On 9/19/2010 9:40, Christoph HÃger wrote: > Hi all, > > since I keep offlineimap the same version for the latest stable (that I > got my hands on) + devel versions, my fedpkg workflow looks like: > > 1. master> build package > > 2. f14> git pull origin master&& fedpkg push&& fedpkg build ... > > 3. f13> git pull origin master&& fedpkg push&& fedpkg build ... > > I think this is how git should be used in that case (no difference > between all branches). But this means taking the release number from > step 1 through all steps (e.g. building offlineimap-6.2.0-2.fc14 without > ever building offlineimap-6.2.0-2.fc14). Is that ok in terms of fedora > package policies? It sounds like you're doing more work than you have to if all the branches are the same. Why not just run ``git merge'' on the other branches? Here's what I generally do: * git checkout master; <edit files>; git commit ...; git push * git checkout f14; git merge master; git push * git checkout f13; git merge master; git push * (...and so on for the other branches) * koji build --nowait dist-f15 'git://pkgs.fedoraproject.org/<pkgname>?#199d3785f44e4f4d0005e2670b1e58179b290c95' * koji build --nowait dist-f14-updates-candidate 'git://pkgs.fedoraproject.org/<pkgname>?#199d3785f44e4f4d0005e2670b1e58179b290c95' * (...and so on for the other branches) Koji doesn't care what branch a given hash comes from since either you or fedpkg will specify a target when calling it, so having several branches at the same commit is not a problem. -- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/devel
