Re: [RFC] Cleaning up die() error messages

> "H. Peter Anvin" <hpa@zytor.com> writes:
> 
>>I thought about this, and probably the sanest way is to wrap malloc() 
>>with something that creates a linked list of allocations.  If we abort, 
>>we can unwind the linked list and free all allocations.
> 
> You can free() the allocated memory with something like that,
> but it is likely that the aborted function would have already
> created some linked structure out of those memory blocks, and
> cleaning up that you would need the real exception handling
> wouldn't you?
>