gmp-4.1.2 integer overflow

[Kevin Ryde]

Jason <jasonmoxham at btclick.com> writes:
>
> Perhaps a definition of "practical limit" 

Well, you usually give up half your address space to code, then some
to stack, then presumably you'll have at least a few numbers of
whatever big size when doing a realistic calculation.  Which makes a
peak of maybe 1/8 or 1/16 of 4Gbytes.

> Please excuse my ignorance on this point but doesn't integer overflow of a 
> buffer size calculation create a buffer overflow and therefore a possible 
> security risk  ?

Probably, but such things haven't had a great deal of attention.  For
sensible inputs there are no problems we know of, of course.