Daniel Wood <dwood@salesforce.com> writes:
> Does the original version of my stress test not repro the problem on 9.2?
[ tries it ... ] No, it doesn't, or at least the MTBF is a couple orders
of magnitude better than on 9.3.
Another odd thing (seen with my short version as well as your original)
is that 9.3/HEAD run the test case enormously faster than 9.2 and 9.1
do. The older versions seem to spend a lot of time sleeping, which
I don't understand.
> Why does LockAcquireExtended() test for "nLocks == 0" in the "if
> (dontWait)" block before calling RemoveLocalLock()?
Looks like a useless test to me --- we wouldn't be here at all if nLocks
had been positive to start with, and there's nothing in between that
could raise the count. On the other hand, removing a LOCALLOCK that
did have positive count would be disastrous. Maybe what would be
more appropriate is an Assert(nLocks == 0) in RemoveLocalLock().
regards, tom lane
