July 2004 - tz - lists.iana.org

RE: asctime.c
by Olson, Arthur David (NIH/NCI) July 27, 2004

July 27, 2004

From: Robert Elz [mailto:kre@munnari.OZ.AU] Sent: Tuesday, July 27, 2004 11:49 AM To: Olson, Arthur David (NIH/NCI) Cc: Tz Subject: Re: asctime.c ... | /* | ** Big enough for something such as | ** ??? ???-2147483648 -2147483648:-2147483648:-2147483648 | -2147483648\n | ** (two three-character abbreviations, five strings denoting | integers, | ** three explicit spaces, two explicit colons, a newline, | ** and a trailing ASCII nul). | */ | char result[2 * 3 + 5 * INT_STRLEN_MAXIMUM(int) + | 3 + 2 + 1 + 1]; > That buffer is big enough for that string, but are we sure that string > is the longest that can ever happen? The year is after all printed > using %ld - so there must at least be the potential for that one to be a long, which might be -9223372036854775808. > Beyond that, C doesn't actually promise that "int" is limited to 32 > bits does it? Given that, the 2147483648 numbers are just speculation. We're using "INT_STRLEN_MAXIMUM(int)" to figure out how long the output for an integer value might be. The macro figures out the length based on "sizeof int", so even if an int is more than 32 bits things will come out right. While we're using a %ld format to print out out the year, what we're printing is the sum of an int (tm_year) and 1900 (TM_YEAR_BASE), so we know the result is int sized. --ado

1 0

Where to download the database
by Craig Orput July 26, 2004

July 26, 2004

Your site is very confusing to me. Where do I download the actual time zone database, in text version? Regards Craig Orput Vice President - Logistics Engineering OIA SCM 17230 NE Sacramento Street Portland, OR 97230 Office 503.872.8470 craig.orput(a)oiascm.com http://www.oiascm.com/

1 0

Re: another attempt to change the clocks in the UK
by Peter Ilieve July 26, 2004

July 26, 2004

Back on 9 June when I mentioned Nigel Beard's Lighter Evenings Bill I wrote: > I haven't found the text of the Bill itself yet. It will probably > turn up on the Parliament website next week. It was finally published on Wednesday. PDF and HTML versions can be found at <http://www.publications.parliament.uk/pa/cm200304/cmbills/115/2004115.htm>. It is a short Bill of only four sections. It changes the time for general purposes to one hour in advance of GMT. It amends the Scotland Act 1998 to make time zones and summer time arrangements a devolved rather than a reserved matter, which will allow the Scottish parliament to keep Scotland on GMT if they so wish. It makes some amendments to the Summer Time Act 1972 which I presume are to allow Northern Ireland, the Channel Islands and the Isle of Man to do their own thing as well (I don't have a copy of the 1972 Act handy to check this). By pure coincidence a relative of mine works as a researcher for Nigel Beard and has spent the last couple of months working on this. What seems to have happened is that Nigel became convinced by the safety arguments for change and decided to do something about it. At this stage he didn't know about John Butterfill's `Extra Daylight' Bill or Lord Archer's similar measure that he introduced during his doomed campaign for London mayoral candidate. Finding all this out was left to his researcher. She produced a briefing which Nigel's ten minute rule speech was based on. In turn, she then passed the specification for the proposed change on to the parliamentary draughtsmen, who add the legalese so it will pass muster as a Bill. The next stage is the Bill's second reading, which is set for 15 October. This is where the first vote on the Bill will take place. If the vote is lost that's it. If it passes the Bill goes to the next, committee, stage. If you look at the Public Bill List on the Parliament website (<http://www.publications.parliament.uk/pa/cm/cmpblist/cmpblist.htm>) you will see that there are literally dozens of Bills down for second reading on 15 October. The second reading is supposed to be the first detailed consideration of a Bill. You can judge for yourselves what sort of scrutiny each of these dozens of Bills will get. My view remains that this Bill has almost no chance of becoming law. Ten minute rule Bills come even further down the pecking order than private members' Bills like John Butterfill's one, and his failed. Bills like this only really stand a chance if they have government support, as the government can give them parliamentary time. I have spoken to someone at the DTI (the department now dealing with time related stuff) and they told me that the government is not in favour of this change, as it would be disruptive to business, transport and tourism, particularly if the UK no longer used a single time zone. Peter Ilieve peter(a)aldie.co.uk

2 1

Re: asctime.c
by Olson, Arthur David (NIH/NCI) July 23, 2004

July 23, 2004

Code in "asctime" such as... { long y; y = timeptr->tm_year + (long) TM_YEAR_BASE; if (y >= -999 && y <= 9999) (void) sprintf(buf, "%.3s %.3s%3d %02d:%02d:%02d %ld\n", wn, mn, timeptr->tm_mday, timeptr->tm_hour, timeptr->tm_min, timeptr->tm_sec, y); else (void) sprintf(buf, "%.3s %.3s%3d %ld\n", wn, mn, timeptr->tm_mday, y); } ...never returns NULL and never (when applied to a "struct tm" derived from a 64-bit time_t) overflows a 26-character buffer passed to asctime. It does lose time-of-day information in the distant past and distant future. --ado

3 2

Asctime.c
by Olson, Arthur David (NIH/NCI) July 21, 2004

July 21, 2004

At the end of this message is a revised version of asctime.c. There are a couple of standard-related matters. 1. The POSIX standard calls for a struct tm that includes an "int tm_year" member. If time_t is 64 bits and int is 32 bits, tm_year won't be able to represent all the years associated with time_t values. Should the standard address this issue? 2. The standard reads... The asctime_r() function shall convert the broken-down time in the structure pointed to by tm into a string (of the same form as that returned by asctime()) that is placed in the user-supplied buffer pointed to by buf (which shall contain at least 26 bytes) and then return buf. ...but a 26-byte buffer won't be adequate if a five-digit (or more) year is involved. Should the standard address this issue? Should "asctime_r" assume that the fed-in buffer is at most 26 bytes and avoid overflowing it? --ado /* ** This file is in the public domain, so clarified as of ** 1996-06-05 by Arthur David Olson (arthur_david_olson(a)nih.gov) */ #ifndef lint #ifndef NOID static char elsieid[] = "@(#)asctime.c 7.11"; #endif /* !defined NOID */ #endif /* !defined lint */ /*LINTLIBRARY*/ #include "private.h" #include "tzfile.h" /* ** A la ISO/IEC 9945-1, ANSI/IEEE Std 1003.1, 2004 Edition. */ char * asctime_r(timeptr, buf) register const struct tm * timeptr; char * buf; { static const char wday_name[][3] = { "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat" }; static const char mon_name[][3] = { "Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" }; register const char * wn; register const char * mn; if (timeptr->tm_wday < 0 || timeptr->tm_wday >= DAYSPERWEEK) wn = "???"; else wn = wday_name[timeptr->tm_wday]; if (timeptr->tm_mon < 0 || timeptr->tm_mon >= MONSPERYEAR) mn = "???"; else mn = mon_name[timeptr->tm_mon]; /* ** The format used in the (2004) standard is ** "%.3s %.3s%3d %.2d:%.2d:%.2d %d\n" ** Use "%02d", as it is a bit more portable than "%.2d". ** Drop each .3 since they're superfluous given how we set wn and mn. */ (void) sprintf(buf, "%s %s%3d %02d:%02d:%02d %ld\n", wn, mn, timeptr->tm_mday, timeptr->tm_hour, timeptr->tm_min, timeptr->tm_sec, timeptr->tm_year + (long) TM_YEAR_BASE); return buf; } /* ** A la ISO/IEC 9945-1, ANSI/IEEE Std 1003.1, 2004 Edition, ** with core dump avoidance. */ char * asctime(timeptr) register const struct tm * timeptr; { /* ** Big enough for something such as ** ??? ???-2147483648 -2147483648:-2147483648:-2147483648 -2147483648\n ** (two three-character abbreviations, five strings denoting integers, ** three explicit spaces, two explicit colons, a newline, ** and a trailing ASCII nul). */ static char result[3 * 2 + 5 * INT_STRLEN_MAXIMUM(int) + 3 + 2 + 1 + 1]; return asctime_r(timeptr, result); }

