Re: checked integer arithmetic

[Bruno Haible]

Paul Eggert wrote:
> Come to think of it, I suppose we should change xalloc_oversized to 
> report an overflow if the resulting size would be greater than 
> PTRDIFF_MAX. That should catch more potential problems in Gnulib and in 
> Gnulib-using code.
> ...
> Here is an example of why arrays larger than PTRDIFF_MAX bytes can cause 
> real problems.

So, an attempt to allocate a memory chunk of 2.5 GB size on 32-bit Linux/x86
is ill-fated already? That is, SIZE_MAX = 4 GB - 1 gives us a fake illusion
that allocations > 2 GB would be OK, and in fact they are not OK, so it's
in fact ssize_t or ptrdiff_t that matters?!

Bruno