Re: tracking area problems on 10.4.11

FROM : Wesley Smith
DATE : Wed Jul 02 20:48:57 2008

> The rect parameter here is in the coordinate system of the receiver,
> so this makes no sense. Instead pass [[mWindow contentView] bounds].


Thanks!!  Hit the nail on the head.  Now it actually works.

> By passing the window's frame, you're taking the window's location in
> screen coordinates and then acting as though it's a meaningful
> location in view coordinates, which probably puts the tracking area
> completely outside the visible area of the window and is therefore
> meaningless.
>
> On another note, this tracking rect stuff is really the kind of thing
> you should be sending to 'self'. It becomes overly complicated to try
> to have some other object manage a view's tracking rect. Instead make
> a custom view and have it manage its own tracking rects. This also
> allows you to do nice things like set up the tracking rects in
> -resetCursorRects, which is the only sane way to do them.


I realize the design is a bit odd, but the circumstances are also
unusual.  The object setting this up is a C++ class bound to a
scripting environment.  If I posted the entire .mm file it would make
more sense.  Thanks for the help.
best,
wes