On 12/19/05, svntest@jaa.iki.fi <svntest@jaa.iki.fi> wrote:
> At least one test FAILED, checking /svntest/obj-sh/tests.log
> FAIL: svnsync_tests.py 1: copy and modify
> FAIL: svnsync_tests.py 2: copy from previous version and modify
> FAIL: svnsync_tests.py 3: copy from previous version
> FAIL: svnsync_tests.py 4: modified in place
> FAIL: svnsync_tests.py 5: tag empty trunk
> FAIL: svnsync_tests.py 6: tag trunk containing a sub-directory
> FAIL: svnsync_tests.py 7: tag trunk containing a file (#2)
> FAIL: svnsync_tests.py 8: tag trunk containing a file
> FAIL: svnsync_tests.py 9: tag with a modified file
> At least one test was SKIPPED, checking /svntest/obj-sh/tests.log
All these are complaining about ProcessTerminatedBySignal, which
implies a crash of some sort, but I can't reproduce it here. Some 64
bit only thing?
Anyway, I am seeing some failures running these tests over DAV, but
it's not crashes, it's an artifact of ra_dav requiring intermediate
directories to be opened before you can create a new subdirectory of
them. I'll figure out the correct fix tomorrow.
FWIW, this and a previous problem DAV had with reqiring files to be
correctly closed are starting to make me wonder if the best solution
is to make the underlying svn_repos_replay2 more correct. Currently
it skips any editor calls that aren't explicitly needed for svnadmin
dump, perhaps that needs to change for it to be correctly used in the
more general case. Or alternatively I can just add more smarts to the
libsvn_ra_dav impl of replay, since the other RA layers seem perfectly
fine with this sort of omission. I need to think some more about
this.
-garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Dec 20 06:52:25 2005