- tz - lists.iana.org

[PROPOSED] Avoid time_t + int overflow
by Paul Eggert Nov. 22, 2024

Nov. 22, 2024

* localtime.c (increment_overflow_time_int): New function. (time2sub): Use it (instead of relying on possibly-undefined behavior) to check for time_t + int overflow. --- localtime.c | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-) diff --git a/localtime.c b/localtime.c index 07ce41b8..14c356e1 100644 --- a/localtime.c +++ b/localtime.c @@ -1882,6 +1882,21 @@ increment_overflow_iinntt(iinntt *lp, int m) #endif } +static bool +increment_overflow_time_int(time_t *tp, int j) +{ +#ifdef ckd_add + return ckd_add(tp, *tp, j); +#else + if (j < 0 + ? (TYPE_SIGNED(time_t) ? *tp < TIME_T_MIN - j : *tp <= -1 - j) + : TIME_T_MAX - j < *tp) + return true; + *tp += j; + return false; +#endif +} + static bool increment_overflow_time(time_t *tp, int_fast32_t j) { @@ -2155,10 +2170,8 @@ time2sub(struct tm *const tmp, return WRONG; } label: - newt = t + saved_seconds; - if ((newt < t) != (saved_seconds < 0)) + if (increment_overflow_time_int(&t, saved_seconds)) return WRONG; - t = newt; if (funcp(sp, &t, offset, tmp)) *okayp = true; return t; -- 2.47.0

1 0

Dead code in strftime()?
by Tom Lane Nov. 22, 2024

Nov. 22, 2024

strftime() in the 2024b tzcode distribution includes if (!p) { errno = EOVERFLOW; return 0; } AFAICS this is dead code, except maybe in the case that the call passes s = NULL, maxsize = 0, which doesn't seem like a case that the code intends to support (and EOVERFLOW would be a rather odd choice of error code if that's the idea). newstrftime.3.txt offers a clue about the intent: This function may fail if: [EOVERFLOW] The format includes an %s conversion and the number of seconds since the Epoch cannot be represented in a time_t. But the %s conversion code never returns zero. There's a suggestive comment there: /* If mktime fails, %s expands to the value of (time_t) -1 as a failure marker; this is better in practice than strftime failing. */ So it looks to me like somebody changed the error handling in the %s stanza but didn't bother to remove the now-dead path in the caller nor update the man page. Am I missing something? regards, tom lane

3 13

[PROPOSED 1/6] Improve zdump overflow checking
by Paul Eggert Nov. 22, 2024

Nov. 22, 2024

This also pacifies -Wsuggest-attribute=pure on x86 on platforms that lack tm_gmtoff and tm_zone. * zdump.c (xstrsize): Remove; no longer used. (tzalloc, saveabbr): Use plain strlen here, since the string sizes have already been checked. (main): Limit ‘longest’ to INT_MAX - 2 so that a printf with it and two spaces does not flake out. Check for arglen overflow once here, rather than on every alloc. --- zdump.c | 24 ++++++++---------------- 1 file changed, 8 insertions(+), 16 deletions(-) diff --git a/zdump.c b/zdump.c index e8178733..b668022d 100644 --- a/zdump.c +++ b/zdump.c @@ -148,17 +148,6 @@ sumsize(ptrdiff_t a, ptrdiff_t b) size_overflow(); } -/* Return the size of of the string STR, including its trailing NUL. - Report an error and exit if this would exceed INDEX_MAX which means - pointer subtraction wouldn't work. */ -static ptrdiff_t -xstrsize(char const *str) -{ - size_t len = strlen(str); - if (len < INDEX_MAX) - return len + 1; - size_overflow(); -} /* Return a pointer to a newly allocated buffer of size SIZE, exiting on failure. SIZE should be positive. */ @@ -266,7 +255,7 @@ tzalloc(char const *val) static ptrdiff_t fakeenv0size; void *freeable = NULL; char **env = fakeenv, **initial_environ; - ptrdiff_t valsize = xstrsize(val); + ptrdiff_t valsize = strlen(val) + 1; if (fakeenv0size < valsize) { char **e = environ, **to; ptrdiff_t initial_nenvptrs = 1; /* Counting the trailing NULL pointer. */ @@ -425,7 +414,7 @@ saveabbr(char **buf, ptrdiff_t *bufalloc, struct tm const *tmp) if (HAVE_LOCALTIME_RZ) return ab; else { - ptrdiff_t absize = xstrsize(ab); + ptrdiff_t absize = strlen(ab) + 1; if (*bufalloc < absize) { free(*buf); @@ -487,6 +476,7 @@ main(int argc, char *argv[]) register time_t cuthitime; time_t now; bool iflag = false; + size_t arglenmax = 0; cutlotime = absolute_min_time; cuthitime = absolute_max_time; @@ -586,12 +576,14 @@ main(int argc, char *argv[]) now = time(NULL); now |= !now; } - longest = 0; for (i = optind; i < argc; i++) { size_t arglen = strlen(argv[i]); - if (longest < arglen) - longest = min(arglen, INT_MAX); + if (arglenmax < arglen) + arglenmax = arglen; } + if (!HAVE_SETENV && INDEX_MAX <= arglenmax) + size_overflow(); + longest = min(arglenmax, INT_MAX - 2); for (i = optind; i < argc; ++i) { timezone_t tz = tzalloc(argv[i]); -- 2.47.0

