[PATCH, libffi]: Fix cls_longdouble_va.c failure on ix64, alpha

[Uros Bizjak]

On Thu, Dec 3, 2009 at 2:42 PM, Dave Korn
<dave.korn.cygwin@googlemail.com> wrote:

>> According to [1]: The syntax for a format placeholder is
>> "%[parameter][flags][width][.precision][length]type".
>>
>> 2009-12-03  Uros Bizjak  <ubizjak@gmail.com>
>>
>>       * testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
>
>  I thought "Didn't we just do this one already?", and sure enough, this is a dup!
>
>    http://gcc.gnu.org/ml/gcc-patches/2009-11/msg01560.html

Yeah, and we both failed to notice that the test now also passes on x86 targets.

Fixed by attached patch.

2009-12-03  Uros Bizjak  <ubizjak@gmail.com>

	* testsuite/libffi.call/cls_longdouble_va.c (main): Fix format string.
	Remove xfails for x86 linux targets.

Uros.
-------------- next part --------------
Index: testsuite/libffi.call/cls_longdouble_va.c
===================================================================
--- testsuite/libffi.call/cls_longdouble_va.c	(revision 154941)
+++ testsuite/libffi.call/cls_longdouble_va.c	(working copy)
@@ -42,9 +42,9 @@ int main (void)
 	args[2] = NULL;
 
 	ffi_call(&cif, FFI_FN(printf), &res, args);
-	// { dg-output "7.0" { xfail i*86-*-linux-* x86_64-*-linux-* } }
+	// { dg-output "7.0" }
 	printf("res: %d\n", (int) res);
-	// { dg-output "\nres: 4" { xfail i*86-*-linux-* x86_64-*-linux-* } }
+	// { dg-output "\nres: 4" }
 
 	CHECK(ffi_prep_closure_loc(pcl, &cif, cls_longdouble_va_fn, NULL, code) == FFI_OK);