(ccing Mike - should've done that earlier) 2013/5/7 David Kastrup <[email protected]>: > Janek Warchoł <[email protected]> writes: >> What i'm most surprised about is the commit message. There was a >> reasonable, quite helpful description in the Rietveld review (see >> https://codereview.appspot.com/7516048/) after i asked for >> explanations. But it seems that Mike used previous Rietveld >> description as his commit message. I don't quite see what kind of >> workflow could lead to such results, but apparently Mike's one does. >> Maybe it's time to move to gerrit for codereviews - from what i've >> read it seems to help with such issues... > > Short of last-minute rewordings it should, as the reviews with Gerrit > are a git path including the commit messages. Many of Mike's commit > messages appear like they are written from scratch at the time of the > commit, sometimes not containing anything that could help matching them > to the corresponding issue/review. Gerrit would not help against that, > but it may be harder to pick a workflow where one would do this.
Also, I remember seeing a setup in which merging the actual changes was done through gerrit - not by the author of the patch in his private repo. This should - to some extent - guarantee that what gets pushed is identical to what was reviewed. Janek _______________________________________________ bug-lilypond mailing list [email protected] https://lists.gnu.org/mailman/listinfo/bug-lilypond
