> >Hey guys, we have a working clipboard. I had to fix a few more things >than I anticipated during my system problems, but it now passes all of >its tests. The same can't be said for the R5 implementation. 5 of my >test cases would crash in R5 if I didn't use the TEST_R5 stuff. The >first case for the LocalCount function only passes some of the time (it >read unitialized memory). In the event that somebody passes a NULL >pointer to the constructor, my implementation attaches to the system >clipboard rather than blowing up in your face like R5. I suppose I >should document that behavior, but I am not sure that I want to >encourage people to attach to the system clipboard that way. Nice work. Reminds me of how it seems we're making a better BeOS than BeOS. --DW