[patch] KDevelop-3.4 Debugger: Disassembler-Outputview broken

Vladimir Prus ghost at cs.msu.su
Wed Nov 8 09:43:41 UTC 2006


On Wednesday 08 November 2006 12:35, Robert Gruber wrote:
> Hi,
>
> On Wed, 8 Nov 2006 11:54:13 Vladimir Prus wrote:
> > Hi Robert,
> > first of all, thanks for the patch!
> >
> > I think it's good except for one design issue. Together with switch to
> > MI, I've tried to make the architecture less centralized. So, view
> > widgets would
> > issue gdb commands themself, and not emit signals that go to
> > gdbcontroller.
> >
> > For the case of disassmbler view, the best approach is to define
> > slotEvent in
> > dissassbler widget, in that slot emit the necessary MI command, and
> > handle the MI reply by a method in disassembler widget.
> >
> > That way you won't need the slotDisassemble in gdbcontroller, and won't
> > need
> > the rawGDBDisassmble signal either.
> >
> > You can look at variablewidget.cpp to see how slotEvent can decide if it
> > should emit a command or not -- but basically, it should emit a command
> > on 'program_state_changed' event, and probably on
> > 'thread_or_frame_changed'
> > event.
> >
> > How does that sound?
> >
> > - Volodya
>
> sounds good :-)
> I'll change my patch that way and commit it (maybe tomorrow).
>
> Just to make sure, this change will include removing slotDisassemble() and
> signal rawGDBDisassemble() from both the GDBController class and it's
> abstract base class DbgController. Is this ok?

Yes, that's fine.

Thanks,
Volodya




More information about the KDevelop-devel mailing list