On Thu, 20 Oct 2011, Robert Elz wrote:
Here are proposed changes to the time zone package. The change to Bahia Brazil has already occurred, and the others are due to occur within 10 days (the weekend after next) so updates have become urgent (yet again)...
[...]
Please test these changes (particularly those of you using 64 bit time_t systems).
I have tested these changes on a NetBSD system with 64-bit time_t. I find the following differences in the output from "zdump -v ${ZONENAME}". America/Bahia gains two DST transitions per year, from now until further notice: America/Bahia Sun Feb 16 02:00:00 2003 UTC = Sat Feb 15 23:00:00 2003 BRT isdst=0 +America/Bahia Sun Oct 16 02:59:59 2011 UTC = Sat Oct 15 23:59:59 2011 BRT isdst=0 +America/Bahia Sun Oct 16 03:00:00 2011 UTC = Sun Oct 16 01:00:00 2011 BRST isdst=1 +America/Bahia Sun Feb 26 01:59:59 2012 UTC = Sat Feb 25 23:59:59 2012 BRST isdst=1 +America/Bahia Sun Feb 26 02:00:00 2012 UTC = Sat Feb 25 23:00:00 2012 BRT isdst=0 [...] The Europe/Kiev clock change that already occurred on 27 March 2011 is retroactively re-labeled as a transition into DST, instead of a change in the UTC offset used for standard time. The displayed local time on 27 March 2011 remains the same, but the zone name and isdst flag were "FET isdst=0" in tzdata2011l, and change to "EEST isdst=1" in the new data: Europe/Kiev Sun Mar 27 00:59:59 2011 UTC = Sun Mar 27 02:59:59 2011 EET isdst=0 -Europe/Kiev Sun Mar 27 01:00:00 2011 UTC = Sun Mar 27 04:00:00 2011 FET isdst=0 +Europe/Kiev Sun Mar 27 01:00:00 2011 UTC = Sun Mar 27 04:00:00 2011 EEST isdst=1 Europe/Kiev gets two DST transitions per year from now until further notice: +Europe/Kiev Sun Oct 30 00:59:59 2011 UTC = Sun Oct 30 03:59:59 2011 EEST isdst=1 +Europe/Kiev Sun Oct 30 01:00:00 2011 UTC = Sun Oct 30 03:00:00 2011 EET isdst=0 +Europe/Kiev Sun Mar 25 00:59:59 2012 UTC = Sun Mar 25 02:59:59 2012 EET isdst=0 +Europe/Kiev Sun Mar 25 01:00:00 2012 UTC = Sun Mar 25 04:00:00 2012 EEST isdst=1 [...] In tzdata2011l, Europe/Tiraspol was a link to Europe/Chisinau. In the new data, Europe/Tiraspol is a separate zone. This change affects times up to and including 31 December 1879, because the longitute used for LMT has changed from that of Chisinau to that of Tiraspol: -Europe/Tiraspol Wed Dec 31 22:04:39 1879 UTC = Wed Dec 31 23:59:59 1879 LMT isdst=0 -Europe/Tiraspol Wed Dec 31 22:04:40 1879 UTC = Wed Dec 31 23:59:40 1879 CMT isdst=0 +Europe/Tiraspol Wed Dec 31 22:01:27 1879 UTC = Wed Dec 31 23:59:59 1879 LMT isdst=0 +Europe/Tiraspol Wed Dec 31 22:01:28 1879 UTC = Wed Dec 31 23:56:28 1879 CMT isdst=0 The Europe/Tiraspol clock change that already occurred on 27 March 2011 is retroactively re-labeled as a change in the UTC offset for standard time, instead of a transition into DST. The displayed local time on 27 March 2011 remains the same, but the zone name and isdst flag were "EEST isdst=1" in tzdata2011l, and change to "FET isdst=0" in the new data. Also, all future DST transitions for Europe/Tiraspol are removed in the new data: Europe/Tiraspol Sun Mar 27 00:59:59 2011 UTC = Sun Mar 27 02:59:59 2011 EET isdst=0 -Europe/Tiraspol Sun Mar 27 01:00:00 2011 UTC = Sun Mar 27 04:00:00 2011 EEST isdst=1 +Europe/Tiraspol Sun Mar 27 01:00:00 2011 UTC = Sun Mar 27 04:00:00 2011 FET isdst=0 -Europe/Tiraspol Sun Oct 30 00:59:59 2011 UTC = Sun Oct 30 03:59:59 2011 EEST isdst=1 -Europe/Tiraspol Sun Oct 30 01:00:00 2011 UTC = Sun Oct 30 03:00:00 2011 EET isdst=0 -Europe/Tiraspol Sun Mar 25 00:59:59 2012 UTC = Sun Mar 25 02:59:59 2012 EET isdst=0 [...] I also noticed a mistake in a comment in the "europe" input file: +# The substance of this change (reinstatement of the Tiraspol entry) +# is from a patch from Petr Machata (2011-11-17) That date should be 2011-10-17. --apb (Alan Barrett)