- tz - lists.iana.org

EST is an hour out in zoneinfo
by Chris Bruner Quadro June 21, 2006

June 21, 2006

EST should give the same results as Canada/Eastern With timezone data of 2006g and glibc of 2.4-r3 created with gentoo Linux. https://bugs.gentoo.org/show_bug.cgi?id=137459 Let me know if you need any additional information. Chris Bruner Compulife Software Inc. ==============================================

1 0

FW: zic leapseconds bug
by Olson, Arthur David (NIH/NCI) [E] June 13, 2006

June 13, 2006

David Robinson is not on the time zone mailing list; direct replies appropriately. --ado -----Original Message----- From: David Robinson [mailto:drtr@apache.org] Sent: Tuesday, June 13, 2006 4:04 PM To: Olson, Arthur David (NIH/NCI) [E] Cc: tz(a)lecserver.nci.nih.gov; bww(a)acm.org; eggert(a)twinsun.com Subject: RE: zic leapseconds bug I think I chose an example that was rather confusing - mixing the timezone and leapsecond transitions; apologies for that. Try this instead. (start with /usr/local/etc/zoneinfo having the leapsecond version of the tzfiles) % cat <<FOO > xlondon # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone Europe/London-Test 0:00 EU GMT/BST Rule EU 1995 only - May 21 2:00s 1:00 BST Rule EU 1995 only - Oct 1 2:00s 0 GMT # Note this line; 2 seconds after midnigt Rule EU 1996 only - Jan 1 0:00:02z 1:00 BST Rule EU 1996 only - Jan 2 2:00z 0 GMT FOO % zic -L leapseconds -y yearistype.sh -d /usr/local/etc/zoneinfo zlondon % ./zdump.exe -v Europe/London-Test | fgrep 1996 Europe/London-Test Mon Jan 1 00:00:00 1996 UTC = Mon Jan 1 00:00:00 1996 GMT isdst=0 Europe/London-Test Mon Jan 1 00:00:00 1996 UTC = Mon Jan 1 00:00:00 1996 GMT isdst=0 Europe/London-Test Mon Jan 1 00:00:01 1996 UTC = Mon Jan 1 01:00:01 1996 BST isdst=1 Europe/London-Test Tue Jan 2 01:59:59 1996 UTC = Tue Jan 2 02:59:59 1996 BST isdst=1 Europe/London-Test Tue Jan 2 02:00:00 1996 UTC = Tue Jan 2 02:00:00 1996 GMT isdst=0 See that the BST transition is now 1 second after midnight. David On Tue, 13 Jun 2006, Olson, Arthur David (NIH/NCI) [E] wrote: >> You can't see the problem with zdump, as the transition is not >> visible > to it. But in Europe/London, there was a transition on 1996-01-01 > 00:00:00. >> Here's my dump of the zoneinfo file for the transition & one row > either side of it... > > Here's what I get at elsie; note that the 1995/1996 leap second > transition does seem to be visible to zdump. > > --ado > > Script started on Tue 13 Jun 2006 01:12:41 PM EDT elsie$ cd ~/src/tz > elsie$ make install TOPDIR=$PWD/tmp REDO=right_only >&- elsie$ ./zdump > -v Europe/London | grep 199[56] Europe/London Sun Mar 26 00:59:59 > 1995 UTC = Sun Mar 26 00:59:59 1995 GMT isdst=0 Europe/London Sun Mar > 26 01:00:00 1995 UTC = Sun Mar 26 02:00:00 1995 BST isdst=1 > Europe/London Sun Oct 22 00:59:59 1995 UTC = Sun Oct 22 01:59:59 1995 > BST isdst=1 Europe/London Sun Oct 22 01:00:00 1995 UTC = Sun Oct 22 > 01:00:00 1995 GMT isdst=0 Europe/London Sun Dec 31 23:59:60 1995 UTC > = Sun Dec 31 23:59:60 1995 GMT isdst=0 Europe/London Mon Jan 1 > 00:00:00 1996 UTC = Mon Jan 1 00:00:00 1996 GMT isdst=0 Europe/London > Sun Mar 31 00:59:59 1996 UTC = Sun Mar 31 00:59:59 1996 GMT isdst=0 > Europe/London Sun Mar 31 01:00:00 1996 UTC = Sun Mar 31 02:00:00 1996 > BST isdst=1 Europe/London Sun Oct 27 00:59:59 1996 UTC = Sun Oct 27 > 01:59:59 1996 BST isdst=1 Europe/London Sun Oct 27 01:00:00 1996 UTC > = Sun Oct 27 01:00:00 1996 GMT isdst=0 elsie$ exit > > script done on Tue 13 Jun 2006 01:13:11 PM EDT > > >

