- tz - lists.iana.org

[colsen@usa.net: conf/1150: /usr/src/share/zoneinfo/north]
by Garrett Wollman April 19, 1996

April 19, 1996

This error is present in tzdata96e on elsie... -GAWollman ------- start of forwarded message (RFC 934 encapsulation) ------- Message-Id: <199604191457.IAA00442(a)mach5.compus.com> From: colsen(a)usa.net Sender: owner-bugs(a)freebsd.org To: FreeBSD-gnats-submit(a)freebsd.org Subject: conf/1150: /usr/src/share/zoneinfo/northamerica file for Denver is wrong Date: Fri, 19 Apr 1996 08:57:58 -0600 (MDT) >Number: 1150 >Category: conf >Synopsis: /usr/src/share/zoneinfo/northamerica file for Denver is wrong >Confidential: no >Severity: serious >Priority: medium >Responsible: freebsd-bugs >State: open >Class: sw-bug >Submitter-Id: current-users >Arrival-Date: Fri Apr 19 08:00:01 PDT 1996 >Last-Modified: >Originator: Christopher B. Olsen >Organization: COMPUS Services Corp. >Release: FreeBSD 2.2-CURRENT i386 >Environment: Tracking current. 32Mb, 166Mhz Pentium, 4Gb MICROP SCSI-2 Fast/Wide, BusLogic BT-956 >Description: The northamerica timezone file contains a bogus description for the Mountain time zone. >How-To-Repeat: link /etc/localtime -> /usr/share/zoneinfo/Denver. And check the time. you will get EDT or EST. >Fix: *** northamerica Fri Apr 19 09:42:41 1996 - --- northamerica.bak Fri Apr 19 09:42:41 1996 *************** *** 162,172 **** Rule Denver 1965 1966 - Oct lastSun 2:00 0 S # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone America/Denver -6:59:56 - LMT 1883 Nov 18 12:00 ! -7:00 US M%sT 1920 ! -7:00 Denver M%sT 1942 ! -7:00 US M%sT 1946 ! -7:00 Denver M%sT 1967 ! -7:00 US M%sT # ZONE-DESCR America Denver United_States Mountain time (except Arizona) # US Pacific time, represented by Los Angeles - --- 162,172 ---- Rule Denver 1965 1966 - Oct lastSun 2:00 0 S # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone America/Denver -6:59:56 - LMT 1883 Nov 18 12:00 ! -7:00 US E%sT 1920 ! -7:00 Denver E%sT 1942 ! -7:00 US E%sT 1946 ! -7:00 Denver E%sT 1967 ! -7:00 US E%sT # ZONE-DESCR America Denver United_States Mountain time (except Arizona) # US Pacific time, represented by Los Angeles >Audit-Trail: >Unformatted: ------- end -------

1 0

2 questions about the list
by Chris Carrier April 17, 1996

April 17, 1996

I would like to ask a couple of questions about this list. About how many people are on it? Also, I see it is from nci.nih.gov. Exactly who runs it? NIH sounds like National Institute of Health, but why would they be running a time zone mailing list? (Sleep deprivation perhaps :) )

1 0

DST and accidents
by Chris Carrier April 6, 1996

April 6, 1996

On 04-Apr-96 20:52 PST Arthur David Olson (ado(a)elsie.nci.nih.gov) wrote about: DST in the news >Page A9 of the Thursday, April 4, 1996 Washington Post has a story headlined >"Daylight Savings May Be Motorists' Loss," .. about a letter to >the New England Journal of Medicine from Stanley Coren of the University of >British Columbia. The article notes that "Moving clocks forward an hour... >causes more traffic accidents the next day, according to [Coren's] study." Hmmm. I remember when the 1986 amendment to the Uniform Time Act was passed which started DST earlier, one of the arguments in favor was that it would REDUCE traffic accidents in the month of April. I was opposed to the 1986 amendment extending DST, even though I don't mind DST in April. I opposed it because I would like to see DST end not later than the last Sunday in September, and to have the time zones brought closer to where they astronomically ought to be to prevent the post-8 a.m. sunrises parts of the country are plagued with in October. Since the amendment didn't touch this (and in fact, the bill that passed the House would have made it worse, extending DST to the 1st Sunday in November) I was against it.

1 0

Re: timezone for Mexico
by Alan.Perry＠Eng.Sun.COM April 5, 1996

April 5, 1996

> From mallen(a)kin.cyborg.com.mx Fri Apr 5 13:50:34 1996 > > I just checked and this year will have a "horario de verano" from > april first to october 27 (both sundays) in central Mexico and the change > will be made at two am in the next day. > > I don't know if this will remain unchanged in the next years as it is > stupid to change time in places with latitudes between the tropics as the > length of the day does not change much. It seems that people in FIDE know > little about Geography/Astronomy. Since this came from a Presidential Decree, I would say that it is the Mexican Administration that knows little about Geography/Astronomy. I am unfamiliar with the parties involved, so I will reserve judgment. I think the Mexico entry in northamerica should be changed to refer to the Decree in the 4 Jan Mexico national newspaper (that I posted to this list, but have since deleted) rather than the FIDE comment. alan

1 0

timezone for Mexico
by Guillermo Mallen F. April 5, 1996

April 5, 1996

I just checked and this year will have a "horario de verano" from april first to october 27 (both sundays) in central Mexico and the change will be made at two am in the next day. I don't know if this will remain unchanged in the next years as it is stupid to change time in places with latitudes between the tropics as the length of the day does not change much. It seems that people in FIDE know little about Geography/Astronomy. Regards Guillermo M. Mallen mallen(a)kin.cyborg.com.mx

1 0

DST in the news
by ado April 5, 1996

April 5, 1996

Page A9 of the Thursday, April 4, 1996 Washington Post has a story headlined "Daylight Savings May Be Motorists' Loss," a Reuter dispatch about a letter to the New England Journal of Medicine from Stanley Coren of the University of British Columbia. The article notes that "Moving clocks forward an hour... causes more traffic accidents the next day, according to [Coren's] study." --ado

