Re: bonobo-activation patch ... (fwd)

[Maciej Stachowiak <mjs noisehavoc org>]

On 16Nov2001 02:36AM (-0500), Michael Meeks wrote:
> 
> Hi Maciej,
> 
>         Gack - I just found this in my inbox - and realized, that I
> committed without reading it properly; sigh - sorry.
> 
>         So - there was only 1 issue really:
> 
> > >         CORBA_exception_free (&ev);
> > > 
> > > -       return 0;
> > > +        if (bonobo_activation_debug_shutdown ()) {
> > > +                return 0;
> > > +        } else {
> > > +                return 1;
> > > +        }
> > >  }
> > 
> >
> > Hmm, if we're going to return error values I would rather reserve it
> > for failing to activate, rather than potential leaks. So I don't agree
> > with this part of the change, unless there is some reason for it I'm 
> > missing.   
> 
>         Well - currently it means that when you run 'make distcheck' it is
> clear that there are no reference leaks in bonobo-activation at all. If   
> you change something in the future - you can be sure that you havn't
> introduced a leak when you 'make distcheck' - in the same way that you can
> be sure you havn't introduced an "it doesn't activate" regression.

As I recall, the instance I complained about in
bonobo-activation-client, which is not used in the regression tests,
as far as I know, but is meant to be used by outside code.

It is possible I misread the patch

 - Maciej