- tz - lists.iana.org

Re: [tz] TZDB patches, UTF-8, and Microsoft Exchange
by Brian Inglis Dec. 4, 2022

Dec. 4, 2022

On 2022-12-04 12:48, Fred Gleason via tz wrote: > On Dec 2, 2022, at 17:21, Derick Rethans via tz <tz(a)iana.org > <mailto:tz@iana.org>> wrote: > >> This problem would go away if you'd except change requests through GitHub's >> pull request feature. > > I for one heartily second this suggestion. While I understand that this can be a > sensitive issue for some long-time contributors here, given the composition of > TZDB’s audience (veteran coders plus the occasional technically > non-sophisticated user), I think anything that simplifies the overall workflow > while making the process more transparent to non-coders would be a net win. > > >> It's quite a common and modern method to propose changes to open source >> software, that I'd expect many would already be familiar with. > > I agree. The GitHub pull request process has become a sort of _lingua franca_ in > the code development realm, arguably more so than any other single methodology > for managing distributed open source development. In that context, the current > _status quo_ is rather *anti*-user friendly in the sense that many coders, when > first encountering the TZDB GitHub repo think that they already know how to > generate and submit patches; only later discovering that they in fact don’t, > since what is arguably the central workflow metaphor of what makes GitHub useful > has (seemingly inexplicably) been switched off. > > On the basis of the patches I’ve seen go by on this listserv over the past few > years, it would also seem that “easing/reducing maintenance burden” has become a > major goal for TZDB. Given that, I would respectfully argue that moving to a > PR-based workflow would do more towards that end than any number of minor code > polishing changes. > > > On Dec 2, 2022, at 17:43, Paul Eggert via tz <tz(a)iana.org <mailto:tz@iana.org>> > wrote: > >> Yes, but GitHub has its own problems. I'd rather stick with email, just as >> many other projects use email for this sort of thing. > > Well, this one at least does. FWIW, none of the other FOSS projects whose dev > lists I follow use the traditional patches-via-email workflow any more. While by > no means all of them use GitHub, what they all do use is some sort of system > that provides automated patch generation/testing/merging along with a GUI option > that allows non-coder types some level of visibility into the process. > > >> Also, I've been thinking of moving the development repository off GitHub for >> other reasons, and I don't want to rely on GitHub-specific features. > > Would you care to share with us what some of those “other reasons” might be? At a guess, I'd imagine for a multi-GNU project contributor, even using a proprietary vendor locked-in (and potentially future deprecated or locked-out) platform might be an issue (possibly queried only in person or private DM), as it is for many other open source contributors, where even mentioning GitHub evokes a spit-blah *NEVER* response, and moving to GH would result in losing some contributions. For some open-source projects, GH is used only as a public and backup repo for their primary dev workflows, as are sv.gnu.org and sv.non-gnu.org in some cases. A number are using in-house and/or public GitLab instances, and others have surfaced their email with public-inbox like https://lore.kernel.org/ where issues and patches can be handled using workflow tools like https://git.kernel.org/pub/scm/utils/b4/b4.git It would be good if the IETF, IANA, and/or ICANN made their email available in such searchable archives, an area where tz is currently lacking. -- 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 retirer but when there is no more to cut -- Antoine de Saint-Exupéry

1 0

[PATCH] Fix America/Nuuk timezone change
by Thomas M. Steenholdt Dec. 4, 2022

Dec. 4, 2022

Fix the upcoming move of America/Nuuk from normaltime -03 to -02, according to the passed bill. --- europe | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-) diff --git a/europe b/europe index a3547b4f..71ea3ab2 100644 --- a/europe +++ b/europe @@ -1126,6 +1126,21 @@ Zone Atlantic/Faroe -0:27:04 - LMT 1908 Jan 11 # Tórshavn # "bekymringer" URL says the intent is no transition March 25, that # Greenland will not go back to winter time in fall 2023, and that # only America/Nuuk is affected (though further changes may occur). +# +# From Thomas M. Steenholdt (2022-12-02): +# Processing of the bill is documented here: +# https://ina.gl/samlinger/oversigt-over-samlinger/samling/dagsordener/dagsor… +# Notably: +# - The bill to move America/Nuuk from UTC-03 to UTC-02 passed. +# - The bill to stop observing DST did not (Greenland will stop observing DST +# when EU does). +# Details on the implementation are here (section 6): +# https://ina.gl/dvd/EM%202022/pdf/media/2553529/pkt17_em2022_tidens_bestemme… +# This is how the change will be implemented: +# 1. The shift *to* DST in 2023 happens as normal. +# 2. The shift *from* DST in 2023 happens as normal, but coincides with the +# shift to UTC-02 normaltime (people will not change their clocks here). +# 3. After this, DST is still observed, but as -02/-01 instead of -03/-02. # Rule NAME FROM TO - IN ON AT SAVE LETTER/S Rule Thule 1991 1992 - Mar lastSun 2:00 1:00 D @@ -1149,8 +1164,8 @@ Zone America/Scoresbysund -1:27:52 - LMT 1916 Jul 28 # Ittoqqortoormiit -1:00 EU -01/+00 Zone America/Nuuk -3:26:56 - LMT 1916 Jul 28 # Godthåb -3:00 - -03 1980 Apr 6 2:00 - -3:00 EU -03/-02 2023 Mar 25 22:00 - -2:00 - -02 + -3:00 EU -03/-02 2023 Oct 28 23:00 + -2:00 EU -02/-01 Zone America/Thule -4:35:08 - LMT 1916 Jul 28 # Pituffik -4:00 Thule A%sT -- 2.34.1

5 7

[PROPOSED] Default to assuming C99 or later
by Paul Eggert Dec. 2, 2022

Dec. 2, 2022

By default, assume the compiler and library support C99 or later. Continue to port to C89 platforms only if the new macro PORT_TO_C89 is nonzero. This new macro is a transitional aid, and is planned to be removed in a future version that always assumes C99 or later. * Makefile, NEWS: Mention this. * private.h (PORT_TO_C89): New macro. (SUPPORT_C89): Move up. (true, false, bool) [!PORT_TO_C89]: Assume C99 or later. If !PORT_TO_C89 simply include <inttypes.h> rather than fooling around with conditional stdint.h, conditional inttypes.h, and lots of other conditional definitions. (restrict) [!PORT_TO_C89]: Do not define. * zdump.c (HAVE_SNPRINTF) [!PORT_TO_C89]: Now true. --- Makefile | 14 ++++++++------ NEWS | 11 ++++++----- private.h | 35 +++++++++++++++++++++++++---------- zdump.c | 2 +- 4 files changed, 40 insertions(+), 22 deletions(-) diff --git a/Makefile b/Makefile index 63f3b5ca..8bee50d8 100644 --- a/Makefile +++ b/Makefile @@ -220,7 +220,7 @@ LDLIBS= # -DHAVE_INCOMPATIBLE_CTIME_R if your system's time.h declares # ctime_r and asctime_r incompatibly with the POSIX standard # (Solaris when _POSIX_PTHREAD_SEMANTICS is not defined). -# -DHAVE_INTTYPES_H=0 if <inttypes.h> does not work* +# -DHAVE_INTTYPES_H=0 if <inttypes.h> does not work*+ # -DHAVE_LINK=0 if your system lacks a link function # -DHAVE_LOCALTIME_R=0 if your system lacks a localtime_r function # -DHAVE_LOCALTIME_RZ=0 if you do not want zdump to use localtime_rz @@ -229,24 +229,24 @@ LDLIBS= # -DHAVE_POSIX_DECLS=0 if your system's include files do not declare # functions like 'link' or variables like 'tzname' required by POSIX # -DHAVE_SETENV=0 if your system lacks the setenv function -# -DHAVE_SNPRINTF=0 if your system lacks the snprintf function +# -DHAVE_SNPRINTF=0 if your system lacks the snprintf function+ # -DHAVE_STDCKDINT_H=0 if neither <stdckdint.h> nor substitutes like # __builtin_add_overflow work* -# -DHAVE_STDINT_H=0 if <stdint.h> does not work* +# -DHAVE_STDINT_H=0 if <stdint.h> does not work*+ # -DHAVE_STRFTIME_L if <time.h> declares locale_t and strftime_l # -DHAVE_STRDUP=0 if your system lacks the strdup function -# -DHAVE_STRTOLL=0 if your system lacks the strtoll function +# -DHAVE_STRTOLL=0 if your system lacks the strtoll function+ # -DHAVE_SYMLINK=0 if your system lacks the symlink function # -DHAVE_SYS_STAT_H=0 if <sys/stat.h> does not work* # -DHAVE_TZSET=0 if your system lacks a tzset function # -DHAVE_UNISTD_H=0 if <unistd.h> does not work* # -DHAVE_UTMPX_H=0 if <utmpx.h> does not work* # -Dlocale_t=XXX if your system uses XXX instead of locale_t +# -DPORT_TO_C89 if tzcode should also run on C89 platforms+ # -DRESERVE_STD_EXT_IDS if your platform reserves standard identifiers # with external linkage, e.g., applications cannot define 'localtime'. # -Dssize_t=long on hosts like MS-Windows that lack ssize_t -# -DSUPPORT_C89 if you build or run tzcode on a C89 platform; this option -# is obsolescent and is planned to be removed when C99+ is required. +# -DSUPPORT_C89 if the tzcode library should support C89 callers+ # -DSUPPRESS_TZDIR to not prepend TZDIR to file names; this has # security implications and is not recommended for general use # -DTHREAD_SAFE to make localtime.c thread-safe, as POSIX requires; @@ -275,6 +275,8 @@ LDLIBS= # $(GCC_DEBUG_FLAGS) if you are using recent GCC and want lots of checking # # * Options marked "*" can be omitted if your compiler is C23 compatible. +# * Options marked "+" are obsolescent and are planned to be removed +# once the code assumes C99 or later. # # Select instrumentation via "make GCC_INSTRUMENT='whatever'". GCC_INSTRUMENT = \ diff --git a/NEWS b/NEWS index 25dcc875..91ec5056 100644 --- a/NEWS +++ b/NEWS @@ -3,15 +3,16 @@ News for the tz database Unreleased, experimental changes Briefly: - tzcode no longer supports C89 unless built with -DSUPPORT_C89 + The code now defaults to C99 or later. Fix use of C23 attributes. Changes to code - tzcode no longer supports C89 by default. To build or run in a - C89 environment, compile with -DSUPPORT_C89, a transitional aid - that is planned to be removed in a near-future version, when C99 - or later will be required. + The code by default is now designed for C99 or later. To build in + a C89 environment, compile with -DPORT_TO_C89. To support C89 + callers of the tzcode library, compile with -DSUPPORT_C89. The + two new macros are transitional aids planned to be removed in a + future version, when C99 or later will be required. On C23-compatible platforms tzcode no longer uses syntax like 'static [[noreturn]] void usage(void);'. Instead, it uses diff --git a/private.h b/private.h index 999c1648..1d6ddf73 100644 --- a/private.h +++ b/private.h @@ -17,12 +17,25 @@ ** Thank you! */ +/* PORT_TO_C89 means the code should work even if the underlying + compiler and library support only C89. SUPPORT_C89 means the + tzcode library should support C89 callers in addition to the usual + support for C99-and-later callers. These macros are obsolescent, + and the plan is to remove them along with any code needed only when + they are nonzero. */ +#ifndef PORT_TO_C89 +# define PORT_TO_C89 0 +#endif +#ifndef SUPPORT_C89 +# define SUPPORT_C89 0 +#endif + #ifndef __STDC_VERSION__ # define __STDC_VERSION__ 0 #endif /* Define true, false and bool if they don't work out of the box. */ -#if __STDC_VERSION__ < 199901 +#if PORT_TO_C89 && __STDC_VERSION__ < 199901 # define true 1 # define false 0 # define bool int @@ -100,10 +113,6 @@ # define HAVE_STRDUP 1 #endif -#ifndef HAVE_STRTOLL -# define HAVE_STRTOLL 1 -#endif - #ifndef HAVE_SYMLINK # define HAVE_SYMLINK 1 #endif /* !defined HAVE_SYMLINK */ @@ -185,6 +194,9 @@ #include <stddef.h> #include <string.h> +#if !PORT_TO_C89 +# include <inttypes.h> +#endif #include <limits.h> /* for CHAR_BIT et al. */ #include <stdlib.h> @@ -254,6 +266,8 @@ # define R_OK 4 #endif /* !defined R_OK */ +#if PORT_TO_C89 + /* ** Define HAVE_STDINT_H's default value here, rather than at the ** start, since __GLIBC__ and INTMAX_MAX's values depend on @@ -334,6 +348,9 @@ typedef int int_fast32_t; #ifndef INTMAX_MAX # ifdef LLONG_MAX typedef long long intmax_t; +# ifndef HAVE_STRTOLL +# define HAVE_STRTOLL true +# endif # if HAVE_STRTOLL # define strtoimax strtoll # endif @@ -396,6 +413,8 @@ typedef unsigned long uintmax_t; # define SIZE_MAX ((size_t) -1) #endif +#endif /* PORT_TO_C89 */ + /* Support ckd_add, ckd_sub, ckd_mul on C23 or recent-enough GCC-like hosts, unless compiled with -DHAVE_STDCKDINT_H=0 or with pre-C23 EDG. */ #if !defined HAVE_STDCKDINT_H && defined __has_include @@ -520,7 +539,7 @@ typedef unsigned long uintmax_t; # endif #endif -#if __STDC_VERSION__ < 199901 && !defined restrict +#if PORT_TO_C89 && __STDC_VERSION__ < 199901 && !defined restrict # define restrict /* empty */ #endif @@ -704,10 +723,6 @@ extern int daylight; extern long altzone; #endif -#ifndef SUPPORT_C89 -# define SUPPORT_C89 0 -#endif - /* ** The STD_INSPIRED functions are similar, but most also need ** declarations if time_tz is defined. diff --git a/zdump.c b/zdump.c index bac4a67a..5173c9ac 100644 --- a/zdump.c +++ b/zdump.c @@ -15,7 +15,7 @@ #include <stdio.h> #ifndef HAVE_SNPRINTF -# define HAVE_SNPRINTF (199901 <= __STDC_VERSION__) +# define HAVE_SNPRINTF (!PORT_TO_C89 || 199901 <= __STDC_VERSION__) #endif #ifndef HAVE_LOCALTIME_R -- 2.37.2

