Re: AC_HEADER_STDBOOL fails with clang

[Eric Blake]

On 08/22/2010 08:59 PM, Anders Kaseorg wrote:
> But now we have clang.  clang does not (and apparently will not [2])
> support this extension, so although it has a working<stdbool.h>, the test
> incorrectly fails with CC=clang:

Thanks for the report.

> @@ -655,8 +655,7 @@ AC_DEFUN([AC_HEADER_STDBOOL],
>             this with GCC, where it should work, to detect more
>             quickly whether someone messes up the test in the
>             future.  */
> -        char digs[] = "0123456789";
> -        int xlcbug = 1 / (&(digs + 5)[-2 + (bool) 1] ==&digs[4] ? 1 : -1);
> +        int xlcbug = 1 / (&((char *)0 + 5)[-2 + (bool) 1] ==&((char *)0)[4] ? 
> 1 : -1);
>   #     endif

That is a conditional check, inside:

#       if defined __xlc__ || defined __GNUC__

So does clang define __GNUC__?  That's a lie, if it does not support the 
same extensions as gcc.

Is there another preprocessor macro that we can reliably use to detect 
if clang is performing the compilation, to skip this code in that case 
(since we know that clang neither has the AIX bug nor the gcc extension)?

> By the way, I note also that the test also fails with CC=gcc CFLAGS='-Wall
> -Werror', although for a different reason (error: the address of 's' will
> always evaluate as 'true').

We already know that using -Werror during configure is asking for 
problems, and our recommendation has long been that users should test 
whether -Werror exists and works independently of blindly sticking it in 
CFLAGS.  There's no way we can reasonably make autoconf be 100% friendly 
with the use of -Werror behavior across all versions of gcc.

At any rate, thanks for catching the error that we meant to use division 
by 0 (rather than our typical use of negative array size bounds) as the 
cause of compilation error on a buggy compiler; you are correct that the 
(1 / (cond ? 1 : -1)) needs to be fixed to (1 / (cond ? 1 : 0)), whether 
or not we also get confirmation that changing the expression to 
something that clang likes will still reject the buggy AIX compiler.

--
Eric Blake   address@hidden    +1-801-349-2682
Libvirt virtualization library http://libvirt.org