RUN-IN-MACLISP-SWITCH vs OPEN-CODE-MAP-SWITCH

[Knowledge-Based Editor <KBE at MIT-AI>]

In System 78.50, ZMail 38.5, microcode 836, on Lisp Machine Twenty-two:

When OPEN-CODE-MAP-SWITCH is T and RUN-IN-MACLISP-SWITCH is T, all MAP things
seem to get a diagnostic about PROG tags (in this case, T) not being supported
in Maclisp. They also get a message about ((COMPILER:MAP-RESULT)), initialization
in a PROG not being supported either. Presumably, if the thing which does the
expansion would put an (INHIBIT-STYLE-WARNINGS ...) around the (PROG T ...) and
the (RETURN-FROM T ...) if there is one, these message would not appear. They sure
had me scared for a while tho' trying to figure out what they were about... I thought
for a while it was a problem I had caused in my macro package somewhere.

Oddly, this problem does not occur when Control-Greek-E (in ZWEI) or the COMPILE 
function are used. Only when M-X Compile Buffer (in ZWEI) or QC-FILE is used do
they appear. That made the problem doubly difficult to track down because my simple
test cases didn't have the problem when I was compiling them a function at a time.
It seems to me that it is a bug that the diagnostics given for the two are not
consistent.

-kmp