This is the mail archive of the
mailing list for the GCC project.
Re: New PO Template file for `gcc'
- From: martin at v dot loewis dot de (Martin v. Löwis)
- To: "Joseph S. Myers" <jsm28 at cam dot ac dot uk>
- Cc: Philipp Thomas <pthomas at suse dot de>, Karl Eichwalder <keichwa at gmx dot net>, Translation Project Robot <translation at IRO dot UMontreal dot CA>, gcc-patches at gcc dot gnu dot org
- Date: 27 May 2003 13:57:52 +0200
- Subject: Re: New PO Template file for `gcc'
- References: <200304251904.h3PJ4bZU010167@vor.iro.umontreal.ca><Pine.LNX.email@example.com><firstname.lastname@example.org><Pine.LNX.email@example.com><20030526170152.GC4124@paradies.suse.de><Pine.LNX.firstname.lastname@example.org>
"Joseph S. Myers" <email@example.com> writes:
> It would be possible to regenerate gcc.pot automatically before each
> snapshot (Zack had a script to do this, but more work was needed). But
> we've been told that the Translation Project does not want automatic
> weekly submission of new .pot files for snapshots, so for this to be
> useful in keeping translations up to date, translators would need to
> interact with GCC some other way.
Indeed. My personal impression is that a few untranslated messages
won't bother anybody. Instead, gcc translations are still in a state
where only few teams have started translating the catalog, and even
fewer had ever completed the catalog (especially since the catalog
kept significantly growing for quite some time).
So I propose that, for a foreseeable future, gcc translations focus on
a single branch, which would be the 3.3 branch at the
moment. Translators can improve the translations as they find time,
and the GCC team produces releases when they see the need and find the
Any tighter integration is IMO unnecessary. If there were n+1 catalogs
per 3.x branch (one prerelease, one catalog per released subminor),
translators will keep being reminded that work needs to be done.
A bigger concern, to me, is the quality of the translations. As gcc is
such a huge thing, many translations will be incorrect. It would be
good if more automatic checks could be added somehow, e.g. for the
proprietary printf directives used in error functions. I know that
Debian had to remove the entire French catalog because it could make
the compiler crash. If anybody could provide code that performs checks
beyond the checks that msgfmt performs, I'd happily integrate it with
the TP robot, so that translations get rejected by the TP already.