Our behavior is within the bounds of the spec:

http://www.stud.uni-karlsruhe.de/~unk6/clim-spec/29-2.html#_6495 http://www.stud.uni-karlsruhe.de/~unk6/clim-spec/29-2.html#_6487

However, perhaps we should emulate the lispworks behavior of not requiring the user to explicitly bind the realizer.

Christophe's bugs are unrelated (one involves failure to embed nested gadgets due to a problem in the order w-o-a-g initializes things, the other I never tracked down).

Hmm, well you've got me there (sort of). I'm pretty sure that as I reduced code from the real program that first showed that problem to the minimal code in my email I did have with-look-and-feel-realization in play with no apparent difference. I probably had some cockpit trouble at that point.

Still no cigar though. Now I get a display visible, but no gadget.

So, back to the debugging pit. Stay tuned.

Paul