[Bf-translations-dev] Major update in translation tools

[Bastien Montagne]

Hi,

I’ve spent the last days working on the i18n stuff both py tools and 
leaning i18n usage in Blender code (I’ll talk about that on bf-committers).

I intend to commit this very soon (likely tomorrow thursday), unless 
some serious objections arise. :)

Here is a summary of the changes in scripts (see also the whole new 
README.txt file for details: 
http://wiki.blender.org/index.php/User:Mont29/i18n ):
* Rewrote all scripts:
** They are now under /trunk/po/tools dir.
** They use a common settings.py settings source.
** Commonly used functions are gathered in utils.py (notably those 
reading/parsing and writing po messages).
** All scripts now have an args management system (argparse module), 
with a minimal help message (-h).
* Nearly all common operations can now be done with two tools:
** update_branches.py to update /branches po's with latest messages.
** update_trunk.py to update /trunk po's and mo's with /branches po's 
translated enough.
* Lower-level scripts remain present, for more specific needs.
* merge_po.py now accepts several sources po's to merge into dst one 
(and it removes fuzzy flag when needed).
* check_po.py is now a report-only tool outputing some nice stats and 
info about current po's.

Another important topic: languages needing post-processing of their po's 
(currently, arabic and persian). I renamed the scripts to 
<language>_to_utf.py (and fixed a bug regarding escaped \" and \t in 
messages...). So now, the <language>.po file *must* be the raw 
(unprocessed) one.
Then, the import_po_from_branches.py script (and hence the 
update_trunk.py one) will use that <language>_to_utf.py script, if 
present, to automatically process the raw <language>.po file and put the 
processed version under /trunk/po.

I hope and think all this will hugely ease to use and manage the 
bf-translation repo!

Best regards,
Bastien