Re: AppIcon patch revisited

[Richard Frith-Macdonald]

On Saturday, February 10, 2001, at 12:33 PM, Stefan Boehringer wrote:

> Hello Richard, 
>  
> I think that there is some good chance to realize Dan's proposal without 
> breaking the 
> layering of the GNUstep API. 
> However this is at the cost of extending the existing backend API. There 
> should be 
> precisely the chance to forward events back to the backend. Then a miniwindow 
> could 
> forward back unneeded mouse down events. 

I'd rather avoid the app developer needing to provide workarounds like using 
some
extended API to send back events.

> In my point of view there is no way around doing some extra work to get mini 
> windows straight. 
> Having a window fully functional but at the same time dockable and correctly 
> tiled needs 
> special treatment somewhere in the libraries. I'm not convinced that it is 
> WMakers part to 
> handle all those complications, since it seems to be quite GNUstep specific 
> etc., and we 
> (the GNUstep libraries) have good chances to handle the case gracefully. 
> Unfortunatly I'm not familier with the backend stuff, so I just want to point 
> out this 
> possibility. 

I believe that if Window Maker simply treated appicons and miniwindows like 
other windows,
the problem would not exist.  As far as I can see, the only way in which it 
should treat
those windows any differently (according to ICCCM) is that it should handle
mapping/unmapping for them, and the gnustep backend library should refrain from 
mapping/unmapping them itsself.