RE: RA-DAV tests acting *very* strangely.

> From: Philip Martin
> Sent: Friday, March 07, 2003 1:40 PM

[...]
> I have just run the regression tests over DAV and they completed
> successfully. I am using Subversion r5237, Apache/APR/APRUTIL HEAD
> from yesterday and BDB 4.0.14 on Linux.
>
> I am having one problem which I think is related to the cutting edge
> Apache as it has only appeared since I updated Apache yesterday. I
> see delays where the client is waiting for Apache and Apache doesn't
> appear to be using any CPU. It's usually a checkout command that
> blocks, I have attached gdb to the client and the stack show it
> blocked in select called from svn_ra_dav__get_props, from
> svn_ra_dav__get_latest_revnum.

Ok, now this is getting weirder by the moment :(

I was just in the process of reopening and adding this to issue 995
(after a conversation I had with CMike right before he called it
 quits for today, or yesterday, if you will):

"The fix for this issue, rev 5202, uncovered a bug for which details can
 be found here:

   http://subversion.tigris.org/servlets/ReadMsg?list=dev&msgNo=32737

 After a night worth of debugging from Mike and a little bit of input from
 myself, without resolve, it seems wise to revert rev 5202 and not delay
 the release of 0.19.

 I've created a branch with the patch applied, so someone can take a
 fresh look at it:

   http://svn.collab.net/repos/svn/branches/issue-995-dev/

 Moving this issue to the next milestone."

I haven't taken these steps yet and with this report from Philip I'm not
going to.

Doing the lengthy davcheck myself:

Subversion rev 5237, Apache 2.0.44 (bundled APR and APR-UTIL), Python 2.2.2

All tests ...PASS

Mike, could this be something local to your system?

Sander

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Mar 7 14:32:03 2003