3 3

FW: Definition of time_t changed from signed to unsigned...
by Olson, Arthur David (NIH/NCI) July 20, 2004

July 20, 2004

Chris Richards is not on the time zone mailing list; direct replies appropriately. --ado -----Original Message----- From: Chris Richards [mailto:cj_richards@hotmail.com] Sent: Friday, July 09, 2004 9:53 AM To: tz(a)lecserver.nci.nih.gov Subject: Definition of time_t changed from signed to unsigned... Hello, I am in the process of upgrading the kernel version on a platform that I work on. I noticed that the time_t definition has been changed in the newer version of the OS. It used to be a simple "long". But it is now an "unsigned long". Basically it changed from a signed to an unsigned definition. The problem is that the timezone subsystem seems to expect time_t to be signed. There are a number of functions that only work if time_t is signed. E.g. function localsub() in file localtime.c. My question is: am I just using an obsolete version of the timezone source or should I just change the time_t typedef in the OS? I quickly scanned the latest source of localtime.c and did not see any changes regarding to how time_t is declared. Thanks in advance, Chris. _________________________________________________________________ MSN Toolbar provides one-click access to Hotmail from any Web page - FREE download! http://toolbar.msn.click-url.com/go/onm00200413ave/direct/01/

6 8

Tzcode2004b.tar.gz and tzdata2004b.tar.gz
by Olson, Arthur David (NIH/NCI) July 19, 2004

July 19, 2004

The files... ftp://elsie.nci.nih.gov/pub/tzcode2004b.tar.gz ...and... ftp://elsie.nci.nih.gov/pub/tzdata2004b.tar.gz ...are now available. These contain the data changes circulated last week on the time zone mailing list by Paul Eggert--most importantly, they include updates for Argentina. I wanted to get these out quickly since they effect present-day operations down south. I'm holding off (briefly, I hope) on incorporating code changes until I can do a more thorough review than I have time for just now. --ado

1 0

RE: Definition of time_t changed from signed to unsigned...
by Olson, Arthur David (NIH/NCI) July 15, 2004

July 15, 2004

> -----Original Message----- > From: Chris Richards [mailto:cj_richards@hotmail.com] > Sent: Tuesday, July 13, 2004 12:42 PM > To: Olson, Arthur David (NIH/NCI) > Subject: RE: Definition of time_t changed from signed to unsigned... > > > Hi, > > Actually, could this be due to me using a tz.cfg database that was compiled > in an environment where time_t was typedef'ed differently to the environment > which loads and uses the tz.cfg? It certainly could be due to using a signed-system-generated data file on an unsigned-time_t system. Signed time values from before 1970 are negative numbers; on an unsigned system they'll look like large positive numbers. This means that for present-day time_t values the "t < sp->ats[0]" test will be satisfied, the software will assume that it's dealing with a time predating the first transition that's recorded, and the software will apply "default" rules. "localtime.c" could be jiggered to ignore pre-1970 transitions if (1) time_t is unsigned and (2) a data file has been generated on a signed system (which can be determined by seeing if the transition times are correctly sorted from the perspective of the system on which the software is being run). Given the notes on this mailing list about the inadvisability of unsigned time_t's, such jiggering is of questionable value. --ado

