[Patch] Unterminated .eh_frame when using --eh-frame-hdr

[Richard Henderson]

On Tue, Nov 23, 2004 at 06:21:30PM +0000, Richard Sandiford wrote:
>   (1) Is it worth keeping the definition of __EH_FRAME_BEGIN__ now that
>       no code refers to it by name?  I decided to in this patch because
>       I've found it useful for debugging in the past.

I think that doesn't matter much.  Given that __EH_FRAME_END__ must
stay to provide a terminating zero, it seems nicely symmetrical to
keep the beginning symbol as well.

>   (2) Is it worth keeping HAVE_DWARF_UNWIND_INFO just to guard these
>       definitions?  I couldn't see what harm they would do to targets
>       that don't use .eh_frame, and since the aim is to reduce the
>       number of configuration variants, it seemed easier to define
>       them unconditionally.

I agree.

> Patch tested on mips64-linux-gnu.  Please install if OK.  Let me know
> what to change if not ;)

Patch looks good to me and to diffstat (224 deletions, 0 additions).
Uli, what do you think?


r~