[IronPython] Thread local storage in IronPython 2.6

[Dino Viehland]

> Yes. Recalculations can be interrupted which uses Thread.Abort (the
> interrupt can happen during arbitrary user code so it is essentially
> asynchronous).

Ok, that could be the source of the leaks.  At least it's the only
thing that pops out at me while reviewing the code.  I'm not sure that
we'll actually harden the code against thread abort for 2.6.0 (this is
really tricky to do) but I can certainly make it so that it won't leak 
when it actually happens and open a bug to harden it later.