1 0

Re: Dust settled on Argentina?
by Paul Eggert July 13, 2004

July 13, 2004

"Olson, Arthur David (NIH/NCI)" <olsona(a)dc37a.nci.nih.gov> writes: > I haven't seen another patch. Should I go ahead and promulgate the > 2004-06-15 patch, or should I wait? Sorry, I dropped the ball on that one. Here's a revised version: it incorporates the discussion to earlier revisions of all the patches I've posted. I also updated the tz-link.htm web page to reflect foreign web site changes. * Patches relevant to current time stamps. - Major reorganizaton for Argentina. . New zones America/Argentina/Tucuman, America/Argentina/La_Rioja, America/Argentina/San_Juan, America/Argentina/ComodRivadavia, America/Argentina/Rio_Gallegos, America/Argentina/Ushuaia reflecting this year's changes. (Thanks to Mariano Absatz and Gwillim Law for their dogged reporting of this info.) . America/Buenos_Aires renamed to America/Argentina/Buenos_Aires, and similarly for America/Cordoba, America/Jujuy, America/Catamarca, America/Mendoza. Backward-compatible links retained so that old TZ settings still work. - Georgia switched to Moscow time on June 27. (Thanks to Joseph S. Myers for this info.) * Patch relevant to historical time stamps. - Where new zones were created, I recorded Shanks's opinions of what happened to Argentina in 1991. * Patch due to ISO 3166 change. - The ISO added a country code AX for the Aaland Islands, so I added a new link Europe/Mariehamn for AX. * Documentation patches. - Add URL for PyTZ. - Update URL for Tru64. - Remove dangling URL to <http://www.hilink.com.au/times/>. - Mention that Java has only a subset of the Olson database. - Move ICU localization reference to "Time Notation" section. - Update names for Astrodienst, worldtimezone.com, INMS. * Some proposed source-code patches for 64-bit hosts. - Use 'long' when computing years, not 'int', to avoid spurious overflow on 64-bit hosts. - Use our (more generous) asctime implementation than relying on the system asctime, so that we can print years before 1900 and after 9999 (even on HP-UX). - When localtime/gmtime fails, print the raw time_t value as an integer than the less-informative "NULL". - Don't attempt to print out all the 64-bit time_t space, as that will take nearly forever; instead, default to printing a few decades into the future, and don't bother to print times before about 1833. This suffices for all the current tz data and is good enough for POSIX tz values until the year 2106. We can revisit this when we add true 64-bit support for tz data. diff -pru tz-2004aa/asia tz/asia --- tz-2004aa/asia 2004-05-27 09:00:30.000000000 -0700 +++ tz/asia 2004-07-13 13:16:45.000000000 -0700 @@ -335,6 +335,16 @@ Link Asia/Nicosia Europe/Nicosia # Instead of putting back clocks at the end of October, Georgia # will stay on daylight savings time this winter to save energy, # President Eduard Shevardnadze decreed Wednesday. +# +# From the BBC via Joseph S. Myers (2004-06-27): +# +# Georgia moved closer to Western Europe on Sunday... The former Soviet +# republic has changed its time zone back to that of Moscow. As a result it +# is now just four hours ahead of Greenwich Mean Time, rather than five hours +# ahead. The switch was decreed by the pro-Western president of Georgia, +# Mikhail Saakashvili, who said the change was partly prompted by the process +# of integration into Europe. + # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone Asia/Tbilisi 2:59:16 - LMT 1880 2:59:16 - TBMT 1924 May 2 # Tbilisi Mean Time @@ -345,7 +355,8 @@ Zone Asia/Tbilisi 2:59:16 - LMT 1880 3:00 E-EurAsia GE%sT 1994 Sep lastSun 4:00 E-EurAsia GE%sT 1996 Oct lastSun 4:00 1:00 GEST 1997 Mar lastSun - 4:00 E-EurAsia GE%sT + 4:00 E-EurAsia GE%sT 2004 Jun 27 + 3:00 RussiaAsia GE%sT # East Timor diff -pru tz-2004aa/backward tz/backward --- tz-2004aa/backward 2004-05-27 09:00:30.000000000 -0700 +++ tz/backward 2004-06-14 12:30:06.845824000 -0700 @@ -4,9 +4,14 @@ # and their old names. Many names changed in late 1993. Link America/Adak America/Atka +Link America/Argentina/Buenos_Aires America/Buenos_Aires +Link America/Argentina/Catamarca America/Catamarca +Link America/Argentina/Cordoba America/Cordoba Link America/Tijuana America/Ensenada Link America/Indianapolis America/Fort_Wayne +Link America/Argentina/Jujuy America/Jujuy Link America/Indiana/Knox America/Knox_IN +Link America/Argentina/Mendoza America/Mendoza Link America/Rio_Branco America/Porto_Acre Link America/Cordoba America/Rosario Link America/St_Thomas America/Virgin diff -pru tz-2004aa/europe tz/europe --- tz-2004aa/europe 2004-05-27 09:00:30.000000000 -0700 +++ tz/europe 2004-06-12 18:37:32.547326000 -0700 @@ -921,6 +921,10 @@ Zone Europe/Helsinki 1:39:52 - LMT 1878 2:00 Finland EE%sT 1981 Mar 29 2:00 2:00 EU EE%sT +# Aaland Is +Link Europe/Helsinki Europe/Mariehamn + + # France # From Ciro Discepolo (2000-12-20): diff -pru tz-2004aa/iso3166.tab tz/iso3166.tab --- tz-2004aa/iso3166.tab 2003-12-15 06:36:37.000000000 -0800 +++ tz/iso3166.tab 2004-06-13 23:50:45.657245000 -0700 @@ -2,11 +2,11 @@ # # @(#)iso3166.tab 1.14 # -# From Paul Eggert <eggert(a)twinsun.com> (2003-02-04): +# From Paul Eggert <eggert(a)twinsun.com> (2004-06-14): # # This file contains a table with the following columns: # 1. ISO 3166-1 alpha-2 country code, current as of -# ISO 3166-1 Newsletter No. V-8 (2003-07-23). See: +# ISO 3166-1 Newsletter No. V-10 (2004-04-26). See: # <a href="http://www.iso.org/iso/en/prods-services/iso3166ma/index.html"> # ISO 3166 Maintenance agency (ISO 3166/MA) # </a>. @@ -36,6 +36,7 @@ AS Samoa (American) AT Austria AU Australia AW Aruba +AX Aaland Islands AZ Azerbaijan BA Bosnia & Herzegovina BB Barbados diff -pru tz-2004aa/southamerica tz/southamerica --- tz-2004aa/southamerica 2004-05-27 09:00:30.000000000 -0700 +++ tz/southamerica 2004-06-22 21:39:13.762385000 -0700 @@ -155,17 +155,47 @@ Rule Arg 2000 only - Mar Sun>=1 0:00 0 - # </a> says that standard time in Argentina from 1894-10-31 # to 1920-05-01 was -4:16:48.25. Go with this more-precise value # over Shanks. +# +# From Mariano Absatz (2004-06-05): +# These media articles from a major newspaper mostly cover the current state: +# http://www.lanacion.com.ar/04/05/27/de_604825.asp +# http://www.lanacion.com.ar/04/05/28/de_605203.asp +# +# The following eight (8) provinces pulled clocks back to UTC-04:00 at +# midnight Monday May 31st. (that is, the night between 05/31 and 06/01). +# Apparently, all nine provinces would go back to UTC-03:00 at the same +# time in October 17th. +# +# Catamarca, Chubut, La Rioja, San Juan, San Luis, Santa Cruz, +# Tierra del Fuego, Tucuman. +# +# From Mariano Absatz (2004-06-14): +# ... this weekend, the