bug#57974: 2022-08-30; Enhancing help facility for function doc string

bug#57974: 2022-08-30; Enhancing help facility for function doc string

[Ikumi Keita]

Oops, sorry. I sent to wrong address. I reassigned this from "auctex" to
"emacs".

>>>>> Ikumi Keita <ikumi@ikumi.que.jp> writes:
> This is a wishlist for record rather than a bug report. See
> <URL:https://lists.gnu.org/r/auctex-devel/2022-09/msg00046.html> for the
> background.

> When a Elisp function A has :override advice B, help facility (C-h f)
> displays doc string A, not B. It only indicates
> This function has :override advice: THE-NAME-OF-B
> modestly.

> This is inconvenient, especially when A is a (major or minor) mode
> function and the user typed C-h m in that mode buffer.

> It would be worth considering to enhance the help facility to display
> doc string of B instead when A is overrided by advice.

> Regards,
> Ikumi Keita
> #StandWithUkraine #StopWarInUkraine

bug#57974: 2022-08-30; Enhancing help facility for function doc string

[Lars Ingebrigtsen]

Ikumi Keita <ikumi@ikumi.que.jp> writes:

> This is a wishlist for record rather than a bug report. See
> <URL:https://lists.gnu.org/r/auctex-devel/2022-09/msg00046.html> for the
> background.
>
> When a Elisp function A has :override advice B, help facility (C-h f)
> displays doc string A, not B. It only indicates
> This function has :override advice: THE-NAME-OF-B
> modestly.
>
> This is inconvenient, especially when A is a (major or minor) mode
> function and the user typed C-h m in that mode buffer.
>
> It would be worth considering to enhance the help facility to display
> doc string of B instead when A is overrided by advice.

I'm not quite sure I follow your meaning here.

Do you have a complete recipe, starting from "emacs -Q", that displays
the problem?

bug#57974: 2022-08-30; Enhancing help facility for function doc string

[Ikumi Keita]

Hi Lars,

>>>>> Lars Ingebrigtsen <larsi@gnus.org> writes:
> I'm not quite sure I follow your meaning here.

> Do you have a complete recipe, starting from "emacs -Q", that displays
> the problem?

1. Save the following file as /tmp/foo.el
----------------------------------------------------------------------
(defun func-A ()
  "doc string of A"
  nil)

(defun func-B ()
  "doc string of B"
  t)

(advice-add 'func-A :override #'func-B)
----------------------------------------------------------------------
2. emacs -Q -l /tmp/foo.el
3. Type C-h f func-A RET
4. Then *Help* buffer pops up and shows
----------------------------------------------------------------------
func-A is a Lisp function in ‘/tmp/foo.el’.

(func-A)

doc string of A

This function has :override advice: ‘func-B’.
----------------------------------------------------------------------
The function A behaves as if it were the function B by the nature of
:override advice, so the doc string of A is almost useless. The actual
behavior is described by the doc string of B.

Unfortunately, when A is a mode function and has a long doc string,
users can easily look over the line
This function has :override advice: ‘func-B’.
far downward in the *Help* buffer when they see the output of C-h m.
That results in that they miss the "true" mode description provided by
the doc string of B.

Stephan suggested that such situation can be improved as
,----
| Maybe the help facility should be more careful in how it builds the
| docstring when a function has an `:override` advice, indeed.
`----
, in https://lists.gnu.org/r/auctex-devel/2022-09/msg00046.html

Regards,
Ikumi Keita
#StandWithUkraine #StopWarInUkraine

bug#57974: 2022-08-30; Enhancing help facility for function doc string

[Lars Ingebrigtsen]

Ikumi Keita <ikumi@ikumi.que.jp> writes:

