Re: interactive-p

Re: interactive-p

[Lars Magne Ingebrigtsen]

Richard Stallman <rms@gnu.org> writes:

> Could you please check for bad uses of interactive-p in these files?

I'm not quite sure what counts as bad uses.  Functions that change
their behavior surprisingly based on whether they're called
interactively or not?

> Please ack when it's done.
>
> gnus/gnus-art.el

The usage here (in one function) seems to make sense.

> gnus/message.el

Look ok here, too.  (Only used in XEmacs, by the way.)

> gnus/pgg.el

Here it's used throughout the package, which seems odd.  I've Cc'd
this message to the Gnus mailing list; perhaps people there have
input on why pgg.el does it this way.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: interactive-p

[Simon Josefsson]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Richard Stallman <rms@gnu.org> writes:
>
>> Could you please check for bad uses of interactive-p in these files?
>
> I'm not quite sure what counts as bad uses.  Functions that change
> their behavior surprisingly based on whether they're called
> interactively or not?
...
>> gnus/pgg.el
>
> Here it's used throughout the package, which seems odd.  I've Cc'd
> this message to the Gnus mailing list; perhaps people there have
> input on why pgg.el does it this way.

IIUC, it do this to be able to have a nice user interface when used
interactively, e.g., switch to the output buffer after decrypting
something, but at the same be usable from lisp code that do not want
pgg to do anything with the user interface.

Is that bad?

The number of uses can be reduced, but I'm not sure I see why that is
better.

Re: interactive-p

[Richard Stallman]

    I'm not quite sure what counts as bad uses.  Functions that change
    their behavior surprisingly based on whether they're called
    interactively or not?

Please look at the latest commands.texi; it explains what is good
and bad usage of interactive-p.  I updated it a week ago.

Re: interactive-p

[Katsumi Yamaoka]

>>>>> In <E1CS5QB-00083T-Ee@fencepost.gnu.org> Richard Stallman wrote:

>     I'm not quite sure what counts as bad uses.  Functions that change
>     their behavior surprisingly based on whether they're called
>     interactively or not?

> Please look at the latest commands.texi; it explains what is good
> and bad usage of interactive-p.  I updated it a week ago.

The "Interective Call" section of the Elisp manual mentions a
good example.  We've fixed bad usages of interactive-p in the
emacs-w3m modules about two years ago.  At that time, TSUCHIYA
Masatoshi showed another example as follows:

(defun foo ()
  (interactive)
  (message (if (interactive-p) "Interactrive" "Non Interactrive")))

(let ((ad-default-compilation-action 'never))
  (defadvice foo
    (around foo-wrapper activate)
    ad-do-it))

(symbol-function 'foo)
 => (lambda nil
      "$ad-doc: foo$"
      (interactive)
      (let (ad-return-value)
	(setq ad-return-value (ad-Orig-foo))
	ad-return-value))

Because of this, the foo command will never say "Interactrive".