Province of Tucuman decided it'd go back to UTC-03:00 +# yesterday midnight (that is, at 24:00 Saturday 12th), since the people's +# annoyance with the change is much higher than the power savings obtained.... +# +# From Gwillim Law (2004-06-14): +# http://www.lanacion.com.ar/04/06/10/de_609078.asp ... +# "The time change in Tierra del Fuego was a conflicted decision from +# the start. The government had decreed that the measure would take +# effect on June 1, but a normative error forced the new time to begin +# three days earlier, from a Saturday to a Sunday.... +# Our understanding was that the change was originally scheduled to take place +# on June 1 at 00:00 in Chubut, Santa Cruz, Tierra del Fuego (and some other +# provinces). Sunday was May 30, only two days earlier. So the article +# contains a contradiction. I would give more credence to the Saturday/Sunday +# date than the "three days earlier" phrase, and conclude that Tierra del +# Fuego set its clocks back at 2004-05-30 00:00. # Unless otherwise specified, data are from Shanks through 1992, from # the IATA otherwise. As noted below, Shanks says that -# America/Cordoba split into 7 subregions during 1991/1992, but we +# America/Cordoba split into 6 subregions during 1991/1992, but we # haven't verified this yet so for now we'll keep it a single region. # # Zone NAME GMTOFF RULES FORMAT [UNTIL] # -# Buenos Aires (BA), Capital Federal (CF), Santa Cruz (SC), -# Tierra del Fuego, Antartida e Islas del Atlantico Sur (TF) -Zone America/Buenos_Aires -3:53:48 - LMT 1894 Oct 31 +# Buenos Aires (BA), Capital Federal (CF), +Zone America/Argentina/Buenos_Aires -3:53:48 - LMT 1894 Oct 31 -4:16:48 - CMT 1920 May # Cordoba Mean Time -4:00 - ART 1930 Dec -4:00 Arg AR%sT 1969 Oct 5 @@ -174,24 +204,19 @@ Zone America/Buenos_Aires -3:53:48 - LMT -3:00 - ART # # Santa Fe (SF), Entre Rios (ER), Corrientes (CN), Misiones (MN), Chaco (CC), -# Formosa (FM), Salta (SA), Tucuman (TM), Santiago del Estero (SE), -# Cordoba (CB), La Rioja (LR), San Juan (SJ), San Luis (SL), La Pampa (LP), -# Neuquen (NQ), Rio Negro (RN), Chubut (CH) +# Formosa (FM), Salta (SA), Santiago del Estero (SE), Cordoba (CB), +# San Luis (SL), La Pampa (LP), Neuquen (NQ), Rio Negro (RN) # # Shanks also makes the following claims, which we haven't verified: # - Formosa switched to -3:00 on 1991-01-07. -# - La Rioja and San Juan switched to -4:00 on 1991-03-01 -# and then to -3:00 on 1991-05-07. # - Misiones switched to -3:00 on 1990-12-29. # - Chaco switched to -3:00 on 1991-01-04. # - San Luis switched to -4:00 on 1990-03-14, then to -3:00 on 1990-10-15, # then to -4:00 on 1991-03-01, then to -3:00 on 1991-06-01. # - Santiago del Estero switched to -4:00 on 1991-04-01, # then to -3:00 on 1991-04-26. -# If we need to add Zones for these areas, we may need to have a subdirectory -# for Argentina, as e.g. "America/San_Luis" is too ambiguious. # -Zone America/Cordoba -4:16:48 - LMT 1894 Oct 31 +Zone America/Argentina/Cordoba -4:16:48 - LMT 1894 Oct 31 -4:16:48 - CMT 1920 May -4:00 - ART 1930 Dec -4:00 Arg AR%sT 1969 Oct 5 @@ -201,8 +226,47 @@ Zone America/Cordoba -4:16:48 - LMT 1894 -4:00 Arg AR%sT 2000 Mar 3 -3:00 - ART # +# Tucuman (TM) +Zone America/Argentina/Tucuman -4:20:52 - LMT 1894 Oct 31 + -4:16:48 - CMT 1920 May + -4:00 - ART 1930 Dec + -4:00 Arg AR%sT 1969 Oct 5 + -3:00 Arg AR%sT 1991 Mar 3 + -4:00 - WART 1991 Oct 20 + -3:00 Arg AR%sT 1999 Oct 3 + -4:00 Arg AR%sT 2000 Mar 3 + -3:00 - ART 2004 Jun 1 + -4:00 - WART 2004 Jun 13 + -3:00 - ART +# +# La Rioja (LR) +Zone America/Argentina/La_Rioja -4:27:24 - LMT 1894 Oct 31 + -4:16:48 - CMT 1920 May + -4:00 - ART 1930 Dec + -4:00 Arg AR%sT 1969 Oct 5 + -3:00 Arg AR%sT 1991 Mar 1 + -4:00 - WART 1991 May 7 + -3:00 Arg AR%sT 1999 Oct 3 + -4:00 Arg AR%sT 2000 Mar 3 + -3:00 - ART 2004 Jun 1 + -4:00 - WART 2004 Jun 20 + -3:00 - ART +# +# San Juan (SJ) +Zone America/Argentina/San_Juan -4:34:04 - LMT 1894 Oct 31 + -4:16:48 - CMT 1920 May + -4:00 - ART 1930 Dec + -4:00 Arg AR%sT 1969 Oct 5 + -3:00 Arg AR%sT 1991 Mar 1 + -4:00 - WART 1991 May 7 + -3:00 Arg AR%sT 1999 Oct 3 + -4:00 Arg AR%sT 2000 Mar 3 + -3:00 - ART 2004 Jun 1 + -4:00 - WART 2004 Oct 17 + -3:00 - ART +# # Jujuy (JY) -Zone America/Jujuy -4:21:12 - LMT 1894 Oct 31 +Zone America/Argentina/Jujuy -4:21:12 - LMT 1894 Oct 31 -4:16:48 - CMT 1920 May -4:00 - ART 1930 Dec -4:00 Arg AR%sT 1969 Oct 5 @@ -216,7 +280,7 @@ Zone America/Jujuy -4:21:12 - LMT 1894 O -3:00 - ART # # Catamarca (CT) -Zone America/Catamarca -4:23:08 - LMT 1894 Oct 31 +Zone America/Argentina/Catamarca -4:23:08 - LMT 1894 Oct 31 -4:16:48 - CMT 1920 May -4:00 - ART 1930 Dec -4:00 Arg AR%sT 1969 Oct 5 @@ -224,10 +288,12 @@ Zone America/Catamarca -4:23:08 - LMT 18 -4:00 - WART 1991 Oct 20 -3:00 Arg AR%sT 1999 Oct 3 -4:00 Arg AR%sT 2000 Mar 3 + -3:00 - ART 2004 Jun 1 + -4:00 - WART 2004 Jun 20 -3:00 - ART # # Mendoza (MZ) -Zone America/Mendoza -4:35:16 - LMT 1894 Oct 31 +Zone America/Argentina/Mendoza -4:35:16 - LMT 1894 Oct 31 -4:16:48 - CMT 1920 May -4:00 - ART 1930 Dec -4:00 Arg AR%sT 1969 Oct 5 @@ -242,6 +308,42 @@ Zone America/Mendoza -4:35:16 - LMT 1894 -3:00 - ART 2004 May 23 -4:00 - WART 2004 Oct 17 -3:00 - ART +# +# Chubut (CH) +# The name "Comodoro Rivadavia" exceeds the 14-byte POSIX limit. +Zone America/Argentina/ComodRivadavia -4:30:00 - LMT 1894 Oct 31 + -4:16:48 - CMT 1920 May + -4:00 - ART 1930 Dec + -4:00 Arg AR%sT 1969 Oct 5 + -3:00 Arg AR%sT 1991 Mar 3 + -4:00 - WART 1991 Oct 20 + -3:00 Arg AR%sT 1999 Oct 3 + -4:00 Arg AR%sT 2000 Mar 3 + -3:00 - ART 2004 Jun 1 + -4:00 - WART 2004 Jun 20 + -3:00 - ART +# +# Santa Cruz (SC) +Zone America/Argentina/Rio_Gallegos -4:36:52 - LMT 1894 Oct 31 + -4:16:48 - CMT 1920 May # Cordoba Mean Time + -4:00 - ART 1930 Dec + -4:00 Arg AR%sT 1969 Oct 5 + -3:00 Arg AR%sT 1999 Oct 3 + -4:00 Arg AR%sT 2000 Mar 3 + -3:00 - ART 2004 Jun 1 + -4:00 - WART 2004 Jun 20 + -3:00 - ART +# +# Tierra del Fuego, Antartida e Islas del Atlantico Sur (TF) +Zone America/Argentina/Ushuaia -4:33:12 - LMT 1894 Oct 31 + -4:16:48 - CMT 1920 May # Cordoba Mean Time + -4:00 - ART 1930 Dec + -4:00 Arg AR%sT 1969 Oct 5 + -3:00 Arg AR%sT 1999 Oct 3 + -4:00 Arg AR%sT 2000 Mar 3 + -3:00 - ART 2004 May 30 + -4:00 - WART 2004 Jun 20 + -3:00 - ART # Aruba # Zone NAME GMTOFF RULES FORMAT [UNTIL] diff -pru tz-2004aa/tz-link.htm tz/tz-link.htm --- tz-2004aa/tz-link.htm 2004-05-27 09:00:30.000000000 -0700 +++ tz/tz-link.htm 2004-07-13 13:21:17.651768000 -0700 @@ -9,7 +9,7 @@ <meta http-equiv="Content-type" content='text/html; charset="US-ASCII"' /> <meta name="DC.Creator" content="Eggert, Paul" /> <meta name="DC.Contributor" content="Olson, Arthur David" /> -<meta name="DC.Date" content="2004-05-24" /> +<meta name="DC.Date" content="2004-07-13" /> <meta name="DC.Description" content="Sources of information about time zones and daylight saving time" /> <meta name="DC.Identifier" content="http://www.twinsun.com/tz/tz-link.htm" /> @@ -47,7 +47,7 @@ including <a href="http://www.apple.com/macosx/">Mac OS X</a>, <a href="http://h71000.www7.hp.com/">OpenVMS</a>, <a href="http://wwws.sun.com/software/solaris/">Solaris</a>, -<a href="http://www.tru64unix.compaq.com/">Tru64</a>, and +<a href="http://h30097.www3.hp.com/">Tru64</a>, and <a href="http://www.sco.com/products/unixware/">UnixWare</a>.