Ed Leafe wrote:
> On Jul 31, 2007, at 3:35 PM, Paul McNett wrote:
>
>> The objection is that we are gaining the benefit of Rosetta (definite
>> benefit, mind you) but losing the ability to edit the source .po
>> directly if we prefer.
>>
>> The .pot files are the source for the .po files, but the .po files are
>> still source.
>
> Here's what the upload page says:
>
> "Here you can upload either a single PO template (.pot) or a tar file
> containing a PO template and a set of PO files (.tar, .tar.gz
> or .tar.bz2). The files you upload will be imported into Launchpad
> shortly."
>
> So maybe you can do what you want. I can't find any more info on this.
That would appear to be the way to do it. We use our local tools to
generate the .pot and the resulting .po's (which take into account
existing translations and merge in any new strings from the .pot). Then
we roll up the .po's and the .pot and send to Rosetta. I can't see any
reason for Rosetta to accept the individual .po's other than to preserve
the translated strings inside them, so that must be the point.
Does the Rosetta upload/download process lend itself to straightforward
automation, like nightly or something? Each night I guess we'd do:
1) svn up dabo
2) msgfmt.py to get the .pot and .po's updated.
3) tarball it and upload to Rosetta
4) download from rosetta
5) svn ci dabo
The download from Rosetta happens right after the upload; the idea there
is that we are downloading anything that changed that day merged with
the most recent .pot and .po's from Dabo. No information loss from
either end; 2-way capability; completely automated; +1.
--
pkm ~ http://paulmcnett.com
©2007 Paul McNett |
