On Fri, Nov 12, 2010 at 07:41:39PM -0500, Seth Robertson wrote: > Not having inspected the code, it certainly appears as if the "Local > branch configured" and "Local ref configured" information is only > accidentally correct, but since the normal configuration is the case > in which it is accurate, no-one noticed the problem. Funny. :) But no, lots of people set up local branches as downstream to no branch or to some distinct remote branch. So first I guess it might be useful to explain the "push matching" behavior. Background: suppose the upstream repository has three branches: master next topic/better-frotz Now I have an idea for a new feature, so I do git checkout -b topic/xyzzy origin/master which forks a new branch set up to pull from master. Now I go about usual work, making changes to master, next, topic/better-frotz, and so on, and at the end I am on my private topic/xyzzy branch. Everything looks good, so it is time to push: git push origin This is a shortcut for git push origin : which pushes all local branches for which an upstream branch with the same name exists. "Wait!" you might ask. "Why push all branches?" It is partly historical. A lot of people have public repositories with multiple branches, and this is a behavior that has been found to be very convenient for that. "Why only branches for which an upstream branch exists?" Some work might be private. If you actually want to push everything, use "git push --mirror". "Why branches with the same name? Why not to the branch each branch pulls from?" Each branch forked from master does not necessarily contain changes that are suitable for inclusion in master immediately. In general, the upstream for each branch is often more stable rather than equally stable to it, so automatic pushing upstream does not always make sense. However, if you want the push-to-where-you-pull-from behavior, just add [push] default = tracking to your ~/.gitconfig. See the documentation for branch.<name>.merge in git-config(1) for details on that. For your test example: rather than sample output, could you describe next to each command what you expect to happen and how that differs from what happens instead? This would make it easier to find the lurking bugs. Hope that helps, Jonathan -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html