I think Mercurial also misses the mark for uploading patch sets which might evolve / change over a few iterations. The workflow usually looks like one user is working on only one set of changes at a time, and in order to show them to upstream they want to publish them somewhere. If these changes get edited several time, and you use a traditional Mercurial repository you end up with lots of zombie tips which you absolutely don't care about, and unless you 'hg strip' very assiduously, they keep reinfecting repositories.
Yes, I know it's not a supported workflow. No, sometimes, this is what I want to do.
I don't think there's a direct equivalent, but you can get the ability to reset to the previous state using bundles. hg help bundle, hg help unbundle. Whenever you do an amend or a rebase, it tells you where it saved the bundle
3 comments
[ 4.8 ms ] story [ 13.0 ms ] threadYes, I know it's not a supported workflow. No, sometimes, this is what I want to do.