This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug testsuite/51059] FAIL: gcc.misc-tests/gcov-14.c (test for excess errors) on powerpc-apple-darwin9
- From: "nathan at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Thu, 10 Nov 2011 07:38:26 +0000
- Subject: [Bug testsuite/51059] FAIL: gcc.misc-tests/gcov-14.c (test for excess errors) on powerpc-apple-darwin9
- Auto-submitted: auto-generated
- References: <bug-51059-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51059
--- Comment #1 from Nathan Sidwell <nathan at gcc dot gnu.org> 2011-11-10 07:38:26 UTC ---
There's something strange about the semantics of weak & inline on darwin (or a
bug in the darwin linker?).
Here's the bits of the testcase:
extern int __attribute__ ((weak)) Foo ();
extern __inline int Foo ()
{
return 0; /* count(-) */
}
The program doesn't call 'Foo', but uses its address. The intent is to make
sure that there's no body of 'Foo' emitted in the executable (or at least
coverage data indicates there's no body). As the external declaration of 'Foo'
is weak, I expect the value of the 'Foo' symbol to be 0.
So, I think this test is either invalid for Darwin (and should be skipped), or
exposing an orthogonal bug (and should be xfailed). Someone who knows darwin
better than me needs to comment.