Hi Leute,
auf der pilotprocess-ML kam Heute die folgenden Mail. Da ich das für
eine gute Sache halte, schaut sie (und die darin enthaltenen links
mal an). Ich denke, wir sollten uns daran ausgiebig
beteiligen ... ;)
Also, wer ist mit dabei?
Bis dann
Thomas.
Ach ja ... Ich hab' nur was den Header zusammengekürzt, damit die
Mail nicht so ultralang wird ... ;)
---------- Forwarded Message ----------
<schnipp>
Date: Sun, 30 Apr 2006 05:36:14 +0100
From: Damien Donlon <[EMAIL PROTECTED]>
To: [email protected],
Pavel =?UTF-8?Q?Jan=C3=ADk?= <[EMAIL PROTECTED]>,
[EMAIL PROTECTED],
[EMAIL PROTECTED],
=?UTF-8?Q?J=C3=B6rg?= Jahnke <[EMAIL PROTECTED]>
Subject: [l10n process] PLEASE READ : OpenOffice.org Translation
<schnapp>
Hi Everyone,
Last month Rafaella Braconi mentioned that we want to test a new
process for contributing translated content for OpenOffice.org. In
this process a CVS server provides up to date PO files which
reflect the translation content of the OpenOffice.org source code.
This process is up and running now for testing & we would greatly
appreciate your input and suggestions on how we can go about
improving it. Until we have completed testing and ensured that
everything works ok, no translated content will go into official
OpenOffice.org builds.
For the purpose of testing we are providing content in the
following languages :
de cs fr it nl
(Pavel. I have provided cs just so as you can compare with your own
statistics & content).
These 2 documents will provide enough information to get you
started :
Introduction :
http://86.43.75.142/oo-translation/index.html
Overview of the Process :
http://86.43.75.142/oo-translation/translation-howto.html
You can view translation statistics generated on the source code
yesterday here :
http://86.43.75.142/oo-translation/stats/HEAD/index.html
The test deployment machine on which the CVS server resides is low
specification so please be patient if the cvs checkouts are a
little slow.
I hope this doesn't sound two much like homework but I have also
attached a document with some suggested exercises which would help
to get quickly familiarised with the process as it works at
present. Doing them would allow you to give more detailed feedback
& suggestions.
Please email the alias or me personally with your feedback &
questions If you see any issues with this methodology for doing
OO.org translation then I want to here about it. All going well,
the next steps are to implement your feedback and move the content
off the test machine and on to OpenOffice.org.
Best Regards,
Damien
-------------------------------------------------------
--
It used to be the fun was in
The capture and kill.
In another place and time
I did it all for thrills.
-- Lust to Love
Suggested Exercises
====================
[1] Note the untranslated wordcount for you language on the Translation
Status pages. Identify a file with untranslated content using the pages.
Check that file out of CVS. Translated all or some of the untranslated
content (you can grep or find for OO-TRANSLATABLE-MSG in the file to
get to the untranslated messages). Once you have translated the messages
commit the file back to CVS. Check the untranslated wordcount for you
language the next day. It should be lower reflecting your translation
work. Recheck the file out of CVS. Note now that the messages you have
translated are commented OO-REVIEWABLE-MESSAGE.
[2] Deliberately break a PO file that you have checked out in some way.
Commit the file back to CVS. Check the next day to see whether the
file appears in the Broken Files list. If you check the directory which
contained the file out of CVS you will note that there is also a file
present called <your filename>.po.broken. Fix the file and commit it
back to CVS. THe file should disappear off the Broken Files list the
next day.
Note: I am particularly interested to get peoples thoughts on treatment
of broken files. At the moment a copy of any broken files is copied into
CVS when found by the server side scripts. THe copy has the suffix file
suffix .broken. The thought behind this is that it allows contributors
to quickly and easily identify broken files and fix them but should the
PO file be removed to just leave the .broken file?
[3] If using Linux and you have msgfmt installed then do a
test compile on all the PO files for your language as follows using a
terminal :
for pofile in `find HEAD/<your language> -name "*.po" -print`
do
msgfmt --check --strict -o /tmp/out $pofile
done >compile-output.txt 2>&1
Read the newly generated compile-output.txt file to check for errors.
Do the same as above, substituting the msgmft with the following if
you have pofilter installed :
pofilter --openoffice $pofile
[4] Add a new PO file to the PO CVS server giving it any name you like.
If this file has no equivalent in the extracted OO.org sources then it
will appear in the last section of the Broken Files list the next day.
Note : This is a methodology for identifying obsolete PO files in CVS that no
longer have an equivalent in the OO.org sources. You might ask why not
automatically remove such files? The answer is that a file might not
have an equivalent in the OO sources not because it is obsolete but because
the source code has been restructured and the source code simly moved. In
such instances the translations are still needed and this is why we dont
want to automatically delete the PO thereby losing them.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
