Re: Incorrect behaviour in NSWindowController...

[Fred Kiefer]

Gregory John Casamento wrote:
> --- Fred Kiefer <address@hidden> wrote:
> > Could you please expalin in more detail, what the possible problem here 
> > is? From this mail it just looks like trying to mimik a bug of MacOSX, 
> > which surely is not worthwhile and probably not what you intend to achive.
> > As this is a rather old line of code, I would not be to surprised to 
> > learn that it is now obsolete. 
>
>
> I don't want to imitate a "bug" of Mac OS X, I want to make certain that it's
> not possible to have sloppy code on GNUstep which might correctly fail
> otherwise.
>
>
> > But I would like to understand it first.
>
> The idea here is, apparently, to retain the window if it's closed (if it's
> owned by a window controller and has isReleasedWhenClosed set to YES) so that
> it doesn't cause a segfault.   It's the age old question of how far do we go to
> prevent the developer from shooting himself in the foot?

This explains things just as far as I can understand them by looking at 
the code. The question is rather, would the current behaviour of GNUstep 
produce a memory leak or does it just make sure things work smoothly?

Could you provide a short example for the correct usage of this feature 
and describe the results on Cocoa as opposed to GNUstep? Perhaps based 
on Ink.app (which currently doesn't use this feature)? I often think, 
that showing code explains things best.

Fred