https://bugs.kde.org/show_bug.cgi?id=352395
--- Comment #14 from Philippe Waroquiers <philippe.waroqui...@skynet.be> --- (In reply to Austin English from comment #13) > (In reply to Austin English from comment #12) > > How commonly are tarballs created/used for random SVN revisions rather than > > releases? > > > > To be frank, I'm not very experienced with makefiles, and the current > > changes have already been a pain. I don't particularly feel like revamping > > it a 6th time for a corner case that doesn't currently work to begin with. > > It seems to be currently broken, for that matter: > $ make dist > ... Yes, I understand that this change has already implied several reworks, following several comments, and that this does not converge very quickly. Sorry for that. For what concerns the broken make dist: this might be something in your setup, related to the documentation building procedure, which is not straightforward: many packages are needed to build the doc. I just tried, and make dist works for me. Otherwise, to my knowledge, make dist is run for a set of SVN versions for the (often several) release candidates and for the final release. So, as far as I understand, we need to? should? produce a vgversion.h as part of the make dist, as this ensures the build from a dist tarball is similar to a 'non dist' build. IMO, identifying tarballs precisely is as (or more?) important than identifying 'in a corner' SVN versions, as release candidates and dist have more chances to be replicated here and there. I will take a look at this whenever I have some time, and see if I can help -- You are receiving this mail because: You are watching all bug changes.
