This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
Re: [RFA] stack.c: Silently pop dummy frame after `return' from stop in dummy
- To: msnyder at redhat dot com (Michael Snyder)
- Subject: Re: [RFA] stack.c: Silently pop dummy frame after `return' from stop in dummy
- From: "Peter.Schauer" <Peter dot Schauer at regent dot e-technik dot tu-muenchen dot de>
- Date: Thu, 28 Sep 2000 9:39:59 MET DST
- Cc: gdb-patches at sourceware dot cygnus dot com
> If I understand, the change would be that, if I'm debugging in a
> function called from GDB and I say "return", instead of finding
> myself in the dummy "function called from GDB" I will find myself
> all the way back in the original frame #0 that was current
> before the fabricated function call.
Yes. Finding oneself in the dummy "function called from GDB" after a return
was the subject of a recent bug report, which triggered all this.
> This seems desireable, but since I'm already accustomed to the
> old behavior, I'd like to hear if anybody else has any feelings
> about it pro or con.
>
> You say the behavior for "finish" has already been changed?
No, the behaviour of finish was not changed.
On many targets I have access to (AIX, Digital Unix, Solaris x86, Linux x86),
finish after a stop in a call dummy takes you all the way back to the original
frame that was current before the fabricated function call.
It doesn't happen on sparcs, which is the motivation for the proposed change #1
in sparc-tdep.c, as I think that GDB should behave consistent across targets
if possible.
--
Peter Schauer pes@regent.e-technik.tu-muenchen.de