[Mulgara-dev] troublesome bug

Paul Gearon gearon at ieee.org
Fri Aug 1 01:18:27 UTC 2008


Hi all,

I'm having some issues running tests at the moment, and I'm wondering
if anyone else is seeing something similar, or if they have ideas for
me.

For a little while now, there's been a test in
ItqlInterpreterBeanUnitTest that occasionally fails due to a race.
Ronald has tracked this down, and says they he'll be getting to it
soon, so I'm not worried about that one.

Also, on and off I've had some tests fail because they couldn't
connect to a running server. This seems to be related to sockets on an
Apple, as I've occasionally had inconsistent connections with other
applications as well.

But just today I'm seeing a new one. Every couple of times through the
tests I have a small set of tests all fail with a
ClassNotFoundException for org.mulgara.resolver.Database. If I run the
tests again, then they all pass, and the next time I see the problem
it happens with a completely different set of tests. The jars and
classpaths are not changing, so I can't see why this would happen.

One thing I should point out is that this is on my MacBookPro, and it
has a TimeMachine backup. Because the tests take so long, and the
backup kicks off every hour (when I forget to switch it off), they are
often interfering with each other. This makes each task take
*significantly* longer. I believe that every failed test had one of
these contentions happening. Even so, I can't think of how it could
result in a classpath not containing a required file.

Any suggestions? I have a lot of work I want to check in, but I'm
scared to check in something with random errors.

Paul



More information about the Mulgara-dev mailing list