This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PING^2] RFC [PATCH] BZ#1077902: New API gettimezone
- From: P J P <pj dot pandit at yahoo dot co dot in>
- To: "libc-alpha at sourceware dot org" <libc-alpha at sourceware dot org>
- Date: Tue, 29 Apr 2014 20:41:08 +0800 (SGT)
- Subject: Re: [PING^2] RFC [PATCH] BZ#1077902: New API gettimezone
- Authentication-results: sourceware.org; auth=none
- References: <1396499286 dot 85118 dot YahooMailNeo at web192405 dot mail dot sg3 dot yahoo dot com> <1397301884 dot 32837 dot YahooMailNeo at web192402 dot mail dot sg3 dot yahoo dot com> <534971E4 dot 6060001 at cs dot ucla dot edu> <53497633 dot 6060804 at redhat dot com> <1397324033 dot 69177 dot YahooMailNeo at web192403 dot mail dot sg3 dot yahoo dot com> <5349A4B0 dot 2070206 at redhat dot com> <1397375798 dot 36419 dot YahooMailNeo at web192401 dot mail dot sg3 dot yahoo dot com> <1397414803 dot 70882 dot YahooMailNeo at web192403 dot mail dot sg3 dot yahoo dot com> <534B8A9F dot 8030806 at redhat dot com> <1397469748 dot 42212 dot YahooMailNeo at web192405 dot mail dot sg3 dot yahoo dot com> <1398146221 dot 72442 dot YahooMailNeo at web192403 dot mail dot sg3 dot yahoo dot com> <1398755742 dot 94004 dot YahooMailNeo at web192405 dot mail dot sg3 dot yahoo dot com> <535F74EE dot 8010002 at redhat dot com>
- Reply-to: P J P <pj dot pandit at yahoo dot co dot in>
Hello,
> On Tuesday, 29 April 2014 3:16 PM, Carlos O'Donell <carlos@redhat.com> wrote:
> The current implementation isn't robust in that the POSIX TZ at the end
> of the zone file might not be enough to define the timezone completely.
True. TZ variable does not include all the entries from a time zone file. Nonetheless, the TZ string at the end of a zone file does work well because it defines the required offsets from UTC for both standard and daylight savings time. When it does not define such offsets, the default of 1 hour is defined and documented. And for cases wherein the TZ string is absent from a zone file, fall-back to UTC is well reasonable too.
IMO, TZ variable is not expected to hold the 'complete' time zone definition from a zone file. Its syntax does not allow for that:
std offset dst [offset],start[/time],end[/time]
It could be treated as a minimal time zone definition to be used in circumstances wherein the complete elaborate definition is not accessible. In that sense, the current patch does offer an acceptable working solution.
> You need to set TZ to a file e.g. TZ=Country/City where possible and
> assume in your use-case that the chroot has an updated set of zones
> installed that match. You want to use the TZ file format because it allows
> a correct implementation by using all the data in the zone files. The POSIX
> TZ is insufficient to be correct all the time.
Agreed. But the 'TZ=Region/City' definition requires that the said file is accessible from the chroot(2) jail. It is not always feasible.
> Can you sketch out the tzalloc-style API and propse it?
Sure! I'll look into it and get back asap. Is such API expected to make the 'complete' time zone definition available to its consumers?
Thank you.
---
Regards
-Prasad
http://feedmug.com