Paul Eggert wrote:
Tiberiu Covaci <tcovaci@bea.com> writes:
According to unofficial sources in Turkey, and Worldclock representant, the right value would be 1:00u. Please advice how to proceed to make the data into the next release of tzdata.
Sorry, I don't see why a change would be needed to the tz data. For example, here is the output of 'zdump -c 2006,2008 -v Asia/Istanbul'. The transitions are at 01:00 UTC, just as you wrote.
Asia/Istanbul Sun Mar 26 00:59:59 2006 UTC = Sun Mar 26 02:59:59 2006 EET isdst=0 Asia/Istanbul Sun Mar 26 01:00:00 2006 UTC = Sun Mar 26 04:00:00 2006 EEST isdst=1 [...]
The right time seems to be 1:00s (not u), so they are not on the EU rules, just almost. I have not yet had time to research when this started, but based on old news reports, it seems that has been so at least since 1997, two examples (1997 and 2006): http://www.byegm.gov.tr/YAYINLARIMIZ/CHR/ING97/03/97X03X25.TXT (search for summer time) http://www.hurriyet.com.tr/english/5305826.asp It is on my TODO-list to research this further, to correct old timestamps. Best regards, Steffen (timeanddate.com)