1 0

RE: zic leapseconds bug
by Olson, Arthur David (NIH/NCI) [E] June 13, 2006

June 13, 2006

> You can't see the problem with zdump, as the transition is not visible to it. But in Europe/London, there was a transition on 1996-01-01 00:00:00. > Here's my dump of the zoneinfo file for the transition & one row either side of it... Here's what I get at elsie; note that the 1995/1996 leap second transition does seem to be visible to zdump. --ado Script started on Tue 13 Jun 2006 01:12:41 PM EDT elsie$ cd ~/src/tz elsie$ make install TOPDIR=$PWD/tmp REDO=right_only >&- elsie$ ./zdump -v Europe/London | grep 199[56] Europe/London Sun Mar 26 00:59:59 1995 UTC = Sun Mar 26 00:59:59 1995 GMT isdst=0 Europe/London Sun Mar 26 01:00:00 1995 UTC = Sun Mar 26 02:00:00 1995 BST isdst=1 Europe/London Sun Oct 22 00:59:59 1995 UTC = Sun Oct 22 01:59:59 1995 BST isdst=1 Europe/London Sun Oct 22 01:00:00 1995 UTC = Sun Oct 22 01:00:00 1995 GMT isdst=0 Europe/London Sun Dec 31 23:59:60 1995 UTC = Sun Dec 31 23:59:60 1995 GMT isdst=0 Europe/London Mon Jan 1 00:00:00 1996 UTC = Mon Jan 1 00:00:00 1996 GMT isdst=0 Europe/London Sun Mar 31 00:59:59 1996 UTC = Sun Mar 31 00:59:59 1996 GMT isdst=0 Europe/London Sun Mar 31 01:00:00 1996 UTC = Sun Mar 31 02:00:00 1996 BST isdst=1 Europe/London Sun Oct 27 00:59:59 1996 UTC = Sun Oct 27 01:59:59 1996 BST isdst=1 Europe/London Sun Oct 27 01:00:00 1996 UTC = Sun Oct 27 01:00:00 1996 GMT isdst=0 elsie$ exit script done on Tue 13 Jun 2006 01:13:11 PM EDT

