On Wed, 23 Sep 2020 13:48:21 +0530, Pulkit Goyal wrote:
> # HG changeset patch
> # User Pulkit Goyal <7895pul...@gmail.com>
> # Date 1599120846 -19800
> # Thu Sep 03 13:44:06 2020 +0530
> # Node ID 30332428d8b0233053742be0c27bcb3e6499ed15
> # Parent 62404f1a66ec45ca819d411746138894dcab6e91
> # EXP-Topic merge-newnode
> merge: store commitinfo if mergetool resolved a dc or cd conflict
> diff --git a/mercurial/merge.py b/mercurial/merge.py
> --- a/mercurial/merge.py
> +++ b/mercurial/merge.py
> @@ -1608,6 +1608,27 @@ def applyupdates(
> if not complete:
> numupdates += 1
> tocomplete.append((f, args, msg))
> + else:
> + action = mresult.getfile(f)[0]
> + if action in (
> + mergestatemod.ACTION_DELETED_CHANGED,
> + mergestatemod.ACTION_CHANGED_DELETED,
> + ):
> + # the merge was conflicting and either the mergetool
> + # choosed by the user or the user themselves on the
> prompt
> + # made a choice.
> + # If the changed file was kept, at commit we end up using
> + # the same filelog. This is not completely correct
> behavior
> + # as the new file post merge also represents that a
> + # conflicting change-delete merge was resolved in it's
> + # favor. Hence we will like to create a new filenode for
> + # that. Let's store this in mergestate extras
> + repo.ui.debug(
> + b"file %s had %s conflicts and mergetool resolved
> it, storing info in commitinfo\n"
> + % (f, action)
> + )
> +
> + ms.addcommitinfo(f, {b'MERGE_REMOVAL_CANDIDATE': b'yes'})
^^^^^^^^^^^^^^^^^^^^^^^
Nit: appears that the naming convention of extras is lisp-case.
I feel this is wrong place to record dc/cd resolution. Does it cover the other
cases such as manual resolution by "hg resolve -m"? If new filenode must be
created on dc/cd, I think it's better to record that the file was dc/cd state,
not the dc/cd state was resolved by premerge.
_______________________________________________
Mercurial-devel mailing list
Mercurial-devel@mercurial-scm.org
https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel
