On May 4, 3:46 pm, Cactus <rieman...@gmail.com> wrote:
> On May 4, 2:55 pm, Marc <marc.gli...@gmail.com> wrote:
>
> > On May 3, 1:00 am, Bill Hart <goodwillh...@googlemail.com> wrote:
>
> > > I think the unsigned long int is correct this time. I see no reason
> > > why the remove function should return a number of bits.
>
> > Well, if the factor being removed is 2, the number of times you can
> > factor it out of a number with n+1 bits like 2^n is...
>
> It looks like another document/code inconsistency that also exists in
> GMP 5.0:
>
> GMP:
> doc: mp_bitcnt_t mpz_remove (mpz t rop, mpz t op, mpz t f)
> dec: __GMP_DECLSPEC unsigned long int mpz_remove __GMP_PROTO
> ((mpz_ptr, mpz_srcptr, mpz_srcptr));
> code: mp_bitcnt_t mpz_remove (mpz_ptr dest, mpz_srcptr src, mpz_srcptr
> f)
>
> MPIR:
> doc: mp_bitcnt_t mpz_remove (mpz t rop, mpz t op, mpz t f)
> dec: __GMP_DECLSPEC unsigned long int mpz_remove __GMP_PROTO
> ((mpz_ptr, mpz_srcptr, mpz_srcptr));
> code: unsigned long int mpz_remove (mpz_ptr dest, mpz_srcptr src,
> mpz_srcptr f)
>
> GMP also has a similar function in mpn that MPIR doesn't have:
>
> mp_bitcnt_t mpn_remove (mp_ptr wp, mp_size_t *wn, mp_ptr up, mp_size_t
> un, mp_ptr vp, mp_size_t vn, mp_bitcnt_t cap)
>
> but its not a part of the public interface.
>
> Brian
It turns out that there were several inconsistencies introduced into
the GMP code when the mp_bitcnt_t type was introduced.
Chris picked up two of these but there was a third one as well.
The GMP folk have now picked this up and have corrected all of them in
the last few days.
In order to maintain maximum interface compatibility between GMP and
MPIR I will, unless anyone objects, add these three changes to the
MPIR repository. But I cannot do anything about the MPIR
documentation as I don't have the tools needed to do this.
They all consist of a subtitution of the mp_bitcnt_t type for the
unsigned long type either as a function parameter or as a return
value.
Brian
--
You received this message because you are subscribed to the Google Groups
"mpir-devel" group.
To post to this group, send email to mpir-devel@googlegroups.com.
To unsubscribe from this group, send email to
mpir-devel+unsubscr...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/mpir-devel?hl=en.
