[tz] [PATCH] Updates for Chile 2013
goldsmit at apple.com
Thu Feb 21 21:14:51 UTC 2013
> This makes sense, I didn’t notice that the 2012 rule actually applies.
I also agree with extending the 2012 rule indefinitely.
On Feb 21, 2013, at 4:12 AM, Petr Machata <pmachata at redhat.com> wrote:
> Robert Elz <kre at munnari.OZ.AU> writes:
>> Date: Wed, 20 Feb 2013 22:55:04 +0100
>> From: Petr Machata <pmachata at redhat.com>
>> Message-ID: <m2ehgaljwl.fsf at redhat.com>
>> | I decided to apply the transition dates to single years only--the DST
>> | policy is reevaluated every year anyway.
>> I think that's a mistake - you're guaranteeing that the data is going
>> to be wrong, rather than simply making it likely.
>> I'd suggest that the data be changed as ...
>> --- southamerica.was 2012-07-25 21:13:46.000000000 +0700
>> +++ southamerica 2013-02-21 07:37:16.000000000 +0700
>> @@ -1277,10 +1277,8 @@
>> Rule Chile 2010 only - Apr Sun>=1 3:00u 0 -
>> Rule Chile 2011 only - May Sun>=2 3:00u 0 -
>> Rule Chile 2011 only - Aug Sun>=16 4:00u 1:00 S
>> -Rule Chile 2012 only - Apr Sun>=23 3:00u 0 -
>> -Rule Chile 2012 only - Sep Sun>=2 4:00u 1:00 S
>> -Rule Chile 2013 max - Mar Sun>=9 3:00u 0 -
>> -Rule Chile 2013 max - Oct Sun>=9 4:00u 1:00 S
>> +Rule Chile 2012 max - Apr Sun>=23 3:00u 0 -
>> +Rule Chile 2012 max - Sep Sun>=2 4:00u 1:00 S
>> # IATA SSIM anomalies: (1992-02) says 1992-03-14;
>> # (1996-09) says 1998-03-08. Ignore these.
>> # Zone NAME GMTOFF RULES FORMAT [UNTIL]
>> That is, it turns out that the rules we inserted for 2012 apply for 2013
>> as well (except we were conservative last year, and assumed the change then
>> was just a one year thing, and put the older dates back for 2013).
>> So, the patch above just assumes that the 2012/2013 strategy will continue
>> forwards, which it might ... it is also possible that it won't, and we will
>> need to change it again next year (or even again later this year) but
>> some kind of reasonable guess is better than simply assuming that summer time
>> is going to start next September, and then never end.
> This makes sense, I didn't notice that the 2012 rule actually applies.
More information about the tz