Thanks John,

I don't disagree about the utility of cl-flet or cl-labels.

Because 'cl is documented in an independent document it is unclear to me
how that example would fit into the current editorial structure of the *GNU
Emacs Lisp Reference Manual*.

Obviously if 'cl-flet and 'cl-labels were merged into GNU Emacs Lisp, the
example would work to illustrate a method for writing a locally defined
functions. With the added benefit of avoiding the more cumbersome 'funcall
structure.

Ben




On Sun, Jan 29, 2017 at 1:20 PM, John Wiegley <jwiegley@gmail.com> wrote:

> >>>>> "br" == ben rudgers <ben.rudgers@gmail.com> writes:
>
> br> The manual does not provide an example showing a locally scoped
> function.
> br> Though the behavior can potentially be inferred from the documentation
> of
> br> functions, an example showing the relationship between =let= and
> =funcall=
> br> might be of practical help for someone who has not made a deep dive
> into
> br> the manual.
>
> I think it would be better to encourage the use of `cl-flet' or
> `cl-labels',
> as though macros are specifically intended for binding local function
> definitions, and will automatically handle whether the function should be a
> `lambda' or a `closure', depending on the setting of `lexical-binding'.
>
> Your approach may have merit in terms of pedagogy, but I'd prefer to nudge
> users toward more idiomatic usage from the beginning.
>
> @example
> (require 'cl-macs)
> (defun foo (x)
>   (cl-flet ((bar (y) (+ x y)))
>     (bar 3)))
> (foo 4)
>      @result{} 7
> @end example
>
> --
> John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
> http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2
>