1 5

(no subject)
by Cool Gab19 Nov. 8, 2024

Nov. 8, 2024

can you add Philippine Daylight Time on your Database I think it was missing from the list of the time zones please reply immediately if solved thanks...

4 8

[PROPOSED] * NEWS, europe: Belated thanks for Ittoqqortoormiit 2024 fix
by Paul Eggert Nov. 8, 2024

Nov. 8, 2024

--- NEWS | 2 +- europe | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/NEWS b/NEWS index a109ae9d..8381a197 100644 --- a/NEWS +++ b/NEWS @@ -233,7 +233,7 @@ Release 2023d - 2023-12-21 20:02:24 -0800 changing its time zone from -01/+00 to -02/-01 at the same moment as the spring-forward transition. Its clocks will therefore not spring forward as previously scheduled. The time zone change - reverts to its common practice before 1981. + reverts to its common practice before 1981. (Thanks to Jule Dabars.) Fix predictions for DST transitions in Palestine in 2072-2075, correcting a typo introduced in 2023a. (Thanks to Jule Dabars.) diff --git a/europe b/europe index f9063949..df334fc2 100644 --- a/europe +++ b/europe @@ -1147,7 +1147,7 @@ Zone Atlantic/Faroe -0:27:04 - LMT 1908 Jan 11 # Tórshavn # However, Greenland will change to Daylight Saving Time again in 2024 # and onwards. -# From a contributor who wishes to remain anonymous for now (2023-10-29): +# From Jule Dabars (2023-10-29): # https://www.dr.dk/nyheder/seneste/i-nat-skal-uret-stilles-en-time-tilbage-m… # with a link to that page: # https://naalakkersuisut.gl/Nyheder/2023/10/2710_sommertid -- 2.43.0

1 0

[PROPOSED] * NEWS: Belated thanks for Palestine 2072-2075 fix
by Paul Eggert Nov. 7, 2024

Nov. 7, 2024

--- NEWS | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/NEWS b/NEWS index 8c315643..a109ae9d 100644 --- a/NEWS +++ b/NEWS @@ -236,7 +236,7 @@ Release 2023d - 2023-12-21 20:02:24 -0800 reverts to its common practice before 1981. Fix predictions for DST transitions in Palestine in 2072-2075, - correcting a typo introduced in 2023a. + correcting a typo introduced in 2023a. (Thanks to Jule Dabars.) Changes to past and future timestamps -- 2.43.0

1 0

[PATCH 1/2] Switch from RFC 8536 to 9636 for documentation
by Paul Eggert Nov. 3, 2024

