # Al Dawhah / Doha 4:00 - %z 1972 Jun 3:00 - %z The comment is not in my file on Debian though in `/usr/share/zoneinfo/tzdata.zi` from *tzdata* 2025c-3. Z Asia/Qatar 3:26:8 - LMT 1920 4 - %z 1972 Jun 3 - %z Of course, I do not know, if the comment would change anything. Would it be possible to also add Qatar’s capital to the database. Or is there another way to make it easier for people to find the right time zone entry? Kind regards, Paul [1]: https://gitlab.gnome.org/GNOME/gnome-control-center/-/issues/3656 [2]: https://github.com/eggert/tz/blob/d5f71488fd1ec4a2f7166ceeac2024a677daaaef/…

3 2

[PROPOSED 0/5] Fix some unlikely undefined behaviors
by Paul Eggert Jan. 12, 2026

Jan. 12, 2026

This patchest fixes some bugs in localtime.c when TZif files specify a UT offset equal to -2**31. Such an offset overflows when negated on a 32-bit platform, and this leads to undefined behavior – typically wraparound arithmetic, but the C standard says anything can happen. The patchset also changes zic.c to reject attempts to create such an offset, adjusts some nearby code to be a bit cleaner and more portable. Paul Eggert (5): Omit unnecessary L suffixes zic no longer generates utoff == -2**31 Make sure 2**31 - 1 is signed Port TWOS_COMPLEMENT to signed-magnitude hosts Disallow UT offsets equal to -2**31 NEWS | 4 ++++ localtime.c | 48 ++++++++++++++++++++++++++++++++---------------- private.h | 7 +++++-- strftime.c | 5 +++-- zic.c | 7 ++++--- 5 files changed, 48 insertions(+), 23 deletions(-) -- 2.51.0

1 5

[PROPOSED] Improve .gitignore patterns
by Paul Eggert Jan. 10, 2026

Jan. 10, 2026

* .gitignore: Ignore all subdirectories. Also ignore *.bak, *.core, *.out, *.t, core, and core.*. Also ignore .*, except for .gitignore itself. --- .gitignore | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) diff --git a/.gitignore b/.gitignore index e7bb0416..5a140b46 100644 --- a/.gitignore +++ b/.gitignore @@ -1,26 +1,38 @@ # Files intentionally not tracked by Git. # This file is in the public domain. +# +# Ignore all subdirectories. +*/ +# +# Ignore non-directories that are the output of builds, +# or are backups or nonces unlikely to be added to the repository. *.a *.asc +*.bak *.ck -*.ckd/ +*.core *.diff -*.dir/ *.i *.o *.orig +*.out *.patch *.pdf *.ps *.rej *.s +*.t *.tar *.tar.* *.txt *.tzs *.zi *~ +.* +!.gitignore ChangeLog +core +core.* date leapseconds tzdir.h -- 2.51.0

1 0

Whither /usr/share/zoneinfo/right?
by Paul Gilmartin Jan. 8, 2026

Jan. 8, 2026

ISTR on Linux long ago, a subdirectory, /usr/share/zoneinfo/right, which used arguments not POSIX time but TAI. Was that a local hack, or was it removed, possibly because of uncertainty of future leap seconds? -- Thanks, gil

4 5

Fwd: Bulletin C number 71
by Tim Parenti Jan. 7, 2026

Jan. 7, 2026

No leap second 2026-06-30, per below. See attached patch. -- Tim Parenti ---------- Forwarded message --------- From: IERS EOP Product Center <iers.eoppc(a)obspm.fr> Date: Tue, 6 Jan 2026 at 06:50 Subject: Bulletin C number 71 To: <bulc.iers(a)obspm.fr> INTERNATIONAL EARTH ROTATION AND REFERENCE SYSTEMS SERVICE (IERS) SERVICE INTERNATIONAL DE LA ROTATION TERRESTRE ET DES SYSTEMES DE REFERENCE SERVICE DE LA ROTATION TERRESTRE DE L'IERS OBSERVATOIRE DE PARIS 61, Av. de l'Observatoire 75014 PARIS (France) Tel. : +33 1 40 51 23 35 e-mail : services.iers(a)obspm.fr http://hpiers.obspm.fr/eop-pc Paris, 06 January 2026 Bulletin C 71 To authorities responsible for the measurement and distribution of time INFORMATION ON UTC - TAI NO leap second will be introduced at the end of June 2026. The difference between Coordinated Universal Time UTC and the International Atomic Time TAI is : from 2017 January 1, 0h UTC, until further notice : UTC-TAI = -37 s Leap seconds can be introduced in UTC at the end of the months of December or June, depending on the evolution of UT1-TAI. Bulletin C is mailed every six months, either to announce a time step in UTC, or to confirm that there will be no time step at the next possible date. Christian BIZOUARD Director Earth Orientation Center of IERS Observatoire de Paris, France

2 1

static_assert macro
by Christos Zoulas Jan. 7, 2026

Jan. 7, 2026

Hello, The new tzcode uses static_assert() with no message which is a c2x extension. clang does not like that by default (without specifying -std=gnu2x). /home/source/ab/HEAD-llvm/src/lib/libc/time/private.h:1017:52: error: '_Static_assert' with no message is a C2x extension [-Werror,-Wc2x-extensions] || TIME_T_MAX >> (TYPE_BIT(time_t) - 2) == 1); ^ , "" Best, christos

3 4

tzcode and tzdata tar base make checks fail
by Brian Inglis Dec. 31, 2025

Dec. 31, 2025

Cygwin uses separate packages of tzcode programs and tzdata files verbatim to support standard local time zones in preference to limited native data. Running make check on the separate tar file fails in both cases. When I took over the packages, I provided basic checks, but reviewed and changed the builds this release, and would like to reenable the built in checks you provide and mentioned recently. The attached logs show the results of having the package build scripts run `make check` from the package build directory which is set up with symlinks to each of the tar sources. Is there anything obviously missing or incorrect with `make check` invocations which could cause the observed failures, or things that may need changed? -- Take care. Thanks, Brian Inglis Calgary, Alberta, Canada La perfection est atteinte Perfection is achieved non pas lorsqu'il n'y a plus rien à ajouter not when there is no more to add mais lorsqu'il n'y a plus rien à retrancher but when there is no more to cut -- Antoine de Saint-Exupéry

2 6

Time zones on Antarctica expedition
by Brooks Harris Dec. 30, 2025

Dec. 30, 2025

Those who have a New York Times subscription may find this entertaining: Does Anybody Really Know What Time It Is? The clocks aboard our icebreaker will be changed several times en route to Antarctica. It’s one of many things that make the expedition feel otherworldly. https://www.nytimes.com/2025/12/29/climate/antarctica-time-zone.html

1 0

[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