Potential Lua implementation for Guile?

Potential Lua implementation for Guile?

[No Itisnt]

I want to apply for GSOC this year. Since Guile has recently obtained
a compilation/language framework, I was thinking that an
implementation of the Lua language for Guile, under the auspices of
the GNU Project, would be a good fit.

A little background on Lua: It has a reputation as an "extensible,
embeddable" programming language. It inherits a lot from Scheme,
including closures and tail call optimization, but unfortunately not
macros :( I would say it definitely meets the criteria of "ubiquitous
intelligent language for extension." It's probably the most popular
"extension language" out there. At maybe ~10000 lines of C code, it's
not an unreasonable summer project either.

Personally, I am interested in learning more about parsing and
compilation. I'd also like to brush up on my Scheme.

Moreover, I think the addition of Lua, similarly to Emacs Lisp, would
put Guile in a better position to subvert existing applications that
use Lua as their scripting language (a lot!).

Let me know what you think.

Re: Potential Lua implementation for Guile?

[Neil Jerram]

No Itisnt <theseaisinhere@gmail.com> writes:

> I want to apply for GSOC this year. Since Guile has recently obtained
> a compilation/language framework, I was thinking that an
> implementation of the Lua language for Guile, under the auspices of
> the GNU Project, would be a good fit.

Yes, I think that's a great idea.

> Moreover, I think the addition of Lua, similarly to Emacs Lisp, would
> put Guile in a better position to subvert existing applications that
> use Lua as their scripting language (a lot!).
>
> Let me know what you think.

I think Guile finally seems to be approaching its original dream.

     Neil

Re: Potential Lua implementation for Guile?

[Daniel Kraft]

Neil Jerram wrote:
> No Itisnt <theseaisinhere@gmail.com> writes:
> 
>> I want to apply for GSOC this year. Since Guile has recently obtained
>> a compilation/language framework, I was thinking that an
>> implementation of the Lua language for Guile, under the auspices of
>> the GNU Project, would be a good fit.
> 
> Yes, I think that's a great idea.

I can only agree here :)  While I've never had contact with Lua so far, 
I really like the idea of implementing further support (especially for 
such languages that are -- as you write and also my impression is -- 
especially suited or even designed as extension languages)!

As I did work on Elisp last summer of code, my experience is that this 
is a very nice project ;)  (Speaking of Elisp, I finally have to find 
some time again to look into the missing pieces...  Hopefully soon, but 
at least during the summer there should again be enough free time.)

Good luck!

Daniel

Re: Potential Lua implementation for Guile?

[Ludovic Courtès]

Hi,

Neil Jerram <neil@ossau.uklinux.net> writes:

> No Itisnt <theseaisinhere@gmail.com> writes:
>
>> I want to apply for GSOC this year. Since Guile has recently obtained
>> a compilation/language framework, I was thinking that an
>> implementation of the Lua language for Guile, under the auspices of
>> the GNU Project, would be a good fit.
>
> Yes, I think that's a great idea.

+1

>> Moreover, I think the addition of Lua, similarly to Emacs Lisp, would
>> put Guile in a better position to subvert existing applications that
>> use Lua as their scripting language (a lot!).
>>
>> Let me know what you think.
>
> I think Guile finally seems to be approaching its original dream.

Yes!

Thanks,
Ludo’.

Re: Potential Lua implementation for Guile?

[Andy Wingo]

Hi!

Sorry for the delay in replying. I was fretting!

On Tue 09 Mar 2010 03:24, No Itisnt <theseaisinhere@gmail.com> writes:

> I want to apply for GSOC this year.

Fantastic!

See I actually looked at Lua when going to write a first second language
(?) for the VM, but decided on JS, thinking I couldn't do Lua well
enough. Then there have been these Lambda the Ultimate threads like
http://lambda-the-ultimate.org/node/3851 that made me think that if
Guile implements Lua, no matter who does it, we'll be stomped by Mike
Pall's LuaJIT!

And that's totally true, I think; I can't imagine ever being quite as
fast as what he does. Respect.

But, I do think Lua would be a useful language to have for application
extension by users. So this could be useful.

> Since Guile has recently obtained
> a compilation/language framework, I was thinking that an
> implementation of the Lua language for Guile, under the auspices of
> the GNU Project, would be a good fit.

So my grudging thought is, "OK". But if I could steer you to finish some
JS things or optimize some Elisp code or work on a Waddell-style inliner
or a Scheme native compiler, these are also interesting and useful
projects :-)

> Let me know what you think.

I think it could be a beautiful summer!

One thing I should mention up front though -- it sounds like you've got
the chops, but as summer approaches, it would be nice to start seeing
some code. The GNU project might not be accepted by the Evil
Empi^H^H^HGoogle this year. Guile might get no positions. Guile might
get two positions. Who knows? But if this is what you'd like to do this
summer, your case (and that of anyone else who's interested) will be
made stronger with plans and code.

Happy hacking!

Andy
-- 
http://wingolog.org/

Re: Potential Lua implementation for Guile?

[No Itisnt]

> So my grudging thought is, "OK". But if I could steer you to finish some
> JS things or optimize some Elisp code or work on a Waddell-style inliner
> or a Scheme native compiler, these are also interesting and useful
> projects :-)

I'm pretty focused on Lua for the summer project at the moment -- I
think it's good since it's kind of self-contained. Compilation is a
pretty nebulous topic, and one that I don't know much about.

> One thing I should mention up front though -- it sounds like you've got
> the chops, but as summer approaches, it would be nice to start seeing
> some code. The GNU project might not be accepted by the Evil
> Empi^H^H^HGoogle this year. Guile might get no positions. Guile might
> get two positions. Who knows? But if this is what you'd like to do this
> summer, your case (and that of anyone else who's interested) will be
> made stronger with plans and code.

Of course. What does the JS implementation need? I signed my FSF
paperwork today, hopefully that will cover future contributions.