Re: [PATCH] accept null selector for respondsTo...

[Willem Rein Oudshoorn]

Jeff Teunissen <deek@d2dc.net> writes:


  [REMOVED PART about behaviour of GNUstep base with respect
   of null selectors].

> And when the new behavior is clearly wrong (as in this case)?

I mean that we NEVER intend to remove the lenient behaviour. 
> 
> I can see a runtime default for optionally choosing the Mac OS
> X-compatible behavior

That is the whole idea

> the new behavior is clearly wrong from both a historical reference and
> on the principle of being generous in accepting potentially-spurious
> input.

I agree wholeheartedely with you.  For me the MacOSX flag is unneeded.
But I see value in trying to track Apple's changes to the API.
So if this means that setting a runtime default turns on MacOSX
behaviour, which brings some improvements and some wrong behaviour,
that is fine with me.

Wim Oudshoorn.