Re: Updating po files on make dist

mn 2003-05-26 klockan 22.44 skrev Bruno Haible:
> Christian Rose writes in
> :
> > How to update an existing translation
> > ...
> > 9. Commit everything (don't forget to add ChangeLog-style commit comment):
> >    cvs commit
> I think this step is incomplete. It should read as follows:
>   9. Check if the POT file was updated by the translators meanwhile.
>      cvs update *.pot
>      If this command reports "U" or "P", the POT file has changed, and
>      you need to update your PO file accordingly:
>      mv xy.po xy.po.tmp
>      cvs update xy.po
>      msgmerge -o xy.po xy.po.tmp *.pot

As Abel already pointed out, we do not keep pot files inside GNOME CVS.
They would in 99.99% of the cases be out of date with the sources
anyway, would be an additional source for cvs conflicts, and there's no
good reason to keep them in CVS either -- pot files are generated files
and translators can update their translations directly from the sources
anyway, thanks to intltool-update, or fetch them from the status pages
on the web when needed.

So any module that doesn't have *.pot in it's po/.cvsignore is buggy in
that regard, and so would also the instructions be if we instructed
translators to commit pot files. :-)

> Furthermore, you can improve another step in this tutorial:
> > How to add a new translation
> > ...
> > 4. Rename the pot file to xy.po:
> >    mv GConf2.pot xy.po
> The 'msginit' program has been created to streamline this step for the
> translators, filling in most of the header entry. So you can write
> this step as follows:
>   4. Create the initial xy.po:
>      msginit -l xy

I'm reluctant to add that, at least yet. msginit seems pretty much tied
to the Translation Project's listing of teams, and our set of teams
doesn't exactly match those by the TP, so msginit in its current state
won't behave well for all GNOME translators. Also, I'd prefer if msginit
left the charset field uninitialized ("CHARSET") or as "UTF-8" (as we
mandate the use of UTF-8 in GNOME) rather than setting it to "ASCII", as
that character set makes a horrible default.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]