Re: What's the spec for emacs lisp virtual machine ?

[Emmy Noether]

On Jul 23, 10:57 am, Barry Margolin <bar...@alum.mit.edu> wrote:
> In article
> <61a26ed5-80eb-4175-b7c4-162a1ce21...@e5g2000yqn.googlegroups.com>,
>  Fren Zeee <frenz...@gmail.com> wrote:
>
> > portability does not require VM, it only requires C while emacs
> > interprets the lisp code. a diff command in C is portable over
> > computers because its written in C. not because of a VM.
>
> An interpreter *is* a VM.  The "machine language" of this VM is the
> language being interpreted.
>
> Maybe what you're really asking is whether we need byte code?  That's
> just an optimization.  Compiling to machine code is even more of an
> optimization, but then the result is not portable.  So a byte code VM is
> a compromise between source code interpretation and machine code
> compilation.

This clarified for OP the confusion Pascal sowed by bringing in C as a
VM. IMHO OP wanted explanation of the VM by a state diagram, some
architectural design or code walkthrough.

Its written on wiki that the effort to make a portable compiler
emitting machine code is more than the effort to write a VM to enable
portability of the compiled byte-code. Only someone familiar with
Pascal p-code (probably the first appearance of byte-code) can explain
this point, or one who has thoroughly read and understood the files
Pascal mentioned. He probably believes he does as he shares the name
with the lingo ;)