Question: BFD_FAIL

[Thiemo Seufer]

Ian Lance Taylor wrote:
> msnyder@sonic.net writes:
> 
> > Are we meant to assume that BFD_FAIL does not return?
> 
> That would be an unwise assumption.  As you have seen, it does
> return.
> 
> > void
> > _bfd_set_gp_value (bfd *abfd, bfd_vma v)
> > {
> >   if (! abfd)
> >     BFD_FAIL ();
> >   if (abfd->format != bfd_object)
> >     return;
> 
> Yep, looks wrong.  Came in here:
> 
> 2003-05-17  Thiemo Seufer <seufer@csv.ica.uni-stuttgart.de>
> 
> 	* bfd.c (_bfd_get_gp_value): Prevent illegal access for abfd null
> 	pointers.
> 	(_bfd_set_gp_value): Likewise.

I think that's the correct fix. Ok?


Thiemo


2007-07-25  Thiemo Seufer  <ths@mips.com>

	* bfd.c (_bfd_set_gp_value): Use abort instead of BFD_FAIL.

Index: bfd.c
===================================================================
RCS file: /cvs/src/src/bfd/bfd.c,v
retrieving revision 1.96
diff -u -p -r1.96 bfd.c
--- bfd.c	3 Jul 2007 14:26:39 -0000	1.96
+++ bfd.c	25 Jul 2007 20:35:27 -0000
@@ -1020,7 +1020,7 @@ void
 _bfd_set_gp_value (bfd *abfd, bfd_vma v)
 {
   if (! abfd)
-    BFD_FAIL ();
+    abort ();
   if (abfd->format != bfd_object)
     return;