Nov. 3, 2024

RFC 9636 is now published, and it obsoletes RFC 8536 for the Time Zone information format (TZif). --- Makefile | 2 +- NEWS | 4 ++++ localtime.c | 2 +- theory.html | 4 ++-- tz-link.html | 4 ++-- tzfile.5 | 12 ++++++------ tzfile.h | 2 +- zic.c | 4 ++-- 8 files changed, 19 insertions(+), 15 deletions(-) diff --git a/Makefile b/Makefile index 57d1d672..6e623eb0 100644 --- a/Makefile +++ b/Makefile @@ -137,7 +137,7 @@ TIME_T_ALTERNATIVES_TAIL = int_least32_t.ck uint_least32_t.ck \ uint_least64_t.ck # What kind of TZif data files to generate. (TZif is the binary time -# zone data format that zic generates; see Internet RFC 8536.) +# zone data format that zic generates; see Internet RFC 9636.) # If you want only POSIX time, with time values interpreted as # seconds since the epoch (not counting leap seconds), use # REDO= posix_only diff --git a/NEWS b/NEWS index adaa4b60..8c315643 100644 --- a/NEWS +++ b/NEWS @@ -24,6 +24,10 @@ Unreleased, experimental changes 1977, and May and July 1990; and removing incorrect transitions in March and September 1978. + Changes to documentation + + Cite Internet RFC 9636, which obsoletes RFC 8536 for TZif format. + Release 2024b - 2024-09-04 12:27:47 -0700 diff --git a/localtime.c b/localtime.c index 7ae9ce5e..f331473e 100644 --- a/localtime.c +++ b/localtime.c @@ -583,7 +583,7 @@ tzloadbody(char const *name, struct state *sp, bool doextend, correction must differ from the previous one's by 1 second or less, except that the first correction can be any value; these requirements are more generous than - RFC 8536, to allow future RFC extensions. */ + RFC 9636, to allow future RFC extensions. */ if (! (i == 0 || (prevcorr < corr ? corr == prevcorr + 1 diff --git a/theory.html b/theory.html index b1717e7d..4d72ee2e 100644 --- a/theory.html +++ b/theory.html @@ -1115,8 +1115,8 @@ However POSIX.1-2024, like earlier POSIX editions, has some limitations: the name of a file from which time-related information is read. The file's format is <dfn><abbr>TZif</abbr></dfn>, a timezone information format that contains binary data; see - <a href="https://datatracker.ietf.org/doc/html/8536">Internet - <abbr>RFC</abbr> 8536</a>. + <a href="https://datatracker.ietf.org/doc/html/9636">Internet + <abbr>RFC</abbr> 9636</a>. The daylight saving time rules to be used for a particular timezone are encoded in the <abbr>TZif</abbr> file; the format of the file allows <abbr>US</abbr>, diff --git a/tz-link.html b/tz-link.html index be2aae54..0853bd7d 100644 --- a/tz-link.html +++ b/tz-link.html @@ -194,9 +194,9 @@ After obtaining the code and data files, see the The code lets you compile the <code><abbr>tz</abbr></code> source files into machine-readable binary files, one for each location. The binary files are in a special format specified by -<a href="https://datatracker.ietf.org/doc/html/8536">The +<a href="https://datatracker.ietf.org/doc/html/9636">The Time Zone Information Format (<abbr>TZif</abbr>)</a> -(Internet <abbr title="Request For Comments">RFC</abbr> 8536). +(Internet <abbr title="Request For Comments">RFC</abbr> 9636). The code also lets you read a <abbr>TZif</abbr> file and interpret timestamps for that location.</p> diff --git a/tzfile.5 b/tzfile.5 index f2529b01..b699bd26 100644 --- a/tzfile.5 +++ b/tzfile.5 @@ -17,7 +17,7 @@ The timezone information files used by .BR tzset (3) are typically found under a directory with a name like .IR /usr/share/zoneinfo . -These files use the format described in Internet RFC 8536. +These files use the format described in Internet RFC 9636. Each file is a sequence of 8-bit bytes. In a file, a binary integer is represented by a sequence of one or more bytes in network order (bigendian, or high-order byte first), @@ -380,7 +380,7 @@ for permanent Atlantic Standard Time (\-04). .IP \(bu Some readers designed for version 2 or 3 and that require strict -conformance to RFC 8536 reject version 4 files whose leap second +conformance to RFC 9636 reject version 4 files whose leap second tables are truncated at the start or end in expiration times. .IP \(bu Some readers ignore the footer, and instead predict future @@ -514,9 +514,9 @@ of one hour, or of 15 minutes, or of 1 minute. .PP Olson A, Eggert P, Murchison K. The Time Zone Information Format (TZif). 2019 Feb. -.UR https://\:datatracker.ietf.org/\:doc/\:html/\:rfc8536 -Internet RFC 8536 +.UR https://\:datatracker.ietf.org/\:doc/\:html/\:rfc9636 +Internet RFC 9636 .UE -.UR https://\:doi.org/\:10.17487/\:RFC8536 -doi:10.17487/RFC8536 +.UR https://\:doi.org/\:10.17487/\:RFC9636 +doi:10.17487/RFC9636 .UE . diff --git a/tzfile.h b/tzfile.h index b1541466..f8c6c8c5 100644 --- a/tzfile.h +++ b/tzfile.h @@ -26,7 +26,7 @@ #endif /* !defined TZDEFRULES */ -/* See Internet RFC 8536 for more details about the following format. */ +/* See Internet RFC 9636 for more details about the following format. */ /* ** Each file begins with. . . diff --git a/zic.c b/zic.c index cf8e79df..6da35270 100644 --- a/zic.c +++ b/zic.c @@ -2761,7 +2761,7 @@ writezone(const char *const name, const char *const string, char version, if (thisleapexpiry) { /* Append a no-op leap correction indicating when the leap second table expires. Although this does not conform to - Internet RFC 8536, most clients seem to accept this and + Internet RFC 9636, most clients seem to accept this and the plan is to amend the RFC to allow this in version 4 TZif files. */ puttzcodepass(leapexpires, fp, pass); @@ -3007,7 +3007,7 @@ stringzone(char *result, struct zone const *zpfirst, ptrdiff_t zonecount) result[0] = '\0'; - /* Internet RFC 8536 section 5.1 says to use an empty TZ string if + /* Internet RFC 9636 section 6.1 says to use an empty TZ string if future timestamps are truncated. */ if (hi_time < max_time) return -1; -- 2.43.0

1 1

Ukraine to abolish DST?
by Steffen Thorsen Oct. 26, 2024

Oct. 26, 2024

Ukraine seems to be getting rid of DST, according to https://universemagazine.com/en/the-verkhovna-rada-canceled-daylight-saving… "During a meeting of the Verkhovna Rada of Ukraine on June 16, daylight saving time was canceled. According to MP Yaroslav Zheleznyak, 261 MPs voted in favor of the bill. Now it must be signed by the President of Ukraine, and then it will come into force." which is again referencing (in Ukrainan) https://lb.ua/society/2024/07/16/624453_verhovna_rada_skasuvala_perehid.html It seems they will plan to set back time this autumn to winter(standard) time, and then stay on that. Best regards, Steffen Thorsen - timeanddate.com

5 8

Brazil may reinstate daylight saving time soonish
by Paul Eggert Oct. 17, 2024

Oct. 17, 2024

A Reuters report[1] says that energy authorities in Brazil have approved a plan to bring back daylight saving time, to save electrical energy during a major drought that has hurt hydroelectric power generation[2]. As far as I know authorities have not decided or published details of the plan, such as which states would be involved and dates and times of transitions. However, one English-language report says that DST would not start until after the election period ends on October 27[3]. This is not a done deal, as any change would need approval by President Luiz Inácio Lula da Silva. That being said, Brazil's DST was abolished by Lula's predecessor and political opponent Jair Bolsonaro, and Lula did an informal Twitter poll shortly after his election in which 66% of respondents voted for DST[4], so presumably Lula would favor restoring DST. [1]: Rodrigo Viga Gaier and Leticia Fucuchima. Brazil energy authorities approve returning to daylight saving time as drought bites. Reuters. 2024-09-20. <https://www.reuters.com/world/americas/brazil-likely-reinstate-daylight-sav…> [2]: Daylight Saving Time Won’t Fix Brazil’s Energy Crisis. Latin American Post. 2024-09-23. <https://latinamericanpost.com/analysis-en/daylight-saving-time-wont-fix-bra…> [3]: Leonardo Vieceli. Brazil's National Electric Operator Recommends Daylight Saving Time. Folha de S.Paulo. 2024-09-20. <https://www1.folha.uol.com.br/internacional/en/business/2024/09/brazils-nat…> [4]: Amanda Audi. Will Brazil bring back daylight saving time? The Brazilian report. 2024-01-02. <https://brazilian.report/society/2024/01/02/lula-daylight-saving-time/>

1 1

[PROPOSED] Port better to glibc when used internally there
by Paul Eggert Oct. 16, 2024

Oct. 16, 2024

I ran into this problem when compiling an experimental version of glibc, which uses tzcode internally. * private.h (static_assert): Override any definition a nonconforming implementation may provide. --- private.h | 1 + 1 file changed, 1 insertion(+) diff --git a/private.h b/private.h index c3304104..4b035ed7 100644 --- a/private.h +++ b/private.h @@ -51,6 +51,7 @@ #endif #if __STDC_VERSION__ < 202311 +# undef static_assert # define static_assert(cond) extern int static_assert_check[(cond) ? 1 : -1] #endif -- 2.43.0

1 0