Re: closure within closure, incorrect (?) byte-compiler warnings

[Emanuel Berg]

Stefan Monnier via Users list for the GNU Emacs text editor wrote:

>>   (let ((next 1))
>>     (defun australian-survivor (ep)
>>       (interactive (list (read-number prompt next)))
>
> IIRC the current compiler code doesn't support this kind of
> "interactive spec nested within a lexical scope".

I did a little experiment (last) and the byte compiler
complains when there is a closure within a closure, now it
just gives the "reference to free variable" and not the
"Unused lexical variable" warning I got last time [1] -
I don't know why, but it's still not true ... no complains
when there is just a closure with `interactive'.

> Your code makes sense, but AFAIK it may not always work
> correctly :-(

You mean its worse than just the incorrect message?

Will it work when just evaluated, not byte-compiled?

And "not always work", how can that be?

> You might want to `M-x report-emacs-bug` (and if so, please
> put me in the `X-Debbugs-Cc:`).

Sure ... soon.

(declare-function set-num    nil)
(declare-function set-number nil)
(declare-function get-num    nil)

(let ((number 0))
  (defun set-num (num)
    (interactive "nNumber: ")
    (setq number num) )

  (defun set-number (&optional num)
    (interactive (list (read-number "Number: " 0)))
    (set-num num) )

  (defun get-num ()
    (interactive)
    (message "%f" number) ))

(declare-function set-num-x    nil)
(declare-function set-number-x nil)
(declare-function get-num-x    nil)

(let ((caption "Number-X: "))
  (let ((number-x 0))
    (defun set-num-x (num-x)
      (interactive "nNumber-X: ")
      (setq number-x num-x) )

    (defun set-number-x (&optional num-x)
      (interactive (list (read-number caption 0)))
      (set-num-x num-x) )

    (defun get-num-x ()
      (interactive)
      (message "%s%f" caption number-x) )))

;; geh.el:43:39: Warning: reference to free variable ‘caption’

[1] https://dataswamp.org/~incal/emacs-init/w3m/w3m-survivor.el

-- 
underground experts united
https://dataswamp.org/~incal