Boost C++ Libraries: Ticket #605: Support different DST rules in timezone db based on years

quatras — Mon, 24 Apr 2006 16:15:46 GMT

Logged In: YES
user_id=994937
It might be easier, in terms of sorting the .csv file, to
combine the effective year range with the zone name, such as:
    "America/New_York:0000-2006", ...
This would have the advantage of (a) keeping all entries for
the same zone together, (b) not adding a new column to the
table, and (c) not impacting time zones whose rules do not
change.

quatras — Fri, 28 Apr 2006 19:53:50 GMT

Logged In: YES
user_id=994937
After further thought, I believe a better way to do would be
to add a code at the end of the ID string that identifies
where the timezone is located, in general terms, so that a
secondary table can be checked for year ranges.  This has
the advantage of requiring only one entry instead of
multiples.  In the New York example, the string would become
"America/New_York;US".  The lookup routines would have to
stop the comparison at the ';'.  Then, the "US" value would
be used to look up US rules for the year in question.  I
would call the string "US" a 'group' code, to be followed by
a 'region' code.  For example, Hawaii has no DST; its code
could be ";US/HI".  The European Union has common DST
handling, so the string for Paris would be
"Europe/Paris;EU/FR".  The other major 'group' is Canada.  A
sample timezone file with these values is attached.

component changed; severity set

Daryle Walker — Fri, 03 Aug 2007 12:05:23 GMT

component None → date_time
severity → Problem

type, summary changed

Katie Chan — Sun, 28 Nov 2010 06:27:44 GMT

type Support Requests → Feature Requests
summary support new 2007 DST rules for timezone db → Support different DST rules in timezone db based on years