> The function A behaves as if it were the function B by the nature of
> :override advice, so the doc string of A is almost useless. The actual
> behavior is described by the doc string of B.
>
> Unfortunately, when A is a mode function and has a long doc string,
> users can easily look over the line
> This function has :override advice: ‘func-B’.
> far downward in the *Help* buffer when they see the output of C-h m.
> That results in that they miss the "true" mode description provided by
> the doc string of B.
>
> Stephan suggested that such situation can be improved as
> ,----
> | Maybe the help facility should be more careful in how it builds the
> | docstring when a function has an `:override` advice, indeed.
> `----
> , in https://lists.gnu.org/r/auctex-devel/2022-09/msg00046.html

Right, makes sense.  But there can be more than one advice -- would
moving just the :overrides to the front of the doc string solve the
problem?

bug#57974: 2022-08-30; Enhancing help facility for function doc string

[Stefan Monnier via Bug reports for GNU Emacs, the Swiss army knife of text editors]

> Right, makes sense.  But there can be more than one advice -- would
> moving just the :overrides to the front of the doc string solve the
> problem?

It used to be there, so maybe that's a good solution, yes.


        Stefan

bug#57974: 2022-08-30; Enhancing help facility for function doc string

[Ikumi Keita]

>>>>> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>> Right, makes sense.  But there can be more than one advice -- would
>> moving just the :overrides to the front of the doc string solve the
>> problem?

> It used to be there, so maybe that's a good solution, yes.

More radical approach would also be an option, i.e., to display doc
string of B instead of A; but I don't strongly insist.

Bye,
Ikumi Keita
#StandWithUkraine #StopWarInUkraine

bug#57974: 2022-08-30; Enhancing help facility for function doc string

[Stefan Monnier via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Ikumi Keita [2022-09-23 00:22:06] wrote:
>>>>>> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>>> Right, makes sense.  But there can be more than one advice -- would
>>> moving just the :overrides to the front of the doc string solve the
>>> problem?
>> It used to be there, so maybe that's a good solution, yes.
> More radical approach would also be an option, i.e., to display doc
> string of B instead of A; but I don't strongly insist.

I don't have a strong opinion either.  I think the issue has more to do
about how to implement it (the info about advice is provided via the
same mechanism used for all kinds of other auxiliary info and is not
designed to distinguish those things that are more or less important,
it has a very limited flexibility right now).


        Stefan

bug#57974: 2022-08-30; Enhancing help facility for function doc string

[Lars Ingebrigtsen]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> I don't have a strong opinion either.  I think the issue has more to do
> about how to implement it (the info about advice is provided via the
> same mechanism used for all kinds of other auxiliary info and is not
> designed to distinguish those things that are more or less important,
> it has a very limited flexibility right now).

Today, I think the display is something along the lines of:

---
func-A is a Lisp function in /tmp/foo.el.

(func-A)

doc string of A

This function has :after advice: func-D.
This function has :before advice: func-C.
This function has :override advice: func-B.
---

Moving the :override line to the start would be easy enough, but should
perhaps also be rephrased to something like:

---
This function has :override advice: func-B, so func-A isn't run at all,
and the documentation below is possibly irrelevant.  See the
documentation to func-B instead.
---

bug#57974: 2022-08-30; Enhancing help facility for function doc string

[Ikumi Keita]

Hi Lars,

>>>>> Lars Ingebrigtsen <larsi@gnus.org> writes:
> Moving the :override line to the start would be easy enough, but should
> perhaps also be rephrased to something like:

> ---
> This function has :override advice: func-B, so func-A isn't run at all,
> and the documentation below is possibly irrelevant.  See the
> documentation to func-B instead.
> ---

Thanks, I think the proposed stuff is organized well enough.

Regards,
Ikumi Keita
#StandWithUkraine #StopWarInUkraine

bug#57974: 2022-08-30; Enhancing help facility for function doc string

[Lars Ingebrigtsen]

Ikumi Keita <ikumi@ikumi.que.jp> writes:

> Thanks, I think the proposed stuff is organized well enough.

I've now done this in Emacs 29.