Re: MacApp with Mac CL

[liberte@ncsa.uiuc.edu (Daniel LaLiberte)]

(Sorry about the previous duplicate posting - gnus is so slow, I thought it
had died!)

Jonathan (vitro!pierce@uu.psi.com) says:
> Even if FF-LOAD could handle C++, MACAPP 3.0 still depends on
> loading object pascal binaries which FF-LOAD cannot handle.

Oh, darn!  However, quoting from the MACL version 1.2 manual (1989), "Any
compiler which produces object files of the MPW format may be used."
But I guess that doesnt say any code that such a compiler produces may
be used.

Others wondered whether loaded code could have its own global variables.
But the same manual says, "Each call to ff-load produces a distinct foreign
function environment, with its own global space, function code, etc."  So
I would presume so, but does it work?

Daniel Ranson (ranson@lannion.cnet.fr) suggests: 
> You might consider a MCL engine with a MacApp front-end, but I don't
> think it will be better than a MCL-only design.

This is an intriging idea.  Can MCL be used as an extension language
for another program, such as one using MacApp?  This would be a fine
solution, provided I could still get some of the advantages of the
common lisp programming environment at runtime.  Alternatively can MCL
produce a library of machine code linkable with another application?

Dan LaLiberte
National Center for Supercomputing Applications
liberte@ncsa.uiuc.edu