1 0

[PROPOSED] Port alignof to Oracle Developer Studio
by Paul Eggert Dec. 2, 2022

Dec. 2, 2022

* zic.c (alignof): Don’t include <stdalign.h> when compiling with Oracle Developer Studio, as release 12.6 lacks <stdalign.h> even though it #defines __STDC_VERSION__ to 201112L. Most likely this circa-2017 release, which #defines __SUNPRO_C to be 0x5150, is the last release, so assume that all releases lack <stdalign.h>. --- zic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/zic.c b/zic.c index f143fcef..5612805f 100644 --- a/zic.c +++ b/zic.c @@ -67,7 +67,7 @@ enum { FORMAT_LEN_GROWTH_BOUND = 5 }; #endif /* The minimum alignment of a type, for pre-C23 platforms. */ -#if __STDC_VERSION__ < 201112 +#if __STDC_VERSION__ < 201112 || defined __SUNPRO_C # define alignof(type) offsetof(struct { char a; type b; }, b) #elif __STDC_VERSION__ < 202311 # include <stdalign.h> -- 2.37.2

2 2

1981/2 Singapore+Malaysia switch from GMT+0730 to GMT+0800
by Robert Elz Dec. 1, 2022

Dec. 1, 2022

This press release was discovered, and brought up in another forum by Geoff Clare. It conflicts with the data in the Asia/Singapore and Asia/Kuala-Lumpur zones. https://corporate.nas.gov.sg/media/collections-and-research/time-zone-adjus… That puts the change as being from 23:29:59 Dec 31, 1981 -> 00:00:00 Jan 1, 1982 whereas tzdata (going back to quite ancient versions) has the transition 23:59:59 Dec 31, 1981 -> 00:30:00 Jan 1, 1982 The version from the web page makes more sense to me (as in that is how I would have done that transition) but that does not mean that it is necessarily correct. Nothing there shows a copy of the actual gazette notification by the President (of Singapore in that case, but I think we can safely assume that it was done the same as in Malaysia) which would be the actual specification, but that 2015 press release says it happened as described that it would happen in the 1981 press release which is included there. The "asia" file just says of this transition: 7:30 - +0730 1982 Jan 1 8:00 - +08 which is correct in a sense for either version of when the transition occurred, and perhaps raises a general question of when a step forward takes place at the boundary between two days, does it more often delete the last part of the previous day or the first part of the next? I can find nothing in my tz list archives that says anything at all about this transition - the earliest message with any specific content about Singapore I could see was a Jan 2004 message from Mok Ly Yng (forwarded to the list by ado@), but that's really about earlier transitions - this one was already in the data (just with SGT instead of +0730 and +08). There's also a long Sept 2014 discussion about a zdump bug that this particular transition, as defined by zic, revealed, and a shorter discussion about the SGT removal in Sept 2018 - but that's all I managed to find about Singapore. While not about the Singapore/Malaysia zones themselves, even older data that appears on the list (ie: looking at diffs patches, and updated zone files) shows that in 1996 this transition was believed to have happened in May 1982 (which seems to be from Shanks). That altered in Dec 2003, in a change described as: * Fix several Malaysia and Singapore entries before 1983. Among other things, they enjoyed 20-minute DST from 1933 to 1936, and we have more-accurate dates for WWII transitions. (Thanks to Mok Ly Yng for this info.) which notably does not mention the change moving from May 1982 -> 1 Jan 1982 but that change was made in that version. Nothing appears to have happened since. I suspect it would be correct to change the first of those lines to be 7:30 - +0730 23:30 1981 Dec 30 but I am no authority on that. kre

5 13

[PROPOSED] Fix C23-related conformance bug
by Paul Eggert Dec. 1, 2022

Dec. 1, 2022

Problem reported by Houge Langley for ‘gcc -std=gnu99’ in: https://bugs.gentoo.org/show_bug.cgi?id=883719 * NEWS: Mention this. * date.c, localtime.c, private.h, zdump.c, zic.c: Use ATTRIBUTE_* at the start of function declarations, not later (such as after the keyword ‘static’). This is required for strict conformance to C23. --- NEWS | 6 ++++++ date.c | 2 +- localtime.c | 4 ++-- private.h | 6 +++--- zdump.c | 12 ++++++------ zic.c | 34 +++++++++++++++++----------------- 6 files changed, 35 insertions(+), 29 deletions(-) diff --git a/NEWS b/NEWS index 9cb175e7..25dcc875 100644 --- a/NEWS +++ b/NEWS @@ -4,6 +4,7 @@ Unreleased, experimental changes Briefly: tzcode no longer supports C89 unless built with -DSUPPORT_C89 + Fix use of C23 attributes. Changes to code @@ -12,6 +13,11 @@ Unreleased, experimental changes that is planned to be removed in a near-future version, when C99 or later will be required. + On C23-compatible platforms tzcode no longer uses syntax like + 'static [[noreturn]] void usage(void);'. Instead, it uses + '[[noreturn]] static void usage(void);' as strict C23 requires. + (Problem reported by Houge Langley.) + Release 2022g - 2022-11-29 08:58:31 -0800 diff --git a/date.c b/date.c index 11c5e5fe..97df6ab0 100644 --- a/date.c +++ b/date.c @@ -42,7 +42,7 @@ static void display(const char *, time_t); static void dogmt(void); static void errensure(void); static void timeout(FILE *, const char *, const struct tm *); -static ATTRIBUTE_NORETURN void usage(void); +ATTRIBUTE_NORETURN static void usage(void); int main(const int argc, char *argv[]) diff --git a/localtime.c b/localtime.c index f8a0a6eb..f1a8598b 100644 --- a/localtime.c +++ b/localtime.c @@ -843,7 +843,7 @@ is_digit(char c) ** Return a pointer to that character. */ -static ATTRIBUTE_REPRODUCIBLE const char * +ATTRIBUTE_REPRODUCIBLE static const char * getzname(register const char *strp) { register char c; @@ -864,7 +864,7 @@ getzname(register const char *strp) ** We don't do any checking here; checking is done later in common-case code. */ -static ATTRIBUTE_REPRODUCIBLE const char * +ATTRIBUTE_REPRODUCIBLE static const char * getqzname(register const char *strp, const int delim) { register int c; diff --git a/private.h b/private.h index c22354ac..999c1648 100644 --- a/private.h +++ b/private.h @@ -646,7 +646,7 @@ DEPRECATED_IN_C23 char *asctime(struct tm const *); char *asctime_r(struct tm const *restrict, char *restrict); DEPRECATED_IN_C23 char *ctime(time_t const *); char *ctime_r(time_t const *, char *); -double difftime(time_t, time_t) ATTRIBUTE_UNSEQUENCED; +ATTRIBUTE_UNSEQUENCED double difftime(time_t, time_t); size_t strftime(char *restrict, size_t, char const *restrict, struct tm const *restrict); # if HAVE_STRFTIME_L @@ -765,10 +765,10 @@ timezone_t tzalloc(char const *); void tzfree(timezone_t); # if STD_INSPIRED # if TZ_TIME_T || !defined posix2time_z -time_t posix2time_z(timezone_t, time_t) ATTRIBUTE_REPRODUCIBLE; +ATTRIBUTE_REPRODUCIBLE time_t posix2time_z(timezone_t, time_t); # endif # if TZ_TIME_T || !defined time2posix_z -time_t time2posix_z(timezone_t, time_t) ATTRIBUTE_REPRODUCIBLE; +ATTRIBUTE_REPRODUCIBLE time_t time2posix_z(timezone_t, time_t); # endif # endif #endif diff --git a/zdump.c b/zdump.c index 04953dcf..bac4a67a 100644 --- a/zdump.c +++ b/zdump.c @@ -89,7 +89,7 @@ static bool warned; static bool errout; static char const *abbr(struct tm const *); -static intmax_t delta(struct tm *, struct tm *) ATTRIBUTE_REPRODUCIBLE; +ATTRIBUTE_REPRODUCIBLE static intmax_t delta(struct tm *, struct tm *); static void dumptime(struct tm const *); static time_t hunt(timezone_t, time_t, time_t, bool); static void show(timezone_t, char *, time_t, bool); @@ -97,7 +97,7 @@ static void showextrema(timezone_t, char *, time_t, struct tm *, time_t); static void showtrans(char const *, struct tm const *, time_t, char const *, char const *); static const char *tformat(void); -static time_t yeartot(intmax_t) ATTRIBUTE_REPRODUCIBLE; +ATTRIBUTE_REPRODUCIBLE static time_t yeartot(intmax_t); /* Is C an ASCII digit? */ static bool @@ -125,7 +125,7 @@ is_alpha(char a) } } -static ATTRIBUTE_NORETURN void +ATTRIBUTE_NORETURN static void size_overflow(void) { fprintf(stderr, _("%s: size overflow\n"), progname); @@ -134,7 +134,7 @@ size_overflow(void) /* Return A + B, exiting if the result would overflow either ptrdiff_t or size_t. */ -static ATTRIBUTE_REPRODUCIBLE ptrdiff_t +ATTRIBUTE_REPRODUCIBLE static ptrdiff_t sumsize(size_t a, size_t b) { #ifdef ckd_add @@ -151,7 +151,7 @@ sumsize(size_t a, size_t b) /* Return a pointer to a newly allocated buffer of size SIZE, exiting on failure. SIZE should be nonzero. */ -static void * ATTRIBUTE_MALLOC +ATTRIBUTE_MALLOC static void * xmalloc(size_t size) { void *p = malloc(size); @@ -916,7 +916,7 @@ showextrema(timezone_t tz, char *zone, time_t lo, struct tm *lotmp, time_t hi) # include <stdarg.h> /* A substitute for snprintf that is good enough for zdump. */ -static int ATTRIBUTE_FORMAT((printf, 3, 4)) +ATTRIBUTE_FORMAT((printf, 3, 4)) static int my_snprintf(char *s, size_t size, char const *format, ...) { int n; diff --git a/zic.c b/zic.c index 892414af..f143fcef 100644 --- a/zic.c +++ b/zic.c @@ -459,20 +459,20 @@ static char roll[TZ_MAX_LEAPS]; ** Memory allocation. */ -static ATTRIBUTE_NORETURN void +ATTRIBUTE_NORETURN static void memory_exhausted(const char *msg) { fprintf(stderr, _("%s: Memory exhausted: %s\n"), progname, msg); exit(EXIT_FAILURE); } -static ATTRIBUTE_NORETURN void +ATTRIBUTE_NORETURN static void size_overflow(void) { memory_exhausted(_("size overflow")); } -static ATTRIBUTE_REPRODUCIBLE ptrdiff_t +ATTRIBUTE_REPRODUCIBLE static ptrdiff_t size_sum(size_t a, size_t b) { #ifdef ckd_add @@ -487,7 +487,7 @@ size_sum(size_t a, size_t b) size_overflow(); } -static ATTRIBUTE_REPRODUCIBLE ptrdiff_t +ATTRIBUTE_REPRODUCIBLE static ptrdiff_t size_product(ptrdiff_t nitems, ptrdiff_t itemsize) { #ifdef ckd_mul @@ -502,7 +502,7 @@ size_product(ptrdiff_t nitems, ptrdiff_t itemsize) size_overflow(); } -static ATTRIBUTE_REPRODUCIBLE ptrdiff_t +ATTRIBUTE_REPRODUCIBLE static ptrdiff_t align_to(ptrdiff_t size, ptrdiff_t alignment) { ptrdiff_t lo_bits = alignment - 1, sum = size_sum(size, lo_bits); @@ -526,7 +526,7 @@ memcheck(void *ptr) return ptr; } -static void * ATTRIBUTE_MALLOC +ATTRIBUTE_MALLOC static void * emalloc(size_t size) { return memcheck(malloc(size)); @@ -538,7 +538,7 @@ erealloc(void *ptr, size_t size) return memcheck(realloc(ptr, size)); } -static char * ATTRIBUTE_MALLOC +ATTRIBUTE_MALLOC static char * estrdup(char const *str) { return memcheck(strdup(str)); @@ -608,7 +608,7 @@ eat(int fnum, lineno num) eats(fnum, num, 0, -1); } -static void ATTRIBUTE_FORMAT((printf, 1, 0)) +ATTRIBUTE_FORMAT((printf, 1, 0)) static void verror(const char *const string, va_list args) { /* @@ -626,7 +626,7 @@ verror(const char *const string, va_list args) fprintf(stderr, "\n"); } -static void ATTRIBUTE_FORMAT((printf, 1, 2)) +ATTRIBUTE_FORMAT((printf, 1, 2)) static void error(const char *const string, ...) { va_list args; @@ -636,7 +636,7 @@ error(const char *const string, ...) errors = true; } -static void ATTRIBUTE_FORMAT((printf, 1, 2)) +ATTRIBUTE_FORMAT((printf, 1, 2)) static void warning(const char *const string, ...) { va_list args; @@ -666,7 +666,7 @@ close_file(FILE *stream, char const *dir, char const *name, } } -static ATTRIBUTE_NORETURN void +ATTRIBUTE_NORETURN static void usage(FILE *stream, int status) { fprintf(stream, @@ -3597,7 +3597,7 @@ lowerit(char a) } /* case-insensitive equality */ -static ATTRIBUTE_REPRODUCIBLE bool +ATTRIBUTE_REPRODUCIBLE static bool ciequal(register const char *ap, register const char *bp) { while (lowerit(*ap) == lowerit(*bp++)) @@ -3606,7 +3606,7 @@ ciequal(register const char *ap, register const char *bp) return false; } -static ATTRIBUTE_REPRODUCIBLE bool +ATTRIBUTE_REPRODUCIBLE static bool itsabbr(register const char *abbr, register const char *word) { if (lowerit(*abbr) != lowerit(*word)) @@ -3622,7 +3622,7 @@ itsabbr(register const char *abbr, register const char *word) /* Return true if ABBR is an initial prefix of WORD, ignoring ASCII case. */ -static ATTRIBUTE_REPRODUCIBLE bool +ATTRIBUTE_REPRODUCIBLE static bool ciprefix(char const *abbr, char const *word) { do @@ -3725,14 +3725,14 @@ getfields(char *cp, char **array, int arrayelts) return nsubs; } -static ATTRIBUTE_NORETURN void +ATTRIBUTE_NORETURN static void time_overflow(void) { error(_("time overflow")); exit(EXIT_FAILURE); } -static ATTRIBUTE_REPRODUCIBLE zic_t +ATTRIBUTE_REPRODUCIBLE static zic_t oadd(zic_t t1, zic_t t2) { #ifdef ckd_add @@ -3746,7 +3746,7 @@ oadd(zic_t t1, zic_t t2) time_overflow(); } -static ATTRIBUTE_REPRODUCIBLE zic_t +ATTRIBUTE_REPRODUCIBLE static zic_t tadd(zic_t t1, zic_t t2) { #ifdef ckd_add -- 2.37.2

2 1

[PROPOSED] Simplify zdump average-taking
by Paul Eggert Dec. 1, 2022

Dec. 1, 2022

* zdump.c (hunt): Simplify slightly, so that one more instruction can be int width rather than time_t width if that’s what the compiler prefers. No need to mention C89 in the comment, as C89 support is going away soon. (The code still happens to work in C89, for what it’s worth.) --- zdump.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) diff --git a/zdump.c b/zdump.c index a020102..04953dc 100644 --- a/zdump.c +++ b/zdump.c @@ -732,13 +732,9 @@ hunt(timezone_t tz, time_t lot, time_t hit, bool only_ok) for ( ; ; ) { /* T = average of LOT and HIT, rounding down. - Avoid overflow, even on oddball C89 platforms - where / rounds down and TIME_T_MIN == -TIME_T_MAX - so lot / 2 + hit / 2 might overflow. */ - time_t t = (lot / 2 - - ((lot % 2 + hit % 2) < 0) - + ((lot % 2 + hit % 2) == 2) - + hit / 2); + Avoid overflow. */ + int rem_sum = lot % 2 + hit % 2; + time_t t = (rem_sum == 2) - (rem_sum < 0) + lot / 2 + hit / 2; if (t == lot) break; tm_ok = my_localtime_rz(tz, &t, &tm) != NULL; -- 2.38.1

3 5

Re: [tz] New change in Mexico
by Brian Inglis Nov. 30, 2022

Nov. 30, 2022

On 2022-11-29 22:37, Tim Parenti wrote: > On Tue, 29 Nov 2022 at 13:29, Paul Eggert wrote: > I expect I didn't recognize > that as including Ciudad Juárez as opposed to one of the plain Juárezes > which also exist in Mexico. > On Tue, 29 Nov 2022 at 19:50, Heitor David Pinto wrote: > The municipalities have well-defined boundaries, see this map: > https://cuentame.inegi.org.mx/mapas/pdf/entidades/div_municipal/chihmpios.p… > Ah, that seems to be part of the issue — while the word "municipality" is often > used to refer to an individual settlement in US English, a "municipio" in Mexico > (and similar words in many other countries) is a second-level administrative > division, which is more akin to, say, a county in much of the US or a civil > parish in much of Britain. That is, these lists are not of individual > settlements which are far apart (although they often share names with such), but > rather of well-defined regions — administrative subdivisions of states — which, > as Heitor points out, cover the entire border. > https://en.wikipedia.org/wiki/Municipio <https://en.wikipedia.org/wiki/Municipio> > https://en.wikipedia.org/wiki/Municipalities_of_Mexico > <https://en.wikipedia.org/wiki/Municipalities_of_Mexico> > While "municipio" does, admittedly, seem to be best translated to English as > "municipality", since the English word has many meanings, it's maybe best to be > more faithful to the original Spanish use here for clarity. We've left > "municipio" untranslated in nearby commentary for Mexico before, so I've applied > the attached to the development repository to hopefully alleviate some further > confusion for others who might come to these comments unaware of some of the > details of Mexico's political subdivisions. While I don't necessarily expect > many more one-off adjustments to the new arrangement in the immediate future, if > that does come to pass, it might behoove us to further standardize our > commentary for the various Mexican zones (in a similar manner as was done for > the many Russian zones in 2014). > Anyway, that puts the dividing line between the two resultant zones along the > border between the municipio of Guadalupe to the northwest and the municipios of > Coyame del Sotol and Ojinaga to the southeast …which meets the US border between > Barrio Alto and Presidio de Pilares, about 15 km south-southeast of my > (comparatively uninformed) guess, which I suppose wasn't all too bad for a > ~900-km-long border. Same in most European countries and Canadian provinces: https://en.wikipedia.org/wiki/Local_government_in_Canada#Regional_municipal… "Depending on the province, this second tier may be called a county, regional municipality, regional district or regional county municipality." There are also some district, island, mountain resort, resort, rural, and specialized municipalities, each of which may have varying geographic extents as an administrative chunk of a province. -- 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 retirer but when there is no more to cut -- Antoine de Saint-Exupéry

1 0

[PROPOSED 1/4] Always define STD_INSPIRED
by Paul Eggert Nov. 30, 2022

Nov. 30, 2022

* private.h (STD_INSPIRED): Default to 0. All uses changed. This is purely refactoring, so that STD_INSPIRED is more like other macros in this file. --- Makefile | 2 +- localtime.c | 10 +++++----- private.h | 7 +++++-- theory.html | 2 +- 4 files changed, 12 insertions(+), 9 deletions(-) diff --git a/Makefile b/Makefile index 2c09f94..f0cbf7f 100644 --- a/Makefile +++ b/Makefile @@ -363,7 +363,7 @@ GCC_DEBUG_FLAGS = -DGCC_LINT -g3 -O3 -fno-common \ # -DNETBSD_INSPIRED=0 # to the end of the "CFLAGS=" line. Otherwise, the functions # "localtime_rz", "mktime_z", "tzalloc", and "tzfree" are added to the -# time library, and if STD_INSPIRED is also defined the functions +# time library, and if STD_INSPIRED is also defined to nonzero the functions # "posix2time_z" and "time2posix_z" are added as well. # The functions ending in "_z" (or "_rz") are like their unsuffixed # (or suffixed-by-"_r") counterparts, except with an extra first diff --git a/localtime.c b/localtime.c index 1d22d35..462d054 100644 --- a/localtime.c +++ b/localtime.c @@ -1697,7 +1697,7 @@ gmtime(const time_t *timep) return gmtime_r(timep, &tm); } -#ifdef STD_INSPIRED +#if STD_INSPIRED struct tm * offtime(const time_t *timep, long offset) @@ -1706,7 +1706,7 @@ offtime(const time_t *timep, long offset) return gmtsub(gmtptr, timep, offset, &tm); } -#endif /* defined STD_INSPIRED */ +#endif /* ** Return the number of leap years through the end of the given year @@ -2325,7 +2325,7 @@ mktime(struct tm *tmp) return t; } -#ifdef STD_INSPIRED +#if STD_INSPIRED time_t timelocal(struct tm *tmp) { @@ -2377,7 +2377,7 @@ leapcorr(struct state const *sp, time_t t) ** XXX--is the below the right way to conditionalize?? */ -#ifdef STD_INSPIRED +#if STD_INSPIRED /* NETBSD_INSPIRED_EXTERN functions are exported to callers if NETBSD_INSPIRED is defined, and are private otherwise. */ @@ -2462,7 +2462,7 @@ posix2time(time_t t) return t; } -#endif /* defined STD_INSPIRED */ +#endif /* STD_INSPIRED */ #if TZ_TIME_T diff --git a/private.h b/private.h index 9cb9c24..b5fdf5a 100644 --- a/private.h +++ b/private.h @@ -691,7 +691,10 @@ extern long altzone; ** declarations if time_tz is defined. */ -#ifdef STD_INSPIRED +#ifndef STD_INSPIRED +# define STD_INSPIRED 0 +#endif +#if STD_INSPIRED # if TZ_TIME_T || !defined offtime struct tm *offtime(time_t const *, long); # endif @@ -738,7 +741,7 @@ struct tm *localtime_rz(timezone_t restrict, time_t const *restrict, time_t mktime_z(timezone_t restrict, struct tm *restrict); timezone_t tzalloc(char const *); void tzfree(timezone_t); -# ifdef STD_INSPIRED +# if STD_INSPIRED # if TZ_TIME_T || !defined posix2time_z time_t posix2time_z(timezone_t, time_t) ATTRIBUTE_REPRODUCIBLE; # endif diff --git a/theory.html b/theory.html index 75e347f..f2b798b 100644 --- a/theory.html +++ b/theory.html @@ -1173,7 +1173,7 @@ The vestigial <abbr>API</abbr>s are: </li> <li> The functions that are conditionally compiled - if <code>STD_INSPIRED</code> is defined should, at this point, be + if <code>STD_INSPIRED</code> is nonzero should, at this point, be looked on primarily as food for thought. They are not in any sense "standard compatible" – some are not, in fact, specified in <em>any</em> standard. -- 2.38.1

1 3

New change in Mexico
by Heitor David Pinto Nov. 30, 2022

Nov. 30, 2022

Hello, Mexico is planning to change time zones again. Last month, the whole state of Chihuahua moved to UTC-6 without DST, causing a time change across the border with the US for part of the year. Now the northern municipalities want to have the same time zone as the respective neighboring cities in the US, for example Juárez in UTC-7 with DST, matching El Paso, and Ojinaga in UTC-6 with DST, matching Presidio. The congress of Chihuahua and the federal Chamber and Senate all approved the change, following the combined procedure specified in the new time zone law, and now it awaits the president's approval. The change should occur on the day after the official publication of the law, again on short notice.https://politica.expansion.mx/congreso/2022/11/24/aprueban-que-ayuntamientos-fronterizos-de-chihuahua-tengan-horario-de-verano But there is another problem. The actual law that was passed says that Juárez and Ojinaga will observe UTC-7 and UTC-6 during the DST period, not outside of it. I think that this was an oversight as a result of trying to pass the law through all the chambers very quickly, but I can't believe that no one noticed it. There is a legislative procedure in Mexico called "fe de erratas" to correct obvious mistakes without having to pass the law again, but in this case the correction would have to alter multiple sections so I'm not sure if it will be possible. Either way, I expect confusion in the next days.http://gaceta.diputados.gob.mx/PDF/65/2022/nov/20221124-VII.pdf Regards,Dave

5 12

Re: [tz] [PROPOSED 2/5] Pacify gcc -std=c89 -pedantic with new attributes
by Brian Inglis Nov. 29, 2022

Nov. 29, 2022

On 2022-11-28 12:47, Paul Eggert via tz wrote: > On 2022-11-28 05:40, Philip Newton wrote: >> "have has" is a bit ugly.... > > Yes, I originally used the name "HAVE___HAS_C_ATTRIBUTE", since it tests whether > __has_c_attribute exists and that's the naming convention, but I changed it to > HAVE_HAS_C_ATTRIBUTE as I worried that the three adjacent underscores would be > confusing. If people prefer HAVE___HAS_C_ATTRIBUTE we can easily change to that. > The longer name would be more consistent. "4.2.8 __has_c_attribute The special operator __has_c_attribute (operand) may be used in ‘#if’ and ‘#elif’ expressions in C code to test whether the attribute referenced by its operand is recognized by GCC in attributes using the ‘[[]]’ syntax." How about HAVE_C23_ATTRIBUTE_CHECK, HAVE_C2X_ATTRIBUTE_CHECK, or something similarly more meaningful and less confusing? -- 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 retirer but when there is no more to cut -- Antoine de Saint-Exupéry

2 1

[tz-announce] 2022g release of tz code and data available
by Paul Eggert via tz-announce Nov. 29, 2022

