Hi all, During an investigation about customer brazilian problem I downloaded your latest tz lib from "ftp://elsie.nci.nih.gov/pub/". I build and install the library on Linux Redhat 9. I created a simple "my_date.c" that perform a mktime into the range 00:00 - 14 Feb 2004 to 23:59 - 15 Feb 2004. The DST from Brasil Sao_Paulo is the following: file southamerica: Rule Brazil 2000 2001 - Oct Sun>=8 0:00 1:00 S Rule Brazil 2001 max - Feb Sun>=15 0:00 0 - Rule Brazil 2002 only - Nov 3 0:00 1:00 S Rule Brazil 2003 max - Oct Sun>=15 0:00 1:00 S So for the DST about Feb Sun>=15 at 0:00 I expected that the DST occurs at 0:00 15 Feb, but I saw that it occurs at 23:49 14 Feb as you can see from the output program. I set the TZ to TZ=America/Sao_Paulo ./my_date .. .. .. date =<day=14 hours=23 min=45> diff=-60 epoque=1076809501 DST=<1> date =<day=14 hours=23 min=46> diff=-60 epoque=1076809561 DST=<1> date =<day=14 hours=23 min=47> diff=-60 epoque=1076809621 DST=<1> date =<day=14 hours=23 min=48> diff=-60 epoque=1076809681 DST=<1> date =<day=14 hours=23 min=49> diff=-3660 epoque=1076813341 DST=<0> <-------- DST occurs here date =<day=14 hours=23 min=50> diff=-60 epoque=1076813401 DST=<0> date =<day=14 hours=23 min=51> diff=-60 epoque=1076813461 DST=<0> date =<day=14 hours=23 min=52> diff=-60 epoque=1076813521 DST=<0> date =<day=14 hours=23 min=53> diff=-60 epoque=1076813581 DST=<0> date =<day=14 hours=23 min=54> diff=-60 epoque=1076813641 DST=<0> .. .. I think that this is a bug because the DST is for 15 Feb at 0:00 and NOT 23:49 14 Feb. Do you have some suggestions or do you already knows about this problem? This is my "my_date.c" (See attached file: my_date.c) Thanks Stefano