Re: Is Elisp really that slow?

[Emanuel Berg]

Stefan Monnier wrote:

>> So from a personal POV, I would be
>> interesting/fun to write some of my Elisp
>> into C, and if I get good enough at it,
>> maybe I could even do it for the project.
>
> FWIW, compiling existing Elisp to C code is
> not really interesting in the sense that if
> it exists as Elisp code it's usually because
> it's not very speed-sensitive.

Yeah, I suspected that much. Still, some Elisp
I'm sure I can find somewhere that is
worth translating?

If not, I'll just translate some of my own
stuff. Here, I wonder how to determine what to
optimize, tho? Obviously it should be something
simple that is used often and can benefit from
a speed-up. Is there an automated way to answer
any of those three questions? Or do you
actually have to use your _brain_ and not just
your fingers to do computer work?

Or maybe I'll write something all new in C...

> What is more interesting is to take existing
> C code and see if it can be rewritten into
> a kind of Elisp (or other language that has
> similar dynamic properties, most importantly
> the ability to replace code at run-time) that
> can be compiled back to efficient code.

That's beyond my capabilities at the moment,
I'm afraid...

> After all, the main problem with the part of
> Emacs that's written in C is that you can't
> change it from your ~/.emacs, contrary to all
> the Elisp code.

Is _that_ the problem with Emacs' C? I thought
there wasn't any problem with Emacs' C, we just
needed more to make the program faster?
For simple, fundamental stuff that people don't
configure anyway because of their
low-level-of-complexity nature? You're saying
all of that is already done?

-- 
underground experts united
http://user.it.uu.se/~embe8573
https://dataswamp.org/~incal