t...@gmplib.org (Torbjörn Granlund) writes:
> Emmanuel Thomé <emmanuel.th...@inria.fr> writes:
>
> I've personally always used the mpz_ptr and mpz_srcptr in functions that
> take mpz arguments, taking the liberty to use this undocumented type.
>
> I suppose we could declare these. I'd like to hear from the rest of the
> GMP "core team".
My fairly strong opinion is that we should document them. They are
important for user code that needs to handle pointers to mpz_t values.
One simple use case:
void foo (mp_srcptr a, mp_srcptr b)
{
if (mpz_cmp (a, b) < 0)
{
mp_srcptr t = a;
a = b;
b = t;
}
... code that requires a >= b ...
}
This gets a lot more awkward if the function has to be declared using
documented types, as
void foo (const mpz_t a, const mpz_t b)
Simplest alternative I see would be to copy into local non-const mpz_t
variables and use mpz_swap, and that's an extra copy I don't want to
force users to make.
Functions returning pointers to an mpz_t is another important case
(mpz_roinit_n is one (unusual) example, spotted by Emmanuel). mpq_numref
and mpq_denref too, if they were implemented as (inline) functions
rather than macros.
I don't think need for handling pointers is that common in GMP user
code, but not so obscure that it should be ignored and unsupported.
Regards,
/Niels
--
Niels Möller. PGP-encrypted email is preferred. Keyid 368C6677.
Internet email is subject to wholesale government surveillance.
_______________________________________________
gmp-devel mailing list
gmp-devel@gmplib.org
https://gmplib.org/mailman/listinfo/gmp-devel
