🚀 go-pugleaf

In article <11810@sail.LABS.TEK.COM> toma@sail.LABS.TEK.COM (Tom Almy) writes:
   The biggest improvement would come from a XLISP compiler, since "interpreter
   overhead" accounts for a large percentage of execution time. A threaded LISP
   interpreter would probably pick up most of the gain, and would be portable.

How about a "byte code compiler", like in Xscheme? One problem is
that after some more expansion, XLisp won't fit into a vanilla PC.

What is the reason why 2.1d uses the address of a static variable as
NIL, instead of a  NULL pointer? One would expect comparisons agains NULL
to be faster and more compact (I realise this certainly does not make much
difference in the speed of the end result, but I am just wondering why this
choice).

Btw, has anyone succeeded in compiling 2.1d with Microsoft C 5.0 so
that the result actually works? I have tried various switch combinations
but the result always dies in some random spot, typically when handling
a load. Debugging is greatly hampered by the fact that the program is too
large to fit into Codeview. Or am I the last person on the net still
trying to use version 5.0...
--
Erkki Ruohtula     / Nokia Telecommunications
eru@tele.nokia.fi / P.O. Box 33 SF-02601 Espoo, Finland
(My private opinions, of course.)