Giuseppe Scrivano wrote: > Micah Cowan <mi...@cowan.name> writes: > >> I think the main thing with migrating to Savannah is that there are a >> lot of "developers" who would have commit access, where there's really >> only a couple people in there who are active developers. So you may want >> to prune that list (see who's actually in the changelog in the first >> place or something?) before using Savannah for repo access. > > Thanks, I will take care of this.
Let me know when you do this, and I'll remove the mainline from hg.addictivecode.org. I'll leave some of the others around; jbuty is still relevant at least, just needs updates and merging to mainline. >> Of course, git is already required (for maintainers, not developers) to >> update a local copy of gnulib to run gnulib-tool --update with: Wget >> uses gnulib, but checks in the results. I did this so I could control >> the stability of the gnulib we were using, and not have any surprising >> things change out from under me. > > I think it will be cleaner to use gnulib in the same way as other > projects are doing it, not checking in the results but using a > "bootstrap" script. To force a specific revision of gnulib, a git > submodule can be used. Yeah, but then you need to communicate which particular submodule the other folks should use (and they all have to maintain git and gnulib now too, though I suppose that problem is lessened if wget is moved to git; but then there's the issue of Windows support). -- Micah J. Cowan http://micah.cowan.name/
