May Day PCL (Rev 2) with AKCL 1.505

[Jeff Dalton <jeff@aiai.edinburgh.ac.uk>]

I just compiled it and was puzzled by a few things.

  * During the compile-pcl process, KCL ran out of value stack
    when "Loading the binary of defs...".  I looked at a backtrace,
    and there wasn't much there.  SO I tried to get it to use less
    value stack (thinking it might _almost_ fit) by compiling defsys.
    Amazingly enough, this worked.  Is this a sign that the value
    stack ought to be larger?

  * There's no "test.lisp" in the PCL tarfile.  Is this an oversight
    or is there a good reason?  (I note that the test.lisp from the 
    Victoria Day release doesn't work any more.)  Is there a test.lisp
    in "Rev 3"? 

  * Why is the compiler sometimes called when I define a class or
    when I instantiate one?  (And at other times.)  Is this the
    expected behavior?

  * It takes a noticeable time to load PCL, significantly more than
    for the Victoria Day version.  Is this what other people have
    experienced?

-- Jeff    

PS  Is there any reason to get Rev 3?

Jeff Dalton,                      JANET: J.Dalton@uk.ac.ed             
AI Applications Institute,        ARPA:  J.Dalton%uk.ac.ed@nsfnet-relay.ac.uk
Edinburgh University.             UUCP:  ...!ukc!ed.ac.uk!J.Dalton