</p> <p> Each location in the database represents a national region where all @@ -104,14 +104,9 @@ is a text-based point-and-click interfac throughout the world.</li> <li>Fancier web interfaces, roughly in ascending order of complexity, include: <ul> -<li><a href="http://www.hilink.com.au/times/">Local Times Around the -World</a></li> <li><a href="http://www.convertit.com/Go/ConvertIt/World_Time/Current_Time.ASP">Current Time in 1000 Places</a></li> <li><a href="http://timezoneconverter.com/">Time Zone Converter</a></li> </ul></li> -<li><a href="http://www.holidayfestival.com/">The Worldwide Holiday -& Festival Site</a> lists DST-related clock changes along with -holidays.</li> <li><a href="http://www.timeanddate.com/worldclock/">The World Clock - Time Zones</a> is a web interface to a time zone database derived from @@ -152,9 +147,8 @@ contains a script <code>parse_olson</cod <code>tz</code> source into <a href="http://www.perl.org/">Perl</a> modules. It is part of the Perl <a href="http://datetime.perl.org/">DateTime Project</a>, which is freely -available under both the GPL and the Perl <a -href="http://www.perl.com/language/misc/Artistic.html">Artistic -License</a>. DateTime::TimeZone also contains a script +available under both the GPL and the Perl Artistic +License. DateTime::TimeZone also contains a script <code>tests_from_zdump</code> that generates test cases for each clock transition in the <code>tz</code> database.</li> <li><a href="http://oss.software.ibm.com/icu/">International Components for @@ -166,6 +160,10 @@ and time API</a> contains a class <code>org.joda.time.tz.ZoneInfoCompiler</code> that compiles <code>tz</code> source into a Joda-specific binary format. Joda Time is freely available under a BSD-style license.</li> +<li><a href="http://sourceforge.net/projects/pytz/">PyTZ - Python Time +Zone Library</a> compiles <code>tz</code> source into +<a href="http://www.python.org/">Python</a>. +It is freely available under a BSD-style license.</li> </ul> <h2>Other <code>tz</code> binary file readers</h2> <ul> @@ -180,15 +178,14 @@ and is widely used in GNU/Linux systems. is a <code>tz</code> binary file reader written in Java. It is freely available under the GNU LGPL.</li> <li><a href="http://s.keim.free.fr/tz/doc.html">Python time zones</a> -is a <code>tz</code> binary file reader written in <a -href="http://www.python.org/">Python</a>. It is freely available -under a BSD-style license.</li> +is a <code>tz</code> binary file reader written in Python. +It is freely available under a BSD-style license.</li> </ul> <h2>Other <code>tz</code>-based time zone conversion software</h2> <ul> <li><a href="http://java.sun.com/">Sun Java</a> releases since 1.4 -contain a copy of a recent <samp>tz</samp> database in a Java-specific -format.</li> +contain a copy of a subset of a recent <samp>tz</samp> database in a +Java-specific format.</li> <li><a href="http://www1.tip.nl/~t876506/AboutTimeZonesHC.html">HyperCard time zones calculator</a> is a HyperCard stack.</li> @@ -198,8 +195,8 @@ Microsoft Windows program.</li> </ul> <h2>Other time zone databases</h2> <ul> -<li><a href="http://www.astro.com/cgi-bin/atlw3/aq.cgi?lang=e">Atlas Query -- Astrodienst</a> is Astrodienst's Web version of Shanks's +<li><a href="http://www.astro.com/cgi/aq.cgi?lang=e">Atlas Query</a> +is Astrodienst's Web version of Shanks's excellent time zone history atlases published in both <a href="http://astrocom.com/software/pcatlas.php">computer</a> and <a href="http://astrocom.com/books/xrefa.php#SHANKS">book</a> form by <a @@ -217,7 +214,7 @@ for the <code>usno*</code> files in the <li><a href="http://www.airportcitycodes.com/aaa/">Airlines, Airplanes and Airports</a> lists current standard times for thousands of airports around the world. This seems to be derived from -the <a href="http://www.iata.org/sked/publications/">Standard +the <a href="http://www.iata.org/ps/publications/9179.htm">Standard Schedules Information Manual (SSIM)</a> of the the <a href="http://www.iata.org/">International Air Transport Association</a>, @@ -238,8 +235,8 @@ recent editions. The pictorial quality is good, but the maps do not indicate summer time, and parts of the data are a few years out of date.</li> -<li><a href="http://worldtimezone.com/">World timezones map with -current time</a> +<li><a href="http://worldtimezone.com/">Current time around the world +and standard time zones map of the world</a> has several fancy time zone maps; it covers Russia particularly well. The maps' pictorial quality is not quite as good as the CIA's but the maps are more up to date.</li> @@ -252,9 +249,7 @@ boundaries between time zones within cou <li>Manifold.net's <a href="http://www.manifold.net/download/freemaps.html">Free Maps and GIS Data</a> includes a Manifold-format map of world time zone -boundaries distributed under the GPL. The GeoCommunity's <a -href="http://software.geocomm.com/data/intl_timezones.html">International -Time Zones</a> publishes the same data in other formats.</li> +boundaries distributed under the GPL.</li> <li>The US Geological Survey's National Atlas of the United States publishes the <a href="http://www.nationalatlas.gov/timeznm.html">Time Zones of the United States</a> in the public domain.</li> @@ -304,7 +299,7 @@ Portuguese)</a>.</dd> <dd>The Institute for National Measurement Standards publishes current and some older information about <a href="http://inms-ienm.nrc-cnrc.gc.ca/time_services/daylight_savings_e.html">Time -Zones and Daylight Saving Time</a>.</dd> +Zones & Daylight Saving Time</a>.</dd> <dt>Chile</dt> <dd>WebExhibits publishes a <a href="http://webexhibits.org/daylightsaving/chile.html" @@ -420,6 +415,12 @@ protocols.</li> <a href="http://www.exit109.com/~ghealton/y2k/yrexamples.html">The Best of Dates, the Worst of Dates</a> covers many problems encountered by software developers when handling dates and time stamps.</li> +<li><a +href="http://oss.software.ibm.com/cvs/icu/~checkout~/icuhtml/design/formatting/ti…">Time +Zone Localization</a> is a proposed mechanism for localizing time zone +labels and abbreviations; for example, one might use it to specify +Russian translations for "Eastern European Summer Time", "EEST", and +<code>Europe/Bucharest</code>.