On Fri, Jan 23, 2026 at 08:33:02PM +0000, Gavin Smith wrote: > > We have this implementation constraint that link those two > > subjects together, but then it could make sense to accept not to have > > translations in output documents if it means having to install > > translations for error messages. > > That doesn't make sense - if users don't want translated messages from > programs, that's completely different to strings being translated in > output files. > > Users might not want to be able to get messages in Swedish so they > don't install a Swedish locale. But they should still be able to > get strings in the output of texi2any translated to Swedish if the > @documentlanguage says it's in that language. (For example, they may > be processing manuals for upload to a website for download by others.)
Maybe I do not understand what you are saying, but it seems to me that you are saying the same thing as I do. My point is that it is normal that it comes as a surprise to users that (at least one) locale for error messages should be installed in order to have output document strings translated. So, I fully agree that translated messages from programs is completly different to strings being translated in output files, but in practice, both are linked in the current implementation. Indeed, one must switch to an error messages translation locale (any locale but C or POSIX) to be able to have strings in document translated. -- Pat