1 0

Re: EU daylight time
by Peter Ilieve April 4, 1996

April 4, 1996

On Apr 3, 9:51pm, Chris Carrier wrote: > Subject: EU daylight time > I am given to understand that the countries of the European Union > have decided to extend their 'summer-time period' until the last > Sunday in October. I would like to ask: > > Can individual EU countries choose to opt out of daylight time > entirely, as American states can opt out of DST under our own > Uniform Time Act of 1966, as amended 1972? I don't believe so, but I am no lawyer. The legal authority for the summer time dates is an EC Directive. This is supposed to have the force of law and EU member states are required to enact national legislation to put directives into effect. Countries are sometimes slow in doing this in other areas but I don't know of any countries not doing so where summer time is concerned. The process of making a directive involves the European Commission agreeing on the terms of it. It is then put to the Council of Ministers for final approval. Both the Commission and the Council are made up of appointees of the national governments. The Commissioners are supposed to ignore their contriy's national interests and work for the EU as a whole but the Council members can and do vote on things based on their own national interest. As a result of all this it is unlikely that a summer time directive would emerge that any country wouldn't agree with. We are now moving towards a system of `qualified majority voting' in the Council so in theory a country's national interest could be out-voted, but I don't know if the summer time directives are covered by this procedure. Countries would not have much incentive to opt out of this summer time end date change as a survey was conducted a few years ago as part of the seventh EC directive setting process and it found the people of Europe to be in favour of moving the date by a large majority. Peter Ilieve peter(a)aldie.co.uk

1 0

EU daylight time
by Chris Carrier April 4, 1996

April 4, 1996

I am given to understand that the countries of the European Union have decided to extend their 'summer-time period' until the last Sunday in October. I would like to ask: Can individual EU countries choose to opt out of daylight time entirely, as American states can opt out of DST under our own Uniform Time Act of 1966, as amended 1972? Chris Carrier

1 0

No summer time in PRC
by cml1001＠cus.cam.ac.uk March 31, 1996

March 31, 1996

Hello. P.R. of China has abolished summer time for a few years. Perhaps this is because summer time is not used in Hong Kong and Macau, so summer time in China caused much confusion. -- William - Have a good time all the time :-) ---------------------------------------------------------------------- URL: http://club.eng.cam.ac.uk/~92cml

1 0

Re: patch for Europe/Lisbon bug in tzdata96d
by ado March 19, 1996

March 19, 1996

The latest change to portugal is incorporated in tzdata96e.tar.gz, available for anonymous ftp from elsie.nci.nih.gov (165.112.73.1) in the ~ftp/pub directory: -rw-r--r-- 1 ado 73436 Mar 19 17:37 tzdata96e.tar.gz Timezone caballeros are warned to be braced for our busy April season. --ado

1 0

Portugal time
by Carlo E. Prelz March 19, 1996

March 19, 1996

Hi. You already know that, again, Portugal goes back to West European Time. From the "europe" file in tzdata96d.tgz: --8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<-- 1:00 EC MET%s 1996 Mar 31 1:00u # Martin Bruckmann <martin(a)ua.pt> (1996-02-29) reports via Peter Ilieve # that Portugal is reverting to 0:00 by not moving its clocks this spring. # The new Prime Minister was fed up with getting up in the dark in the winter. 0:00 - "WET DST" 1996 Oct 27 1:00u 0:00 EC WET%s --8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<-- Well, only 1 1/2 weeks are separating us from March 31, and I was doing some tests (on Linux, very recent kernel - more info on request). Well, from Oct. 27 all will be OK for sure, but the rule for the intermediate period does not set the dst flag - probably because the transition to the new rule happens at the same moment as the DST one. I could not find a way to set the dst flag in rules - the only way to have the rule behave correctly is to write: 1:00 - "WET DST" 1996 Oct 27 1:00u (1:00 instead of 0:00). This way, the time is right. The dst flag is still off, and that could crack programs that use it. Do you know of a better way? Am I doing something wrong? TIA Carlo -- * Se la Strada e la sua Virtu' non fossero state messe da parte, * K * Carlo E. Prelz - fluido(a)marktest.pt che bisogno ci sarebbe * di parlare tanto di amore e di rettitudine? (Chuang-Tzu)

2 1

tzdata96d.tar.gz
by ado March 19, 1996

March 19, 1996

tzdata96d.tar.gz is now available for anonymous ftp from elsie.nci.nih.gov (165.112.73.1) in the ~ftp/pub directory: -rw-r--r-- 1 ado 73472 Mar 18 20:53 tzdata96d.tar.gz It reflects the proposals recently circulated on the tz list by Paul Eggert, in particular the Portugal change that comes into play at the end of this month. --ado

1 0

Error in Timezone data
by Alan Eugene Davis March 16, 1996

March 16, 1996

Fernando Alegre <alegre(a)mars.superlink.net> Package Maintainer, Debian Linux Package timezone-7.46-0.deb cc: tz(a)elsie.nci.nih.gov (Upstream) In the debian linux timezone package, in the file australasia, in the source, there may be an error. I hesitate to make such an assertion, but I am certain, having lived there for the past decade, that "Truk" (now properly known as Chuuk), at about 7°N 151°E, is in the time zone GMT+10, and not +11 as indicated in the australisia file. You may also be interested that the capital of Chuuk ( a State of the Federated States of Micronesia, since I think 1986) is Weno (properly, Wené, but usually spelled Weno). The approximate Latitude and Longitude are (probably for my island of Toon, not for Wené, these ought to be checked): 7° 19.8' N 151° 34.8' E ---------------------From the file australasia---------------- # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone Pacific/Yap 9:12:32 - LMT 1901 # Colonia 9:00 - PLT 1969 Oct 10:00 - GST Zone Pacific/Truk 10:07:08 - LMT 1901 10:00 - GST 1978 Oct 11:00 - NCST <------------WRONG? Zone Pacific/Ponape 10:32:52 - LMT 1901 # Kolonia 11:00 - NCST Zone Pacific/Kosrae 10:51:56 - LMT 1901 11:00 - NCST 1969 Oct 12:00 - NZST --------------------------------------------------------------------------------- As of 1995, I was the only Linux user in Chuuk. I would be very much interested in learning of information that Chuuk Lagoon is in the GMT+11 timezone. I don't think it is, however. Alan Davis adavis(a)saba.kuentos.guam.net

