On Thu, Sep 24, 2026 at 08:30:28AM +0300, Eli Zaretskii wrote: > > Date: Wed, 23 Sep 2026 21:19:09 +0200 > > From: Patrice Dumas <[email protected]> > > > > It is something for the emacs developers, the already use -o, so they > > could set @setfilename to the base file name, which is more expected > > than what they do. > > How do you expect us to do that, when we want to support a wide range > of Texinfo versions with the same Texinfo sources and Makefile's?
This is not the issue here. Unless I missed something, how you call makeinfo in C or texi2any already allows to have the correct output for a wide range of Texinfo versions with the same Texinfo sources and Makefile's. The issue here is what happens when you call manually like makeinfo emacs-xxx.texi > And why impose such a task on use in the first place? There's nothing > particularly wrong with the current behavior, even if it looks > surprising at first sight: it is deterministic and easily explained. You know better than me what is better for the emacs project, but my feeling is that having the Info file created in the directory specified within the source file itself is quite unexpected. I would expect the output file to be created in the directory I run the command in, unless I pass a specific command line option. > And it has been with us for a long enough time. That may be a good reason to keep some support in texi2any, but it is not a good reason to keep using it if it can be unexpected/confusing. -- Pat
