[Ur] Could we prevent most segmentation faults caused by stack overflows this way?

[Adam Chlipala]

Marc Weber wrote:
> Would something like this work reasonable?
> Its using pthread to get stack size (this was used to tell the ruby
> garbage collector which memory to scan. After telling about it the
> ruby gc related segfaults were gone)
>    

I would certainly be more comfortable using an established 
stack-overflow-catching library that has been tested for portability.  
Does anyone know of such a thing?