On Wed, 2008-11-26 at 06:56 +0100, Neels J Hofmeyr wrote:
> Hyrum K. Wright wrote:
> > XFAIL: merge_tests.py 19: merge should skip over unversioned obstructions
>
> The output expects foo to be `Skipped', but the real output has a tree-conflict:
>
> [[[
> CMD: svn merge -r1:2
> file:///arch/elego/svn/trunk/subversion/tests/cmdline/svn-test-work/repositories/merge_tests-19/A/B/F
> svn-test-work/working_copies/merge_tests-19/A/C --config-dir
> /arch/elego/svn/trunk/subversion/tests/cmdline/svn-test-work/local_tmp/config
> --password rayjandom --no-auth-cache --username jrandom <TIME = 0.056066>
> --- Merging r2 into 'svn-test-work/working_copies/merge_tests-19/A/C':
> A svn-test-work/working_copies/merge_tests-19/A/C/Q
> A svn-test-work/working_copies/merge_tests-19/A/C/Q/bar
> C svn-test-work/working_copies/merge_tests-19/A/C/foo
> Summary of conflicts:
> Tree conflicts: 1
> ]]]
>
> C is an empty dir and receives the merge that adds some files. Only that
> there exists an unversioned file `foo' in C, which obstructs the addition of
> `foo' by the merge.
>
> It looks to me like this should *not* be a tree-conflict! `foo' is
> obstructed, it has nothing to do with conflicting tree changes. If foo was
> locally added, that would be different.
>
> I think we should make all the obstructions by unversioned files/dirs skip
> as tested by this test. Would that be a problem?
I agree. That is the right thing to do. (Issue #3161.) Steve did this
for us in the "update" code, but the "merge" code still contains some
cases where an obstruction is detected as a tree conflict.
I have just seen your patch to help do this, in the "tree-conflict:
double-tree-conflict bug!" thread. I'll take a look later.
- Julian
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-11-26 13:28:25 CET