Re: Addendum to previous report

[Kevin Ryde]

Alexander Perlis <address@hidden> writes:
>
> But when I divide an intint by an int using mpn_tdiv_qr, I'm running
> into difficulty: again I'm always passing "2 limbs" as the size of
> the number by which I am dividing, and this appears to be the
> problem. A brief examination of the mpn_tdiv_qr source indicates that
> it assumes the top limb is nonzero.

Yes, that's so.

> Hi. My previous documentation bug report concerned mpn_tdiv_qr
> requiring the top limb of the divisor to be nonzero. Indeed, the
> documentation at the top of the source code tdiv_qr.c states this,
> but the TeX documentation does not. Conclusion: the previous "bug
> report" is a "documentation bug".

Yes, thanks, that'll be corrected for the next release.  The same
applies to mpn_gcd and mpn_sqrtrem.

> I ran into a similar issue with mpn_get_str, which appears to have a
> similar requirement.

Thanks, I wasn't aware of that one.  Yep, power-of-2 bases will need a
non-zero high limb for the count_leading_zeros, though a cursory
glance suggests non-power-of-2 bases might be ok.