</li> <li> Alphabetic time zone abbreviations should not be used as unique identifiers for UTC offsets as they are ambiguous in practice. For diff -pru tz-2004aa/zone.tab tz/zone.tab --- tz-2004aa/zone.tab 2003-10-06 06:32:21.000000000 -0700 +++ tz/zone.tab 2004-06-22 21:35:35.301478000 -0700 @@ -41,11 +41,17 @@ AQ -6617+11031 Antarctica/Casey Casey St AQ -7824+10654 Antarctica/Vostok Vostok Station, S Magnetic Pole AQ -6640+14001 Antarctica/DumontDUrville Dumont-d'Urville Base, Terre Adelie AQ -690022+0393524 Antarctica/Syowa Syowa Station, E Ongul I -AR -3436-05827 America/Buenos_Aires E Argentina (BA, DF, SC, TF) -AR -3124-06411 America/Cordoba most locations (CB,CC,CH,CN,ER,FM,LP,LR,MN,NQ,RN,SA,SE,SF,SJ,SL,TM) -AR -2411-06518 America/Jujuy Jujuy (JY) -AR -2828-06547 America/Catamarca Catamarca (CT) -AR -3253-06849 America/Mendoza Mendoza (MZ) +AR -3436-05827 America/Argentina/Buenos_Aires Buenos Aires (BA, CF) +AR -3124-06411 America/Argentina/Cordoba most locations (CB, CC, CN, ER, FM, LP, MN, NQ, RN, SA, SE, SF, SL) +AR -2411-06518 America/Argentina/Jujuy Jujuy (JY) +AR -3124-06411 America/Argentina/Tucuman Tucuman (TM) +AR -2828-06547 America/Argentina/Catamarca Catamarca (CT) +AR -2926-06651 America/Argentina/La_Rioja La Rioja (LR) +AR -3132-06831 America/Argentina/San_Juan San Juan (SJ) +AR -3253-06849 America/Argentina/Mendoza Mendoza (MZ) +AR -4552-06730 America/Argentina/ComodRivadavia Chubut (CH) +AR -5138-06913 America/Argentina/Rio_Gallegos Santa Cruz (SC) +AR -5448-06818 America/Argentina/Ushuaia Tierra del Fuego (TF) AS -1416-17042 Pacific/Pago_Pago AT +4813+01620 Europe/Vienna AU -3133+15905 Australia/Lord_Howe Lord Howe Island @@ -59,6 +65,7 @@ AU -3455+13835 Australia/Adelaide South AU -1228+13050 Australia/Darwin Northern Territory AU -3157+11551 Australia/Perth Western Australia AW +1230-06858 America/Aruba +AX +6006+01957 Europe/Mariehamn AZ +4023+04951 Asia/Baku BA +4352+01825 Europe/Sarajevo BB +1306-05937 America/Barbados diff -pru tz-2004aa/asctime.c tz/asctime.c --- tz-2004aa/asctime.c 1998-05-28 06:56:06.000000000 -0700 +++ tz/asctime.c 2004-06-14 21:23:43.000000000 -0700 @@ -41,14 +41,14 @@ char * buf; else mn = mon_name[timeptr->tm_mon]; /* ** The X3J11-suggested format is - ** "%.3s %.3s%3d %02.2d:%02.2d:%02.2d %d\n" - ** Since the .2 in 02.2d is ignored, we drop it. + ** "%.3s %.3s%3d %.2d:%.2d:%.2d %d\n" + ** Use "%02d", as it is a bit more portable than "%.2d". */ - (void) sprintf(buf, "%.3s %.3s%3d %02d:%02d:%02d %d\n", + (void) sprintf(buf, "%.3s %.3s%3d %02d:%02d:%02d %ld\n", wn, mn, timeptr->tm_mday, timeptr->tm_hour, timeptr->tm_min, timeptr->tm_sec, - TM_YEAR_BASE + timeptr->tm_year); + timeptr->tm_year + (long) TM_YEAR_BASE); return buf; } diff -pru tz-2004aa/date.c tz/date.c --- tz-2004aa/date.c 2001-10-09 10:31:31.000000000 -0700 +++ tz/date.c 2004-06-14 22:44:08.703969000 -0700 @@ -630,8 +630,12 @@ const time_t t; time_t outt; tm = *localtime(&t); - cent = (tm.tm_year + TM_YEAR_BASE) / 100; - year_in_cent = (tm.tm_year + TM_YEAR_BASE) - cent * 100; + cent = tm.tm_year / 100 + TM_YEAR_BASE / 100; + year_in_cent = tm.tm_year % 100; + if (year_int_cent < 0) { + cent--; + year_in_cent += 100; + } month = tm.tm_mon + 1; day = tm.tm_mday; hour = tm.tm_hour; diff -pru tz-2004aa/localtime.c tz/localtime.c --- tz-2004aa/localtime.c 2003-12-15 06:36:35.000000000 -0800 +++ tz/localtime.c 2004-06-14 23:11:00.519250000 -0700 @@ -134,8 +134,11 @@ static void gmtsub P((const time_t * ti static void localsub P((const time_t * timep, long offset, struct tm * tmp)); static int increment_overflow P((int * number, int delta)); +static int long_increment_overflow P((long * number, int delta)); static int normalize_overflow P((int * tensptr, int * unitsptr, int base)); +static int long_normalize_overflow P((long * tensptr, + int * unitsptr, int base)); static void settzname P((void)); static time_t time1 P((struct tm * tmp, void(*funcp) P((const time_t *, @@ -1149,7 +1152,7 @@ register struct tm * const tmp; register const struct lsinfo * lp; register long days; register long rem; - register int y; + register long y; register int yleap; register const int * ip; register long corr; @@ -1218,7 +1221,7 @@ register struct tm * const tmp; y = EPOCH_YEAR; #define LEAPS_THRU_END_OF(y) ((y) / 4 - (y) / 100 + (y) / 400) while (days < 0 || days >= (long) year_lengths[yleap = isleap(y)]) { - register int newy; + register long newy; newy = y + days / DAYSPERNYEAR; if (days < 0) @@ -1294,6 +1297,18 @@ int delta; } static int +long_increment_overflow(number, delta) +long * number; +int delta; +{ + long number0; + + number0 = *number; + *number += delta; + return (*number < number0) != (delta < 0); +} + +static int normalize_overflow(tensptr, unitsptr, base) int * const tensptr; int * const unitsptr; @@ -1309,6 +1324,21 @@ const int base; } static int +long_normalize_overflow(tensptr, unitsptr, base) +long * const tensptr; +int * const unitsptr; +const int base; +{ + register int tensdelta; + + tensdelta = (*unitsptr >= 0) ? + (*unitsptr / base) : + (-1 - (-1 - *unitsptr) / base); + *unitsptr -= tensdelta * base; + return long_increment_overflow(tensptr, tensdelta); +} + +static int tmcomp(atmp, btmp) register const struct tm * const atmp; register const struct tm * const btmp; @@ -1336,6 +1366,8 @@ const int do_norm_secs; register int dir; register int bits; register int i, j ; + register long li; + long y; register int saved_seconds; time_t newt; time_t t; @@ -1352,42 +1384,46 @@ const int do_norm_secs; return WRONG; if (normalize_overflow(&yourtm.tm_mday, &yourtm.tm_hour, HOURSPERDAY)) return WRONG; - if (normalize_overflow(&yourtm.tm_year, &yourtm.tm_mon, MONSPERYEAR)) + y = yourtm.tm_year; + if (long_normalize_overflow(&y, &yourtm.tm_mon, MONSPERYEAR)) return WRONG; /* - ** Turn yourtm.tm_year into an actual year number for now. + ** Turn y into an actual year number for now. ** It is converted back to an offset from TM_YEAR_BASE later. */ - if (increment_overflow(&yourtm.tm_year, TM_YEAR_BASE)) + if (long_increment_overflow(&y, TM_YEAR_BASE)) return WRONG; while (yourtm.tm_mday <= 0) { - if (increment_overflow(&yourtm.tm_year, -1)) + if (long_increment_overflow(&y, -1)) return WRONG; - i = yourtm.tm_year + (1 < yourtm.tm_mon); - yourtm.tm_mday += year_lengths[isleap(i)]; + li = y + (1 < yourtm.tm_mon); + yourtm.tm_mday += year_lengths[isleap(li)]; } while (yourtm.tm_mday > DAYSPERLYEAR) { - i = yourtm.tm_year + (1 < yourtm.tm_mon); - yourtm.tm_mday -= year_lengths[isleap(i)]; - if (increment_overflow(&yourtm.tm_year, 