JIT breaks on feather3

[Mark Glines]

Mark Glines wrote:
> I've done some work on this, but don't have it working yet.  I have 
> created trac #18 to track this issue: 
> https://trac.parrot.org/parrot/ticket/18

After some advice on #parrot from chromatic and particle, I've put 
everything together and posted some patches to the above trac ticket.

The idea is that ManagedStruct should stash away a pointer to a free() 
function, similar to how the Pointer PMC stashes away a pointer to a 
mark() function.  Since PMC_data() is already used (for the struct being 
managed), I put the free() pointer in PMC_struct_val(), just like how 
the Pointer PMC does it.

However, it doesn't quite work - I've got a debugging print statement in 
there that prints out the value, and what it prints looks correct, but 
then it proceeds to call a bogus address and segfaults anyway.  It works 
fine under gdb, just fails when run on the comand line, and I can't 
figure out why.

Mark