Another bug (rotatef should return NIL)

[quiroz@cs.rochester.edu]

Rotatef returns the old value of the first subform, against the
specification in p. 99.

Although using the guaranteed NIL return to effect conditional
control would most likely be in bad taste, it would still be legal
CL, so this change is incompatible with the standard.

I have a fix to load in the USER package, but I would rather fix the
macro in the LISP package.  I haven't done that yet because I don't
know if the implementation depends on the currently erroneous
behavior.

=Cesar
--------
Cesar Augusto  Quiroz Gonzalez

Department of Computer Science     {allegra|seismo}!rochester!quiroz
University of Rochester            or
Rochester,  NY 14627               quiroz@cs.rochester.edu