Why a defmacro that calls a defun?

Why a defmacro that calls a defun?

[Eric Abrahamsen]

I'm poking through Gnus, and seeing a fair number of places with code
that looks like this:

(defmacro nnoo-define-skeleton (backend)
  `(eval-and-compile
     (nnoo-define-skeleton-1 ',backend)))

(defun nnoo-define-skeleton-1 (backend)
  ...)

What does this actually do? All I can see happening is that a quote mark
is added before the BACKEND argument in the macro. What is happening
here?

Thanks,
Eric

Re: Why a defmacro that calls a defun?

[Barry Margolin]

In article <mailman.668.1506986797.27995.help-gnu-emacs@gnu.org>,
 Eric Abrahamsen <eric@ericabrahamsen.net> wrote:

> I'm poking through Gnus, and seeing a fair number of places with code
> that looks like this:
> 
> (defmacro nnoo-define-skeleton (backend)
>   `(eval-and-compile
>      (nnoo-define-skeleton-1 ',backend)))
> 
> (defun nnoo-define-skeleton-1 (backend)
>   ...)
> 
> What does this actually do? All I can see happening is that a quote mark
> is added before the BACKEND argument in the macro. What is happening
> here?

It just causes the function to be called at compile time as well as run 
time, due to the use of eval-and-compile.

-- 
Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***

Re: Why a defmacro that calls a defun?

[Eric Abrahamsen]

Barry Margolin <barmar@alum.mit.edu> writes:

> In article <mailman.668.1506986797.27995.help-gnu-emacs@gnu.org>,
>  Eric Abrahamsen <eric@ericabrahamsen.net> wrote:
>
>> I'm poking through Gnus, and seeing a fair number of places with code
>> that looks like this:
>> 
>> (defmacro nnoo-define-skeleton (backend)
>>   `(eval-and-compile
>>      (nnoo-define-skeleton-1 ',backend)))
>> 
>> (defun nnoo-define-skeleton-1 (backend)
>>   ...)
>> 
>> What does this actually do? All I can see happening is that a quote mark
>> is added before the BACKEND argument in the macro. What is happening
>> here?
>
> It just causes the function to be called at compile time as well as run 
> time, due to the use of eval-and-compile.

Huh, thanks for that. I guess it's obvious, but it was not clicking for me.