The following test fail on all phase enabled version > --- a/tests/test-phases-exchange.t > +++ b/tests/test-phases-exchange.t > @@ -94,10 +94,23 @@ push from alpha to beta should update ph > | > o 1 public a-B - 548a3d25dbf0 > | > o 0 public a-A - 054250a37db4 > > + $ hg push -r 2 ../beta > + pushing to ../beta > + searching for changes > + no changes found > + [1] > + $ hgph > + @ 3 draft a-D - b555f63b6063 > + | > + o 2 public a-C - 54acac6f23ab > + | > + o 1 public a-B - 548a3d25dbf0 > + | > + o 0 public a-A - 054250a37db4 > $ hg push ../beta > pushing to ../beta > searching for changes > no changes found > [1] The phase on 2 is not changeset and 2 stay draft. The push below without --rev works as expectable.
Confirmed on 2.6.3 too.
Fixed by http://selenic.com/repo/hg/rev/b396032e1bb8 Pierre-Yves David <pierre-yves.david@logilab.fr> phase: properly compute ancestors of --rev on push (issue3786) Now that discovery is working on unfiltered changeset, I had a good occasion to look at that bug again. This let me realise that a trivial node vs rev comparision was the cause of this two years old bugs? Happy second birthday phases! (please test the fix)