Nov. 29, 2022

The 2022g release of the tz code and data is available. This release contains the following changes: Briefly: The northern edge of Chihuahua changes to US timekeeping. Much of Greenland stops changing clocks after March 2023. Fix some pre-1996 timestamps in northern Canada. C89 is now deprecated; please use C99 or later. Portability fixes for AIX, libintl, MS-Windows, musl, z/OS In C code, use more C23 features if available. C23 timegm now supported by default Fixes for unlikely integer overflows Changes to future timestamps In the Mexican state of Chihuahua, the border strip near the US will change to agree with nearby US locations on 2022-11-30. The strip's western part, represented by Ciudad Juárez, switches from -06 all year to -07/-06 with US DST rules, like El Paso, TX. The eastern part, represented by Ojinaga, will observe US DST next year, like Presidio, TX. (Thanks to Heitor David Pinto.) A new Zone America/Ciudad_Juarez splits from America/Ojinaga. Much of Greenland, represented by America/Nuuk, stops observing winter time after March 2023, so its daylight saving time becomes standard time. (Thanks to Jonas Nyrup and Jürgen Appel.) Changes to past timestamps Changes for pre-1996 northern Canada (thanks to Chris Walton): Merge America/Iqaluit and America/Pangnirtung into the former, with a backward compatibility link for the latter name. There is no good evidence the two locations differ since 1970. This change affects pre-1996 America/Pangnirtung timestamps. Cambridge Bay, Inuvik, Iqaluit, Rankin Inlet, Resolute and Yellowknife did not observe DST in 1965, and did observe DST from 1972 through 1979. Whitehorse moved from -09 to -08 on 1966-02-27, not 1967-05-28. Colombia's 1993 fallback was 02-06 24:00, not 04-04 00:00. (Thanks to Alois Treindl.) Singapore's 1981-12-31 change was at 16:00 UTC (23:30 local time), not 24:00 local time. (Thanks to Geoff Clare via Robert Elz.) Changes to code Although tzcode still works with C89, bugs found in recent routine maintenance indicate that bitrot has set in and that in practice C89 is no longer used to build tzcode. As it is a maintenance burden, support for C89 is planned to be removed soon. Instead, please use compilers compatible with C99, C11, C17, or C23. timegm, which tzcode implemented in 1989, will finally be standardized 34 years later as part of C23, so timegm is now supported even if STD_INSPIRED is not defined. Fix bug in zdump's tzalloc emulation on hosts that lack tm_zone. (Problem reported by Đoàn Trần Công Danh.) Fix bug in zic on hosts where malloc(0) yields NULL on success. (Problem reported by Tim McBrayer for AIX 6.1.) Fix zic configuration to avoid linkage failures on some platforms. (Problems reported by Gilmore Davidson and Igor Ivanov.) Work around MS-Windows nmake incompatibility with POSIX. (Problem reported by Manuela Friedrich.) Port mktime and strftime to debugging platforms where accessing uninitialized data has undefined behavior (strftime problem reported by Robert Elz). Check more carefully for unlikely integer overflows, preferring C23 <stdckdint.h> to overflow checking by hand, as the latter has had obscure bugs. Changes to build procedure New Makefile rule check_mild that skips checking whether Link lines are in the file 'backward'. (Inspired by a suggestion from Stephen Colebourne.) Here are links to the release files: https://www.iana.org/time-zones/repository/releases/tzcode2022g.tar.gz https://www.iana.org/time-zones/repository/releases/tzdata2022g.tar.gz https://www.iana.org/time-zones/repository/releases/tzdb-2022g.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 4e604e41115182c3ab1185bab4c888822511348e dated 2022-11-29 08:58:31 -0800 and tagged '2022g' in the development GitHub repository at <https://github.com/eggert/tz>. Here are the SHA-512 checksums for the release files: 5188edd8d05238a88da734cf78fadfb57640d4db3e7a0a9dddd70e3071e16b6eebc2e2ab17109b7dafafae68abfbc857df481cfdc3ffe63f7eb1569ea0b5997a tzcode2022g.tar.gz 7f79394295e00e3a24ebdbf9af3bc454a65f432a93b517e7e96c7f9db9949f6f5fdae9892a9d3789ff44ae0eb1bfe4744d36976b4624659af951d26414f94e65 tzdata2022g.tar.gz f471046189f519de5735ac2d8c3edb27cbe925247b06f44634e700e5e4453ec5f715d85256fc74d300bcdaa070a7600fcc054327f2dfe743ab3c0fe404ff83c1 tzdb-2022g.tar.lz Here are the GPG digital signatures for the release files: -----BEGIN PGP SIGNATURE----- iQIzBAABCgAdFiEEfjeSqdis99YzvBWI7ZfpDmKqfjQFAmOGOvAACgkQ7ZfpDmKq fjSI4w/+PkVPCILI34+QeXIm6LLPhEPLxtb7q3rvfoBXk/qyottcPJvQXSAsqX/c Hus6lCNK0COyICVn48rf3ss+GNxLev/ZTMVLUBkH8WrgSe+5lPDjmaZjkbzIGPs8 L9TqE60K8DyeLNnZPf4Pkpx6o5r7uO3iQSN5qFyj5H82mDJ4vcQLZSNvDIA4+sRm +qmsBcn6Gk3EU3u3/YrxhuEY0ymUrJvm5LSG62CqLwA1Pd+REF70Ir5LdO4yJoUS blWe/CLUcnRfzMEK9/NGa1z3GtkisPYBCPnYJ93m2n4ASvFQ+R3t7NCEY2W+96ce 3kSPIX5W2g0/gcDTi1bxdLnQkO4RxtSQnlCjCae58dQNUgOa2X8mdRniRjjYC059 JY35VUzmX9wCPpcuq0iMbpOEWXeWga0ZDRmlZtZCwPKxkw+AhoJQVp5jxGNa4kwZ CZDjQga3ObvN8aQqMi9TpabJ4m2ZdN1yHQ7OTN80N67tHYMAbW0aY3J0JRIVjTHo P+8Y6FaCLzJA73BgnR7faiILXIPtXE54X+koo8x2Uwtrl8/wHsaPTippq96ts6Y6 3Zdr3EAQgAPhYtUgtHoIbLFBLjmyL5sUXOTwG/gMkUYXVkDpeBnc+BHrkv0Sjaw7 Vzz5HthhnSblP9bLB3KPTbbM0t+aypsSOYOSZWnIK8ANI9BdrNQ= =tSyt -----END PGP SIGNATURE----- -----BEGIN PGP SIGNATURE----- iQIzBAABCgAdFiEEfjeSqdis99YzvBWI7ZfpDmKqfjQFAmOGOvAACgkQ7ZfpDmKq fjSedhAAu8XkzqB8NmXDYfEurBCUPe7zCBxBe7yqdt5W+tDA9snOtab3d8Oq7Lsb vNZfh4qHIxNbEYpoXOc0Qce3XKXGgyKDmZeFYQJ0ziqoZXetNLQEmg7T/H2m3GHl eo7K3w2hvyBB26FuDtimE5pzuJo9qJLXDp2SbOAbKa1w5HibyvZBOfcQozZ4CBmE mBc8LzyFY3zOhiy7CGYskqOzatNa0Eid95RkK2RXXTnN965N1qdyOz3z9SkJFH9h MR84VGYBmk0UGdPGmp1imkGmIrO/8AFny19ECCe4F74o2DFSPzANOaEh+i3fIP3d GNR96la9GGopmna+VNQt6zxdw4DHRoNkAupnJ2NVM0lbAMOG0mmwCFkCzW4gzTVm 7NdAtZyH+HjiKYPdQNIHbsvl0K1qIo1yXrb4jTd8eXaxA65msCOOwjqWk7a8ZjrJ NrLNmGeLow9svofEAeLpc9BUgwF90JGIlb+tPYovyj/xSU9vhQl1ZEmFYYdaN781 rGZ5jKtwBZWoAxW1Copf8S3jcxleSVMFo6MTEU+K+svZLd8SH5BnETa0fYWSTzKA uh2Y4V6pFe0/JVVC15mYkytbP6TC6prRmi44jRHtVRDtizGd7FFxYacIApkhBVi2 Qb8Qy+oZySwL+dbRlu+bA9nHIdACtxRouVEWpAecCBLaK3kBb7I= =6g+2 -----END PGP SIGNATURE----- -----BEGIN PGP SIGNATURE----- iQIzBAABCgAdFiEEfjeSqdis99YzvBWI7ZfpDmKqfjQFAmOGOvIACgkQ7ZfpDmKq fjT6/A/5AQdjRCEQngohVPQcxWE57FhHN+3Yq72iCbi64MhnL8sR+fHcni8CRb4f SIxi/tXeLdKJVV8+w7TBeQIDBDHrWXla2NSTHNcGREdp/wMoWagxBzAx3g2qQ7eT Qz6JhG4pHlOGK3u0/RESuCnhS/XmKTYj6BeDNc33LQ4iB0Lh+o9O/rz6jQyO3Y3L 84QMNJ4Edq/pOz9d9T1/cWO1j8Cn4Mled9WdCyOqZnhzRcJdbi/gDydC5aJIj1Fx 5OUrRwblIRVxMbK+kZA4D0lHR56KZb7NsaFGfsnW0m5zHUpXoGNYaFelKlWoRblO OEneM34IgNxsLhe9J9r/epSWB/14PmaY0zdXlNmKqELa7zYiJkiAwucPWtpxkPJG q6WPt4Pk4bWcud9esdgG7vjBz3ffrz2hUrmSMxuwLPHLuPRLrBoeMAmecFixGpea BUmR+XCPV9m19zcnvshBFGuLTNwFZSWBlUB9lWIocXc9cIMo4W/gNdUhnjb9L0+z VtAJzws1VcRV5zpkanwt+K50xjUlunxcRgLbLY2oMCpeDQyxH6DTnUr8RdJeMmvM ZPJYekxG0W8/eEjDOUFIQqvOtK1sQVhCZhjeciDW7to/K1p7DJy+g5g7lJn5KK4B Oz8vqmhdvdoIAK5uezPILpyJT7UpdUhmZ1WwQRqhk6G3jXPigQc= =Kyhr -----END PGP SIGNATURE-----

1 0

America/Nuuk changes from UTC-2 to UTC-2
by Jonas Nyrup Nov. 29, 2022

Nov. 29, 2022

On last Saturday in October 2023 when DST ends America/Nuuk will switch from -03/-02 to -02/-01 https://sermitsiaq.ag/forslagtidsforskel-danmark-mindskes-sommertid-beholdes describes the proposed bill about switching from UTC-3 to UTC-2 and that is not about abolishing DST as they do not want to make a decision on that matter before the rest of EU. https://sermitsiaq.ag/groenland-skifte-tidszone-trods-bekymringer writes that on November 24th 2022 the Parliament of Greenland approved the bill and that *the bill* will take effect on March 25th 2023 at 22:00 when DST begins. Both articles translates well with Google Translate. Inatsisartut = the Parliament of Greenland Naalakkersuisut = the government of Greenland I assume the official document will be published on https://naalakkersuisut.gl/kundgoerelser/2022?sc_lang=da Best regards Jonas Nyrup

2 2

[PROPOSED 0/5] Finish off support for C23 standard attributes
by Paul Eggert Nov. 28, 2022

Nov. 28, 2022

The attached patches update tzcode to use the remaining C23 standard attributes that look useful for tzcode. I don't see any tzcode use for the C23 attributes 'deprecated' and 'nodiscard'. Paul Eggert (5): Use C23 [[noreturn]] if available Pacify gcc -std=c89 -pedantic with new attributes Use C23 [[reproducible]] if available. Use C23 [[unsequenced]] if available Use C23 [[fallthrough]] if available Makefile | 2 +- date.c | 2 +- localtime.c | 4 +-- private.h | 77 ++++++++++++++++++++++++++++++++++++++++++++--------- zdump.c | 10 +++---- zic.c | 32 +++++++++++----------- 6 files changed, 90 insertions(+), 37 deletions(-) -- 2.37.2

