Re: [Chicken-users] Error 70, what does it mean?

[Michael Erdmann]

Felix Winkelmann wrote:
> Michael Erdmann wrote:
>
> > Felix Winkelmann wrote:
> >
> > > Michael Erdmann wrote:
> > >
> > > > Hallo, i am trying to compile slib, and i get the following strange
> > > > error:
> > > >
> > > > chicken -syntax fluidlet.scm
> > > > compiling `fluidlet.scm' ...
> > > > Error: #(syntax-object fluid-let ((#f top) shift #(ribcage 
> > > > (#(import-token *top*)) () ())))
> > > >
> > > > make: *** [fluidlet.c] Error 70
> > > > address@hidden:~/scheme/slib>
> > > >
> > > >
> > > > Error 70, what doe it mean?
> > >
> > > 70 is just the standard error return code (EX_SOFTWARE in
> > > sysexits.h).
> > >
> > > Thw actual problem appears to be that you are compiling
> > > SLIBs version of the fluid-let macro in "high-level macro"
> > > mode (with syntax-rules/syntax-case macros enabled).
> > > You don't really need that, since Chicken already supports
> > > fluid-let, though.
> > > I can't tell you exactly what happens here, probably the
> > > definition of the macro does something fishy.
> > > Could you show me the actual source-code?
> > Well i have managed to come around this problem, but actually
> > i am not sure what i have done. First of all i did run chicken
> >
> >
> > address@hidden:~/scheme/slib> csc -c  fluidlet.scm
> > Error: syntax error in line 22 - malformed expression `(clauses . body)'
> > *** Shell command terminated with exit status 1: 
> > /usr/local/bin/chicken fluidlet.scm -output-file fluidlet.c -quiet
> >
> > which was leading me to the dot in the argument list, which
> > i found strange as well. I have deleted this '.' and the component
> > compiles :-)  Bit i am not sure if this is correct!
>
> > (defmacro fluid-let (clauses . body)
> >   (let ((ids (map car clauses))
> >     (new-tmps (map (lambda (x) (gentemp)) clauses))
> >     (old-tmps (map (lambda (x) (gentemp)) clauses)))
> >     `(let (,@(map list new-tmps (map cadr clauses))
> >        ,@(map list old-tmps (map (lambda (x) #f) clauses)))
> >        (dynamic-wind
> >        (lambda ()
> >          ,@(map (lambda (ot id) `(set! ,ot ,id))
> >             old-tmps ids)
> >          ,@(map (lambda (id nt) `(set! ,id ,nt))
> >             ids new-tmps))
> >        (lambda () ,@body)
> >        (lambda ()
> >          ,@(map (lambda (nt id) `(set! ,nt ,id))
> >             new-tmps ids)
> >          ,@(map (lambda (id ot) `(set! ,id ,ot))
> >             ids old-tmps))))))
>
> `defmacro' is not supported by Chicken, but you can implement it
> with
>
> (define-macro (defmacro name llist . body)
>   `(define-macro (,name ,@llist) ,@body) )
Thanks :-(. May be i have to read more in detail what is different
or not supported. defmacro is scheme or an extension?

> The 'require' of dynamic wind is not necessary, and will probably
> break chicken. The compiled form of fluidlet.scm will also
> not work with other files, since macros only exist at compile-time,
> not at run-time (usually).
> Last but not least, you don't have to do this, since Chicken already
> *has* fluid-let. :-)
>
> Are you trying to compile SLIB? I recommend not to use the infrastructure
> and just grab those files you are interested in and massage them into
> a form that can be used with Chicken. SLIB has several nasty suprises
> for Scheme implementations that use separate compilation. If you need
> specific packages of SLIB, just tell me and I will try to port them
> to Chicken, if possible.
Thanls, not in the moment, but it seems slib seems to be used
quite regulary. I have stoped by attemps to recompile schmooz and
soft_scm for a moment since the effort seems the be considereable.

I am quite firstrate regarding the reuse of code, but proably
i need to gain some additional experience in it.

Michael

> cheers,
> felix