Subject RE: [IBO] mysterious ntddl.breakpoint on ib_session.ReleaseSessionHookRef
Author G. Nau
Am 12 Oct 2006 um 9:21 hat Jason L. Wharton geschrieben:
Jason, I out-commented the ib_session

finalization
// ReleaseSessionHookRef; // Make sure hooks are valid at least this far
into it.

and the debugger break is gone.

So, does it indicate a problem within IBO or does it seems to be somewhere
else?
Does anybody has seen this before?

Regards
Gunther

> This seems like appropriate behavior to me.
>
> Jason Wharton
>
> > -----Original Message-----
> > From: IBObjects@yahoogroups.com [mailto:IBObjects@yahoogroups.com] On
> > Behalf Of G. Nau
> > Sent: Wednesday, October 11, 2006 3:43 PM
> > To: iBObjects@yahoogroups.com
> > Subject: [IBO] mysterious ntddl.breakpoint on
> > ib_session.ReleaseSessionHookRef
> >
> > D7Prof, IBO 4.5.Ai
> >
> > I'm observing my IDE jumping into the CPU debugger on program exit (even
> > on a button click with halt(0) on the main form).
> > The final application outside delphi also continues to live as a zombie
> > process on program exit (no window visible, but still existing in
> > Taskmanager).
> > The CPU debugger says:
> > ntdll.DbgBreakPoint: int3
> > ret
> >
> > The debug stack says:
> > ReleaseSessionHookRef <<<<!!!!!!!
> > FinalizeUnits
> > @Halt0
> > @Halt(???)
> > Tform.Button1Click($336B710)
> > .....
> >
> > So the int3 happens on the ibsession.releaseSessionHookref called from a
> > loop in system.FinalizeUnits; (that's finalizing all included units, which
> > registered finalization procedures)
> >
> > Anybody has any idea about that or seen that before?
> > Or is it a leftover debugger trigger in one of Microsoft latest patches
> > (installed by Windows update service)?
> >
> > Regards
> > Gunther
> >
> >






___________________________________________________________
Der fr�he Vogel f�ngt den Wurm. Hier gelangen Sie zum neuen Yahoo! Mail: http://mail.yahoo.de