2 7

Doubts about a typo fix
by Alejandro Colomar Nov. 26, 2022

Nov. 26, 2022

Hi Paul, I don't undertstand the commit shown below. It breaks the '-' symbol, which now is a hyphen. See the table mentioned yesterday: ┌──────────────────────────────────────────────────────────────────┐ │Keycap Appearance and meaning Special character and meaning │ ├──────────────────────────────────────────────────────────────────┤ │" " neutral double quote \[dq] neutral double quote │ │' ’ closing single quote \[aq] neutral apostrophe │ │- ‐ hyphen \- or \[-] minus sign/Unix dash │ │\ (escape character) \e or \[rs] reverse solidus │ │^ ˆ modifier circumflex \(ha circumflex/caret/“hat” │ │` ‘ opening single quote \(ga grave accent │ │~ ˜ modifier tilde \(ti tilde │ └──────────────────────────────────────────────────────────────────┘ So, at least, it should be (but I believe the initial code was correct): .q "zic \*\-r @$(date +%s)" However, I wonder what that \* is intending to do there (I can see no difference in my screen with or without it). Cheers, Alex --- commit 918e10e8963b3c0d38d3b5fb8ec9cf08ecd03757 Author: Paul Eggert <eggert(a)cs.ucla.edu> Date: Tue Jul 12 06:26:53 2022 -0700 * zic.8: fix minus typo diff --git a/zic.8 b/zic.8 index 0cd0781e..e8816e5b 100644 --- a/zic.8 +++ b/zic.8 @@ -145,7 +145,7 @@ .SH OPTIONS 31-bit signed integers. On platforms with GNU .BR date , -.q "zic \-r @$(date +%s)" +.q "zic \*-r @$(date +%s)" omits data intended for past timestamps. Although this option typically reduces the output file's size, the size can increase due to the need to represent the timestamp range -- <http://www.alejandro-colomar.es/>

5 7

New law about Time and Time zones in Greenland
by Jürgen Appel Nov. 25, 2022

Nov. 25, 2022

Dear all, i was made aware that on Thursday, 2022-11-24, a law proposal concerning time and time zones for Greenland was voted upon in Greenland Parliament: Reference (in Danish and Greenlandic): https://ina.gl/samlinger/oversigt-over-samlinger/samling/dagsordener/ dagsorden.aspx?lang=da&day=24-11-2022 https://ina.gl/media/2553527/pkt17_em2022_tidens_-bestemmelse_lov_da.pdf Translated with www.DeepL.com/Translator and annotated: =====[Proposal]===== Inatsisartut (Parliament of Greenland) Act on the Regulation of Time § 1. The standard time for Greenland shall be UTC -2. Clause 2. Naalakkersuisut (the Cabinet of Greenland) may establish rules that specified areas shall follow a different time zone than that determined in paragraph 1. Clause 3. Naalakkersuisut may establish rules on summer time. § 2. This Inatsisartut Act shall enter into force on 25 March 2023 at 22:00. Clause 2. At the same time, Landstingslov nr. 14 of 19 October 1989 regarding the application of summer time is repealed. Greenland Self-Government, x. xx xxx The President of Naalakkersuisut =================== If I understand this correctly, from the next planned switch to summer time, Greenland will permanently stay at that time, i.e. no switch back to winter time in 2023 will occur. Whether a switch to a shifted summer time will occur in spring 2024 is not completely transparent to me, as in some remarks (https://ina.gl/media/ 2554772/pkt17_em2022_tidens_bestemmelse_bet_2beh_da.pdf) it is written that no changes in the summer time regulations will be applied, and that this will occur only after EU has decided about the same topic. If some danish native speaker, or better yet someone from Greenland could confirm/clarify this, that would be nice. Also concerning "specified areas", i.e. today Ittoqqortoormiit (now UTC -1), Danmarkshavn (UTC +0), and Pituffik/Thule Air Base (UTC-4), I am unsure whether any changes are scheduled or whether the cabinet plans to keep/ reestablish the current rules. Best regards, Jürgen Appel -- Jürgen Appel Senior Scientist Denmark's National Metrology Institute Dansk Fundamental Metrologi, DFM A/S (dfm.dk) Kogle Allé 5 DK-2970 Hørsholm Denmark Mobile: +45 25459049 Email: jap(a)dfm.dk VAT: DK-29217939

2 1

Re: [tz] [PATCH v2 1/4] zic.8: Add public domain notice
by Brian Inglis Nov. 25, 2022

Nov. 25, 2022

On 2022-11-23 11:32, Paul Eggert via tz wrote: > The file already contains a public-domain notice at the bottom, in a > human-readable format that is visible to anybody who looks at the printable > version of the man page. No public domain notices appear in the *.?.txt source files or in the generated man pages. It may be useful to specify that the document, data, program, and source files for all components are in the public domain in the LICEN[CS]E sections. It would also be useful if you also added SPDX licence comments like man-pages after either your line(s) or theirs, such as: .\" SPDX-License-Identifier: LicenseRef-IANA-TZ-Public-Domain [suggested] .\" SPDX-FileCopyrightText: project authors [suggested] and also in the generated output?, as it may save downstreams doing so as licensing requirements are enforced by packaging tools and repos. -- 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 retirer but when there is no more to cut -- Antoine de Saint-Exupéry

2 1

[PROPOSED] Use C23 [[maybe_unused]] if available
by Paul Eggert Nov. 21, 2022

Nov. 21, 2022

* Makefile (GCC_DEBUG_FLAGS): Omit -Wno-unused-parameters, as the code is now safe for gcc -Wunused-parameters. * private.h (ATTRIBUTE_MAYBE_UNUSED): New macro. * localtime.c (gmtsub): * strftime.c (strftime_l) [HAVE_STRFTIME_L]: * zdump.c (gmtoff): Use it to mark possibly-unused parameters. * zdump.c (hunt): * zic.c (getleapdatetime): Omit unused arg. All callers changed. --- Makefile | 2 +- localtime.c | 4 ++-- private.h | 13 +++++++++++++ strftime.c | 2 +- zdump.c | 13 +++++++------ zic.c | 6 +++--- 6 files changed, 27 insertions(+), 13 deletions(-) diff --git a/Makefile b/Makefile index 92d78cb4..b2c6ff6f 100644 --- a/Makefile +++ b/Makefile @@ -300,7 +300,7 @@ GCC_DEBUG_FLAGS = -DGCC_LINT -g3 -O3 -fno-common \ -Wtrampolines -Wundef -Wuninitialized -Wunused-macros -Wuse-after-free=3 \ -Wvariadic-macros -Wvla -Wwrite-strings \ -Wno-address -Wno-format-nonliteral -Wno-sign-compare \ - -Wno-type-limits -Wno-unused-parameter + -Wno-type-limits # # If your system has a "GMT offset" field in its "struct tm"s # (or if you decide to add such a field in your system's "time.h" file), diff --git a/localtime.c b/localtime.c index 7faa5d32..4dd9d259 100644 --- a/localtime.c +++ b/localtime.c @@ -1662,8 +1662,8 @@ localtime_r(const time_t *timep, struct tm *tmp) */ static struct tm * -gmtsub(struct state const *sp, time_t const *timep, int_fast32_t offset, - struct tm *tmp) +gmtsub(ATTRIBUTE_MAYBE_UNUSED struct state const *sp, time_t const *timep, + int_fast32_t offset, struct tm *tmp) { register struct tm * result; diff --git a/private.h b/private.h index 4d40582e..fee18cb2 100644 --- a/private.h +++ b/private.h @@ -433,6 +433,19 @@ typedef unsigned long uintmax_t; # define ATTRIBUTE_FORMAT(spec) /* empty */ #endif +#ifdef __has_c_attribute +# if __has_c_attribute(__maybe_unused__) +# define ATTRIBUTE_MAYBE_UNUSED [[__maybe_unused__]] +# endif +#endif +#ifndef ATTRIBUTE_MAYBE_UNUSED +# if 2 < __GNUC__ + (7 <= __GNUC_MINOR__) +# define ATTRIBUTE_MAYBE_UNUSED __attribute__ ((__unused__)) +# else +# define ATTRIBUTE_MAYBE_UNUSED /* empty */ +# endif +#endif + #if !defined _Noreturn && __STDC_VERSION__ < 201112 # if 2 < __GNUC__ + (8 <= __GNUC_MINOR__) # define _Noreturn __attribute__((__noreturn__)) diff --git a/strftime.c b/strftime.c index 9ea40fec..b23b6101 100644 --- a/strftime.c +++ b/strftime.c @@ -117,7 +117,7 @@ static char * _yconv(int, int, bool, bool, char *, char const *); #if HAVE_STRFTIME_L size_t strftime_l(char *s, size_t maxsize, char const *format, struct tm const *t, - locale_t locale) + ATTRIBUTE_MAYBE_UNUSED locale_t locale) { /* Just call strftime, as only the C locale is supported. */ return strftime(s, maxsize, format, t); diff --git a/zdump.c b/zdump.c index 89bcd25c..de1e1116 100644 --- a/zdump.c +++ b/zdump.c @@ -91,7 +91,7 @@ static bool errout; static char const *abbr(struct tm const *); static intmax_t delta(struct tm *, struct tm *) ATTRIBUTE_PURE; static void dumptime(struct tm const *); -static time_t hunt(timezone_t, char *, time_t, time_t, bool); +static time_t hunt(timezone_t, time_t, time_t, bool); static void show(timezone_t, char *, time_t, bool); static void showextrema(timezone_t, char *, time_t, struct tm *, time_t); static void showtrans(char const *, struct tm const *, time_t, char const *, @@ -627,7 +627,7 @@ main(int argc, char *argv[]) || (ab && (delta(&newtm, &tm) != newt - t || newtm.tm_isdst != tm.tm_isdst || strcmp(abbr(&newtm), ab) != 0))) { - newt = hunt(tz, argv[i], t, newt, false); + newt = hunt(tz, t, newt, false); newtmp = localtime_rz(tz, &newt, &newtm); newtm_ok = newtmp != NULL; if (iflag) @@ -707,7 +707,7 @@ yeartot(intmax_t y) return t; } -/* Search for a discontinuity in timezone TZ with name NAME, in the +/* Search for a discontinuity in timezone TZ, in the timestamps ranging from LOT through HIT. LOT and HIT disagree about some aspect of timezone. If ONLY_OK, search only for definedness changes, i.e., localtime succeeds on one side of the @@ -715,7 +715,7 @@ yeartot(intmax_t y) before the transition from LOT's settings. */ static time_t -hunt(timezone_t tz, char *name, time_t lot, time_t hit, bool only_ok) +hunt(timezone_t tz, time_t lot, time_t hit, bool only_ok) { static char * loab; static ptrdiff_t loabsize; @@ -807,7 +807,8 @@ adjusted_yday(struct tm const *a, struct tm const *b) my_gmtime_r and use its result instead of B. Otherwise, B is the possibly nonnull result of an earlier call to my_gmtime_r. */ static long -gmtoff(struct tm const *a, time_t *t, struct tm const *b) +gmtoff(struct tm const *a, ATTRIBUTE_MAYBE_UNUSED time_t *t, + ATTRIBUTE_MAYBE_UNUSED struct tm const *b) { #ifdef TM_GMTOFF return a->TM_GMTOFF; @@ -878,7 +879,7 @@ static void showextrema(timezone_t tz, char *zone, time_t lo, struct tm *lotmp, time_t hi) { struct tm localtm[2], gmtm[2]; - time_t t, boundary = hunt(tz, zone, lo, hi, true); + time_t t, boundary = hunt(tz, lo, hi, true); bool old = false; hi = (SECSPERDAY < hi - boundary ? boundary + SECSPERDAY diff --git a/zic.c b/zic.c index 8f8088ec..7a7a5094 100644 --- a/zic.c +++ b/zic.c @@ -1970,7 +1970,7 @@ inzsub(char **fields, int nfields, bool iscont) } static zic_t -getleapdatetime(char **fields, int nfields, bool expire_line) +getleapdatetime(char **fields, bool expire_line) { register const char * cp; register const struct lookup * lp; @@ -2048,7 +2048,7 @@ inleap(char **fields, int nfields) if (nfields != LEAP_FIELDS) error(_("wrong number of fields on Leap line")); else { - zic_t t = getleapdatetime(fields, nfields, false); + zic_t t = getleapdatetime(fields, false); if (0 <= t) { struct lookup const *lp = byword(fields[LP_ROLL], leap_types); if (!lp) @@ -2076,7 +2076,7 @@ inexpires(char **fields, int nfields) else if (0 <= leapexpires) error(_("multiple Expires lines")); else - leapexpires = getleapdatetime(fields, nfields, true); + leapexpires = getleapdatetime(fields, true); } static void -- 2.37.2

