You are quite right, Pierre. My apologies. I looked into things, and while I was debugging my PCL/CLOS/MOP problems, I seem to have gaffed the state of the PCL, and THAT was what was really wrong.
Now I'll test better in isolation before emailing. I wish I could trace why that problem appeared, but the code I'm working with is absolutely enormous (port of a large system, Allegro->CMUCL), so I haven't been able to. Sorry for the time-waster, Robert
