Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.emacs > #1610

Re: `list-of' macro snippet [regarding Comprehensions]

From "Pascal J. Bourguignon" <pjb@informatimago.com>
Newsgroups gnu.emacs.sources, gnu.emacs.help, comp.emacs, comp.lang.lisp, comp.lang.functional
Subject Re: `list-of' macro snippet [regarding Comprehensions]
Date 2012-11-03 04:09 +0100
Organization Informatimago
Message-ID <87pq3vmlg5.fsf@informatimago.com> (permalink)
References <44777946-db62-48fa-9e99-2fd06c6d296c@g18g2000vbf.googlegroups.com> <b4135226-2e01-494a-bec7-6a215fdc723d@w2g2000vbc.googlegroups.com> <f9dfcdb2-9d1c-4307-8d1b-588760466ed3@b12g2000vbg.googlegroups.com> <d909bf45-a8b5-4974-8135-a101af3531d6@y6g2000vbb.googlegroups.com>

Cross-posted to 5 groups.

Show all headers | View raw


Rivka Miller <rivkaumiller@gmail.com> writes:

> This is not that hard.
>
> You can take the defmacro  comp out of the defun.
>
> It should not be hard
> ----------------------------------------
> background info in gnu.emacs.help
>
> I spent a few hours poring over and fixed some of the variables and
> backquotes and character codes.
>
> The defmacro is now only nested in one function where it is needed.
> Hope, someone can help get it to work and produce some kind of demo
> examples.
>
> (defun open-bracket (stream ch)
>
>   (defmacro comp ((e &rest qs) l2)
>     (if (null qs) `(cons ,e ,l2)        ; rule A
>       (let ((q1 (car qs))
>             (q (cdr qs)))
>         (if (not(eq (cadr q1) '<-))  ; a generator?
>             `(if ,q1 (comp (,e ,@q),l2) ,l2) ; rule B
>           (let ((v (car q1))            ; rule C
>                 (l1 (third q1))
>                 (h (gentemp "H-"))
>                 (us (gentemp "US-"))
>                 (us1 (gentemp "US1-")))
>             `(labels ((,h (,us)         ; corresponds to a letrec
>                           (if (null ,us) ,l2
>                             (let ((,v (car ,us))
>                                   (,us1 (cdr ,us)))
>                               (comp (,e ,@q) (,h ,us1))))))
>                (,h ,l1)))))))
>
>   (do ((l nil)
>        (c (read stream t nil t)(read stream t nil t)))
>       ((eq c '|]|) `(comp ,(reverse l) ()))
>   (push c l))
> )
>
> (defun closing-bracket (stream ch) '|]|)
>
> (eval-when (compile load eval)
>   (set-macro-character #\[ #'open-bracket)
>   (set-macro-character #\] #'closing-bracket))


Why are you repeating this code here in comp.lang.lisp?

I already told you that it was nonsensical!

When you put the defmacro for inside the defun, it may not have
compile-time effects.  Therefore when you compile a file containing
brackets, the reader macro function will be called, the macro will be
defined, but maybe only into the run-time environment, not into the
compilation environment.  Therefore the compiler may not know that COMP
is a macro, and it may very well signal an error when compiling a
bracket expression!

Why do you insist putting the defmacro inside the defun?

-- 
__Pascal Bourguignon__
http://www.informatimago.com

Back to comp.emacs | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread


Thread

Re: `list-of' macro snippet [regarding Comprehensions] Rivka Miller <rivkaumiller@gmail.com> - 2012-10-28 17:49 -0700
  Re: `list-of' macro snippet [regarding Comprehensions] Rivka Miller <rivkaumiller@gmail.com> - 2012-10-28 19:20 -0700
    Re: `list-of' macro snippet [regarding Comprehensions] Rivka Miller <rivkaumiller@gmail.com> - 2012-11-02 19:06 -0700
      Re: `list-of' macro snippet [regarding Comprehensions] "Pascal J. Bourguignon" <pjb@informatimago.com> - 2012-11-03 04:09 +0100
        Re: `list-of' macro snippet [regarding Comprehensions] Rivka Miller <rivkaumiller@gmail.com> - 2012-11-03 14:57 -0700
          Re: `list-of' macro snippet [regarding Comprehensions] Rivka Miller <rivkaumiller@gmail.com> - 2012-11-07 09:08 -0800
            Re: `list-of' macro snippet [regarding Comprehensions] "Pascal J. Bourguignon" <pjb@informatimago.com> - 2012-11-07 23:30 +0100

csiph-web