Skip to content
Snippets Groups Projects

prevent fold from creating merge commits with pX being an ancestor of pY

Open Anton Shestakov requested to merge topic/stable/fold-merges into branch/stable
1 unresolved thread

Merge request reports

Loading
Loading

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
  • Pierre-Yves David assigned to @av6 and unassigned @marmoute

    assigned to @av6 and unassigned @marmoute

  • Pierre-Yves David added 11 commits

    added 11 commits

    • 428330c9...1b5afe8b - 9 commits from branch branch/stable
    • ecb1eca8f942 - tests: prepare test-fold.t for the upcoming changes
    • b096dc9cfce7 - fold: don't create merge commits with pX being an ancestor of pY (issue6332)

    Compare with previous version

  • Anton Shestakov added 15 commits

    added 15 commits

    • b096dc9c...d55e83fa - 13 commits from branch branch/stable
    • 5670c090f8db - tests: prepare test-fold.t for the upcoming changes
    • 836384d70c36 - fold: don't create merge commits with pX being an ancestor of pY (issue6332)

    Compare with previous version

  • Anton Shestakov added 4 commits

    added 4 commits

    • 31a72d43d20f - 1 commit from branch branch/stable
    • ec6d455130ea - foldcheck: return p1 in addition to p2 for folded commits
    • a357bf87b1ed - tests: prepare test-fold.t for the upcoming changes
    • 7675fbb4de60 - fold: don't create merge commits with pX being an ancestor of pY (issue6332)

    Compare with previous version

  • Anton Shestakov added 2 commits

    added 2 commits

    • eff014af82a9 - fold: don't create merge commits with pX being an ancestor of pY (issue6332)
    • 64c9248f2c3d - fold: allow folding merge commits in a way that avoids issue6332

    Compare with previous version

  • Anton Shestakov assigned to @marmoute and unassigned @av6

    assigned to @marmoute and unassigned @av6

  • Anton Shestakov resolved all threads

    resolved all threads

  • As far as I understand the goal of this commit is to prevent the creation of oedipus-merge through fold. However such oedipus-merge are allow in some case. For example when merging between branch. So the current check seems too harsh as I understand it.

    Moreover I believe a better experience would be to simply drop the extra parent (in the case we could create an illegal oedipus-merge).

    (to make things simpler, they are also people we would like to be able to create oedipus-merge).

    In all case, unless I misunderstood something I don't think we should take this MR as is.

  • Pierre-Yves David assigned to @av6 and unassigned @marmoute

    assigned to @av6 and unassigned @marmoute

    • Author Maintainer

      However such oedipus-merge are allow in some case. For example when merging between branch.

      So if given this:

        o    3: default
        |\
        | o  2: stable
        | |
        o |  1: default
        |/
        o  0: default

      and I fold 1+3, then I should get

      o    4: default
      |\
      | o  1: stable
      |/
      o  0: default

      ?

      What other cases are there?

      Moreover I believe a better experience would be to simply drop the extra parent (in the case we could create an illegal oedipus-merge).

      Dropping one of the parents is what the last patch does.

    • I cannot hink of other case from the top of my head.

      Dropping one of the parents is what the last patch does.

      This is the second time you says this, but I have no idea which patch you are referring to.

    • Author Maintainer

      which patch you are referring to

      The one that does p2 = repo[None].

    • Please register or sign in to reply
  • Anton Shestakov assigned to @marmoute and unassigned @av6

    assigned to @marmoute and unassigned @av6

  • Pierre-Yves David assigned to @av6 and unassigned @marmoute

    assigned to @av6 and unassigned @marmoute

  • Please register or sign in to reply
    Loading