Failures in tests for obj-c++....

[David Ayers]

David Ayers wrote:

> 2005-06-07  David Ayers  <d.ayers@inode.at>
> 
> 	* archive.c, init.c, selector.c: Include hash.h.
> 	* archive.c, class.c, encoding.c, gc.c, hash.c, hash_compat.c,
> 	init.c, misc.c, nil_method.c, objects.c, sarray.c, selector.c,
> 	sendmsg.c, thr-dce.c, thr-decosf1.c, thr-irix.c, thr-mach.c,
> 	thr-objc.c, thr-os2.c, thr-posix.c, thr-pthreads.c, thr-rtems.c,
> 	thr-single.c, thr-solaris.c, thr-vxworks.c, thr-win32.c, thr.c:
> 	Include Objective-C headers with quotes and objc/ directory
> 	prefix.
> 

Committed after testing on i686-pc-linux-gnu.

(Approved here: http://gcc.gnu.org/ml/gcc-patches/2005-06/msg00641.html)

This /should/ fix the problem.  I wonder whether we should also remove
-I$(srcdir)/objc from the INCLUDES in Makefile.in to avoid future usage
of unqualified header files.

Also we should be weary of relying on:
#include "objc/runtime.h"            /* the kitchen sink */
which will in turn:
#include <objc/files.h>
which in theory could still continue to include old headers in setups
like Zem and Christian seem to be encountering as I assume the -B /-I
flags maybe somehow incorrect.

In any case, it would be nice to know if this actually does fix the
problem at hand.  I need some sleep now though.

Cheers,
David Ayers