PR 5303, 5755

[Doug Kwan (關振德)]

Do'h.  I thought HJ's fix have been accepted.  To simplest way to fix
this is to disable to optimization at run-time when splay tree cannot
handle 64 bit bfd_vma.  I will fix that and resubmit then.

-Doug

2008/2/11, Alan Modra <amodra@bigpond.net.au>:
> In http://sourceware.org/ml/binutils/2007-11/msg00144.html I said I'd
> revert Doug Kwan's speed up for bfd_dwarf2_find_nearest_line if 64-bit
> target support on a 32-bit host remained broken.  HJ has made a number
> of attempts to cure the underlying problem in libiberty/splay-tree.c,
> but so far hasn't managed to convince anyone to apply anything.
>
> In http://sourceware.org/ml/binutils/2007-12/msg00027.html Doug said
> there were a couple of ways to fix the problem inside BFD without
> touching libiberty, and offered to make the changes.  No one commented
> on which way was best, and perhaps because of that, no patch was
> forthcoming.  (Either of Doug's fixes is fine by me..)
>
> So..
>         PR 5303, 5755
>         * arange-set.c: Delete.
>         * arange-set.h: Delete.
>         * dwarf2.c: Revert 2007-09-21 changes.
>         * Makefile.am: Likewise.
>         * Makefile.in: Regenerate.
>         * po/SRC-POTFILES.in: Regenerate.
>
> --
> Alan Modra
> Australia Development Lab, IBM
>