On Thu, Sep 22, 2011 at 07:42, Tom Lane <t...@sss.pgh.pa.us> wrote: > pratikchirania <pratik.chira...@hp.com> writes: >> 1. I am in Costa Rica, using Windows Server 2008 R2/Windows 2003 with >> PostgreSQL 8.3/9.0 >> 2. System Date/Time Settings shows "CST/Central America" with UTC-6 as extra >> display >> 4. There is NO DST for CST (Central America) a.k.a America/Costa_Rica to the >> PostgreSQL database through the pg_timezone_names view > > Well, if you want no-DST behavior, this is wrong: > >> show timezone >> "CST6CDT" > > That timezone specifies daylight savings behavior (CDT). As it happens, > it's going to follow the USA rules for when to switch, but any switch > would be wrong for Costa Rica. You need the America/Costa_Rica setting. > > Now having said that, it appears that the reason you got "CST6CDT" by > default is that we map the Windows "Central America Standard Time" and > "Central America Daylight Time" registry strings to that. This seems > clearly wrong. A look at the tzdata "northamerica" file shows that > noplace in Central America other than Mexico has observed DST with any > regularity, and certainly none of them could be said to follow USA DST > rules. > > Mexico is a separate case, because there are separate "Central Standard > Time (Mexico)" and "Central Daylight Time (Mexico)" entries, which we > map to "America/Mexico_City", which seems proper. > > I think we ought to map "Central America Standard Time" to plain CST6. > (Or we could map to one of America/Costa_Rica, America/Guatemala, > America/El_Salvador, etc, but that seems more likely to offend people in > the other countries than provide any additional precision.) I am not > sure what we ought to do with "Central America Daylight Time", but on > the evidence here I wonder whether that setting exists in the wild at > all. > > Magnus, AFAICT from the commit logs, that lookup table was your work to > begin with --- do you remember anything about the reasoning for the > Central America entries?
Hmm. not entirely. I know the initial round was basically all just guesses. Then at some point we added tools/win32tzlist.pl. But IIRC the actual timezones picked were more or less still guestimates. So I think it's just a mistake in that, and should be changed. -- Magnus Hagander Me: http://www.hagander.net/ Work: http://www.redpill-linpro.com/ -- Sent via pgsql-bugs mailing list (pgsql-bugs@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-bugs
