[PD] Re: [PD-announce] k_guile v0.0.1

[Larry Troxler]

On Saturday 10 January 2004 17:05, Michal Seta wrote:
> On Sat, 2004-01-10 at 11:34, Larry Troxler wrote:
> > Yes, very nice. I think this officially obsoletes my pd_scheme library.
> > (It was built on Scheme In One Day, which isn't a full-featured standard
> > Scheme).
>
> Since I think (or did I dream that up?) you said that you would port it
> possibly to guile I was silently waiting for that to happen... :)
>

Well, I'm sorry if you were holding your breath! II'd love to do a lot of 
things but my free time is limited. I'm glad it happened though. Sorry if I 
promised something that I didn't deliver (but I don't think I ever said that 
I was definitely going to do it,  only that it would be nice (?)) Grr. Maybe 
I had an optimisitic moment at one point and said I would port it that 
weekend, or something like that <sigh>.

> > Kjetil, I see you took a different approach than I did,
>
> What I liked in Larry's extern was that I could send arbitrary scheme
> expressions to it...  But I don't know if this is a real advantage,
> certainly helps trying out things...
>
> > For running something like Common Music (Michal, have you tried it yet?)
>
> I didn't really have time to play with it so much.  Not to mention that
> I'm only slowly learning scheme/lisp/CM.  I did manage to load CM and
> call some functions but I haven't actually done anything useful yet.
>

Well that's a good start! Was it straightforward? I'm probably going to try
 this tonight. It truly would be great to have a robust CM running inside PD. 

The missing element I think is that to take full advantage of Common Music (or 
any Lisp environment for that matter) in a PD environment, you would want an 
interactive listener into the interpreter that is running inside PD. In other 
words, in an ideal world, I'd like to be able to load up scheme inside PD, 
and then from Emacs or whatever, be able to evaluate expressions inside that 
interpreter, while PD is running (I imagine through an IP socket). 

> One thing I noticed, though, is that when you load the whole CM system
> into a script and then make changes in the script and reload (even
> completely destroying the object and creating a new one) it in pd, pd
> segfaults reliably (would there be a way to load CM once and then access
> it from other .scm scripts?). 

Hmm, right now I didn't even know that it's possible to reload a script in to 
a running PD object (did you set the script up to reload your source file 
upon receiving a PD message?). I should probably shut up now unti I actually 
try this external myself.

Larry