Benno Schulenberg <coordina...@translationproject.org> skribis: > On 2014-06-09 00:07, Ludovic Courtès wrote: >> Saluton, Benno, >> >> Benno Schulenberg <coordina...@translationproject.org> skribis: >>> As will be clear from the above, I would not want to synchronize but >>> draw a strict dividing line: decide who should translate what, and >>> use those translations everywhere where the strings occur. >> >> Okay, that makes sense. >> >> Honestly, I’m not thrilled by the idea of having to deal with two >> translation sources. > > It's only slightly harder than dealing with a single source. You > automate things. Have a look at the bootstrap script in for example > grep. All you do is send out a POT file (and in this case two, to > two different addresses), and at the time of rolling a release > tarball, the script takes care of fetching and merging any available > translations.
OK. >> Besides, would you suggest using different text domains for package >> meta-data and for strings of the Guix tools? > > Precisely. The ideal would be to have two POT files for Guix, one > with the interface strings, guix.pot, and one with all the package > descriptions, say guix-blurbs.pot. OK, will do. > I would propose to send the first to the TP, and the second to GNU web > translators, as those strings appear also on the web, and Ineiev says > that those texts must be at least reviewed by the web translators. Unfortunately, as I wrote, Guix also has descriptions for non-GNU packages to be translated, which makes this procedure unsuitable. Or would we need a third text domain? Or...? Help welcome! Ludo’.
