Fri, 21 Dec 2001 12:15:25 -0600
Date: 20 Dec 2001 23:24:15 -0900
From: email@example.com (James A. Crippen)
QBBFL: false - xyzzy
I wish I could remember what that was about. The couple mentions I can
find suggest it indicates that a block of special-variable bindings is
associated with the top frame. (BBFL: Binding-Block FLag.)
CAR-SYM-MODE: 1 - error except (CAR NIL) is NIL
CAR-NUM-MODE: 0 - error
CDR-SYM-MODE: 1 - error except (CDR NIL) is NIL
CDR-NUM-MODE: 0 - error
Once upon a time the lispm was a university project, and they
experimented with stuff. This is one of the places that wasn't cleaned
up too well for production: you could actually change the behavior of
CAR and CDR for symbols and numbers, making it error or maybe do other
things, as well as what became the official Common Lisp behavior.
Don't remember these.
MAR-MODE: read-trap=false; write-trap=false
MAR mode was equivalent to what some debuggers call "watchpoints"; it
would trap if the specified address were read or written, depending on
the flags. (MAR stands for "Memory Address Register.") It even worked,
sometimes, but we didn't officially support it; it only allowed
watching one address.
PGF is usually an abbreviation for "page fault," but I don't know what
Not sure what the flags are for. "Scavenge" and "transport" are GC
A "sequence break" is a process switch, but I can't remember why one
would be deferred. Maybe it has something to do with inhibiting
Metering was what is usually called "profiling" now, you could collect
scads of information about execution times. "Submetering," introduced
later, was faster and often more useful.
I think this was related to the BREAKON command, which let you trap on
entry to a given function.