1 0

Blather 2000 Conference
by Monty Solomon March 15, 1996

March 15, 1996

FYI. Excerpt from mini-AIR Mar 96. ----------------------------------------------------------- 1996-03-14 Blather 2000 Conference Enough hot air has already been released to make it obvious that Humankind must anticipate accelerating amounts of endless palaver about how New Year's Eve December 31st 1999 isn't REALLY the End of the 20th Century and how January 1st, 2000 isn't REALLY going to be the First Day of the 21st Century. AIR will sponsor a conference on "Blather 2000 -- Strategies to Avoid the Pedantry." We will bring together the World's Leading Thinkers to debate and make one big announcement about what the deal is, put their statement into a widely-publicized Web Page, and then ask the second-rate usage pundits, congenital columnists, and other emitters of Sound 'n' Fury (TM) to just Shut Up About It. If you have anything to say on the matter -- anything that hasn't already been said -- please direct it to the Blather 2000 Conference chairman, Dr. Y. Foo <blather2k(a)mit.edu>

1 0

proposed tz changes for Portugal, Etc/GMT-14, and Mideast/Riyadh89
by Paul Eggert March 12, 1996

March 12, 1996

Here are proposed changes to the tz database: * Portugal is apparently not moving its clocks this spring. * Since Kiritimati is now 14 hours ahead of GMT, there should be an Etc/GMT-14 zone. (I suppose if Cuba goes ahead with its threat to move its clocks ahead by 24 hours, we'll need an Etc/GMT-19 zone...) * Fix a typo in solar89. =================================================================== RCS file: RCS/europe,v retrieving revision 1995.13 retrieving revision 1995.13.1.1 diff -c -r1995.13 -r1995.13.1.1 *** europe 1995/12/21 17:18:29 1995.13 --- europe 1996/03/12 03:16:42 1995.13.1.1 *************** *** 1644,1650 **** # From Rui Pedro Salgueiro <rps(a)inescca.inescc.pt> (November 12, 1992): # Portugal has recently (September, 27) changed timezone # (from WET to MET or CET) to harmonize with EEC. ! 1:00 EC MET%s # We don't know what happened to Madeira or the Azores, # so we'll just use Shanks for now. # Zone NAME GMTOFF RULES FORMAT [UNTIL] --- 1644,1655 ---- # From Rui Pedro Salgueiro <rps(a)inescca.inescc.pt> (November 12, 1992): # Portugal has recently (September, 27) changed timezone # (from WET to MET or CET) to harmonize with EEC. ! 1:00 EC MET%s 1996 Mar 31 1:00u ! # Martin Bruckmann <martin(a)ua.pt> (1996-02-29) reports via Peter Ilieve ! # that Portugal is reverting to 0:00 by not moving its clocks this spring. ! # The new Prime Minister was fed up with getting up in the dark in the winter. ! 0:00 - "WET DST" 1996 Oct 27 1:00u ! 0:00 EC WET%s # We don't know what happened to Madeira or the Azores, # so we'll just use Shanks for now. # Zone NAME GMTOFF RULES FORMAT [UNTIL] =================================================================== RCS file: RCS/etcetera,v retrieving revision 1994.4 retrieving revision 1994.4.1.1 diff -c -r1994.4 -r1994.4.1.1 *** etcetera 1994/02/24 20:42:21 1994.4 --- etcetera 1996/03/12 03:16:42 1994.4.1.1 *************** *** 27,33 **** # zic -l GMT-12 # so we moved the names into the Etc subdirectory. ! Zone Etc/GMT-13 13 - GMT-13 # 12 hours ahead of GMT, plus DST Zone Etc/GMT-12 12 - GMT-12 Zone Etc/GMT-11 11 - GMT-11 Zone Etc/GMT-10 10 - GMT-10 --- 27,34 ---- # zic -l GMT-12 # so we moved the names into the Etc subdirectory. ! Zone Etc/GMT-14 14 - GMT-14 # 14 hours ahead of GMT ! Zone Etc/GMT-13 13 - GMT-13 Zone Etc/GMT-12 12 - GMT-12 Zone Etc/GMT-11 11 - GMT-11 Zone Etc/GMT-10 10 - GMT-10 =================================================================== RCS file: RCS/solar89,v retrieving revision 1996.2 retrieving revision 1996.2.1.1 diff -c -r1996.2 -r1996.2.1.1 *** solar89 1995/10/30 15:07:12 1996.2 --- solar89 1996/03/12 03:16:42 1996.2.1.1 *************** *** 390,393 **** 3:07:04 sol89 ?? 1990 3:07:04 - ?? # For backward compatibility... ! Link Asia/Riyadh88 Mideast/Riyadh88 --- 390,393 ---- 3:07:04 sol89 ?? 1990 3:07:04 - ?? # For backward compatibility... ! Link Asia/Riyadh89 Mideast/Riyadh89

1 0

Time Bomb Still Ticking For Year 2000
by Monty Solomon March 1, 1996

March 1, 1996

Excerpt from Edupage, 15 February 1996 TIME BOMB STILL TICKING FOR YEAR 2000 The Gartner Group predicts that half of all companies affected by the year 2000 date field problem will still be unprepared when the fateful day arrives. "A lot of companies are like deer frozen in the headlights of a big truck coming right at them," says a Gartner analyst. Some industry experts estimate the cost of fixing the problem at $40 million per large corporation, with the global price tag pegged at $400 billion to $600 billion. Many corporations are wondering if their old systems are worth all the trouble: "Do we just fix the millennium bug, or should we take this as an opportunity to put in some new systems?" asks one CIO. (Information Week 5 Feb 96 p30)

