RFA: Remove unused support for target-based exception handling
Fri Dec 7 15:02:00 GMT 2007
On Thu, Dec 06, 2007 at 10:19:11PM -0800, Jim Blandy wrote:
> (I'd been hoping that this would let me delete print_frame_label_vars,
> but when I got back there it turned out I'd misread the code, and
> instead we now use it unconditionally... sigh.)
I suspect you can delete that and still leave things unchanged; you'd
have to check, but I'm reasonably certain that you have to go back to
g++ 2.7 or even earlier to find the generated code that routine is
trying to support.
On Fri, Dec 07, 2007 at 09:29:13AM +0200, Eli Zaretskii wrote:
> > From: Jim Blandy <email@example.com>
> > Date: Thu, 06 Dec 2007 22:19:11 -0800
> > This was kind of like finding something old and nasty in the back of
> > the fridge. :)
> Could you (or someone else) perhaps tell the history of this code?
> Like when it was introduced and for what purposes, and how it became
It was part of the Great HP Merge, in 1998. The changelog entry is
the first 2575 lines of ChangeLog-1998. HP-UX apparently had some OS
facility used for catching C++ exceptions; this was specific to the
aCC compiler, not G++ on HP-UX, and support for that compiler was
removed from GDB at the beginning of this year.
> Also, doesn't the manual needs to be fixed if we remove this?
The feature is still there, and will still work for g++ 3.x / 4.x
compiled code, and any other compiler using the common C++ ABI.
More information about the Gdb-patches