Re: Binutils 2.11.x problems with oracle 9i

["H . J . Lu" <hjl at lucon dot org>]

On Thu, Jun 28, 2001 at 04:11:37PM +0930, Alan Modra wrote:
> On Thu, Jun 28, 2001 at 02:18:23PM +0930, Alan Modra wrote:
> > On Wed, Jun 27, 2001 at 05:31:24PM -0700, H . J . Lu wrote:
> > >Alan, what happened to your patch:
> > > 
> > > http://sources.redhat.com/ml/binutils/2001-02/msg00094.html
> > 
> > I forgot about it. :)  Committing.
> 
> No, I'm not committing.  I remember now why I didn't, and that's
> because both your and my patch are somewhat bogus.  (Mine as
> posted even more so because it was wrong).  Anyway, we would
> end up with something like
> 
>   if (! finfo->info->relocateable
>       && ! finfo->info->allow_shlib_undefined
>       && h->root.type == bfd_link_hash_undefined
>       && (! finfo->info->shared
> 	  || (finfo->info->no_undefined
> 	      && (h->root.u.undef.abfd->flags & DYNAMIC) == 0))
>       && (h->elf_link_hash_flags & ELF_LINK_HASH_REF_DYNAMIC) != 0
>       && (h->elf_link_hash_flags & ELF_LINK_HASH_REF_REGULAR) == 0)
> 
> Considering the shared case, I reckon that
> 
>       (h->root.u.undef.abfd->flags & DYNAMIC) == 0
>       && (h->elf_link_hash_flags & ELF_LINK_HASH_REF_DYNAMIC) != 0
>       && (h->elf_link_hash_flags & ELF_LINK_HASH_REF_REGULAR) == 0
> 
> is always false.  So we really just want to revert one of rth's changes.
> 
> bfd/ChangeLog
> 	* elflink.h (elf_link_output_extsym): Revert 1999-04-10.
> 

It looks good to me. Can you check it in?

Thanks.

H.J.