1 0

Portugal to change timezone again (??)
by peter＠aldie.co.uk Feb. 29, 1996

Feb. 29, 1996

The following article appeared in uk.misc and a few other groups as part of a discussion of time matters. (It is surprising how much confusion there still is about the EU start/end dates.) >From: martin(a)ci.ua.pt (Martin Bruckmann) >Newsgroups: uk.misc,soc.culture.french,rec.travel.air,rec.travel.misc, > rec.travel.europe >Subject: Re: Time difference between Paris & London after clocks change >Date: 27 Feb 1996 15:35:53 GMT >Organization: Universidade de Aveiro, Portugal >Message-ID: <4gv8cp$atu(a)titan.ci.ua.pt> > >... >And, a novelty (or back to how it was before E.U.) : Portugal is >re-adopting its more natural time, i.e. the same as in the U.K. It is >still 1h delay compared to its solar time, but it's already better than >the E.U. imposed Central European Time, which forced little kids to >wake up very early to go to school in the total darkness in Winter... > >-- >All about Portuguese Internet Sites & People on http://sapo.ua.pt I asked for some more info by email and this was the result, quoted with permission: >From: Martin Bruckmann <martin(a)ua.pt> >To: peter(a)aldie.co.uk (Peter Ilieve) >Date: Thu, 29 Feb 1996 10:09:31 +0000 (PWT) > >> [my query] >> Is this true? It was only a few years ago that you changed from GMT to CET. >> I would like to pass this information along to Arthur Olson's timezone >> mailing list. Do you have a reference to the law or whatever it is that >> says you will change back? When will you change? > >Yes, it is true. Actually, we just stayed with "Winter time" instead of >changing to Summer time this year. This has been decided by the new Prime >Minister and his government a month ago or so. Exact legal references ? I >don't have any, nor would I know where to look for them. >What is this "Arthur Olson's timezone mailing list" ? After a bit more clarification it seems that the new Portugese Prime Minister has got fed up with getting up in the dark in the Winter and has decided they won't put the clocks forward this March. I assume they will then put them back in October, reverting to GMT as they were before 1992. I am not an expert on Portugal and although Martin will undoubtedly know more than me I don't think he would consider himself a timezone expert, so best treat this as a warning for now. Does anyone know more? Peter Ilieve peter(a)aldie.co.uk

1 0

tzcode96e.tar.gz
by ado Feb. 29, 1996

Feb. 29, 1996

The file tzcode96e.tar.gz is now available for anonymous ftp from elsie.nci.nih.gov (165.112.73.1) in the ~ftp/pub directory: -rw-r--r-- 1 ado 74253 Feb 29 10:41 tzcode96e.tar.gz This include internationalization changes and the fix to the documentation for strftime. --ado

1 0

newstrftime.3 patch
by J.T. Conklin Feb. 29, 1996

Feb. 29, 1996

A NetBSD user discovered that the descriptions of the %x and %X format specifiers are swapped. --jtc Index: newstrftime.3 =================================================================== RCS file: /home/cvsroot/tz/tzcode/newstrftime.3,v retrieving revision 1.1.1.1 diff -c -r1.1.1.1 newstrftime.3 *** newstrftime.3 1996/01/15 21:48:02 1.1.1.1 --- newstrftime.3 1996/02/29 15:04:55 *************** *** 192,201 **** as a decimal number (0-6). .TP %X ! is replaced by the locale's appropriate date representation. .TP %x ! is replaced by the locale's appropriate time representation. .TP %Y is replaced by the year with century as a decimal number. --- 192,201 ---- as a decimal number (0-6). .TP %X ! is replaced by the locale's appropriate time representation. .TP %x ! is replaced by the locale's appropriate date representation. .TP %Y is replaced by the year with century as a decimal number.

1 0

Timezone names for Greenland
by ohnielse＠fysik.dtu.dk Feb. 27, 1996

Feb. 27, 1996

Dear Sir, In your timezone data file for Europe you define: # -3:00 WGT+DST Western Greenland* # -2:00 MGT+DST Middle Greenland* # -1:00 EGT+DST Eastern Greenland* Zone America/Scoresbysund -1:29:00 - LMT 1916 Jul 28 -2:00 - MGT 1980 Apr 6 2:00 -2:00 M-Eur MGT%s 1981 Mar 29 -1:00 M-Eur EGT%s Zone America/Godthab -3:26:56 - LMT 1916 Jul 28 -3:00 - WGT 1980 Apr 6 2:00 -3:00 M-Eur WGT%s Zone America/Thule -4:35:08 - LMT 1916 Jul 28 -4:00 - AST The naming "America" is politically incorrect. Greenland is a part of Denmark, but has had home rule for 10 or 11 years. If anything, you would have to define the zone names as "Denmark/...." or perhaps "Greenland/....". Furthermore, the town names that you give are Danish names. Greenlanders would undoubtedly insist that the Greenland names be used, so perhaps better namings would be: America/Scoresbysund -> Greenland/Ittoqqortoormiit America/Godthab -> Greenland/Nuuk America/Thule -> Greenland/Dundas I don't know if the timezone names have been fixed by international convention, but the naming would appear controversial :-) With best regards, Dr. Ole H. Nielsen Technical University of Denmark % Ole Holm Nielsen % UNI-C, Building 304 % Technical University of Denmark, DK-2800 Lyngby, Denmark % E-mail: Ole.H.Nielsen(a)uni-c.dk or Ole.H.Nielsen(a)fysik.dtu.dk % WWW URL: http://www.fysik.dtu.dk/persons/ohnielse.html % Telephone: (+45) 35 87 89 65

2 1

Bug in localtime.c
by Gibson_Roger/HEMEL_ENGINEERING_SERVERS＠uniplex.co.uk Feb. 23, 1996

