> I mean that my linked repo didn’t have all changesets from > hg.python.org/cpython. I don’t think it should matter, but I don’t know > why the diff was empty, so I thought this information might help. [...] > See http://bugs.python.org/file21398/c43e264256e4.diff : this > corresponds to a merge I did from default into my branch, > http://hg.python.org/sandbox/merwok/rev/c43e264256e4 . It should have > been a diff corresponding to my branch - cpython’s main repo default > branch head IIUC.
I see what happened: my revspec couldn't cope with the fact that you branched fix5845 off rlcompleter-config. It was looking for the branchpoint of that branch, and taking that as a potential base. However, this didn't give a revision from default, but the latest revision in rlcompleter-config as a candidate base. It then compared that to the second parent of the latest merge into fix5845, which was from the default branch, and tried to compute the maximum. Since these two revisions are not ordered, max() returned some junk revision (I don't understand what Mercurial does in this case). In any case, I have now changed the revspec to use Baptiste's idea, which is really smart, which also allows me to drop the sourcebranch information. Let's see whether it is correct this time. I also added quotes into the branch name, so hyphens should be allowed now. Regards, Martin _______________________________________________ Python-Dev mailing list Python-Dev@python.org http://mail.python.org/mailman/listinfo/python-dev Unsubscribe: http://mail.python.org/mailman/options/python-dev/archive%40mail-archive.com
