Re: vc-resolve-conflicts is not flexible enough for alternate SCM systems

[Stefan Monnier]

>>>>> "Jonathan" == Jonathan Kamens <jik@kamens.brookline.ma.us> writes:
> The supported mechanism for resolving conflicts after merges is not
> flexible enough; it has two problems:
> 1) It can't cope with conflict markers other than those produced by
>    rcsmerge or diff3;

What other markers have you seen ?
(of course, knowing how Perforce uses gratuitously non-standard
diff-hunk-headers, I wouldn't be surprised if they also gratuitously used
non-standard conflict markers).

> 2) It can't cope with SCMs that provide ancestor information in their
>    conflict markers.

Are you sure ?  I seem to remember changing VC to use smerge-mode which
definitely supports ancestor information (I even use a locally-hacked
CVS executable which adds that ancestor info, even tho I sadly never
cleaned up this patch enough to send it to the diff3 and CVS teams).

`vc-resolve-conflicts' should simply be removed and was only kept
because Andre wanted the new VC to also work under Emacs-20 (where
smerge-mode was not available).


        Stefan

PS: Actually, I find the automatic call to smerge-ediff annoying.
    I'd rather let the user call smerge-ediff manually from the Smerge
    menu if she wants to.