[MERGE] remove broken test_32_lock_wait_succeed

[Martin Pool]

Martin Pool has voted +0.
Status is now: Semi-approved
Comment:
Well, I dislike removing tests, but having them intermittently fail is 
also pretty bad.

There are (at least) two problems here:

There is a potential race:the wait_and_unlock thread may not have got 
started before the main thread tries to acquire the lock.  And if the 
machine is slow, or busy, it may timeout before the lock gets released. 
Both of them are in principle fixable, but as you say no one has done it 
yet.

I think I would prefer that you make it raise TestSkipped rather than 
just deleting the whole test though, so that at least it's visible that 
we *should* be testing it.  (And if you agree, paste in the description 
of what's wrong.)

For details, see: 
http://bundlebuggy.aaronbentley.com/request/%3C46659322.7000404%40utoronto.ca%3E