Re: Error messages during load of .scm file

[Sven Hartrumpf <Sven dot Hartrumpf at FernUni-Hagen dot de>]

On 25 May 2003, Jim White <jim@pagesmiths.com> wrote: 
> > Do I have to eliminate the cause of these warnings?
> 
> No, the VerifyError is the problem which says it has something to do 
> with setting a field.
> 
> The indication of being "atInteractiveLevel" should mean that it is code 
> that is being read from the console.  If it is not, then perhaps it is 
> in a file and is a statement that is interpreted directly rather being 
> compiled first.  If that is the case (as it appears from the above), 
> then the only thing I can suggest at the moment is to break down the 
> load of those files to isolate which statement is going wrong.

O.K. I have isolated the following reduced function "make-mf" (which makes
not much sense any more, but shows the error when loaded into Kawa, although
other Schemes can load it):

(define set-union (lambda (compare a b)
  (append a b)
  ))

(define make-mf (lambda (forms base gergen-results)
  (for-each
    (lambda (gergen-result)
      (for-each
        (lambda (form)
          (set-union
            (lambda (a b)
              (let ((result (string<? (car a) (car b))))
                (if result
                  (string<? (cadr a) (cadr b))
                  result)))
            forms
            (list (list
                    (car form)
                    (car form)))))
        gergen-result))
    gergen-results)
  ; store forms
  (call-with-output-file "a"
    (lambda (output-stream)
      (for-each (lambda (form)
                  (newline output-stream))
                forms)))
  ))


The funny thing is that if I replace "forms" in the last line by '("a"),
it can be loaded by Kawa. Is there some (strange) type inferencing going on
for the parameter "forms"?

Greetings
Sven