This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH] PR mi/2086 -break-insert missing error diagnostic
- From: Daniel Jacobowitz <drow at false dot org>
- To: Nick Roberts <nickrob at snap dot net dot nz>
- Cc: gdb-patches at sources dot redhat dot com, Ulrich dot Weigand at de dot ibm dot com
- Date: Fri, 17 Nov 2006 16:15:56 -0500
- Subject: Re: [PATCH] PR mi/2086 -break-insert missing error diagnostic
- References: <17717.17922.450312.898237@kahikatea.snap.net.nz> <20061017213119.GA14010@nevyn.them.org> <17717.20397.279496.222573@kahikatea.snap.net.nz> <20061017215754.GA15210@nevyn.them.org> <17717.27348.977065.613551@kahikatea.snap.net.nz> <20061018153109.GB10436@nevyn.them.org> <17718.19901.569519.210795@kahikatea.snap.net.nz>
On Thu, Oct 19, 2006 at 04:52:29AM +1300, Nick Roberts wrote:
> Daniel Jacobowitz writes:
> > On Wed, Oct 18, 2006 at 12:44:20PM +1300, Nick Roberts wrote:
> > > The name deprecated_set_gdb_event_hooks suggests that they're earmarked for
> > > removal. These hooks only seem to get get in one place so presumably they
> > > could be set up once at initialisation and left with that value.
> >
> > They're also used by insight. This is one of the sucky things about
> > having insight directly linked to GDB.
>
> OK, how about just using the hack in mi_cmd_thread_select for
> mi_cmd_break_insert for the moment.
This is just nasty. I believe I pointed you at Ulrich's analysis of
this problem upthread:
http://sources.redhat.com/ml/gdb/2006-10/msg00021.html
It seems pretty clear to me that the patch which switched things to
return the result of catch_exceptions_with_msg was wrong. The
functions are defined to return an enum gdb_rc. Can't we make
them do that again? Simple, obviously correct.
--
Daniel Jacobowitz
CodeSourcery