20:16:24pfdietzI wrote a targeted random tester for TYPEP on integer types. No more failures found.
20:17:03pfdietz(also on various types in the REAL type hierarchy)
20:35:45Xof"some mop magic" is sort-of defined mop interface
20:36:31Xofif you believe in the mop, you believe in the ability to define an anonymous subclass with no instances, and yet be able to find it later with class-direct-subclasses
20:37:11XofI believe the phrase is "with great power comes great responsibility"
20:41:42phoeXof: got it. That's correct. These references must not be weak.
20:44:36stassatsi wonder if just using SETF instead of calling functions for &keyword processing will solve all my worries
20:45:16stassatsno duplicate variables, no bogus substitutions
20:46:16stassats&keyword processing code is about as convoluted as PCL
20:46:23pfdietzEdge cases though, like repeated keys and :allow-other-keys.
20:46:47stassatspfdietz: it's already handled via setf
20:47:01stassatsbut in the end it takes these variables and call the actual main function
20:51:38stassatsi don't know how to solve it yet, but at least i know how not to solve it