1 0

FW: zic leapseconds bug
by Olson, Arthur David (NIH/NCI) [E] June 13, 2006

June 13, 2006

David Robinson is not on the time zone mailing list; direct replies appropriately. --ado -----Original Message----- From: David Robinson [mailto:drtr@apache.org] Sent: Tuesday, June 13, 2006 1:07 PM To: Olson, Arthur David (NIH/NCI) [E] Cc: tz(a)lecserver.nci.nih.gov Subject: RE: zic leapseconds bug I should have mentioned that I'm writing my own code to read the compiled zoneinfo files, as specified in tzfile.5. You can't see the problem with zdump, as the transition is not visible to it. But in Europe/London, there was a transition on 1996-01-01 00:00:00. Here's my dump of the zoneinfo file for the transition & one row either side of it: tttime gmtoff isdst abbr issdt isgmt 814323619 +0000 0 GMT 1 1 (1995-10-22 01:00:00) 820454419 +0000 0 GMT 0 0 (should be 1996-01-01 00:00:00) 828234020 +3600 1 BST 1 1 (1996-03-31 01:00:00) leaptime leapcorr 773020818 19 (1994-06-30 23:59:60) 820454419 20 (1995-12-31 23:59:60) 867715220 21 (1997-06-30 23:59:60) The transition time is the same as the leap time for the transition resulting from the switch to the EU rules, i.e. it's one second too early. David On Tue, 13 Jun 2006, Olson, Arthur David (NIH/NCI) [E] wrote: > Here's what's happening with leap seconds now, as divined using zdump: > > Script started on Tue 13 Jun 2006 11:56:20 AM EDT > elsie$ cd ~/src/tz > elsie$ make install TOPDIR=$PWD/tmp REDO=right_only >&- > elsie$ ./zdump -v America/New_York | grep 1972 > America/New_York Sun Apr 30 06:59:59 1972 UTC = Sun Apr 30 > 01:59:59 1972 EST isdst=0 > America/New_York Sun Apr 30 07:00:00 1972 UTC = Sun Apr 30 03:00:00 > 1972 EDT isdst=1 > America/New_York Fri Jun 30 23:59:60 1972 UTC = Fri Jun 30 19:59:60 > 1972 EDT isdst=1 > America/New_York Sat Jul 1 00:00:00 1972 UTC = Fri Jun 30 20:00:00 > 1972 EDT isdst=1 > America/New_York Sun Oct 29 05:59:59 1972 UTC = Sun Oct 29 > 01:59:59 1972 EDT isdst=1 > America/New_York Sun Oct 29 06:00:00 1972 UTC = Sun Oct 29 01:00:00 > 1972 EST isdst=0 > America/New_York Sun Dec 31 23:59:60 1972 UTC = Sun Dec 31 18:59:60 > 1972 EST isdst=0 > America/New_York Mon Jan 1 00:00:00 1973 UTC = Sun Dec 31 19:00:00 > 1972 EST isdst=0 > elsie$ exit > > script done on Tue 13 Jun 2006 11:56:54 AM EDT > > What should be different? > > --ado

1 0

zic leapseconds bug
by David Robinson June 13, 2006

June 13, 2006

There appears to be a bug in zic when leapseconds are provided, for any time zone transition near a leapsecond. The bug is that after reading in the Leaps as Posix seconds in the trans[] array, adjleap() is called. This converts these to genuine UTC seconds. However, this array is then used to fixup transition times which are in posix seconds, by for (i = 0; i < timecnt; ++i) { j = leapcnt; while (--j >= 0) if (ats[i] >= trans[j]) { ats[i] = tadd(ats[i], corr[j]); break; } } ats[] is stored in posix seconds, but trans[] is in leap seconds. The effect is that for Europe/London, the transition on 1996-01-01 00:00:00 is actually set to be at 1995-12-31 23:59:60, i.e. on the leap second (820454419). Any transition in the range 1996-01-01 00:00:00 to 00:00:18 is similarly off by one second. The fix is obvious in principle (have a copy of trans[] that isn't leap-second adjusted, and use that in the code above), but not all the uses of trans[] in the code are obvious to me (what are Rolling leapsecs?). I'd also suggest that zic reject any transition in its input with a second=60. It's usually run without a leap file so there's no point in implementing the correct behaviour for that case. Many thanks, David Robinson

2 1

FW: zic leapseconds bug
by Olson, Arthur David (NIH/NCI) [E] June 13, 2006

June 13, 2006

David Robinson is not on the time zone mailing list; direct replies appropriately. --ado -----Original Message----- From: David Robinson [mailto:drtr@apache.org] Sent: Tuesday, June 13, 2006 11:07 AM To: tz(a)lecserver.nci.nih.gov Subject: zic leapseconds bug There appears to be a bug in zic when leapseconds are provided, for any time zone transition near a leapsecond. The bug is that after reading in the Leaps as Posix seconds in the trans[] array, adjleap() is called. This converts these to genuine UTC seconds. However, this array is then used to fixup transition times which are in posix seconds, by for (i = 0; i < timecnt; ++i) { j = leapcnt; while (--j >= 0) if (ats[i] >= trans[j]) { ats[i] = tadd(ats[i], corr[j]); break; } } ats[] is stored in posix seconds, but trans[] is in leap seconds. The effect is that for Europe/London, the transition on 1996-01-01 00:00:00 is actually set to be at 1995-12-31 23:59:60, i.e. on the leap second (820454419). Any transition in the range 1996-01-01 00:00:00 to 00:00:18 is similarly off by one second. The fix is obvious in principle (have a copy of trans[] that isn't leap-second adjusted, and use that in the code above), but not all the uses of trans[] in the code are obvious to me (what are Rolling leapsecs?). I'd also suggest that zic reject any transition in its input with a second=60. It's usually run without a leap file so there's no point in implementing the correct behaviour for that case. Many thanks, David Robinson

