Hi Bruno, At 2022-06-19T16:43:51+0200, Bruno Haible wrote: > Hi Branden, > > Have things worked out for you by now? Is the .tarball-version > workflow clear? Should we document it better in the git-version-gen > script?
Thanks for following up. I haven't had time to dig into this. I've become uncertain about whether groff's build is doing to the right thing with respect to the '.version' file that it creates in Makefile.am (by running git-version-gen). In Git checkouts, this file has a tendency to keep stale version string in it that grows progressively staler, but this doesn't damage the version string that actually gets bundled into build artifacts. Another vexing question is that of builds from the snapshot archives that Savanah produces. > > cgit is not developed by the Savannah admins. Are you suggesting > > that the Savannah admins fork cgit? I think that if GNU doesn't have the infrastructure or personnel to support these, then, yes, the Savannah administrators should fork (or just patch) cgit to the extent necessary to suppress the exposure of these snapshot download links. I had no idea they weren't truly supported. I can't be the only person who assumed that they were. There doesn't seem to be a lot of support on the groff mailing list for this feature, either--at least one fellow developer has questioned why I'm expending the effort to document a procedure for building from snapshot archives, given that two other methods (Git checkouts and distribution archives) already exist. If snapshot archives are an undesired sidecar that came along with cgit, and neither the Savannah admins nor gnulib developers have the cycles to spend to support them, then I likely should abandon my own efforts in this area. It would certainly free me up to work on more interesting development challenges in groff! Thank you again for not letting my questions lie fallow. Regards, Branden
signature.asc
Description: PGP signature
