Re: strange behaviour in loop

OK, thanks for your time on it.

I agree with your observations.


On Wed, 24 Apr 2002, Steven Majewski wrote:

~~~
~~~On Wed, 24 Apr 2002, Frederic Udina wrote:
~~~
~~~> I observed that loop has a strange behaviour:
~~~...
~~~> > (loop foo foo  (incf ii) foo (when (> ii 10000) (return)))
~~~> NIL
~~~>
~~~> doesnt give any error despite foo isn't bound to anything
~~~...
~~~> Why is that? Is it a bug or is something I do not know about loop?
~~~
~~~Loop is supposed to evaluate each form, however in xlcont.c/xloop I
~~~see the code fragment:
~~~
~~~      for (argv = xlargv, argc = xlargc; ; xlargv = argv, xlargc = argc) {
~~~        while (moreargs()) {
~~~          arg = nextarg();
~~~          if (consp(arg))		## <-- !
~~~            xleval(arg);
~~~        }
~~~
~~~So it loops like it only evals conses.
~~~
~~~Thank looks like a bug.
~~~
~~~However, other than your expected error message, is there any other
~~~effect that you could expect from merely evaluating a symbol in that
~~~position ?  I guess you COULD look at it as a 'feature' of the
~~~optimization of omitting NO-OPs.
~~~
~~~Bug or feature ?
~~~
~~~I think I'ld still vote bug: mainly because that allows it to accept
~~~the syntax for the more complex Common Lisp extended LOOP macro
~~~without either an error message or correct evaluation:
~~~
~~~> ( loop for i from 1 to 10 (print 1) (return))
~~~
~~~1
~~~NIL
~~~>
~~~
~~~
~~~-- Steve
~~~
~~~


 Frederic Udina
  ____________________________________________________________________
  <[email protected]> Dept. D'Economia i Empresa   Universitat Pompeu Fabra
                  Ramon Trias Fargas, 25-27      08005 Barcelona SPAIN
  tel. 34- 935421756/1763 fax: 34- 935421746     http://libiya.upf.es/
  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  There are many here among us who feel that life is but a joke  B.Dy.