Using the correct terms LAP vs ELisp Bytecode or something else

Using the correct terms LAP vs ELisp Bytecode or something else

[Rocky Bernstein]

[-- Attachment #1: Type: text/plain, Size: 630 bytes --]

What's the difference between LAP and Elisp Bytecode (if that is the right
term)?

LAP stands for Lisp Assembly Program, so I imagine that refers  to the text
representation. Bytecode however generally refers to a more binary
representation, but I guess often it includes the mnemonics for the code.

If I  want to describe the the assembly instructions, should I refer to it
as LAP instructions or Elisp bytecode instructions or Emacs Lisp bytecode
instructions?

And is it "bytecode' or "byte code"?

Or maybe there aren't any conventions and none of this matters so I can
just use whatever I feel like.

Please advise. Thanks.

[-- Attachment #2: Type: text/html, Size: 743 bytes --]

Re: Using the correct terms LAP vs ELisp Bytecode or something else

[Stefan Monnier]

> What's the difference between LAP and Elisp Bytecode (if that is the right
> term)?
> LAP stands for Lisp Assembly Program, so I imagine that refers  to the text
> representation.

IIRC, LAP code is represented as a list (or is it a vector?) where each
instruction is represented by a symbol with some arguments.  So it's not
a textual representation, but it's a "symbolic" representation that's
fairly easy to manipulate and with many similarities to traditional
assembly: use of labels instead of actual addresses, use of
pseudo-instructions (e.g. labels), and things like that.

> Bytecode however generally refers to a more binary
> representation, but I guess often it includes the mnemonics for the code.

Elisp bytecode is represented as a unibyte string (i.e. an array of
bytes), so yes, it's a lot more binary.  It's the "garbage" that used to
show up in backtraces.

> If I want to describe the the assembly instructions, should I refer to it
> as LAP instructions or Elisp bytecode instructions or Emacs Lisp bytecode
> instructions?

I think you can talk about the "bytecode instruction" and have that
include both the LAP representation of that instruction and its actual
bytecode representation: either the difference doesn't matter, or the
context should make it clear which is meant.

> And is it "bytecode' or "byte code"?

No opinion on that.

> Or maybe there aren't any conventions and none of this matters so I can
> just use whatever I feel like.

Pretty much, yes.


        Stefan

Re: Using the correct terms LAP vs ELisp Bytecode or something else

[Rocky Bernstein]

[-- Attachment #1: Type: text/plain, Size: 860 bytes --]

Thanks for the information...

On Wed, 20 Dec 2017 20:50:28 -0500 Stefan Monnier <monnier@iro.umontreal.ca>
wrote:

>
> > What's the difference between LAP and Elisp Bytecode (if that is the
> right
> > term)?
> > LAP stands for Lisp Assembly Program, so I imagine that refers  to the
> text
> > representation.
>
> IIRC, LAP code is represented as a list (or is it a vector?) where each
> instruction is represented by a symbol with some arguments.  So it's not
> a textual representation, but it's a "symbolic" representation that's
> fairly easy to manipulate and with many similarities to traditional
> assembly: use of labels instead of actual addresses, use of
> pseudo-instructions (e.g. labels), and things like that.
>

This is the first I am hearing about LAP code being something that is
accessible from Lisp. Where are packages that work with LAP?

[-- Attachment #2: Type: text/html, Size: 1285 bytes --]

[SUSPECTED SPAM] Re: Using the correct terms LAP vs ELisp Bytecode or something else

[Stefan Monnier]

> This is the first I am hearing about LAP code being something that is
> accessible from Lisp.  Where are packages that work with LAP?

bytecomp.el, byte-opt.el, and disass.el.  Probably not much more than that.


        Stefan