subversion-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Bert Huijben" <b...@qqmail.nl>
Subject RE: 1.9.0-beta1 up for testing/signing
Date Fri, 06 Mar 2015 15:18:54 GMT


> -----Original Message-----
> From: Philip Martin [mailto:philip.martin@wandisco.com]
> Sent: vrijdag 6 maart 2015 15:57
> To: Branko ─îibej
> Cc: dev@subversion.apache.org
> Subject: Re: 1.9.0-beta1 up for testing/signing
> 
> Philip Martin <philip.martin@wandisco.com> writes:
> 
> > I'm seeing reproducible fails with FSFS_SHARDING=3
> >
> > $ nice make check TESTS=subversion/tests/cmdline/svnadmin_tests.py
> > FSFS_SHARDING=3
> > [1/1] svnadmin_tests.py..................................................FAILURE
> > At least one test was SKIPPED, checking
> /home/pm/sw/subversion/obj/tests.log
> > SKIP:  svnadmin_tests.py 9: 'svnadmin hotcopy' checking db/format file
> > At least one test FAILED, checking /home/pm/sw/subversion/obj/tests.log
> > FAIL:  svnadmin_tests.py 33: detect invalid changed path list entries
> > FAIL:  svnadmin_tests.py 34: detect denormalized names and name collisions
> >
> > The tests don't handle the variations in "svnadmin verify" output caused
> > by different sharding.
> 
> That's just a testsuite bug, I'm also seeing reproducible failures with
> the davautocheck and USE_HTTPV1=1:
> 
> FAIL:  lt-ra-test 6: commit callback failure
> FAIL:  externals_tests.py 57: svn copy --pin-externals wc->repos
> 
> The externals one is a client SEGV on commit.

Easy to reproduce on Windows as it immediately generated a good stacktrace via our crash handler.

Problem is fixed in r1664653.

The problem is 100% related to pinning externals, and I can confirm that it only accepts pre-HTTPv2
dav.

	Bert


Mime
View raw message