Right.

I'm tabling this work for at least a few days for other reasons, but 
I'll proceed using some variation of way 1 you proposed, and a single GC 
per invocation.

Thanks!
Daniel

Andreas Raab wrote:
> Daniel Vainsencher wrote:
> 
>> Another question. My problem is not with the GC per se, just with the 
>> fact that it happens many times when reshaping Behavior. So a trivial 
>> solution avoiding concurrency related problems is to expand the 
>> critical section (have one for each call to CB, instead of one for 
>> each class affected), and perform GC at its end.
> 
> 
> Yes, I thought about this, too. There are severe problems with it since 
> in some situations classes get reshaped multiple times during 
> recompilation (such as when crossing the class/meta class hierarchy) and 
> then things break almost immediately. It is also a risky solution since 
> the broken instances survive much longer and nobody can say for sure 
> what effects any of these might have.
> 
> Cheers,
>  - Andreas
>