Re: [Sbcl-devel] Re: [Sbcl-help] ftype declarations for generic functions

[Nicolas N. <Nic...@iw...>]

William Harold Newman <wil...@ai...> writes:

> > I think for gfs a combination of types would be the most reasonable type
> > declaration, e.g. (or (function (string) fixnum) (function (list) fixnum)).
> > To be practically usable, this should probably adapt dynamically to new
> > method definitions...
> 
> If you are suggesting keeping the FTYPE definition as narrow as 
> the OR of all DEFMETHODs seen so far, I don't think that's going
> to work correctly.
> 
> file1.lisp:
>   (DEFGENERIC FOO (X))
>   (DEFMETHOD FOO ((X INTEGER)) ...)
>   ;; so here FTYPE FOO would be (OR (FUNCTION (INTEGER) *))?
> 
> file2.lisp:
>   (DEFUN FROB (X)
>     (ETYPECASE X 
>       (INTEGER (FOO X))
>       (REAL (LIST (FOO X) (EST-PRECISION X))) ;should not compile to type error
>       ...))
> 
> file3.lisp:
>   (DEFMETHOD FOO ((X SINGLE-FLOAT)) ...)
>   (DEFMETHOD FOO ((X DOUBLE-FLOAT)) ...)
>   ;; Now FTYPE FOO has become
>   ;;   (OR (FUNCTION (INTEGER) *)
>   ;;       (FUNCTION (SINGLE-FLOAT) *)
>   ;;       (FUNCTION (DOUBLE-FLOAT) *))
>   ;; but it's too late to avoid compile-time errors in DEFUN FROB.

Such type declarations should probably only be used (when compiling FROB)
if the generic function has been declared sealed.  BTW[1], were Gerd
Moellmann's PCL improvements ported to SBCL?

Nicolas.

[1] I am on sbcl-help/devel only since about one week and do not use SBCL
yet.