Memory allocators / STL

[Benjamin Kosnik]

> I offer many thanks to the anonymous
> person that radically improved the output of gcc -v to allow for a
> hint of this type of problem to be seen (Alexander?).

Actually, Phil Edwards. Agreed, invaluable.

This problem only showed up on multiple-processor machines, so one way 
systems should always work. 

There's not really anything more that can be done about this, without 
getting a test case from the mysql people who are having problems. I'd 
think it would not be hard to modify the posted test case to produce 
their bug. 

Or the bug doesn't exist, and it's some problem in their code.

-benjamin