1)) + li = y + (1 < yourtm.tm_mon); + yourtm.tm_mday -= year_lengths[isleap(li)]; + if (long_increment_overflow(&y, 1)) return WRONG; } for ( ; ; ) { - i = mon_lengths[isleap(yourtm.tm_year)][yourtm.tm_mon]; + i = mon_lengths[isleap(y)][yourtm.tm_mon]; if (yourtm.tm_mday <= i) break; yourtm.tm_mday -= i; if (++yourtm.tm_mon >= MONSPERYEAR) { yourtm.tm_mon = 0; - if (increment_overflow(&yourtm.tm_year, 1)) + if (long_increment_overflow(&y, 1)) return WRONG; } } - if (increment_overflow(&yourtm.tm_year, -TM_YEAR_BASE)) + if (long_increment_overflow(&y, -TM_YEAR_BASE)) + return WRONG; + yourtm.tm_year = y; + if (yourtm.tm_year != y) return WRONG; if (yourtm.tm_sec >= 0 && yourtm.tm_sec < SECSPERMIN) saved_seconds = 0; - else if (yourtm.tm_year + TM_YEAR_BASE < EPOCH_YEAR) { + else if (y < EPOCH_YEAR - TM_YEAR_BASE) { /* ** We can't set tm_sec to 0, because that might push the ** time below the minimum representable time. diff -pru tz-2004aa/strftime.c tz/strftime.c --- tz-2004aa/strftime.c 2001-10-04 14:01:17.000000000 -0700 +++ tz/strftime.c 2004-06-14 22:37:47.213984000 -0700 @@ -108,6 +108,7 @@ static const struct lc_time_T C_time_loc static char * _add P((const char *, char *, const char *)); static char * _conv P((int, const char *, char *, const char *)); +static char * _lconv P((long, const char *, char *, const char *)); static char * _fmt P((const char *, const struct tm *, char *, const char *, int *)); size_t strftime P((char *, size_t, const char *, const struct tm *)); @@ -210,8 +211,12 @@ label: ** something completely different. ** (ado, 1993-05-24) */ - pt = _conv((t->tm_year + TM_YEAR_BASE) / 100, - "%02d", pt, ptlim); + { + int c = (TM_YEAR_BASE / 100 + + t->tm_year / 100 + - (t->tm_year % 100 < 0)); + pt = _conv(c, "%02d", pt, ptlim); + } continue; case 'c': { @@ -379,12 +384,13 @@ label: ** (ado, 1996-01-02) */ { - int year; + long year; int yday; int wday; int w; - year = t->tm_year + TM_YEAR_BASE; + year = t->tm_year; + year += TM_YEAR_BASE; yday = t->tm_yday; wday = t->tm_wday; for ( ; ; ) { @@ -436,10 +442,13 @@ label: pt = _conv(w, "%02d", pt, ptlim); else if (*format == 'g') { + int g = year % 100; + if (g < 0) + g += 100; *warnp = IN_ALL; - pt = _conv(year % 100, "%02d", + pt = _conv(g, "%02d", pt, ptlim); - } else pt = _conv(year, "%04d", + } else pt = _lconv(year, "%04ld", pt, ptlim); } continue; @@ -476,13 +485,17 @@ label: } continue; case 'y': + { + int y = t->tm_year % 100; + if (y < 0) + y += 100; *warnp = IN_ALL; - pt = _conv((t->tm_year + TM_YEAR_BASE) % 100, - "%02d", pt, ptlim); + pt = _conv(y, "%02d", pt, ptlim); + } continue; case 'Y': - pt = _conv(t->tm_year + TM_YEAR_BASE, "%04d", - pt, ptlim); + pt = _lconv(t->tm_year + (long) TM_YEAR_BASE, + "%04ld", pt, ptlim); continue; case 'Z': #ifdef TM_ZONE @@ -586,6 +599,19 @@ const char * const ptlim; } static char * +_lconv(n, format, pt, ptlim) +const long n; +const char * const format; +char * const pt; +const char * const ptlim; +{ + char buf[INT_STRLEN_MAXIMUM(long) + 1]; + + (void) sprintf(buf, format, n); + return _add(buf, pt, ptlim); +} + +static char * _add(str, pt, ptlim) const char * str; char * pt; diff -pru tz-2004aa/zdump.8 tz/zdump.8 --- tz-2004aa/zdump.8 2003-09-16 04:12:41.000000000 -0700 +++ tz/zdump.8 2004-06-17 13:33:33.238073000 -0700 @@ -40,6 +40,12 @@ otherwise. .TP .BI "\-c " cutoffyear Cut off the verbose output near the start of the given year. +By default, verbose output is cut off a few decades after the +current date and time. .SH "SEE ALSO" newctime(3), tzfile(5), zic(8) +.SH "BUGS" +Time discontinuities are detected via a heuristic that suffices for +all historically valid discontinuities, but one can construct +artificial time zones for which the heuristic fails. .\" @(#)zdump.8 7.4 diff -pru tz-2004aa/zdump.c tz/zdump.c --- tz-2004aa/zdump.c 2003-09-16 04:12:41.000000000 -0700 +++ tz/zdump.c 2004-06-17 14:12:10.425374000 -0700 @@ -6,6 +6,7 @@ static char elsieid[] = "@(#)zdump.c 7.3 ** You can use this code to help in verifying other implementations. */ +#include "limits.h" /* for CHAR_BIT */ #include "stdio.h" /* for stdout, stderr, perror */ #include "string.h" /* for strcpy */ #include "sys/types.h" /* for time_t */ @@ -24,6 +25,25 @@ static char elsieid[] = "@(#)zdump.c 7.3 #define FALSE 0 #endif /* !defined FALSE */ +#ifndef TYPE_BIT +#define TYPE_BIT(type) (sizeof (type) * CHAR_BIT) +#endif /* !defined TYPE_BIT */ + +#ifndef TYPE_SIGNED +#define TYPE_SIGNED(type) (((type) -1) < 0) +#endif /* !defined TYPE_SIGNED */ + +#ifndef INT_STRLEN_MAXIMUM +/* +** 302 / 1000 is log10(2.0) rounded up. +** Subtract one for the sign bit if the type is signed; +** add one for integer division truncation; +** add one more for a minus sign if the type is signed. +*/ +#define INT_STRLEN_MAXIMUM(type) \ + ((TYPE_BIT(type) - TYPE_SIGNED(type)) * 302 / 1000 + 1 + TYPE_SIGNED(type)) +#endif /* !defined INT_STRLEN_MAXIMUM */ + #ifndef EXIT_SUCCESS #define EXIT_SUCCESS 0 #endif /* !defined EXIT_SUCCESS */ @@ -56,10 +76,18 @@ static char elsieid[] = "@(#)zdump.c 7.3 #define TM_YEAR_BASE 1900 #endif /* !defined TM_YEAR_BASE */ +#ifndef DAYSPERWEEK +#define DAYSPERWEEK 7 +#endif /* !defined DAYSPERWEEK */ + #ifndef DAYSPERNYEAR #define DAYSPERNYEAR 365 #endif /* !defined DAYSPERNYEAR */ +#ifndef MONSPERYEAR +#define MONSPERYEAR 12 +#endif /* !defined MONSPERYEAR */ + #ifndef isleap #define isleap(y) ((((y) % 4) == 0 && ((y) % 100) != 0) || ((y) % 400) == 0) #endif /* !defined isleap */ @@ -128,6 +156,9 @@ static long delta P((struct tm * newp, s static time_t hunt P((char * name, time_t lot, time_t hit)); static size_t longest; static char * progname; +static char * my_asctime_r P((const struct tm *, char *)); +static char * my_asctime P((const struct tm *)); +static void showtime P((time_t, const struct tm *)); static void show P((char * zone, time_t t, int v)); int @@ -139,17 +170,19 @@ char * argv[]; register int c; register int vflag; register char * cutoff; - register int cutyear; - register long cuttime; + register long cutyear; + long y; + register time_t lasttime; char ** fakeenv; time_t now; time_t t; time_t newt; time_t hibit; + time_t time_t_min; + time_t time_t_max; struct tm tm; - struct tm newtm; + struct tm * newtm; - INITIALIZE(cuttime); #if HAVE_GETTEXT - 0 (void) setlocale(LC_MESSAGES, ""); #ifdef TZ_DOMAINDIR @@ -176,22 +209,44 @@ _("%s: usage is %s [ --version ] [ -v ] argv[0], argv[0]); (void) exit(EXIT_FAILURE); } + for (hibit = 1; (hibit << 1) != 0; hibit <<= 1) + continue; + time_t_max = ~((time_t) 0); + if (TYPE_SIGNED(time_t)) + time_t_max &= ~hibit; + time_t_min = TYPE_SIGNED(time_t) ? hibit : 0; + (void) time(&now); if (cutoff != NULL) { - int