[Monotone-devel] Re: What does it mean when a revision hastwobranch certs?

[Lapo Luchini]

Craig L. Ching wrote:
> Well, that brings up an interesting question.  So would it also be
> possible for a revision to have two author certs as well?  I mean, if
> you can add certs manually, what sorts of things should I be watching
> for?  I'm working on a repository browser so I'm trying to make sure I
> have all the edge cases covered.  I was going under the assumption that
> a revision would only be on one branch and have one author, but now I
> need to rethink that ;-)


Even if "mtn cert" was not there, there are some kind of revisions that
can fairly often have two authors: clean merges... one person commits a
(small) divergence, two person pull it offline, try "mtn update", need
to do "mtn merge" first, if the merge is clean, they *will* produce the
*same* revision.
When they sync, the 3+3 certs become 5 (the branch cert is identical,
only one is kept).

(if the merge is interactive they could produce two different merges,
but that's no problem: they will most probably be closer one another
than the original to-be-merged-version were)