Re: CL compatibility question

Re: CL compatibility question

[Lawrence Mitchell]

Joe Corneli wrote:

[...]

> In ELISP, not so great.

> (defbashfun taco (beef beans)
>   (car $@))
> (taco t nil)
> ;=> Debugger entered--Lisp error: (wrong-number-of-arguments
> (lambda (name lambda &body body) (\` (defun (\, name) (&rest
> $@) (apply (lambda (\, lambda) (\,@ body)) $@)))) 3)

This is because &body isn't a recognised lambda list keyword in
emacs lisp.  If you want to use full CL-style lambda lists in
DEFUN and DEFMACRO forms, you should do:

(require 'cl)

and use the DEFUN* and DEFMACRO* forms (note star).

(defmacro* defbashfun (name lambda &body body)
  `(defun ,name (&rest $@)
     (apply (lambda ,lambda ,@body) $@)))

(defbashfun taco (beef beans) (car $@))

(taco t nil) => t

Alternately, define defbashfun as:

(defmacro defbashfun (name lambda &rest body)
  ...)

[...]

-- 
Lawrence Mitchell <wence@gmx.li>

Re: CL compatibility question

[Stefan Monnier]

> (defmacro defbashfun (name lambda &body body)
                                    ^^^^^
This does not exist in `defmacro'.
You can use (require 'cl) and then `defmacro*' which supports &body IIRC.


        Stefan

Re: CL compatibility question

[Miles Bader]

Joe Corneli <jcorneli@math.utexas.edu> writes:
> (defmacro defbashfun (name lambda &body body)
>   `(defun ,name (&rest $@)
>      (apply (lambda ,lambda ,@body) $@)))
...
> Why is this?  Are there plans that would change this behavior in the
> future?

Use &rest, not &body (in CL, &body is really just &rest with a hint to
the pretty-printer).

-Miles
-- 
[|nurgle|]  ddt- demonic? so quake will have an evil kinda setting? one that
            will  make every christian in the world foamm at the mouth?
[iddt]      nurg, that's the goal

CL compatibility question

[Joe Corneli]

Frode Vatvedt Fjeld suggested (on comp.lang.lisp, 2004-05-06) the
following approach to building something like Bash's `$@' in LISP:

(defmacro defbashfun (name lambda &body body)
  `(defun ,name (&rest $@)
     (apply (lambda ,lambda ,@body) $@)))

In CLISP, this works great.  

(defbashfun taco (beef beans)
  (car $@))
(taco t nil)
;=> T

In ELISP, not so great.

(defbashfun taco (beef beans)
  (car $@))
(taco t nil)
;=>

Debugger entered--Lisp error: (wrong-number-of-arguments (lambda (name lambda &body body) (\` (defun (\, name) (&rest $@) (apply (lambda (\, lambda) (\,@ body)) $@)))) 3)
  (lambda (name lambda &body body) (\` (defun ... ... ...)))(taco (beef beans) (car $@))
  (defbashfun taco (beef beans) (car $@))
  eval((defbashfun taco (beef beans) (car $@)))
  eval-last-sexp-1(nil)
  eval-last-sexp(nil)
* call-interactively(eval-last-sexp)

Why is this?  Are there plans that would change this behavior in the
future?