bug#28781: 25.3; Override erroneous "incorrect verb voice" from checkdoc

bug#28781: 25.3; Override erroneous "incorrect verb voice" from checkdoc

[Radon Rosborough]

I have the following definition:

    (defmacro straight--functionp (object)
      "Non-nil if OBJECT, an unquoted symbol, evaluates to a function.
    However, if OBJECT evaluates to its own symbol value or t, then
    return nil. This is useful for allowing a function to be called
    with nil, non-nil, or a function object, without worrying about
    the non-nil value being interpreted as a function: just call the
    function with the quoted name of the argument, or use t."
      (let ((object-sym (make-symbol "object")))
        `(let ((,object-sym ,object))
           (and (not (memq ,object-sym '(,object t)))
                (functionp ,object-sym)))))

M-x checkdoc reports the following on the first line:

    Probably "evaluates" should be imperative "evaluate"

This is wrong. How can I suppress this spurious warning, other than by
rewording the docstring? I know that checkdoc has
`checkdoc-symbol-words' for suppressing spurious warnings about unquoted
symbols, but there appears to be nothing similar I can customize in
file-local variables for this case.

bug#28781: 25.3; Override erroneous "incorrect verb voice" from checkdoc

[Eli Zaretskii]

> From: Radon Rosborough <radon.neon@gmail.com>
> Date: Tue, 10 Oct 2017 10:52:18 -0700
> 
> I have the following definition:
> 
>     (defmacro straight--functionp (object)
>       "Non-nil if OBJECT, an unquoted symbol, evaluates to a function.

Try

    "Return non-nil if OBJECT, an unquoted symbol, evaluates to a function."

Or even

    "Return non-nil if OBJECT, evaluates to a function.
  OBJECT should be an unquoted symbol."

bug#28781: 25.3; Override erroneous "incorrect verb voice" from checkdoc

[Radon Rosborough]

> Try
>
>     "Return non-nil if OBJECT, an unquoted symbol, evaluates to a
>     function."

That makes sense. Is it recommended to have all function docstrings
start with a verb? If so, maybe the docstring for `functionp' should
be changed (after all, it was what I modeled this one off of):

    Non-nil if OBJECT is a function.

bug#28781: 25.3; Override erroneous "incorrect verb voice" from checkdoc

[Eli Zaretskii]

> From: Radon Rosborough <radon.neon@gmail.com>
> Date: Tue, 10 Oct 2017 12:44:05 -0700
> Cc: 28781@debbugs.gnu.org
> 
> Is it recommended to have all function docstrings start with a verb?

Yes.  From the ELisp manual, node "Documentation Tips"

     For a function, the first line should briefly answer the question,
     “What does this function do?” For a variable, the first line should
     briefly answer the question, “What does this value mean?”

(It talks about the first line of the doc string.)

> If so, maybe the docstring for `functionp' should be changed (after
> all, it was what I modeled this one off of):
> 
>     Non-nil if OBJECT is a function.

Yes, I think so.

bug#28781: 25.3; Override erroneous "incorrect verb voice" from checkdoc

[Radon Rosborough]

[-- Attachment #1: Type: text/plain, Size: 211 bytes --]

> > If so, maybe the docstring for `functionp' should be changed (after
> > all, it was what I modeled this one off of):
> >
> >     Non-nil if OBJECT is a function.
>
> Yes, I think so.

The patch is attached.

[-- Attachment #2: 0001-Fix-docstring-style-for-functionp.patch --]
[-- Type: application/octet-stream, Size: 787 bytes --]

From 8cf42f2e59935450fbdadbc6dd8b9ccc564da7cc Mon Sep 17 00:00:00 2001
From: Radon Rosborough <radon.neon@gmail.com>
Date: Tue, 10 Oct 2017 13:28:28 -0700
Subject: [PATCH] Fix docstring style for `functionp'

* src/eval.c (functionp): Make docstring start with verb.
---
 src/eval.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/eval.c b/src/eval.c
index acda64e7f0..3f279454b1 100644
--- a/src/eval.c
+++ b/src/eval.c
@@ -2681,7 +2681,7 @@ call8 (Lisp_Object fn, Lisp_Object arg1, Lisp_Object arg2, Lisp_Object arg3,
 }
 
 DEFUN ("functionp", Ffunctionp, Sfunctionp, 1, 1, 0,
-       doc: /* Non-nil if OBJECT is a function.  */)
+       doc: /* Return non-nil if OBJECT is a function.  */)
      (Lisp_Object object)
 {
   if (FUNCTIONP (object))
-- 
2.14.2

bug#28781: 25.3; Override erroneous "incorrect verb voice" from checkdoc

[Eli Zaretskii]

> From: Radon Rosborough <radon.neon@gmail.com>
> Date: Tue, 10 Oct 2017 13:31:33 -0700
> Cc: 28781@debbugs.gnu.org
> 
> 
> [1:text/plain Hide]
> 
> > > If so, maybe the docstring for `functionp' should be changed (after
> > > all, it was what I modeled this one off of):
> > >
> > >     Non-nil if OBJECT is a function.
> >
> > Yes, I think so.
> 
> The patch is attached.

Thanks, pushed.