2 2

[PROPOSED] Prefer <stdckdint.h> for overflow checking
by Paul Eggert Nov. 20, 2022

Nov. 20, 2022

If <stdckdint.h> is available, or if its macros can be implemented easily, use the macros for overflow checking. This should help avoid future integer overflow bugs. * date.c (dogmt, timeout): * localtime.c (localsub, timesub, increment_overflow) (increment_overflow32, increment_overflow_time, time2sub): * zdump.c (sumsize, tzalloc): * zic.c (size_sum, size_product, grow_nitems_alloc, oadd, tadd): Prefer ckd_add etc. to doing overflow checking by hand. * private.h (HAVE_STDCKDINT_H): Default to true if __has_include(<stdckdint.h>). Include <stdckdint.h> if available. (ckd_add, ckd_sub, ckd_mul): Provide substitutes on recent-enough GCC-compatible compilers. --- Makefile | 2 ++ NEWS | 5 ++++- date.c | 10 ++++++++++ localtime.c | 33 +++++++++++++++++++++++++++++++++ private.h | 29 +++++++++++++++++++++++++++++ zdump.c | 12 ++++++++++++ zic.c | 31 +++++++++++++++++++++++++++++++ 7 files changed, 121 insertions(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 5ec0d4a..92d78cb 100644 --- a/Makefile +++ b/Makefile @@ -230,6 +230,8 @@ LDLIBS= # functions like 'link' or variables like 'tzname' required by POSIX # -DHAVE_SETENV=0 if your system lacks the setenv function # -DHAVE_SNPRINTF=0 if your system lacks the snprintf function +# -DHAVE_STDCKDINT_H=0 if neither <stdckdint.h> nor substitutes like +# __builtin_add_overflow work* # -DHAVE_STDINT_H=0 if <stdint.h> does not work* # -DHAVE_STRFTIME_L if <time.h> declares locale_t and strftime_l # -DHAVE_STRDUP=0 if your system lacks the strdup function diff --git a/NEWS b/NEWS index 7c46d3c..a401bee 100644 --- a/NEWS +++ b/NEWS @@ -4,6 +4,7 @@ News for the tz database Fix some pre-1996 timestamps in northern Canada. C89 is now deprecated; please use C99 or later. Portability fixes for AIX, libintl, MS-Windows, musl, z/OS + In C code, use more C23 features if available. C23 timegm now supported by default Fixes for unlikely integer overflows @@ -55,7 +56,9 @@ Unreleased, experimental changes uninitialized data has undefined behavior (strftime problem reported by Robert Elz). - Check more carefully for unlikely integer overflows. + Check more carefully for unlikely integer overflows, preferring + C23 <stdckdint.h> to overflow checking by hand, as the latter has + had obscure bugs. Changes to build procedure diff --git a/date.c b/date.c index cbc0ec1..1319462 100644 --- a/date.c +++ b/date.c @@ -122,8 +122,14 @@ dogmt(void) for (n = 0; environ[n] != NULL; ++n) continue; +#if defined ckd_add && defined ckd_mul + if (!ckd_add(&n, n, 2) && !ckd_mul(&n, n, sizeof *fakeenv) + && n <= SIZE_MAX) + fakeenv = malloc(n); +#else if (n <= min(PTRDIFF_MAX, SIZE_MAX) / sizeof *fakeenv - 2) fakeenv = malloc((n + 2) * sizeof *fakeenv); +#endif if (fakeenv == NULL) { fprintf(stderr, _("date: Memory exhausted\n")); errensure(); @@ -187,8 +193,12 @@ timeout(FILE *fp, char const *format, struct tm const *tmp) ptrdiff_t size = 1024 / 2; for ( ; ; ) { +#ifdef ckd_mul + bool bigger = !ckd_mul(&size, size, 2) && size <= SIZE_MAX; +#else bool bigger = (size <= min(PTRDIFF_MAX, SIZE_MAX) / 2 && (size *= 2, true)); +#endif char *newcp = bigger ? realloc(cp, size) : NULL; if (!newcp) { fprintf(stderr, diff --git a/localtime.c b/localtime.c index 7326567..7faa5d3 100644 --- a/localtime.c +++ b/localtime.c @@ -1565,6 +1565,14 @@ localsub(struct state const *sp, time_t const *timep, int_fast32_t setname, return NULL; /* "cannot happen" */ result = localsub(sp, &newt, setname, tmp); if (result) { +#if defined ckd_add && defined ckd_sub + if (t < sp->ats[0] + ? ckd_sub(&result->tm_year, + result->tm_year, years) + : ckd_add(&result->tm_year, + result->tm_year, years)) + return NULL; +#else register int_fast64_t newy; newy = result->tm_year; @@ -1574,6 +1582,7 @@ localsub(struct state const *sp, time_t const *timep, int_fast32_t setname, if (! (INT_MIN <= newy && newy <= INT_MAX)) return NULL; result->tm_year = newy; +#endif } return result; } @@ -1783,6 +1792,12 @@ timesub(const time_t *timep, int_fast32_t offset, y = newy; } +#ifdef ckd_add + if (ckd_add(&tmp->tm_year, y, -TM_YEAR_BASE)) { + errno = EOVERFLOW; + return NULL; + } +#else if (!TYPE_SIGNED(time_t) && y < TM_YEAR_BASE) { int signed_y = y; tmp->tm_year = signed_y - TM_YEAR_BASE; @@ -1793,6 +1808,7 @@ timesub(const time_t *timep, int_fast32_t offset, errno = EOVERFLOW; return NULL; } +#endif tmp->tm_yday = idays; /* ** The "extra" mods below avoid overflow problems. @@ -1867,6 +1883,9 @@ ctime_r(const time_t *timep, char *buf) static bool increment_overflow(int *ip, int j) { +#ifdef ckd_add + return ckd_add(ip, *ip, j); +#else register int const i = *ip; /* @@ -1879,22 +1898,30 @@ increment_overflow(int *ip, int j) return true; *ip += j; return false; +#endif } static bool increment_overflow32(int_fast32_t *const lp, int const m) { +#ifdef ckd_add + return ckd_add(lp, *lp, m); +#else register int_fast32_t const l = *lp; if ((l >= 0) ? (m > INT_FAST32_MAX - l) : (m < INT_FAST32_MIN - l)) return true; *lp += m; return false; +#endif } static bool increment_overflow_time(time_t *tp, int_fast32_t j) { +#ifdef ckd_add + return ckd_add(tp, *tp, j); +#else /* ** This is like ** 'if (! (TIME_T_MIN <= *tp + j && *tp + j <= TIME_T_MAX)) ...', @@ -1906,6 +1933,7 @@ increment_overflow_time(time_t *tp, int_fast32_t j) return true; *tp += j; return false; +#endif } static bool @@ -2029,11 +2057,16 @@ time2sub(struct tm *const tmp, return WRONG; } } +#ifdef ckd_add + if (ckd_add(&yourtm.tm_year, y, -TM_YEAR_BASE)) + return WRONG; +#else if (increment_overflow32(&y, -TM_YEAR_BASE)) return WRONG; if (! (INT_MIN <= y && y <= INT_MAX)) return WRONG; yourtm.tm_year = y; +#endif if (yourtm.tm_sec >= 0 && yourtm.tm_sec < SECSPERMIN) saved_seconds = 0; else if (yourtm.tm_year < EPOCH_YEAR - TM_YEAR_BASE) { diff --git a/private.h b/private.h index 4315a85..4d40582 100644 --- a/private.h +++ b/private.h @@ -392,6 +392,35 @@ typedef unsigned long uintmax_t; # define SIZE_MAX ((size_t) -1) #endif +/* Support ckd_add, ckd_sub, ckd_mul on C23 or recent-enough GCC-like + hosts, unless compiled with -DHAVE_STDCKDINT_H=0 or with pre-C23 EDG. */ +#if !defined HAVE_STDCKDINT_H && defined __has_include +# if __has_include(<stdckdint.h>) +# define HAVE_STDCKDINT_H true +# endif +#endif +#ifdef HAVE_STDCKDINT_H +# if HAVE_STDCKDINT_H +# include <stdckdint.h> +# endif +#elif defined __EDG__ +/* Do nothing, to work around EDG bug <https://bugs.gnu.org/53256>. */ +#elif defined __has_builtin +# if __has_builtin(__builtin_add_overflow) +# define ckd_add(r, a, b) __builtin_add_overflow(a, b, r) +# endif +# if __has_builtin(__builtin_sub_overflow) +# define ckd_sub(r, a, b) __builtin_sub_overflow(a, b, r) +# endif +# if __has_builtin(__builtin_mul_overflow) +# define ckd_mul(r, a, b) __builtin_mul_overflow(a, b, r) +# endif +#elif 7 <= __GNUC__ +# define ckd_add(r, a, b) __builtin_add_overflow(a, b, r) +# define ckd_sub(r, a, b) __builtin_sub_overflow(a, b, r) +# define ckd_mul(r, a, b) __builtin_mul_overflow(a, b, r) +#endif + #if 3 <= __GNUC__ # define ATTRIBUTE_CONST __attribute__((const)) # define ATTRIBUTE_MALLOC __attribute__((__malloc__)) diff --git a/zdump.c b/zdump.c index 669d8fd..89bcd25 100644 --- a/zdump.c +++ b/zdump.c @@ -137,9 +137,15 @@ size_overflow(void) static ATTRIBUTE_PURE ptrdiff_t sumsize(size_t a, size_t b) { +#ifdef ckd_add + ptrdiff_t sum; + if (!ckd_add(&sum, a, b) && sum <= SIZE_MAX) + return sum; +#else ptrdiff_t sum_max = min(PTRDIFF_MAX, SIZE_MAX); if (a <= sum_max && b <= sum_max - a) return a + b; +#endif size_overflow(); } @@ -254,9 +260,15 @@ tzalloc(char const *val) ptrdiff_t initial_nenvptrs = 1; /* Counting the trailing NULL pointer. */ while (*e++) { +# ifdef ckd_add + if (ckd_add(&initial_nenvptrs, initial_envptrs, 1) + || SIZE_MAX < initial_envptrs) + size_overflow(); +# else if (initial_nenvptrs == min(PTRDIFF_MAX, SIZE_MAX) / sizeof *environ) size_overflow(); initial_nenvptrs++; +# endif } fakeenv0size = sumsize(valsize, valsize); fakeenv0size = max(fakeenv0size, 64); diff --git a/zic.c b/zic.c index a69d4e0..8f8088e 100644 --- a/zic.c +++ b/zic.c @@ -475,18 +475,30 @@ size_overflow(void) static ATTRIBUTE_PURE ptrdiff_t size_sum(size_t a, size_t b) { +#ifdef ckd_add + ptrdiff_t sum; + if (!ckd_add(&sum, a, b) && sum <= SIZE_MAX) + return sum; +#else ptrdiff_t sum_max = min(PTRDIFF_MAX, SIZE_MAX); if (a <= sum_max && b <= sum_max - a) return a + b; +#endif size_overflow(); } static ATTRIBUTE_PURE ptrdiff_t size_product(ptrdiff_t nitems, ptrdiff_t itemsize) { +#ifdef ckd_mul + ptrdiff_t product; + if (!ckd_mul(&product, nitems, itemsize) && product <= SIZE_MAX) + return product; +#else ptrdiff_t nitems_max = min(PTRDIFF_MAX, SIZE_MAX) / itemsize; if (nitems <= nitems_max) return nitems * itemsize; +#endif size_overflow(); } @@ -536,11 +548,18 @@ static ptrdiff_t grow_nitems_alloc(ptrdiff_t *nitems_alloc, ptrdiff_t itemsize) { ptrdiff_t addend = (*nitems_alloc >> 1) + 1; +#if defined ckd_add && defined ckd_mul + ptrdiff_t product; + if (!ckd_add(nitems_alloc, *nitems_alloc, addend) + && !ckd_mul(&product, *nitems_alloc, itemsize) && product <= SIZE_MAX) + return product; +#else ptrdiff_t amax = min(PTRDIFF_MAX, SIZE_MAX); if (*nitems_alloc <= ((amax - 1) / 3 * 2) / itemsize) { *nitems_alloc += addend; return *nitems_alloc * itemsize; } +#endif memory_exhausted(_("integer overflow")); } @@ -3716,16 +3735,28 @@ time_overflow(void) static ATTRIBUTE_PURE zic_t oadd(zic_t t1, zic_t t2) { +#ifdef ckd_add + zic_t sum; + if (!ckd_add(&sum, t1, t2)) + return sum; +#else if (t1 < 0 ? ZIC_MIN - t1 <= t2 : t2 <= ZIC_MAX - t1) return t1 + t2; +#endif time_overflow(); } static ATTRIBUTE_PURE zic_t tadd(zic_t t1, zic_t t2) { +#ifdef ckd_add + zic_t sum; + if (!ckd_add(&sum, t1, t2) && min_time <= sum && sum <= max_time) + return sum; +#else if (t1 < 0 ? min_time - t1 <= t2 : t2 <= max_time - t1) return t1 + t2; +#endif if (t1 == min_time || t1 == max_time) return t1; time_overflow(); -- 2.38.1

