Re: #define CCLO

[Mikael Djurfeldt]

Michael Livshin <address@hidden> writes:

> "Dale P. Smith" <address@hidden> writes:
> 
> > Poking around in the goops code I noticed something, current CVS Guile
> > does not compile with CCLO un#define'ed.
> > 
> > Should more bits of guile be wrapped in #ifdef CCLO, or should the
> > #define just go away?
> 
> the current policy (if I'm not mistaken, I'm kinda out of the loop for 
> the time being) is to replace compiled closures by callable smobs, so
> the define should become irrelevant.

Your right, but the current compiled closures support tail-recursion
(if the cclo procedure is a lambda closure) while the current
applicable smobs do not.  So, we can't throw them out until we're
supporting tail-recursive applicable smobs.

It would be nice if we could support this in a generic way so that it
would even be possible to have tail-recursive primitive applicable
closures and tail-recursive byte-code closures.

> but it's irrelevant anyway, as evidently none of the developers
> bothers to try building with CCLO undefined.

Right.