compilation

compilation

[Bill Schottstaedt]

> is anybody from the Guile community right now hacking
> Hobbit or some other scheme-to-c compiler for Guile?

In a sense, yes (mentioned in an earlier msg to guile-user);
and I'm very interested in this problem in general -- I've
written a partial Common Lisp->C translator (and a very long time
ago, worked (in PDP-10 assembly language...) on the SAIL
compiler).  I tried Hobbit but did not get much speed-up,
but as previously mentioned, my context is heavily numerical.
And I don't remember (this was a while ago and my memory is
unreliable) any support for "foreign" functions; since I
use Guile almost solely as an extension language, I have
to be able to call my own stuff.  If it isn't included
in the optimized code, then there's no point... The other
problem here is that a "quick and dirty hack" like the one
I'm doing now provides 90% of what I need (in a couple
week's work), and I'm swamped (as is everyone) with
other things to do, so a big project to do it right doesn't
resonate. But, maybe I'd get seduced into helping once it got
started.

(My experience with the Lisp->C translator is that you regret
it in the long run: dynamic loading is an incredible pain,
debugging is next-to-impossible, files end up scattered
everywhere, etc.)

_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: compilation

[Rob Browning]

Bill Schottstaedt <bil@ccrma.Stanford.EDU> writes:

>  and I'm swamped (as is everyone) with other things to do, so a big
> project to do it right doesn't resonate. But, maybe I'd get seduced
> into helping once it got started.

 :>

At the moment Marius' arguments that we should clean up guile's
execution model with an eye toward future "compilation" before we dive
back in to any particular strategy sound pretty good to me.

I have this suspicion that if this is done right (with hooks in
suitable places and well defined semantics) then it might make it a
lot easier to try out various approaches.  However, without more
investigation, it's possible that I'm just falling victim to wishful
thinking.

-- 
Rob Browning
rlb @defaultvalue.org, @linuxdevel.com, and @debian.org
Previously @cs.utexas.edu
GPG=1C58 8B2C FB5E 3F64 EA5C  64AE 78FE E5FE F0CB A0AD

_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel