bug#3108: 23.0.92; 2C-command: missing arglist

bug#3108: 23.0.92; 2C-command: missing arglist

[Drew Adams]

1. C-h f 2C-command shows this:
 
2C-command is an interactive autoloaded keymap in `two-column.el'.
 
It is bound to <f2>, C-x 6.
 

Not documented.
 

2. In Emacs 22, it shows this:
 
2C-command is an interactive autoloaded keymap in `two-column'.
It is bound to <f2>, C-x 6.
[Missing arglist.  Please make a bug report.]
 
Not documented.
 

Emacs 22 is more correct here, since the arglist is missing.  Even if
there is no doc string, we should show the signature.
 

In GNU Emacs 23.0.92.1 (i386-mingw-nt5.1.2600)
 of 2009-03-30 on SOFT-MJASON
Windowing system distributor `Microsoft Corp.', version 5.1.2600
configured using `configure --with-gcc (3.4)'
 

bug#3108: 23.0.92; 2C-command: missing arglist

[Andreas Schwab]

"Drew Adams" <drew.adams@oracle.com> writes:

> Emacs 22 is more correct here, since the arglist is missing.  Even if
> there is no doc string, we should show the signature.

2C-command is a keymap, it does not have an arglist.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

bug#3108: 23.0.92; 2C-command: missing arglist

[Drew Adams]

> > Emacs 22 is more correct here, since the arglist is 
> > missing.  Even if there is no doc string, we should show the signature.
> 
> 2C-command is a keymap, it does not have an arglist.

Oops.

Then in that case, shouldn't the help show something like this:
"Prefix command (definition is a keymap associating keystrokes with commands)."

That is shown after the library is loaded, but not before. Why is it that,
before loading (i.e. just from the autoload), help (1) is capable of saying
which (prefix) keys the command is bound to, but (2) is not capable of adding
the line saying that it is a prefix command (keymap etc.)?

Just wondering if that sentence couldn't be included even before loading. Is it
because we cannot know at that time that the keys are prefix keys?

The autoload cookie is this:
;;;###autoload (autoload '2C-command "two-column" () t 'keymap)

That would seem to provide the info that this is a prefix key.

Not very important, but if this could be improved it would help. You can close
the bug (and bug #3109), if you like.

bug#3108: 23.0.92; 2C-command: missing arglist

[Stefan Monnier]

> Then in that case, shouldn't the help show something like this:
> "Prefix command (definition is a keymap associating keystrokes with
> commands)."

It does say it (quoting the text you quoted):

  2C-command is an interactive autoloaded keymap in `two-column.el'.
  ...

Notice the "autoloaded keymap" there.  The "interactive" part is rather
pointless (I can't think of a good meaning for "interactive" when
applied to keymaps).


        Stefan

bug#3108: 23.0.92; 2C-command: missing arglist

[Drew Adams]

> > Then in that case, shouldn't the help show something like this:
> > "Prefix command (definition is a keymap associating keystrokes with
> > commands)."
> 
> It does say it (quoting the text you quoted):
> 
>   2C-command is an interactive autoloaded keymap in `two-column.el'.
>   ...
> 
> Notice the "autoloaded keymap" there.

Yes, I saw that.

It's a minor point, but I still have the question whether the ultimate text
couldn't be available before loading. It seems we have all the requisite
information beforehand. 

The ultimate text is itself just boilerplate, requiring no special knowledge of
the command (keymap) itself. There is nothing to be gained (for this) by loading
the library. Since we know that this is a prefix command (keymap) beforehand, I
would think we could use the ultimate text even before loading the library.

But I won't belabor this - as I said, you can close the bug.