Well, RHEL-5 is now in late phase where only serious problems and security bugs get fixed and this bug doesn't seem too serious to me, since it is even hard reproducible, so hardly hit by customers as well. There is also almost no chance mysql-5.0 gets updated, since upstream doesn't support this version any more. So, closing this bug for now; we can re-open it if something from above changes.
thanks for the evaluation; marking the test not to run on RHEL5 then
(In reply to Karel Volný from comment #2)
> marking the test not to run on RHEL5 then
well, on second thought ... we should have the "expectedness" integrated into our workflows yet, so maybe it'd be more proper to keep this test and just let the failures be waived based on the fact that it links to a bug that was decided not to be fixed in RHEL5 ... let's try and re-enable