1 0

tzdata Copyright
by Wolber, Richard C June 7, 2006

June 7, 2006

Greetings, Who is considered the "copyright holder" of the tzdata package? ..Chuck.. -- Chuck Wolber (Meriadoc Sandybanks) Electronic Flight Bag Crew Information Systems/ Linux Wonk 253.576.1154 "At the back of the Daylight Saving scheme I detect the bony, blue-fingered hand of Puritanism, eager to push people into bed earlier, and get them up earlier, to make them healthy, wealthy and wise in spite of themselves." - Robertson Davies, The diary of Samuel Marchbanks Clarke, Irwin (1947), XIX, Sunday

1 0

Australian time zone abbreviations
by Chuck Soper June 2, 2006

June 2, 2006

I recently found this Australian government web page on time zones: <http://www.australia.gov.au/about-australia-13time> And this government web page lists time zone names and abbreviations: <http://www.bom.gov.au/climate/averages/tables/daysavtm.shtml> I have read the notes in the australasia tzdata file. Should the current official time zone abbreviations be used rather than the popular ones? This would be my preference. Perhaps, the older tz abbreviations without the "A" for Australia could be maintained for dates in the past similar to using PWT (Pacific Wartime) for Pacific Time? Thanks, Chuck

5 4

Bern is the Capital of Switzerland
by Didier CLERC June 1, 2006

June 1, 2006

Hello, Why is the timezone for Switzerland Europe/Zurich and not Europe/Bern ? Bern is the Capitale of Switzerland. Considering Zurich to be the capitale of Switzerland is an offense to most of the Swiss. Thank you for your answer, Didier CLERC

1 0

RE: Possible TZ Data Errors with every 1st / 2nd / 3rd Xday of the month
by Andy Lipscomb June 1, 2006

June 1, 2006

No, in that case at least, the data is correct. The second Saturday of a month can be between the 8th and the 14th, inclusive; this year, the second Saturday of October is the 14th. Since the changeover occurs at 24:00 on that Saturday and Unix doesn't like 24:00, it is coded as 0:00 the following day, which is the 15th this year, and in general can run from the 9th to the 15th. Thus, Sun>=9 is correct. (This is not the "second Sunday," but the "Sunday after the second Saturday." USA readers will see the resemblance to Election Day.) ________________________________ From: Dr Hingston [mailto:dr.hingston@paradise.net.nz] Sent: Wed 31 May 2006 04:31 To: tz(a)lecserver.nci.nih.gov Subject: Possible TZ Data Errors with every 1st / 2nd / 3rd Xday of the month Dear List Members Woops, seems I got the address wrong, my apologies! I hope this is better! Regarding: Rule Chile 1999 max - Oct Sun>=9 4:00u 1:00 S Rule Chile 2000 max - Mar Sun>=9 3:00u 0 - The expression Sun>=9 from the above rules would appear to me to be incorrect, as it suggests the third Sunday in some years. e.g. October this year. Summary: Supreme Decree Number 1,489 (6 October 1970) - This law, enacted during the presidency of Eduardo Frei Montalva, declared that Daylight Saving Time will advance clocks by one hour, beginning the second Saturday of October, and ending the second Saturday of March. Chilean time is administered by the Chilean Navy's Hydrographic and Oceanographic Service (SHOA). On discussion with the Chilean Consulate the above is confirmed, and time for changeover is midnight Saturday (I take to mean 00:00 Sunday)

2 1