> > > Lintian complains about this, because it means that the file is not > > > present in the source package. IMO this is incorrect: generated files > > > should not be present in the source package. There's still discussion > > > (which I intend to restart) if it is acceptable to leave them in after > > > running the clean target, but AFAIK there's no discussion that it is > > > acceptable to remove them. Of course I could be wrong about that. :-) Hell, no. IT IS NOT ACCEPTABLE. POT files are the material needed by translators to work on. Assuming they'll be able to generate them from cryptic gettext commands, run autostuff magic, and weird incantations to get the very basic file they need is just wrong wrong wrong wrong. Please *don't* remove generated POT files from upstream source. Please don't remove debian/po/templates.pot. Please don't change lintian. This is IMHO a developer's disease to have the will of providing "clean" source, "clean" from generated stuff. This is often a good disease but, in that case, this is a very bad one. Please provide us with the material we need and don't hide it under layers of code. Of course, all i18n infrastructure could maybe be adapted in some way and cope with missing templates.pot files. But, actually, we can't assume that all translators will use the i18n infrastructure to get the material. Just like we can't assume that upstream translatros will egt the material they work on from Rosetta, Pootle or whatever fancy tool used to assist l10n work. We already have stuff and packages in Debian that hide POT files (the debconf package comes to mind)....and that's already a PITA. Yes, I know it means that developers have to maintain generated files in their VCS. So what? > > > > Perhaps I'm missing something here, but it looks like both > > translators[1] and the website[2] still expect the templates.pot file to > > be there. Wouldn't it make more sense to update the website and i18n.d.n > > scripts[3] before removing the lintian check? It would make sense to *not* change lintian and packages before that happens. And making that happen with collaboration of i18n people is very very unlikely. Because it is plain wrong to not provide POT files, period.....because it breaks existing practice, that works, for just the sake of being "clean". > Anyway, thanks for the comments, my report suggested to ignore the > translators and break their procedures. This was not my intention, of > course. Thanks for the comment. For sure, CC'ing -i18n in the original bug report would have been a good idea: I could have started to shoot much much earlier..:-)
Attachment:
signature.asc
Description: Digital signature