Re: RIGS bottoming out in rbbr

[Adam Fedor]

On Feb 15, 2005, at 2:13 AM, Matthew D Swank wrote:

> I realize (or least it seems to me) that RIGS is no longer maintained. 
> However, with a few changes the cvs version compiles. I tried using 
> rbbr to look at the class hierarchy generated by the binding, but rbbr 
> would seg-fault whenever it tried access the superclass of NSObject or 
> NSProxy.  I decided to overload the superclass class method for each 
> class to return 'nil' (and later on: 'Object').  nil stopped the 
> seg-faults, but produced other errors when rbbr tried to access nil's 
> metaclass.  Having these two superclass methods return Object seems to 
> work just fine.  My question is: Is this a reasonable solution?  Will 
> I run into any trouble if ruby thinks the two GNUstep root objects 
> derive from Object?

Well, I'm not sure what would happen if a method search travelled up 
beyond NSObject or NSProxy.  How about having it return 
NSObject/NSProxy? Or perhaps that's the infinite loop that is causing 
the seg fault...