[Ecls-list] C-INLINE and LET bindings

[Juan Jose Garcia-Ripoll]

2006/12/30, Marco Gidde <marco.gidde at tiscali.de>:
> while playing around with ECL and CFFI I got a strange compiler error
> that did not occur with e.g. SBCL. I was able to strip it down to the
> following code: [...]
> Both functions are the same except for the superfluous LET in
> C-MALLOC-2. The problem is that the first form compiles while the
> second doesn't. The error message is
>
> > ;;; Compiling ecl.lsp.
> > ;;; Compiling (DEFUN C-MALLOC-2 ...).
> > Not a condition type: C::FOREIGN-DATA
> > No restarts available.
> > Broken at MAKE-CONDITION.

This was an typo in the type-inference code from the compiler. Instead
of the lisp type SI::FOREIGN-DATA, there appeared just FOREIGN-DATA.
Type inference was only activated by the LET form because it had to
analize whether the let-variable could be replaced by just the inline
form.

Fixed in CVS.

Regards,

Juanjo

-- 
Dpto. de Fisica Teorica I, Fac. de CC Fisicas, Universidad Complutense,
Ciudad Universitaria s/n Madrid 2804 (Spain)
http://teorica.fis.ucm.es/~jjgarcia/