[ntp:questions] Kind of OT : GMT timezones in linux

Per Hedeland per at hedeland.org
Wed Jul 11 06:46:36 UTC 2007

In article <m3abu3sy5i.fsf at lugabout.jhcloos.org> James Cloos
<cloos at jhcloos.com> writes:
>>>>>> "JimC" == James Cloos <cloos at jhcloos.com> writes:
>>>>>> "Per" == Per Hedeland <per at hedeland.org> writes:
>Per> [ The quoted message was posted to the mailing list
>Per>   but never appeared in the newsgroup. ]
>(Odd, I'm just read it via the newsgroup; let's see whether posting via
>the newsgroup works better....)

Yes, the posting appeared in the newsgroup about half a day after mine,
~ 2 days after your posting - I guess I should have said "so far" rather
than "never" (though I don't know if my mentioning the fact had some
effect on its appearance:-).

>Per> Um, what standard would that be? A reference would be appreciated (and
>Per> ideally a quote if it isn't available online).
>I discovered it when a change was made to ghostscript's svn related to
>it.  The colon is definitely required in windows, and at that time we
>(I and the committer) found a reference to one of the standards bodies'
>version of the tzset docs (presumably either POSIX or SUS) which agrees
>with the GNU manpage and doze's docs & implementation that the colon is

Well, that's more than a little vague - I don't really doubt what you're
saying, but I would be interested in seeing the actual standards text
(and knowing its origin exactly).

>JimC> Cf: tzset(3), such as at:
>JimC> http://ccrma.stanford.edu/planetccrma/man/man3/tzset.3.html
>Per> Well, that is a Linux man page, and as such of limited information
>Per> value.
>Well, a GNU man page, as it docs glibc.  

You have a point, but given the number of GNU/Hurd systems out there,
and the fact that glibc doesn't come with man pages AFAIK, I think I'll
continue to call it a "Linux man page".

>The described syntax exactly matches the standard.  Hense why I pointed
>there.  I was rushed and didn't search for bsd, soalris, hpux, et al pages.
>In general, to maximimize cross-platform portability, you need the colon.

Maybe, but since I couldn't find one among the major current Unices that
required the colon despite their docs discussing it, I wouldn't be
surprised to find a version that actually *didn't* work with the colon,
standard or not. Of course the file name usage originates with the Olson
time zone package - its docs go to some length in describing the POSIX.1
standard for TZ, and the file name usage as an extension of that, but
the only mention of the colon I could find is: "POSIX.1-1996 says that
an abbreviation cannot start with ':'" - which agrees well with the
implementations I checked, but doesn't per se support your claim. Of
course it shows usage of file names without colon for the TZ setting.

--Per Hedeland
per at hedeland.org

More information about the questions mailing list