KWalletd crashing on cvs head

[Richard Smith]

On Saturday 27 November 2004 17:05, Zack Rusin wrote:
> On Saturday 27 November 2004 12:16, Richard Smith wrote:
> > On Saturday 27 November 2004 16:55, Zack Rusin wrote:
> > > On Saturday 27 November 2004 08:16, Richard Smith wrote:
> > > > The attached patch to kpassdlg.cpp fixes this. OK to commit?
> > >
> > > How come the strengthbar is never initialized in this one?
> >
> > If you mean, how come the strengthbar never used to be initialised
> > (which my patch fixes), I guess whoever wrote the original code just
> > didn't think of it. And then didn't test the dialog in
> > non-NewPassword mode.
>
> Well, that person should be shot, cut and stabbed, not necessarily in
> that order, but that's a different matter.
> My question was "how come the strengthbar is a null object in slotOK"?.
> IIRC all code paths of NewPassword lead through init which sets the
> strengthbar correctly, so how come it's null in slotOK?

Ah, I see. You mean, why bother adding a check for null in slotOK? Well, 
obviously m_strengthBar will have been initialised if we're in NewPassword 
mode and we're in slotOK. Obviously. But what's true now isn't going to be 
true forever. And what's wrong with (slightly) more robust code?
-- 
Thanks,
Richard