Re: [RFA] Patch to fix reverse return from subroutine error

[Hui Zhu <teawater@gmail.com>]

Sorry for late everybody.
Looks like we still have trouble with it,right?

We have plan B  for this issue.  This is a old implement when p record
still have reverse command inside it.

This issue is because we don't know the frame_info of step_frame_id.
Without it, we cannot get the frame_info of high level function with
function get_prev_frame.
Without frame_info, we cannot get the frame_id.
Then we cannot make sure if this inferior return from subroutine.

So we can save this prev_frame_id when we save step_frame_id in reverse mode.
in step_once,
      frame = get_current_frame ();
      tp->step_frame_id = get_frame_id (frame);
Update it to inferior_status.
And update it in handle_inferior_event:
  ecs->event_thread->step_range_start = stop_pc_sal.pc;
  ecs->event_thread->step_range_end = stop_pc_sal.end;
  ecs->event_thread->step_frame_id = get_frame_id (frame);
  ecs->event_thread->current_line = stop_pc_sal.line;
  ecs->event_thread->current_symtab = stop_pc_sal.symtab;

Then, we can use it to make sure if inferior return from subroutine or not.

What do you think about it?


Thanks,
Hui



On Mon, Jun 29, 2009 at 05:08, Daniel Jacobowitz<drow@false.org> wrote:
> On Sun, Jun 28, 2009 at 11:44:37AM -0700, Michael Snyder wrote:
>> Great, I looked up your paper, and found it helpful.
>> Are your gdb patches available online?  Any head start
>> would be a boon...
>
> I think I've posted them before, but I can't remember where.  A copy
> is attached.  The epilogue bits are in patch-4.
>
> --
> Daniel Jacobowitz
> CodeSourcery
>