Re: LIST-TABLES problem on LispWorks

On 8/13/06, Kevin Rosenberg <kevin-HJRc7zDS/[email protected]> wrote:
> Bill Atkins wrote:
> > I got the UFFI tests to pass by changing ensure-char-storable in
> > uffi/src/objects.lisp to:
> >
> > (defmacro ensure-char-storable (obj)
> >  #+(or digitool (and lispworks (not lispworks5.0))) obj
> >  #+(or lispworks allegro lispworks5.0 cmu sbcl scl openmcl) `(char-code
> >  ,obj))
>
> I reviewed the function. Is the below what you meant? Otherwise, in
> your code, lispworks 4.4 and below would be selected in both reader
> conditionals.
>
> (defmacro ensure-char-storable (obj)
>   #+(or digitool (and lispworks (not lispworks5.0))) obj
>   #+(or lispworks5.0 allegro cmu sbcl scl openmcl) `(char-code ,obj))

Yes, my mistake.

> Assuming that Lispworks intended to make this change in their FFI, it
> will be somewhat problematic in the future since each new version of
> lispworks will require being added to the lispworks5.0 conditionals.
>
> Does you version also define lispworks5? If so, at least that should
> be present in all future 5.x releases and would be better than
> lispworks5.0, assuming Lispworks FFI doesn't change again.

It does.  I was looking for :lispworks5 originally and must have missed it.

> > However, the CLSQL problem still exists.  I cleaned out all my fasls,
> > did a make clean in the clsql/uffi directory, and rebuilt, but I still
> > get the same output from LIST-TABLES.  I'll let you know if I  make
> > any progress on the CLSQL problem.
>
> All the foreign string conversion for all backends takes place in
> uffi/clsql-uffi.lisp in the function CONVERT-RAW-FIELD. As you can
> see, for strings that function invokes UFFI:CONVERT-FROM-FOREIGN-STRING.
>
> --
> Kevin Rosenberg
> [email protected]
>


-- 
Bill Atkins