In message <1367705161.54.0.137067178029.issue2550805@psf.upfronthosting.co.za> 
<1367705161.54.0.137067178029.issue2550805@psf.upfronthosting.co.za>,
Bernhard Reiter writes:
>Bernhard Reiter added the comment:
>
>Hmm, the failure in testFilteringRangeGeekInterval was half an hour to might
>night and now, five minutes after midnight it does not happen anymore.
>Maybe it was unrelated.

You just reminded me of something I saw and couldn't reproduce when I
was adding my tx_Source patch.

I had what I think was the same failure where the lists hd different
numbers of elements for sqlite on one test run, but when I did another
test everything passed ok. I just chalked it up so something broke on
my test system.

I did a run of all tests (db and sqlite back ends) about 20 minutes
ago on the current head roundup release and I don't see any failures.