On Wed, 2008-11-12 at 17:22 -0500, Karl Fogel wrote:
> I just ran the tests (FSFS only) over all RA layers, using today's
> head of Subversion (r34160) and Apache HTTPD 2.2.9. All layers
> passed, except for ra_serf (also today's head, r1202), which had four
> failing tests:
>
> FAIL: commit_tests.py 28: commit deletion of out-of-date file or dir
This one gets a '404 Not Found' which looks like it ought to be at least
wrapped by an 'out of date' error message.
> FAIL: update_tests.py 47: tree conflicts on update 1.2
> FAIL: switch_tests.py 32: tree conflicts on switch 1.2
These two fail an assertion in recent tree-conflicts code.
> FAIL: tree_conflict_tests.py 2: up/sw file: del/rpl/mv onto modify
This one's an 'abort'.
As three of these are tree-conflict tests, and the other is in the same
sort of area, I had that much of a look at the test log. The fact that
it all works with Neon says to me that these are all bugs in (RA-)Serf.
Whilst Serf is no doubt heading to be a good thing, I have to be
selective in what I pay attention to, and Serf is off my radar. I just
assume that its client-visible behaviour should match Neon's in almost
every respect, and where it doesn't that's a bug in (RA-)Serf, so I
don't intend to investigate these further myself. However, if somebody
wants to investigate and wants help in understanding or checking the
tree-conflicts aspect of it, I would be glad to help.
I just wanted to say this to give other folks a chance to check we're
"on the same page".
- Julian
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-11-13 12:33:35 CET