🚀 go-pugleaf

In article <3114@enea.se> sommar@enea.se (Erland Sommarskog) writes:
>cat *.el | wc on /usr/gnu/src/emacs/lisp revealed 55.000 lines.
>However, are any of these 55.000 lines are linked into /usr/gnu/
>bin/emacs? 

Many of those files are loaded prior to unexec'ing bin/emacs.  I don't know
how many, though; you could probably determine this by looking at the
makefile or build script.

>	    How many are really kernel-emacs, and how many are
>just the editing interface, and special-packages like electric-c?

I guess the problem is that there is no clear agreement on what is
"kernel-emacs".  To me, part of the definition of Emacs is that the default
binding of meta-B (for instance) is to move the cursor backward by a word;
i.e. I consider the simple word motion commands to be part of the Emacs
kernel.  In GNU Emacs, backward-word is written in Elisp.

I also find your reference to "*just* the editing interface" curious.
Emacs is primarily an editor, so the editing interface is an important part
of it.  For sure, Emacs is flexible enough that extension packages can
bypass the normal editing interface if they need to, but that doesn't mean
that the normal editing interface isn't part of Emacs.

If GNU Emacs isn't written in Lisp, then PDP-10 EMACS wasn't written in
TECO.  All the display management and interactive keyboard dispatching was
built into TECO itself, and EMACS was simply a bunch of TECO macros which
were stored into the keyboard dispatch table.  But these macros are
themselves the difference between TECO and EMACS.


--
Barry Margolin, Thinking Machines Corp.

barmar@think.com
{uunet,harvard}!think!barmar