Feb. 23, 1996

From: Roger Gibson (rxg(a)uniplex.co.uk) I have found a bug in 'localtime.c' for which I am sending a proposed fix. The problem occurs when 'mktime' is called with tm_mday <= 0, tm_mon > 1 and tm_year is a leap year or the year after a leap year. For such a case, the normalization is performed incorrectly, and yields a date which is one day out. The following example program demonstrates this: -------------------------------- #include <stdio.h> #include <time.h> int main() { struct tm tm; char buf[256]; memset(&tm, 0, sizeof tm); tm.tm_year = 96; /* 1996 */ tm.tm_mon = 2; /* March */ tm.tm_mday = 0; tm.tm_hour = 12; mktime(&tm); /* Should give Feb 29 */ strftime(buf, sizeof buf, "%C", &tm); printf("%s\n", buf); return 0; } -------------------------------- To deal with cases like this, the 'time2' function decrements the year, and adds the corresponding number of days to 'tm_mday'. The problem is that the number of days is taken to be that in the year after it has been decremented, whereas it is actually that in the year which contains the end of February which is spanned by the interval. It is necessary to take 'tm_mon' into account in determining this. The same problem occurs in the part of the code which deals with values of 'tm_mday' which exceed the number of days in a year. The version of 'time2' below contains my attempt to fix these problems. It seems to work, but I'd like to know if you concur. -------------------------------- Version information: static char elsieid[] = "@(#)localtime.c 7.44"; (Still present in version 7.55.) -------------------------------- static time_t time2(tmp, funcp, offset, okayp) struct tm * const tmp; void (* const funcp) P((const time_t*, long, struct tm*)); const long offset; int * const okayp; { register const struct state * sp; register int dir; register int bits; register int i, j ; register int saved_seconds; time_t newt; time_t t; struct tm yourtm, mytm; *okayp = FALSE; yourtm = *tmp; if (normalize_overflow(&yourtm.tm_hour, &yourtm.tm_min, MINSPERHOUR)) return WRONG; if (normalize_overflow(&yourtm.tm_mday, &yourtm.tm_hour, HOURSPERDAY)) return WRONG; if (normalize_overflow(&yourtm.tm_year, &yourtm.tm_mon, MONSPERYEAR)) return WRONG; /* ** Turn yourtm.tm_year 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)) return WRONG; while (yourtm.tm_mday <= 0) { if (increment_overflow(&yourtm.tm_year, -1)) return WRONG; if (yourtm.tm_mon >= 2) yourtm.tm_mday += year_lengths[isleap(yourtm.tm_year + 1)]; else yourtm.tm_mday += year_lengths[isleap(yourtm.tm_year)]; } while (yourtm.tm_mday > DAYSPERLYEAR) { if (yourtm.tm_mon >= 2) yourtm.tm_mday -= year_lengths[isleap(yourtm.tm_year + 1)]; else yourtm.tm_mday -= year_lengths[isleap(yourtm.tm_year)]; if (increment_overflow(&yourtm.tm_year, 1)) return WRONG; } for ( ; ; ) { i = mon_lengths[isleap(yourtm.tm_year)][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)) return WRONG; } } if (increment_overflow(&yourtm.tm_year, -TM_YEAR_BASE)) return WRONG; if (yourtm.tm_year + TM_YEAR_BASE < EPOCH_YEAR) { /* ** We can't set tm_sec to 0, because that might push the ** time below the minimum representable time. ** Set tm_sec to 59 instead. ** This assumes that the minimum representable time is ** not in the same minute that a leap second was deleted from, ** which is a safer assumption than using 58 would be. */ if (increment_overflow(&yourtm.tm_sec, 1 - SECSPERMIN)) return WRONG; saved_seconds = yourtm.tm_sec; yourtm.tm_sec = SECSPERMIN - 1; } else { saved_seconds = yourtm.tm_sec; yourtm.tm_sec = 0; } /* ** Calculate the number of magnitude bits in a time_t ** (this works regardless of whether time_t is ** signed or unsigned, though lint complains if unsigned). */ for (bits = 0, t = 1; t > 0; ++bits, t <<= 1) continue; /* ** If time_t is signed, then 0 is the median value, ** if time_t is unsigned, then 1 << bits is median. */ t = (t < 0) ? 0 : ((time_t) 1 << bits); for ( ; ; ) { (*funcp)(&t, offset, &mytm); dir = tmcomp(&mytm, &yourtm); if (dir != 0) { if (bits-- < 0) return WRONG; if (bits < 0) --t; else if (dir > 0) t -= (time_t) 1 << bits; else t += (time_t) 1 << bits; continue; } if (yourtm.tm_isdst < 0 || mytm.tm_isdst == yourtm.tm_isdst) break; /* ** Right time, wrong type. ** Hunt for right time, right type. ** It's okay to guess wrong since the guess ** gets checked. */ /* ** The (void *) casts are the benefit of SunOS 3.3 on Sun 2's. */ sp = (const struct state *) (((void *) funcp == (void *) localsub) ? lclptr : gmtptr); #ifdef ALL_STATE if (sp == NULL) return WRONG; #endif /* defined ALL_STATE */ for (i = 0; i < sp->typecnt; ++i) { if (sp->ttis[i].tt_isdst != yourtm.tm_isdst) continue; for (j = 0; j < sp->typecnt; ++j) { if (sp->ttis[j].tt_isdst == yourtm.tm_isdst) continue; newt = t + sp->ttis[j].tt_gmtoff - sp->ttis[i].tt_gmtoff; (*funcp)(&newt, offset, &mytm); if (tmcomp(&mytm, &yourtm) != 0) continue; if (mytm.tm_isdst != yourtm.tm_isdst) continue; /* ** We have a match. */ t = newt; goto label; } } return WRONG; } label: newt = t + saved_seconds; if ((newt < t) != (saved_seconds < 0)) return WRONG; t = newt; (*funcp)(&t, offset, tmp); *okayp = TRUE; return t; }

1 0

Timezones
by Ravi Narine Feb. 21, 1996

Feb. 21, 1996

Hi there, My name is Ravi Narine, and I am a software developer for Bell Sygma Inc. We are using the Timezone library that you have created. I am wondering is there a program that can take information (ex. city,country,province,state,area code or anything on that means) and return me the timezone that is in. Greatly appreciated Ravi Narine

