On Thu, Feb 12, 2009 at 05:38:56PM -0500, James Y Knight wrote:
> I was really looking foward to 1.6's tree conflict feature. In
> particular I was looking forward to it fixing the long-standing bug
> with merge/revert/merge causing any added-by-merge files to be skipped
> by the second merge. This is by far the most irritating form of tree
> conflict I run into.
>
> The documentation in http://svn.collab.net/repos/svn/trunk/notes/tree-conflicts/
> seems to consider this case as something that should be handled, so
The notes in that directory are largely out-of-date at this point.
They have not been updated in ages.
> I'm a bit surprised to see that it apparently /has not/ been fixed by
> the tree-conflicts work as it currently exists! Did that feature just
> get dropped at some point along the way?
We decided not to treat obstructions as tree conflicts.
See issue #3161 for details:
http://subversion.tigris.org/issues/show_bug.cgi?id=3161
> I'll note that this would be a lot less irritating if issue #3101 was
> fixed. It seems pretty obvious that an unmodified add-with-history
> should be completely removed by an svn revert. If that were fixed,
> that merge sometimes "skips" files would be a lot less common...
>
> Are either of these in the plans for 1.6 final?
Yes, I think you really want issue #3101 fixed. It has not much
to do with tree conflicts. It really is about whether an obstruction
should be created during 'svn revert' of a locally added file with
history.
1.6 is right around the corner now, I don't think we'll add this as
blocker item for the release. The release has already been delayed for
more than a month due to much more serious bugs than this one.
But if a fix was made post-release it could of course be backported
to a 1.6.x point release.
Stefan
Received on 2009-02-13 04:18:33 CET