Re: Emacs rewrite in a maintainable language

[David Kastrup]

John Wiegley <address@hidden> writes:

>>>>>> Ken Raeburn <address@hidden> writes:
>
>>> The main reason I stopped is that it didn't help performance enough. The
>>> slowness imposed by constantly looking up symbol names at global scope, due
>>> to dynamic binding, dominated the profiling results.
>
>> Do you think the introduction of lexical binding support changes the balance
>> much? (Or was your work after lexical binding was added?)
>
> I think it should really help performance, but I haven't measured yet.

I don't think it will really help performance all that much if you
measure it now since the inescapable cost of establishing and resolving
dynamic bindings at every let or function boundary has both made certain
optimizations mostly pointless as well as lending preference to a
programming style using iteration and setq rather than recursive
techniques.

Also, code like

`(lambda (...) ...)

does not magically rewrite itself into a compiled form just because
Elisp now got closures.

So there is a lot of trickle-down into the code base to wait for until
one can really make a statement.  And then the other reference point is
gone.

-- 
David Kastrup