1 0

None
by Roland McGrath Feb. 19, 1996

Feb. 19, 1996

In the GNU C library I use zic.c and zdump.c verbatim from tzcode96d, except for marking all the strings for user messages so they can be translated through message catalogs. Could these changes perhaps be merged so I don't have to re-modify each new version? It is quite simple and unobtrusive: put in private.h or someplace: #ifndef _ #define _(x) x #endif (in my system, _ will be defined as a macro that invokes the translation lookup). Here are the patches to zic.c and zdump.c to make their "messages" use _("messages"). Can I hope to get this merged into the tzcode distribution? Thanks, Roland --- /home/fsf/roland/libc/time/=ado/96d/zic.c Mon Feb 12 11:01:33 1996 +++ /src/libc/time/zic.c Mon Feb 19 18:31:47 1996 @@ -396,10 +396,10 @@ const char * const string; ** zic ... 2>&1 | error -t "*" -v ** on BSD systems. */ - (void) fprintf(stderr, "\"%s\", line %d: %s", + (void) fprintf(stderr, _("\"%s\", line %d: %s"), filename, linenum, string); if (rfilename != NULL) - (void) fprintf(stderr, " (rule from \"%s\", line %d)", + (void) fprintf(stderr, _(" (rule from \"%s\", line %d)"), rfilename, rlinenum); (void) fprintf(stderr, "\n"); ++errors; @@ -408,9 +408,9 @@ const char * const string; static void usage P((void)) { - (void) fprintf(stderr, "%s: usage is %s \ + (void) fprintf(stderr, _("%s: usage is %s \ [ -s ] [ -v ] [ -l localtime ] [ -p posixrules ] [ -d directory ]\n\ -\t[ -L leapseconds ] [ -y yearistype ] [ filename ... ]\n", +\t[ -L leapseconds ] [ -y yearistype ] [ filename ... ]\n"), progname, progname); (void) exit(EXIT_FAILURE); } @@ -444,7 +444,7 @@ char * argv[]; directory = optarg; else { (void) fprintf(stderr, -"%s: More than one -d option specified\n", +_("%s: More than one -d option specified\n"), progname); (void) exit(EXIT_FAILURE); } @@ -454,7 +454,7 @@ char * argv[]; lcltime = optarg; else { (void) fprintf(stderr, -"%s: More than one -l option specified\n", +_("%s: More than one -l option specified\n"), progname); (void) exit(EXIT_FAILURE); } @@ -464,7 +464,7 @@ char * argv[]; psxrules = optarg; else { (void) fprintf(stderr, -"%s: More than one -p option specified\n", +_("%s: More than one -p option specified\n"), progname); (void) exit(EXIT_FAILURE); } @@ -474,7 +474,7 @@ char * argv[]; yitcommand = optarg; else { (void) fprintf(stderr, -"%s: More than one -y option specified\n", +_("%s: More than one -y option specified\n"), progname); (void) exit(EXIT_FAILURE); } @@ -484,7 +484,7 @@ char * argv[]; leapsec = optarg; else { (void) fprintf(stderr, -"%s: More than one -L option specified\n", +_("%s: More than one -L option specified\n"), progname); (void) exit(EXIT_FAILURE); } @@ -567,7 +567,7 @@ const char * const tofile; if (mkdirs(toname) != 0) (void) exit(EXIT_FAILURE); if (link(fromname, toname) != 0) { - (void) fprintf(stderr, "%s: Can't link from %s to ", + (void) fprintf(stderr, _("%s: Can't link from %s to "), progname, fromname); (void) perror(toname); (void) exit(EXIT_FAILURE); @@ -688,7 +688,7 @@ associate P((void)) ** a '%s' in the format is a bad thing. */ if (strchr(zp->z_format, '%') != 0) - error("%s in ruleless zone"); + error(_("%s in ruleless zone")); } } if (errors) @@ -709,10 +709,10 @@ const char * name; char buf[BUFSIZ]; if (strcmp(name, "-") == 0) { - name = "standard input"; + name = _("standard input"); fp = stdin; } else if ((fp = fopen(name, "r")) == NULL) { - (void) fprintf(stderr, "%s: Can't open ", progname); + (void) fprintf(stderr, _("%s: Can't open "), progname); (void) perror(name); (void) exit(EXIT_FAILURE); } @@ -723,7 +723,7 @@ const char * name; break; cp = strchr(buf, '\n'); if (cp == NULL) { - error("line too long"); + error(_("line too long")); (void) exit(EXIT_FAILURE); } *cp = '\0'; @@ -743,7 +743,7 @@ const char * name; } else { lp = byword(fields[0], line_codes); if (lp == NULL) - error("input line of unknown type"); + error(_("input line of unknown type")); else switch ((int) (lp->l_value)) { case LC_RULE: inrule(fields, nfields); @@ -759,14 +759,14 @@ const char * name; case LC_LEAP: if (name != leapsec) (void) fprintf(stderr, -"%s: Leap line in non leap seconds file %s\n", +_("%s: Leap line in non leap seconds file %s\n"), progname, name); else inleap(fields, nfields); wantcont = FALSE; break; default: /* "cannot happen" */ (void) fprintf(stderr, -"%s: panic: Invalid l_value %d\n", +_("%s: panic: Invalid l_value %d\n"), progname, lp->l_value); (void) exit(EXIT_FAILURE); } @@ -774,17 +774,17 @@ const char * name; ifree((char *) fields); } if (ferror(fp)) { - (void) fprintf(stderr, "%s: Error reading ", progname); + (void) fprintf(stderr, _("%s: Error reading "), progname); (void) perror(filename); (void) exit(EXIT_FAILURE); } if (fp != stdin && fclose(fp)) { - (void) fprintf(stderr, "%s: Error closing ", progname); + (void) fprintf(stderr, _("%s: Error closing "), progname); (void) perror(filename); (void) exit(EXIT_FAILURE); } if (wantcont) - error("expected continuation line not found"); + error(_("expected continuation line not found")); } /* @@ -839,16 +839,16 @@ const int nfields; static struct rule r; if (nfields != RULE_FIELDS) { - error("wrong number of fields on Rule line"); + error(_("wrong number of fields on Rule line")); return; } if (*fields[RF_NAME] == '\0') { - error("nameless rule"); + error(_("nameless rule")); return; } r.r_filename = filename; r.r_linenum = linenum; - r.r_stdoff = gethms(fields[RF_STDOFF], "invalid saved time", TRUE); + r.r_stdoff = gethms(fields[RF_STDOFF], _("invalid saved time"), TRUE); rulesub(&r, fields[RF_LOYEAR], fields[RF_HIYEAR], fields[RF_COMMAND], fields[RF_MONTH], fields[RF_DAY], fields[RF_TOD]); r.r_name = ecpyalloc(fields[RF_NAME]); @@ -867,13 +867,13 @@ const int nfields; static