Re: [PATCH] Make AC_COMPUTE_INT public

[Ralf Wildenhues]

Hello Paolo,

Sorry for the delay.  (The next one might be even longer...)

* Paolo Bonzini wrote on Wed, Jun 28, 2006 at 03:06:47PM CEST:
> 
> >First, AC_COMPUTE_INT doesn't need to have the exact same arguments
> >that _AC_COMPUTE_INT has.  If there's some extra argument that it
> >should have for public consumption, we can add it.
> >  
> That's right, and that's what I did.  The two parameters of 
> AC_COMPUTE_INT are now the same as for AC_CACHE_CHECK.  Note that the 
> macros are m4_defun-ed, not AC_DEFUN-ed, to avoid failures in the 
> testsuite from setting an empty variable.

This is not a good reason to use m4_defun over AC_DEFUN.  Rather add
AC_COMPUTE_INT to the ac_exclude_list list in tests/mktests.sh.  I think
this public macro should be AC_DEFUNed, if only for consistency when a
third party AC_DEFUNs this (and maybe to avoid any interface changes
should this macro eventually be AU_DEFUNed).

* Paolo Bonzini wrote on Thu, Jun 29, 2006 at 06:12:13PM CEST:

> --- lib/autoconf/types.m4     19 Jun 2006 18:28:12 -0000      1.39
> +++ lib/autoconf/types.m4     29 Jun 2006 16:04:30 -0000
> @@ -670,20 +670,20 @@ AC_DEFUN([AC_CHECK_SIZEOF],

> +# The cast to long int works around a bug in the HP C Compiler
> +# version HP92453-01 B.11.11.23709.GP, which incorrectly rejects
> +# declarations like `int a3[[(sizeof (unsigned char)) >= 0]];'.

This was present before, but: for nice output, you could write
  [#] declarations like `int a3[[(sizeof (unsigned char)) >= 0]];'.


Other than these minor nits, and since nobody else complained any more,
I don't see a reason against applying your patch.

Cheers,
Ralf