On Tue, Jun 29, 2010 at 10:36 AM, C. Michael Pilato <cmpilato_at_collab.net> wrote:
> On 06/29/2010 10:26 AM, Brandt, Servatius (External) wrote:
>> This is an suggestion to improve the --reintegrate option to perform
>> similar checks on a 2-URL merge as it already does when used with
>> a 1-URL merge.
>
> Ahem. I think we lack this patch:
>
> Index: subversion/svn/merge-cmd.c
> ===================================================================
> --- subversion/svn/merge-cmd.c (revision 958677)
> +++ subversion/svn/merge-cmd.c (working copy)
> @@ -334,6 +334,11 @@
> }
> else
> {
> + if (opt_state->reintegrate)
> + return svn_error_create(SVN_ERR_CL_ARG_PARSING_ERROR, NULL,
> + _("Cannot specify --reintegrate "
> + "with two URLs"));
> +
> err = svn_client_merge3(sourcepath1,
> &first_range_start,
> sourcepath2,
>
> I don't believe we support --reintegrate on a two-URL merge. Unfortunately,
> the client doesn't tell you that -- it just silently falls into a
> non-reintegrate merge codepath.
...from the depths of issue #3648 '2-URL merges incorrectly
reverse-merge mergeinfo for merge target' I raise my head and weakly
utter: "+1 to this patch"
Seriously, I'm strongly opposed to expanding --reintegrate's current
narrow use case. It's been difficult enough to make reintegrate work
correctly with the subtree mergeinfo recording changes in 1.7. IMHO
expanding its purpose in life is going to lead to more complex,
special-case code, that is difficult to maintain...keep in mind this
is coming from the guy who has pretty much never said "no" to making
the merge-tracking code handle every conceivable use-case.
Paul
Received on 2010-06-29 17:38:43 CEST