y; - - cutyear = atoi(cutoff); - cuttime = 0; - for (y = EPOCH_YEAR; y < cutyear; ++y) - cuttime += DAYSPERNYEAR + isleap(y); - cuttime *= SECSPERHOUR * HOURSPERDAY; + cutyear = atol(cutoff); + } else { + /* + ** By default, the cutoff is the maximum time value, + ** or a few decades from now, whichever comes first. + ** This prevents us from generating too much output if + ** time_t is wider than 32 bits. + */ + newtm = gmtime (&now); + if (newtm == NULL) + cutyear = now < 0 ? LONG_MIN : LONG_MAX; + else { + cutyear = newtm->tm_year + (long) TM_YEAR_BASE; + cutyear = 2050 + 100 * ((cutyear - 1930) / 100); + } + } + lasttime = 0; + for (y = EPOCH_YEAR; y < cutyear; ++y) { + newt = lasttime - !lasttime; + newt += (DAYSPERNYEAR + isleap(y)) * SECSPERHOUR * HOURSPERDAY; + if (newt < lasttime) { + lasttime = time_t_max; + break; + } + lasttime = newt; } - (void) time(&now); longest = 0; for (i = optind; i < argc; ++i) if (strlen(argv[i]) > longest) longest = strlen(argv[i]); - for (hibit = 1; (hibit << 1) != 0; hibit <<= 1) - continue; { register int from; register int to; @@ -224,40 +279,47 @@ _("%s: usage is %s [ --version ] [ -v ] /* ** Get lowest value of t. */ - t = hibit; - if (t > 0) /* time_t is unsigned */ - t = 0; + t = time_t_min; show(argv[i], t, TRUE); t += SECSPERHOUR * HOURSPERDAY; show(argv[i], t, TRUE); + /* + ** But don't go below - 2**32 in the following loop, + ** as that predates standard time, would cause us to + ** generate too much output, and could dump core if + ** localtime returns NULL. The earliest known use of + ** standard time is 1837 (Iceland) and - 2**32 is a + ** bit before that. + */ + if ((t + 1) / 2 < -2147483647) { + t = -1 - 2147483647; + t *= 2; + } tm = *localtime(&t); (void) strncpy(buf, abbr(&tm), (sizeof buf) - 1); for ( ; ; ) { - if (cutoff != NULL && t >= cuttime) + if (lasttime < t) break; newt = t + SECSPERHOUR * 12; - if (cutoff != NULL && newt >= cuttime) + if (lasttime < newt) break; if (newt <= t) break; - newtm = *localtime(&newt); - if (delta(&newtm, &tm) != (newt - t) || - newtm.tm_isdst != tm.tm_isdst || - strcmp(abbr(&newtm), buf) != 0) { + newtm = localtime(&newt); + if (newtm == NULL) + break; + if (delta(newtm, &tm) != (newt - t) || + newtm->tm_isdst != tm.tm_isdst || + strcmp(abbr(newtm), buf) != 0) { newt = hunt(argv[i], t, newt); - newtm = *localtime(&newt); - (void) strncpy(buf, abbr(&newtm), + newtm = localtime(&newt); + (void) strncpy(buf, abbr(newtm), (sizeof buf) - 1); } t = newt; - tm = newtm; + tm = *newtm; } - /* - ** Get highest value of t. - */ - t = ~((time_t) 0); - if (t < 0) /* time_t is signed */ - t &= ~hibit; + t = time_t_max; t -= SECSPERHOUR * HOURSPERDAY; show(argv[i], t, TRUE); t += SECSPERHOUR * HOURSPERDAY; @@ -323,7 +385,7 @@ struct tm * oldp; return -delta(oldp, newp); result = 0; for (tmy = oldp->tm_year; tmy < newp->tm_year; ++tmy) - result += DAYSPERNYEAR + isleap(tmy + TM_YEAR_BASE); + result += DAYSPERNYEAR + isleap(tmy + (long) TM_YEAR_BASE); result += newp->tm_yday - oldp->tm_yday; result *= HOURSPERDAY; result += newp->tm_hour - oldp->tm_hour; @@ -334,6 +396,80 @@ struct tm * oldp; return result; } +/* +** my_asctime_r and my_asctime are stolen from asctime.c. Unlike +** Standard C asctime, my_asctime has well-defined behavior even if +** the year is less than -999 or greater than 9999. +*/ + +static char * +my_asctime_r(timeptr, buf) +register const struct tm * timeptr; +char * buf; +{ + static const char wday_name[][3] = { + "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat" + }; + static const char mon_name[][3] = { + "Jan", "Feb", "Mar", "Apr", "May", "Jun", + "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" + }; + register const char * wn; + register const char * mn; + + if (timeptr->tm_wday < 0 || timeptr->tm_wday >= DAYSPERWEEK) + wn = "???"; + else wn = wday_name[timeptr->tm_wday]; + if (timeptr->tm_mon < 0 || timeptr->tm_mon >= MONSPERYEAR) + mn = "???"; + else mn = mon_name[timeptr->tm_mon]; + /* + ** The X3J11-suggested format is + ** "%.3s %.3s%3d %.2d:%.2d:%.2d %d\n" + ** Use "%02d", as it is a bit more portable than "%.2d". + */ + (void) sprintf(buf, "%.3s %.3s%3d %02d:%02d:%02d %ld\n", + wn, mn, + timeptr->tm_mday, timeptr->tm_hour, + timeptr->tm_min, timeptr->tm_sec, + timeptr->tm_year + (long) TM_YEAR_BASE); + return buf; +} + +static char * +my_asctime(timeptr) +register const struct tm * timeptr; +{ + /* + ** Big enough for something such as + ** ??? ???-2147483648 -2147483648:-2147483648:-2147483648 -2147483648\n + ** (two three-character abbreviations, five strings denoting integers, + ** three explicit spaces, two explicit colons, a newline, + ** and a trailing ASCII nul). + */ + static char result[3 * 2 + 5 * INT_STRLEN_MAXIMUM(int) + + 3 + 2 + 1 + 1]; + + return my_asctime_r(timeptr, result); +} + +static void +showtime(t, timeptr) +time_t t; +const struct tm * timeptr; +{ + if (timeptr != NULL) { + const char *p; + for (p = my_asctime (timeptr); *p != '\n'; p++) + putchar (*p); + } else { + if (TYPE_SIGNED(time_t)) + printf ("%ld\n", (long) t); + else + printf ("%lu\n", (unsigned long) t); + } +} + static void show(zone, t, v) char * zone; @@ -343,13 +479,15 @@ int v; struct tm * tmp; (void) printf("%-*s ", (int) longest, zone); - if (v) - (void) printf("%.24s UTC = ", asctime(gmtime(&t))); + if (v) { + showtime(t, gmtime(&t)); + (void) printf(" UTC = "); + } tmp = localtime(&t); - (void) printf("%.24s", asctime(tmp)); + showtime(t, tmp); if (*abbr(tmp) != '\0') (void) printf(" %s", abbr(tmp)); - if (v) { + if (v && tmp != NULL) { (void) printf(" isdst=%d", tmp->tm_isdst); #ifdef TM_GMTOFF (void) printf(" gmtoff=%ld", tmp->TM_GMTOFF); @@ -365,7 +503,7 @@ struct tm * tmp; register char * result; static char nada; - if (tmp->tm_isdst != 0 && tmp->tm_isdst != 1) + if (!tmp || (tmp->tm_isdst != 0 && tmp->tm_isdst != 1)) return &nada; result = tzname[tmp->tm_isdst]; return (result == NULL) ? &nada : result;

1 0

RE: Definition of time_t changed from signed to unsigned...
by Olson, Arthur David (NIH/NCI) July 13, 2004

July 13, 2004

> The problem is that the timezone subsystem seems to expect time_t to be signed. There are a number of functions that only work if time_t is signed. > E.g. function localsub() in file localtime.c. The presence of the comment... /* ** If time_t is signed, then 0 is just above the median, ** assuming two's complement arithmetic. ** If time_t is unsigned, then (1 << bits) is just above the median. */ ...is a reminder that some effort went into ensuring that the stuff in localtime.c would work whether time_t was signed or unsigned. I'd love to get details on what's going wrong with localsub when time_t is unsigned. --ado

1 0