char * buf; if (nfields < ZONE_MINFIELDS || nfields > ZONE_MAXFIELDS) { - error("wrong number of fields on Zone line"); + error(_("wrong number of fields on Zone line")); return FALSE; } if (strcmp(fields[ZF_NAME], TZDEFAULT) == 0 && lcltime != NULL) { buf = erealloc(buf, (int) (132 + strlen(TZDEFAULT))); (void) sprintf(buf, -"\"Zone %s\" line and -l option are mutually exclusive", +_("\"Zone %s\" line and -l option are mutually exclusive"), TZDEFAULT); error(buf); return FALSE; @@ -881,7 +881,7 @@ const int nfields; if (strcmp(fields[ZF_NAME], TZDEFRULES) == 0 && psxrules != NULL) { buf = erealloc(buf, (int) (132 + strlen(TZDEFRULES))); (void) sprintf(buf, -"\"Zone %s\" line and -p option are mutually exclusive", +_("\"Zone %s\" line and -p option are mutually exclusive"), TZDEFRULES); error(buf); return FALSE; @@ -893,7 +893,7 @@ const int nfields; strlen(fields[ZF_NAME]) + strlen(zones[i].z_filename))); (void) sprintf(buf, -"duplicate zone name %s (file \"%s\", line %d)", +_("duplicate zone name %s (file \"%s\", line %d)"), fields[ZF_NAME], zones[i].z_filename, zones[i].z_linenum); @@ -909,7 +909,7 @@ register char ** const fields; const int nfields; { if (nfields < ZONEC_MINFIELDS || nfields > ZONEC_MAXFIELDS) { - error("wrong number of fields on Zone continuation line"); + error(_("wrong number of fields on Zone continuation line")); return FALSE; } return inzsub(fields, nfields, TRUE); @@ -949,10 +949,10 @@ const int iscont; } z.z_filename = filename; z.z_linenum = linenum; - z.z_gmtoff = gethms(fields[i_gmtoff], "invalid GMT offset", TRUE); + z.z_gmtoff = gethms(fields[i_gmtoff], _("invalid GMT offset"), TRUE); if ((cp = strchr(fields[i_format], '%')) != 0) { if (*++cp != 's' || strchr(cp, '%') != 0) { - error("invalid abbreviation format"); + error(_("invalid abbreviation format")); return FALSE; } } @@ -978,8 +978,8 @@ const int iscont; zones[nzones - 1].z_untiltime > min_time && zones[nzones - 1].z_untiltime < max_time && zones[nzones - 1].z_untiltime >= z.z_untiltime) { - error("Zone continuation line end time is not \ -after end time of previous line"); + error(_("Zone continuation line end time is \ +not after end time of previous line")); return FALSE; } } @@ -1006,7 +1006,7 @@ const int nfields; time_t t; if (nfields != LEAP_FIELDS) { - error("wrong number of fields on Leap line"); + error(_("wrong number of fields on Leap line")); return; } dayoff = 0; @@ -1015,7 +1015,7 @@ const int nfields; /* * Leapin' Lizards! */ - error("invalid leaping year"); + error(_("invalid leaping year")); return; } j = EPOCH_YEAR; @@ -1030,7 +1030,7 @@ const int nfields; dayoff = oadd(dayoff, eitol(i)); } if ((lp = byword(fields[LP_MONTH], mon_names)) == NULL) { - error("invalid month name"); + error(_("invalid month name")); return; } month = lp->l_value; @@ -1043,7 +1043,7 @@ const int nfields; cp = fields[LP_DAY]; if (sscanf(cp, scheck(cp, "%d"), &day) != 1 || day <= 0 || day > len_months[isleap(year)][month]) { - error("invalid day of month"); + error(_("invalid day of month")); return; } dayoff = oadd(dayoff, eitol(day - 1)); @@ -1056,7 +1056,7 @@ const int nfields; ** Cheap overflow check. */ if (t / SECSPERDAY != dayoff) { - error("time overflow"); + error(_("time overflow")); return; } tod = gethms(fields[LP_TIME], "invalid time of day", FALSE); @@ -1078,11 +1078,11 @@ const int nfields; positive = TRUE; count = 2; } else { - error("illegal CORRECTION field on Leap line"); + error(_("illegal CORRECTION field on Leap line")); return; } if ((lp = byword(fields[LP_ROLL], leap_types)) == NULL) { - error("illegal Rolling/Stationary field on Leap line"); + error(_("illegal Rolling/Stationary field on Leap line")); return; } leapadd(tadd(t, tod), positive, lp->l_value, count); @@ -1097,15 +1097,15 @@ const int nfields; struct link l; if (nfields != LINK_FIELDS) { - error("wrong number of fields on Link line"); + error(_("wrong number of fields on Link line")); return; } if (*fields[LF_FROM] == '\0') { - error("blank FROM field on Link line"); + error(_("blank FROM field on Link line")); return; } if (*fields[LF_TO] == '\0') { - error("blank TO field on Link line"); + error(_("blank TO field on Link line")); return; } l.l_filename = filename; @@ -1133,7 +1133,7 @@ const char * const timep; register char * ep; if ((lp = byword(monthp, mon_names)) == NULL) { - error("invalid month name"); + error(_("invalid month name")); return; } rp->r_month = lp->l_value; @@ -1177,11 +1177,11 @@ const char * const timep; break; default: /* "cannot happen" */ (void) fprintf(stderr, - "%s: panic: Invalid l_value %d\n", + _("%s: panic: Invalid l_value %d\n"), progname, lp->l_value); (void) exit(EXIT_FAILURE); } else if (sscanf(cp, scheck(cp, "%d"), &rp->r_loyear) != 1) { - error("invalid starting year"); + error(_("invalid starting year")); return; } cp = hiyearp; @@ -1197,22 +1197,22 @@ const char * const timep; break; default: /* "cannot happen" */ (void) fprintf(stderr, - "%s: panic: Invalid l_value %d\n", + _("%s: panic: Invalid l_value %d\n"), progname, lp->l_value); (void) exit(EXIT_FAILURE); } else if (sscanf(cp, scheck(cp, "%d"), &rp->r_hiyear) != 1) { - error("invalid ending year"); + error(_("invalid ending year")); return; } if (rp->r_loyear > rp->r_hiyear) { - error("starting year greater than ending year"); + error(_("starting year greater than ending year")); return; } if (*typep == '\0') rp->r_yrtype = NULL; else { if (rp->r_loyear == rp->r_hiyear) { - error("typed single year"); + error(_("typed single year")); return; } rp->r_yrtype = ecpyalloc(typep); @@ -1242,12 +1242,12 @@ const char * const timep; if (rp->r_dycode != DC_DOM) { *ep++ = 0; if (*ep++ != '=') { - error("invalid day of month"); + error(_("invalid day of month")); ifree(dp); return; } if ((lp = byword(dp, wday_names)) == NULL) { - error("invalid weekday name"); + error(_("invalid weekday name")); ifree(dp); return; } @@ -1256,7 +1256,7 @@ const char * const timep; if (sscanf(ep, scheck(ep, "%d"), &rp->r_dayofmonth) != 1 || rp->r_dayofmonth <= 0 || (rp->r_dayofmonth > len_months[1][rp->r_month])) { - error("invalid day of month"); + error(_("invalid day of month")); ifree(dp); return; } @@ -1303,7 +1303,8 @@ const char * const name; if (mkdirs(fullname) != 0) (void) exit(EXIT_FAILURE); if ((fp = fopen(fullname, "wb")) == NULL) { - (void) fprintf(stderr, "%s: Can't create ", progname); + (void) fprintf(stderr, _("%s: Can't create "), + progname); (void) perror(fullname); (void) exit(EXIT_FAILURE); } @@ -1368,7 +1369,7 @@ const char * const name; for (i = 0; i < typecnt; ++i) (void) putc(ttisgmts[i], fp); if (ferror(fp) || fclose(fp)) { - (void) fprintf(stderr, "%s: Write error on ", progname); + (void) fprintf(stderr, _("%s: Write error on "), progname); (void) perror(fullname); (void) exit(EXIT_FAILURE); } @@ -1586,7 +1587,7 @@ const int type; if (timecnt == 0 && type == 0 && isdsts[0] == 0) return; /* handled by default rule */ if (timecnt >= TZ_MAX_TIMES) { - error("too many transitions?!"); + error(_("too many transitions?!")); (void) exit(EXIT_FAILURE); } ats[timecnt] = starttime; @@ -1620,7 +1621,7 @@ const int ttisgmt; ** many. */ if (typecnt >= TZ_MAX_TYPES) { - error("too many local time types"); + error(_("too many local time types")); (void) exit(EXIT_FAILURE); } gmtoffs[i] = gmtoff; @@ -1648,13 +1649,13 @@ int count; register int i, j; if (leapcnt + (positive ? count : 1) > TZ_MAX_LEAPS) { - error("too many leap seconds"); + error(_("too many leap seconds")); (void) exit(EXIT_FAILURE); } for (i = 0; i < leapcnt; ++i) if (t <= trans[i]) { if (t == trans[i]) { - error("repeated leap second moment"); + error(_("repeated leap second moment")); (void) exit(EXIT_FAILURE); } break; @@ -1704,8 +1705,8 @@ const char * const type; return TRUE; if (result == (1 << 8)) return FALSE; - error("Wild result from command execution"); - (void) fprintf(stderr, "%s: command was '%s', result was %d\n", + error(_("Wild result from command execution")); + (void) fprintf(stderr, _("%s: command was '%s', result was %d\n"), progname, buf, result); for ( ; ; ) (void) exit(EXIT_FAILURE); @@ -1799,7 +1800,7 @@ register char * cp; else while ((*dp = *cp++) != '"') if (*dp != '\0') ++dp; - else error("Odd number of quotation marks"); + else error(_("Odd number of quotation marks")); } while (*cp != '\0' && *cp != '#' && (!isascii(*cp) || !isspace((unsigned char) *cp))); if (isascii(*cp) && isspace((unsigned char) *cp)) @@ -1819,7 +1820,7 @@ const long t2; t = t1 + t2; if ((t2 > 0 && t <= t1) || (t2 < 0 && t >= t1)) { - error("time overflow"); + error(_("time overflow")); (void) exit(EXIT_FAILURE); } return t; @@ -1838,7 +1839,7 @@ const long t2; return min_time; t = t1 + t2; if ((t2 > 0 && t <= t1) || (t2 < 0 && t >= t1)) { - error("time overflow"); + error(_("time overflow")); (void) exit(EXIT_FAILURE); } return t; @@ -1885,7 +1886,7 @@ register const int wantedy; if (rp->r_dycode == DC_DOWLEQ) --i; else { - error("use of 2/29 in non leap-year"); + error(_("use of 2/29 in non leap-year")); (void) exit(EXIT_FAILURE); } } @@ -1919,7 +1920,7 @@ register const int wantedy; --i; } if (i < 0 || i >= len_months[isleap(y)][m]) { - error("no day in month matches rule"); + error(_("no day in month matches rule")); (void) exit(EXIT_FAILURE); } } @@ -1942,7 +1943,7 @@ const char * const string; i = strlen(string) + 1; if (charcnt + i > TZ_MAX_CHARS) { - error("too many, or too long, time zone abbreviations"); + error(_("too many, or too long, time zone abbreviations")); (void) exit(EXIT_FAILURE); } (void) strcpy(&chars[charcnt], string); @@ -1977,7 +1978,7 @@ char * const argname; */ if (mkdir(name, 0755) != 0) { (void) fprintf(stderr, - "%s: Can't create directory ", + _("%s: Can't create directory "), progname); (void) perror(name); ifree(name); @@ -1999,7 +2000,7 @@ const int i; l = i; if ((i < 0 && l >= 0) || (i == 0 && l != 0) || (i > 0 && l <= 0)) { (void) fprintf(stderr, - "%s: %d did not sign extend correctly\n", + _("%s: %d did not sign extend correctly\n"), progname, i); (void) exit(EXIT_FAILURE); } --- /home/fsf/roland/libc/time/=ado/96d/zdump.c Mon Feb 12 11:01:55 1996 +++ /src/libc/time/zdump.c Tue Feb 13 02:45:50 1996 @@ -132,7 +132,7 @@ char * argv[]; if (c != EOF || (optind == argc - 1 && strcmp(argv[optind], "=") == 0)) { (void) fprintf(stderr, -"%s: usage is %s [ -v ] [ -c cutoff ] zonename ...\n", +_("%s: usage is %s [ -v ] [ -c cutoff ] zonename ...\n"), argv[0], argv[0]); (void) exit(EXIT_FAILURE); } @@ -224,9 +224,9 @@ char * argv[]; show(argv[i], t, TRUE);

1 0

Re: mktime() failure vs 31-DEC-1969 23:59:59 GMT
by tomp＠zk3.dec.com Feb. 19, 1996

Feb. 19, 1996

Hi Valery, > My man page on mktime() (Solaris2) clearly says that 'tm_year > of the tm structure must be for year 1970 or later.' This would seem to limit the usefulness of timezone rules prior to 1970 with respect to mktime(). Perhaps they did this for standards compliance? thanks, - Tom

2 3

Re: mktime() failure vs 31-DEC-1969 23:59:59 GMT
by Bradley White Feb. 16, 1996

Feb. 16, 1996

>> Call localtime() on the seemingly-in-error mktime() result, and compare >> the relevant portions of the structs tm. A match implies a non-error. > This doesn't sound very stable, but perhaps there is no other alternative. I'm not sure of what you mean by stable. This is, after all, exactly what mktime() does to determine is has the matching time_t. > Of course, we only need to check this > when the return of mktime() is -1 Exactly what I meant by "Call localtime() on the seemingly-in-error mktime() result."

2 1