This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
Re: C++ testsuite changes
- From: Daniel Jacobowitz <drow at mvista dot com>
- To: Michael Elizabeth Chastain <mec dot gnu at mindspring dot com>
- Cc: gdb at sources dot redhat dot com, kettenis at chello dot nl
- Date: Thu, 1 Jan 2004 16:35:04 -0500
- Subject: Re: C++ testsuite changes
- References: <20040101212247.6D18E4B35A@berman.michael-chastain.com>
On Thu, Jan 01, 2004 at 04:22:47PM -0500, Michael Chastain wrote:
> > I would really prefer it if you didn't rewrite the tests to accomodate
> > the ABI change (a very specific change) and change all sorts of other
> > tests at the same time. It makes it impossible to tell from your
> > patches when you make a change like this one.
>
> Sigh, you're right. I should have done this in several stages,
> where the first stage is lot of gdb_test_multiple with no change
> in output.
>
> I can go back and make it that way if you want. Shall I do that?
At this point I don't think it's worth it. For the remaining testcases
perhaps?
> > Eh... why don't you? It's a feature that we don't print the virtual
> > base pointer in recent gcc/dwarf combinations.
>
> Of course it's acceptable if gdb does *not* print the virtual base
> pointer.
>
> If gdb *does* print a virtual base pointer, do we consider that a
> bug in gcc? Because that's what "XFAIL" means. Or is it a bug in gdb?
> Then I should file a PR for it.
>
> My opinion is that we should just accept it. There's far worse bugs
> in C++ support that aren't getting any attention.
It's a bug in GDB.
I'm currently working on the C++ PRs. Unfortunately the one at the top
of my list also triggers a GCC bug. So it is taking time.
--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer