[PATCH] support for mingw-w64

Ozkan Sezer sezeroz at gmail.com
Fri Sep 11 14:11:08 CEST 2009

On Fri, Sep 11, 2009 at 3:05 PM, Torbjorn Granlund <tg at gmplib.org> wrote:
> nisse at lysator.liu.se (Niels Möller) writes:
>  Torbjorn Granlund <tg at gmplib.org> writes:
>  > * We unfortunately cannot use intptr_t unconditionally, since it is not
>  >   available everywhere we want GMP to work; it is a C99 feature.
>  >
>  >   I wonder if size_t could be used instead?  While perhaps not elegant,
>  >   it should work.  Does anybody see any problem?
>  Another alternative is ptrdiff_t. My understanding is that this is
>  fairly portable in both theory and practice.
> My undertanding is that sizeof(ptrdiff_t) == sizeof(size_t) should hold
> true on any ISO C90 or C99 system.  They only differ in signedness.
> Perhaps some usages of ptrdiff_t could make use of the sign ("which
> sub-object within the objext comnes first in memory?"), but I doubt GMP
> will care.
> I suggest that we either use size_t, or if people think ptrdiff_t make

I am OK with these.

> more sense, use that.  But we might as well use a typedef in gmp-impl.h:
> typedef size_t gmp_intptr_t;

Or, how about doing a AC_CHECK_TYPE(intptr_t, size_t)
in configure.in?  (I didn't test this one bit, though..)

> --
> Torbjörn


More information about the gmp-devel mailing list