New PO Template file for `gcc'

[Joseph S. Myers]

On Mon, 28 Apr 2003, Karl Eichwalder wrote:

> Yes and no.  In general it's useful to let translators update and even
> add new translations for 3.2.x because then (GNU)/Linux distributors can
> pick up translations matching 3.2.x; some of them care about this
> posibility and that's why I always want the stable releases present at
> the TP site.

They will also find that gcc.pot hasn't been regenerated on the 3.2 branch 
for ages, nor have the .po files on 3.2 branch been updated from the TP 
site for ages (as the GCC i18n maintainer hasn't lately been committing 
the files when they are sent to gcc-patches, and no-one else feels it's 
their regular responsibility to do so).

The documentation for how to make GCC releases
<http://gcc.gnu.org/releasing.html> and branches
<http://gcc.gnu.org/branching.html> now explains what the Release Manager
should do to ensure that translations are up to date, but this is a recent
addition to that documentation.  In principle .po files will get updated
on the release branches from the most recent files for that branch at
<http://www2.iro.umontreal.ca/~pinard/po/registry.cgi?domain=gcc>, but
making a release is an error-prone 18-step process at present and the .pot
regeneration and .po file updating needs to be remembered before the
release script is run.

> Maybe you can provide a 3.3 snapshot?  I guess, it would be the easiest
> to install gcc-3.3-b20030428.pot for translation purposes.

gcc.pot hasn't been regenerated on the 3.3 branch since the 2002-12-30
snapshot (nor have the .po files, mainline or branch, been updated since I
last did so in February, though new versions of various .po files have
appeared since then).

> > (The 3.2.3 release was submitted by some automated process that submits 
> > new GNU releases.)
> 
> It was my fault to pick it up...  Thanks for the warning by your side!

I'd presumed (when writing <http://gcc.gnu.org/translation.html> in the
hope it would encourage gcc.pot and the .po files to be kept more up to
date in GCC CVS, and maintainers to understand how to deal with bug
reports about translations) that it was automated (and possibly linked to
the process updating <http://www.gnu.org/server/new-software.html>), since
I wasn't submitting the releases, and I didn't think Zack or the i18n
maintainer would have been given the lack of recent gcc.pot regeneration
on the 3.2 branch, and I'm sure no other GCC maintainer would have been
submitting releases to the TP site.

-- 
Joseph S. Myers
jsm28@cam.ac.uk