gmp-4.1.2 assertion failure

Kevin Ryde user42 at
Thu Feb 12 07:30:46 CET 2004

Jason <jasonmoxham at> writes:
> note this still may not be correct for all cases as
> perhaps (cnt-GMP_NAIL_BITS)*e or bsize*e could overflow ?

Yep, they could, for instance when e==ULONG_MAX.  It'll need to be a
two-limb calculation using umul_ppmm.

More information about the gmp-bugs mailing list