Re: [clisp-list] can mapcan do this?

[Sam S. <sd...@gn...>]

> * In message <200...@te...>
> * On the subject of "[clisp-list] can mapcan do this?"
> * Sent on Thu, 27 Feb 2003 08:25:42 -0500
> * Honorable Andrew Pimlott <an...@pi...> writes:
>
> This blew me away:
> 
>     > (setf c '((PIG2 BLUE4) (PIG1 RED)))
>     ((PIG2 BLUE4) (PIG1 RED))
>     > (mapcan #'(lambda (p) (cdr (assoc p c))) '(PIG1 PIG2)) 
>     (RED BLUE4)
>     > c
>     ((PIG2 RED BLUE4) (PIG1 BLUE4))
> 
> c has been modified, but not in the way I expected: the cdrs of
> '(PIG2 BLUE4) and '(PIG1 RED) have been swapped.  Does this mean
> that the lambda really returns by reference?  And what is the point
> of this?  It seems to me that whichever way you combine the two
> lists, you need three pointers (one for the beginning of the first
> list, one for the beginning of the second, and one to traverse the
> first), so why pick the way that has unintuitive results?

You did not specify which version of CLISP you are using.
I must assume that it is not the current CVS head (which behaves the way
you expect).

> Moreover, the reference I am using
> 
>     http://www.lispworks.com/reference/HyperSpec/Body/f_mapc_.htm#mapcan
> 
> indicates that (mapcar ...) is equivalent to (apply #'nconc (mapcar
> ...)), and
> 
>     > (setf c '((PIG2 BLUE4) (PIG1 RED)))
>     ((PIG2 BLUE4) (PIG1 RED))
>     > (apply #'nconc (mapcar #'(lambda (p) (cdr (assoc p c))) '(PIG1 PIG2)))
>     (RED BLUE4)
>     > c
>     ((PIG2 BLUE4) (PIG1 RED BLUE4))
> 
> So is mapcan allowed to do this?  Or am I wrong in understanding that
> the two must have the same side-effects?

It may be argued both ways.
At any rate, the next release will behave the way you expect.

-- 
Sam Steingold (http://www.podval.org/~sds) running RedHat8 GNU/Linux
<http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/>
<http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html>
PI seconds is a nanocentury