Re: AC_TYPE_UINT8_T and co

[Patrick Welche]

On Wed, May 30, 2007 at 02:48:12PM +0200, Stepan Kasal wrote:
> And yes, it sounds inconsistent.  IMVHO the ac_cv_c_ prefix should be
> changed to ac_cv_type_ here.  What do others think?

I think that

- changing ac_cv_c_ to ac_cv_type_ and
- changing AC_CHECK_TYPES to test for ac_cv_type_ != no rather than = yes

would then make

AC_TYPE_INT64_T
AC_CHECK_TYPES([int64_t])

work as I would expect - if that is what it intended(!) In my ficticious
long long int exists, but not int64_t case:

AC_TYPE_INT64_T doesn't find int64_t, so #defines it to long long int, and
sets ac_cv_type_int64_t='long long int'.

AC_CHECK_TYPES does a AC_CACHE_CHECK for ac_cv_type_int64_t which
is not "no" so HAVE_INT64_T is defined, and I can happily use
#ifdef HAVE_INT64_T in my example.

How does that sound?


Cheers,

Patrick

typepat.diff
Description: Text document