-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 01/25/11 21:17, Eric Blake wrote:
On 01/25/2011 11:45 AM, Zdenek Styblik wrote:
> Hello,
>
> I somehow dared to open 'po/cs.po' file and it made me wonder what's the
> life cycle of these files. Especially this file seems to be somehow
> "shifted", because translations don't even match to their English
> counterpart.
Upstream, the .po files are re-generated with the latest available
translation at every release. And if that is not frequent enough for
you, running 'make dist' will regenerate the libvirt.pot master template
to the current source code strings then re-merge all existing
translation .po files to use that new template (you can also use make -C
po update-po to update just the po directory instead of creating an
entire distribution tarball).
I'm not sure where the best canonical location is for looking for the
latest available translations, nor what schedule is used by the various
translators in providing updated files for libvirt to incorporate
per-release (and it probably differs by language) (GNU projects host
translations on
http://translationproject.org, but libvirt is not a GNU
project). Within the .po files, "shifted" locations are generally not a
problem (locations in the .pot file are more for reference of the
translator when translating a particular build of libvirt); gettext
itself works on string contents rather than source code locations when
To be honest here, I haven't understood a bit of it. On the other hand,
I'm pretty maxed out and should be in bed already.
In the lame terms - modifying cs.po file - bad or good?
Where to put/post changes/diff/whatever.
And I must note I'm not going to use Czech translations by myself nor
planning to do 100% translation on my own time. Yet I can't stand
rubbish that's in the 'po/cs.po'.
On the other hand, if your answer is going to to be like: "What you're
doing is pretty much futile", then pat on back and move on *tired*
As for the plan. Clean up, translate as much as possible, somebody else
can eventually pick it up. I don't even know whom would to revision or
testing it.
actually serving up translations. Gettext also does a pretty decent
job
of fuzzy matching, both to make the translator's job easier
(translations from the previous release that can carry forward to the
current release are reused) and the end user (if the end user's
translation database is older than the installed libvirt, they still get
most strings translated if there wasn't a lot of churn in string
contents in the meantime).
I'm sorta speechless. Nothing against gettext or anything, but almost
everything marked as "fuzzy" is worth of deletion.
For its defense, the last translation is from 2008 or something.
I don't feel like being able to continue in constructive way in this
paragraph :s
> I doubt anybody is using Czech translation for libvirt and to be
honest,
> I would be enormously surprised if someone, anyone, did.
One thing I've learned about i18n is to never be surprised at who is
using a particular translation. I'm sure that someone is using it, or
there wouldn't have been a push to provide the translation file for
inclusion in a libvirt release in the first place.
No Czech speaking man cares, obviously. I'm not saying you are wrong.
Thanks,
Z.
- --
Zdenek Styblik
Net/Linux admin
OS
TurnovFree.net
email: stybla(a)turnovfree.net
jabber: stybla(a)jabber.turnovfree.net
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla -
http://enigmail.mozdev.org/
iEYEARECAAYFAk0/OWQACgkQ8MreUbSH7ilXaACfcNRkFwB44vs5+dFf7ay5FMVt
5p4AoMlOXnL2lf8/D6/SV7ogsNWqiDfY
=4Vrf
-----END PGP SIGNATURE-----