lambda an (interactive) function but not a command

lambda an (interactive) function but not a command

[Emanuel Berg]

In my config file for the Gnus summary
buffer [1] I have this [2]. What I wonder is,
for the "F" key to invoke the
`gnus-summary-followup-inline' function, for
such a simple one-liner, why can't one use
a lambda for that, i.e.

    (lambda (interactive) (gnus-summary-mail-forward 4))

?

Only then I get the error

    (wrong-type-argument commandp ...

Is it because the whole thing is based on
names, and a lambda by definition doesn't have
a name, and so it can just be an
anonymous *function*, not an "anonymous
command" if you will?

TIA


[1] line 72 @ http://user.it.uu.se/~embe8573/emacs-init/gnus/summary.el

[2]

(defun gnus-summary-followup-inline ()
  (interactive)
  (gnus-summary-mail-forward 4) )

;;; KEYS

(let ((the-map gnus-summary-mode-map))
    (disable-super-global-keys the-map)
    (set-vertical-keys         the-map)
    ;; group
    (define-key the-map  "o"  #'gnus-summary-insert-old-articles)
    (define-key the-map  "w"  #'gnus-summary-exit-and-update-group)
    ;; reply
    (define-key the-map  "r"  #'gnus-summary-followup-with-original)
    (define-key the-map  "R"  #'gnus-summary-reply-with-original)
    ;; message
    (define-key the-map  "d"  #'gnus-summary-delete-article)
    (define-key the-map  "D"  #'gnus-summary-mark-as-dormant)
    (define-key the-map  "D"  #'gnus-summary-mark-as-dormant)
    (define-key the-map  "f"  #'gnus-article-fill-cited-article)
    (define-key the-map  "F"  #'gnus-summary-followup-inline)
    (define-key the-map  "U"  #'gnus-summary-uncool-subject)
    (define-key the-map  "z"  #'gnus-score-edit-current-scores)
    (define-key the-map "\r"  #'gnus-summary-show-article)
    ;; mark
    (define-key the-map  "l"  #'gnus-summary-put-mark-as-unread-next)
    (define-key the-map  "s"  #'gnus-summary-mark-as-read-forward)
    )

-- 
underground experts united
http://user.it.uu.se/~embe8573

Re: lambda an (interactive) function but not a command

[Michael Heerdegen]

Emanuel Berg <moasenwood@zoho.eu> writes:

> In my config file for the Gnus summary
> buffer [1] I have this [2]. What I wonder is,
> for the "F" key to invoke the
> `gnus-summary-followup-inline' function, for
> such a simple one-liner, why can't one use
> a lambda for that, i.e.
>
>     (lambda (interactive) (gnus-summary-mail-forward 4))

If you want this to work, you should add an argument list.  An
interactive lambda also needs one.  In the above function, (interactive)
is specified as argument list, so an interactive spec is missing and it
is not a command.

Michael.

Re: lambda an (interactive) function but not a command

[Emanuel Berg]

Michael Heerdegen wrote:

> If you want this to work, you should add an
> argument list. An interactive lambda also
> needs one. In the above function,
> (interactive) is specified as argument list,
> so an interactive spec is missing and it is
> not a command.

You mean like this

    (lambda (interactive) () (gnus-summary-mail-forward 4))

?

If so, it doesn't work

    (wrong-type-argument commandp (lambda
    (interactive) nil
    (gnus-summary-mail-forward 4))) in
    command-execute

Maybe you meant something else?

-- 
underground experts united
http://user.it.uu.se/~embe8573

Re: lambda an (interactive) function but not a command

[Amin Bandali]

How about this?

(lambda () (interactive) (gnus-summary-mail-forward 4))

Re: lambda an (interactive) function but not a command

[Michael Heerdegen]

Emanuel Berg <moasenwood@zoho.eu> writes:

> You mean like this
>
>     (lambda (interactive) () (gnus-summary-mail-forward 4))
>
> ?

No,

  (lambda () (interactive) (gnus-summary-mail-forward 4))

please.  Syntax is analogue to that of `defun': argument list comes
first.  See the docstring of `lambda' :-)


Michael.

Re: lambda an (interactive) function but not a command

[Emanuel Berg]

Michael Heerdegen wrote:

> No,
>
>   (lambda () (interactive) (gnus-summary-mail-forward 4))
>
> please. Syntax is analogue to that of
> `defun': argument list comes first. See the
> docstring of `lambda' :-)

Heh heh, in my Elisp init files, I have used
`lambda' no less than 42 times!

You have to read what I think, not what I
type :)

-- 
underground experts united
http://user.it.uu.se/~embe8573

quoting lambdas (was: Re: lambda an (interactive) function but not a command)

[Emanuel Berg]

Now when we are on the interesting subject of
lambdas, I've heard one shouldn't quote them.
This makes sense as they are code, not data.

But what about the hash-quote or
"function quote", i.e. the #' ? IIUC that tells
the byte-compiler that something is a function,
so the byte-compiler can examine if the
function is defined at that point, and possibly
even do some optimizations - but obviously the
byte-compiler should already know that a lambda
is a function!

So #' with lambdas should be
meaningless, right?

But I suppose if they are meaningless, one
might put them there just as well :)

What's the policy from Emacs High Command?

-- 
underground experts united
http://user.it.uu.se/~embe8573

Re: quoting lambdas

[Michael Heerdegen]

Emanuel Berg <moasenwood@zoho.eu> writes:

> So #' with lambdas should be meaningless, right?
>
> But I suppose if they are meaningless, one might put them there just
> as well :)

Yes both times, AFAIK.

It's even the case that `lambda' is a normal macro that implicitly
sharp-quotes "itself", e.g.

 (macroexpand '(lambda () 1))

  ==> #'(lambda nil 1)

Michael.