> > btw: I'm curious about the "two patches + one more to fix the first two". > > It is not the usual thing to squash patches before distributing them, > > particularly when one of them fixes things in the others? > > On [email protected] we send the patches for the commits that we push > to the master branch. > > Ideally I would have caught the mistake before pushing so I could have 1 > patch per module (sys_stat-h and sys_types-h). But I realized when I was > replying to your bug report after pushing. Oops...
Aha, should have guessed... > The correct single patch can be generated using this command in a Gnulib > checkout though: > > $ git diff > 8ba0c4da1ea978b770d1bbf8e306396ed556205f..d5dcc6c91d3c7c6f293fe906af885831af458dac > > I expect that you do not have it cloned, so I have attached it for > you. Since m4 (atleast from git) puts Gnulib tests in tests/, I expect > you can trim the ChangeLog and .texi files and the patch will work. Oh, thank you. Actually, I had already solved it by applying the three patches and then doing "diff -ru m4-1.4.20.orig m4-1.4.20". Very often, there are more than one way to do that sort of things. I confirm that the non-code parts for the patches do not apply cleanly to current m4 source, so yes, I will probably skip them. Thanks a lot.
