On Mon, 2009-04-06 at 16:36 -0400, Dale Worley wrote:
> On Mon, 2009-04-06 at 16:20 -0400, Scott Lawrence wrote:
> > On Mon, 2009-04-06 at 15:57 -0400, Dale Worley wrote:
> > > On Mon, 2009-04-06 at 15:01 -0400, Scott Lawrence wrote:
> > > > This has been going on for a few weeks, roughly corresponding to when I
> > > > switched to my new laptop on F10 (as if there weren't already enough
> > > > variables).  I've at least occasionally seen the same failures in the
> > > > build server logs.
> > > 
> > > IIRC, I'm seeing timing problems in these tests also, and I'm on FC8.
> > 
> > Unless I hear either strong objections or a volunteer to fix these, I'm
> > going to start disabling them when they crash on me and checking that
> > in.
> 
> Perhaps we should schedule fixing the tests at part of 4.0?

If we have time to work on them anyway, I certainly don't object, but I
don't think we should spend any time on them in favor of
release-blocking items.

I am fairly sure that the problem is with the tests.  These are, I
suspect, and example of a problematic sort of unit test - they actually
start a user agent (or more than one) and send messages back and forth.
This makes them sensitive to port usages and timing issues.  If they are
not very very carefully constructed, it's easy to end up with a flakey
test.

I try to define classes so that that sort of test is not needed - just
synthesize messages and pass them to the active methods of the objects,
but never start the threads and use the network.  Not always possible,
though...

_______________________________________________
sipx-dev mailing list
[email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-dev
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev

Reply via email to