- tz - lists.iana.org

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