New mini-gmp macro GMP_CMP?

[Niels Möller]

"Marco Bodrato" <bodrato at mail.dm.unipi.it> writes:

> Il Mer, 23 Novembre 2016 11:28 pm, Niels Möller ha scritto:
>>   int
>>   mpz_cmp_si (const mpz_t u, long v)
>>   {
>>     if (v >= 0)
>>       return mpz_cmp_ui (u, v);
>>     else if (u->_mp_size >= 0)
>>       return 1;
>>     else
>>       return - mpz_cmpabs_ui (u, GMP_NEG_CAST(mp_limb_t, v));
>>   }
>
> I like this version.

Maybe, I find this use of mpz_cmpabs_ui a bit strange. Something with 

  mpz_cmpabs_ui (u, GMP_ABS_CAST(mp_limb_t, v))

would look more symmetric.

> And, IMO, we shouldn't use a multiplication... ;-)

I think what I intended was a conditional negation, which could be done
with (x ^ sign) - sign (where sign should be either 0 or ~0). But that's
maybe a bit too obscure.

> long int
> mpz_get_si (const mpz_t u)
> {
>   if (u->_mp_size < 0)
>     return (long) (- u->_mp_d[0] | GMP_LIMB_HIGHBIT);
>   else
>     return (long) (mpz_get_ui (u) & ~GMP_LIMB_HIGHBIT);
> }

Looks fine to me. I find the result on overflow a bit odd (but I think
it's the same as gmp). If I understand it correctly, the value v of
mpz_get_si (u), is defined by the conditions

  sgn(v) == sgn(u) and 
  v == u (mod GMP_LIMB_HIGHBIT)

Is that right?

Regards,
/Niels

-- 
Niels Möller. PGP-encrypted email is preferred. Keyid 368C6677.
Internet email is subject to wholesale government surveillance.