Argentina's (no) daylight savings time

Mariano Absatz baby at
Mon Apr 16 22:30:47 UTC 2001


Im new to the list so forgive me if this has already been discused.

I administer a bunch of Solaris and Linux (RedHat 6.x/7.x) machines and a 
month or so ago, all of the Linuxes went one hour early...

It was quite clear that it was related to a timezone problem... the 
problem was I never quite understood the format of the zoneinfo files... 
anyway, I got the latest files from to try 
to solve the problem, but it didn't work...

I just saw that there is another release (2001b), but the bug is still 

I edited the southamerica file by hand not exactly knowing what I'm 
doing, but I kinda fixed it.

The point is that as far as I remember, in the last 3 or 4 years, DST was 
not used (there were a few decrees and I think there was a failed 
national survey, but eventually, DST was not applied). 

Anyway, the "least damage" editing I could think of (not fully 
understanding the zoneinfo file logic) was deleting all the zone data 
regarding the year 2000 rule for all Argentine cities in the file 
(Buenos_Aires, Rosario, Cordoba, Jujuy, Catamarca and Mendoza).

It worked, though I ain't quite sure if it will go wrong in October...

Probably, the rule itself should be erased, though I didn't do it...

Is there a somehow concise explanation of the format of the zoneinfo 
files, other than the comments in the file itself?... in fact, I only 
checked the southamerica file, maybe there is another one with a better 

I actually don't have the laws, decrees and/or Supreme Court roulings 
regarding DST to justify what I am doing, but I DO know what time is it 
now :-) (and the fact that it's 3 hours west from GMT, and that I didn't 
change to DST for more than a couple of years).

If I get to understand the format and relationship betwenn Rules and 
Zones I can correct the Rules/Zones for the couple of last years... I 
actually don't have previous data, and given the conflicts DST generated 
the last time it was used (I can't even remember when that was) I don't 
think there IS a knowledgable source for the past, other than what was 
put in there by other people in the past.


If you are interested this is the diff output:
# diff southamerica.wrong southamerica.better
<                       -4:00   Arg     AR%sT   2000 Mar  3 0:00
<                       -4:00   Arg     AR%sT   2000 Mar  3 0:00
<                       -4:00   Arg     AR%sT   2000 Mar  3 0:00
<                       -4:00   Arg     AR%sT   2000 Mar  3 0:00
<                       -4:00   Arg     AR%sT   2000 Mar  3 0:00
<                       -4:00   Arg     AR%sT   2000 Mar  3 0:00

Mariano Absatz
baby at

More information about the tz mailing list