Merging code to PQM - update

[Vincent Ladeuil]

>>>>> Robert Collins <robertc at robertcollins.net> writes:

Thanks for the summary !

There are two points I haven't seen mentioned here nor in the bugs. In
some cases, avoiding the local merge before submission means we lose
some flexibility:

- we lack the ability to mark a bug as fixed by the merge itself. This
  may be minor but I consider that the bug is fixed when the reviewers
  agree with my submission and I tend to use --fixes quite late in the
  game. I.e. when I don't request the merge to occur from my submitted
  branch and merge locally instead, I use --fixes for this post-merge
  commit and it becomes the one merged by pqm.

- I've seen several submissioins fail due to NEWS conflicts. I'm sure
  some of them could have been avoided if the news_merge plugin was
  active on pqm. Yet , for other cases, the NEWS entry in the submission
  is not in the right release anyway, so may be we need a way to define
  in which release the news entries for a given proposal should be
  inserted.

Otherwise, when things were working properly, the queue feature was
really nice to use.

       Vincent