Re: FW: [Gcl-devel] Floating-point performance of GCL? Bug?

[Nicolas Neuss]

Nicolas Neuss <address@hidden> writes:

> > > I want to proceed as follows:
> > > 
> > > 1. Setup the most ANSI-compatible version of GCL
> > > 2. Setup defsystem.
> > > 3. Setup infix.
> > > 4. Look where I get hung for the first time with compiling Femlisp.
> > > 
> > 
> > CVS head is quite a bit more ansi compliant, but it is much more of a
> > moving target.  Personally, I recommend applications use the stable
> > branch if they can so we can concentrate on cvs head without worrying
> > about breakage.
> 
> OK, I used 2.6.3 now.  I succeeded in compiling and loading both infix and
> mk-defsystem after adding the nickname "COMMON-LISP" to the "LISP" package.
> Now I get stopped with
> 
> (let ((directory (pathname-directory *femlisp-pathname*)))
>   (setf (logical-pathname-translations "FEMLISP")
>       `(("**;*.*.*" 
>          ,(make-pathname :directory `(,@directory :wild-inferiors)
>                          :name :wild :type :wild :version :wild)))))
> 
> because the function LOGICAL-PATHNAME-TRANSLATIONS is not defined.  Is
> there a substitute for this function or should I wait for gcl-2.7.0?
> 
> Yours, Nicolas.

I have proceeded a little further by setting the pathname to be an ordinary
string.  Unfortunately, now MK:OOS yields an error:

>(mk:oos 'femlisp 'compile :minimal-load t :verbose t :compile-during-load t)

Error: MAKE::IGNORE-ERRORS is invalid as a function.
...

I thought that defsystem would work with GCL.  Is this wrong?  At least,
#+GCL appears in the file defsystem.lisp...

Thanks, Nicolas.