Iain Arnell wrote:
>Of course it's technically possible, but policy-wise, everything
>should be built from source.

Ah, of course.  I should really be publishing the real source (build
script and templates) of T:OTZ:D.  Of course, the tarballs that I put
on CPAN do have to include the tzfiles: the essence of the module is to
package that data in a CPAN-usable way.  I can't even do it by packaging
the Olson data and code source files, because the Olson code is Unix-only
and requires a C compiler, either requirement being a big problem for
a CPAN module.

>          Our tzdata package doesn't always strictly follow upstream;
>the maintainer sometimes incorporates proposed changes before the
>official release.

In such cases, T:OTZ:D should not share the files with tzdata, because
T:OTZ:D follows only the canonical Olson releases.  You could still
package T:OTZ:D to share some files with tzdata and have others itself,
if you just want to save space, but the differing update schedules are
a problem.

>                  And we'd need to introduce a strict dependency
>between tzdata and T:OTZ:D - not updating one package without the
>other

I think that would be fine on the installation side, and in packaging
it would be fine if tzdata were packaged automatically, akin to the
way that I package T:OTZ:D for CPAN.  But your handling of tzdata is
evidently more complex than just blindly repackaging upstream, and that
would play havoc with any attempt to tie T:OTZ:D to it.

>I'm still thinking that the simplest and most consistent approach will
>be to generate Data.pm as part of our existing tzdata build to ensure
>that we're always in sync.

I don't think there's any value in keeping T:OTZ:D in synch with
/usr/share/zoneinfo.  Its very purpose is to be consistent across
platforms, independent of the host's native timezone database.
DT:TZ:Local is about matching host behaviour, and with DT:TZ:Tzfile
available it should be fairly easy to get it reading /usr/share/zoneinfo.

-zefram

Reply via email to