1 0

[PROPOSED 0/6] unlikely integer overflow fixes
by Paul Eggert Nov. 20, 2022

Nov. 20, 2022

This patch set fixes some unlikely integer overflow bugs. These patches are relevant only for unusual situations, e.g., objects so large that their sizes exceed PTRDIFF_MAX bytes; or for unusual architectures, e.g., architectures where INT_MAX == UINT_MAX. In most cases I do not know of any practical way to exploit the bug, even if POSIX says the bug is possible. Paul Eggert (6): Don’t assume INT_MAX < UINT_FAST64_MAX Fix size_t overflow check if SIZE_MAX == INT_MAX date: integer overflow fixes Improve integer overflow checking in zdump Fix theoretical integer overflow in zic.c Fix another theoretical zic.c overflow NEWS | 3 ++ date.c | 32 +++++++------------ private.h | 4 +++ zdump.c | 64 +++++++++++++++++++++----------------- zic.c | 92 +++++++++++++++++++++++++++++++++++-------------------- 5 files changed, 113 insertions(+), 82 deletions(-) -- 2.38.1

1 6

[PATCH 1/5] Simplify inputline return value
by Paul Eggert Nov. 20, 2022

Nov. 20, 2022

* zic.c (inputline): Return bool, not ptrdiff_t. All callers changed. --- zic.c | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-) diff --git a/zic.c b/zic.c index 548c1c5..0671889 100644 --- a/zic.c +++ b/zic.c @@ -1575,10 +1575,9 @@ associate(void) /* Read a text line from FP into BUF, which is of size BUFSIZE. Terminate it with a NUL byte instead of a newline. - Return the line's length, not counting the NUL byte. - On EOF, return a negative number. + Return true if successful, false if EOF. On error, report the error and exit. */ -static ptrdiff_t +static bool inputline(FILE *fp, char *buf, ptrdiff_t bufsize) { ptrdiff_t linelen = 0, ch; @@ -1589,7 +1588,7 @@ inputline(FILE *fp, char *buf, ptrdiff_t bufsize) exit(EXIT_FAILURE); } if (linelen == 0) - return -1; + return false; error(_("unterminated line")); exit(EXIT_FAILURE); } @@ -1604,7 +1603,7 @@ inputline(FILE *fp, char *buf, ptrdiff_t bufsize) } } buf[linelen] = '\0'; - return linelen; + return true; } static void @@ -1626,13 +1625,11 @@ infile(int fnum, char const *name) } wantcont = false; for (num = 1; ; ++num) { - ptrdiff_t linelen; char buf[_POSIX2_LINE_MAX]; int nfields; char *fields[MAX_FIELDS]; eat(fnum, num); - linelen = inputline(fp, buf, sizeof buf); - if (linelen < 0) + if (!inputline(fp, buf, sizeof buf)) break; nfields = getfields(buf, fields, sizeof fields / sizeof *fields); -- 2.38.1

1 4

[PROPOSED 0/4] Port to C89, and deprecate C89
by Paul Eggert Nov. 19, 2022

Nov. 19, 2022

While looking into fixing some bugs in tzcode, I noticed some places where existing code wouldn’t work with ancient C89 compilers. I fixed the problems I found in this patch series. Quite possibly other problems remain, problems that I didn’t notice. As nobody has reported C89-related problems for ages, it appears that these old compilers are no longer being used to compile tzcode. So I propose that we remove C89 support, to avoid future maintenance burdens such as the one I recently went through. As part of the transition away from C89 support, the last patch in this series announces that C89 will be removed in some future release. We can keep C89 support for at least one release longer. Paul Eggert (4): Port to C89 gcc -pedantic Port UINT_FAST64_MAX back to C89 Port int_fast64_t to C89 64-bit long * NEWS: Deprecate C89. NEWS | 7 +++++++ private.h | 47 +++++++++++++++++++++++------------------------ 2 files changed, 30 insertions(+), 24 deletions(-) -- 2.38.1

1 4

Re: [tz] Country name change from Turkey to Turkiye
by Brian Inglis Nov. 19, 2022

Nov. 19, 2022

On 2022-11-17 23:09, Bin Li via tz wrote: > Do you have any plans to change Turkey's name? Because it's the official name now. > https://www.npr.org/2022/06/03/1102841197/turkey-changes-its-official-name-… <https://www.npr.org/2022/06/03/1102841197/turkey-changes-its-official-name-…> One of the stupider ISO decisions, as English supports no accents, except in loanwords, and they are just ignored, except possibly in academic publishing. I think it is time native English language speakers push back against and ignore decisions made by international committees of non-native English language speakers, to try and change what native English language speakers use to what the foreign language uses in Latin alphabet transliterations, just as the other 100+ countries who have no "official" transliteration do. -- 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 retirer but when there is no more to cut -- Antoine de Saint-Exupéry

3 2

Country name change from Turkey to Turkiye
by Bin Li Nov. 18, 2022

Nov. 18, 2022

Dear, Do you have any plans to change Turkey's name? Because it's the official name now. https://www.npr.org/2022/06/03/1102841197/turkey-changes-its-official-name-…

6 8

New York Times: Time is Running Out for the Leap Second
by Jonathan Leffler Nov. 15, 2022

Nov. 15, 2022

https://www.nytimes.com/2022/11/14/science/time-leap-second.html -- Jonathan Leffler <jonathan.leffler(a)gmail.com> #include <disclaimer.h> Guardian of DBD::Informix - v2018.1031 - http://dbi.perl.org "Blessed are we who can laugh at ourselves, for we shall never cease to be amused."

1 0