minibuffer and current-local-map

minibuffer and current-local-map

[Stephen Berman]

[-- Attachment #1: Type: text/plain, Size: 2649 bytes --]

The value returned by current-local-map for the minibuffer seems to be
wrong (both in the trunk and in Emacs 22.1), containing partial
reduplication:

(keymap
 (C-tab . file-cache-minibuffer-complete)
 (10 . minibuffer-complete-and-exit)
 (13 . minibuffer-complete-and-exit)
 keymap
 (menu-bar keymap
	   (minibuf "Minibuf" keymap
		    (tab menu-item "Complete" minibuffer-complete
			 ([9]
			  . "  (TAB)")
			 :help "Complete as far as possible")
		    (space menu-item "Complete Word" minibuffer-complete-word
			   ([32]
			    . "  (SPC)")
			   :help "Complete at most one word")
		    (63 menu-item "List Completions" minibuffer-completion-help
			([63]
			 . "  (?)")
			:help "Display all possible completions")
		    "Minibuf" keymap
		    (return menu-item "Enter" exit-minibuffer
			    (nil)
			    :help "Terminate input and exit minibuffer")
		    (quit menu-item "Quit" keyboard-escape-quit
			  ([134217755 27]
			   . "  (M-ESC ESC)")
			  :help "Abort input and exit minibuffer")
		    "Minibuf")
	   keymap
	   (minibuf "Minibuf" keymap
		    (return menu-item "Enter" exit-minibuffer
			    (nil)
			    :help "Terminate input and exit minibuffer")
		    (quit menu-item "Quit" keyboard-escape-quit
			  ([134217755 27]
			   . "  (M-ESC ESC)")
			  :help "Abort input and exit minibuffer")
		    "Minibuf"))
 (27 keymap
     (118 . switch-to-completions)
     keymap
     (114 . previous-matching-history-element)
     (115 . next-matching-history-element)
     (112 . previous-history-element)
     (110 . next-history-element))
 (prior . switch-to-completions)
 (C-tab . file-cache-minibuffer-complete)
 (63 . minibuffer-completion-help)
 (32 . minibuffer-complete-word)
 (9 . minibuffer-complete)
 keymap
 (menu-bar keymap
	   (minibuf "Minibuf" keymap
		    (return menu-item "Enter" exit-minibuffer
			    (nil)
			    :help "Terminate input and exit minibuffer")
		    (quit menu-item "Quit" keyboard-escape-quit
			  ([134217755 27]
			   . "  (M-ESC ESC)")
			  :help "Abort input and exit minibuffer")
		    "Minibuf"))
 (C-tab . file-cache-minibuffer-complete)
 (9 . self-insert-command)
 (up . previous-history-element)
 (prior . previous-history-element)
 (down . next-history-element)
 (next . next-history-element)
 (27 keymap
     (114 . previous-matching-history-element)
     (115 . next-matching-history-element)
     (112 . previous-history-element)
     (110 . next-history-element))
 (10 . exit-minibuffer)
 (13 . exit-minibuffer)
 (7 . abort-recursive-edit))

As a result of this, typing <C-down-mouse-3> in the minibuffer pops up a
confusing context menu, see the attached screen shots.

Steve Berman


[-- Attachment #2: minibuf-menu1 --]
[-- Type: image/png, Size: 44545 bytes --]

[-- Attachment #3: minibuf-menu2 --]
[-- Type: image/png, Size: 33873 bytes --]

[-- Attachment #4: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: minibuffer and current-local-map

[Richard Stallman]

    The value returned by current-local-map for the minibuffer

Inside which command?  What function was used to invoke the
minibuffer?  What map did it specify?

							       seems to be
    wrong (both in the trunk and in Emacs 22.1), containing partial
    reduplication:

Is the duplication present in the map that was passed as an argument
for use in the minibuffer?

Re: minibuffer and current-local-map

[Stephen Berman]

On Wed, 07 Nov 2007 23:42:33 -0500 Richard Stallman <rms@gnu.org> wrote:

>     The value returned by current-local-map for the minibuffer
>
> Inside which command?  What function was used to invoke the
> minibuffer?  What map did it specify?

I noticed this after typing `M-x', thus activating the minibuffer, and
there typing C-down-mouse-3, popping up the confusing context menu seen
in the screen shots attached to my OP.  C-down-mouse-3 invokes
mouse-popup-menubar-stuff, which calls mouse-major-mode-menu, which
calls current-local-map.

> 							       seems to be
>     wrong (both in the trunk and in Emacs 22.1), containing partial
>     reduplication:
>
> Is the duplication present in the map that was passed as an argument
> for use in the minibuffer?

I'm not sure what you mean by "passed as an argument" here.  I saw the
value of current-local-map that I posted when I stepped through
mouse-major-mode-menu with edebug.  I also got this value by doing the
following (is step 4 an instance of passing the map as an argument for
use in the minibuffer?):

1. emacs -Q
2. M-: (setq enable-recursive-minibuffers t) RET
3. M-x
4. M-: (current-local-map) RET

Steve Berman

Re: minibuffer and current-local-map

[Richard Stallman]

    > Is the duplication present in the map that was passed as an argument
    > for use in the minibuffer?

    I'm not sure what you mean by "passed as an argument" here.

Whatever invokes the minibuffer specifies the map to use.  In this
case, the caller is in C code.  Now that we know it was M-x, we can
look in the source and see which map it is.

Re: minibuffer and current-local-map

[Richard Stallman]

I see the duplicate in the normal value of
minibuffer-local-completion-map.  The parent of that map is
minibuffer-local-map, and both of them have menu bar submaps.

I think there is nothing wrong with the keymaps.
When keymap A has B as its parent, keymap A's submap for a certain
prefix should have, as its parent, keymap B's submap for that prefix.
What you see here is that happening for the prefix `menu-bar' and
then again, within that, for the prefix `minibuf'.
Three levels of keymap `minibuffer-local-completion-map',
each inheriting from the corresponding level of the parent,
`minibuffer-local-map'.

The solution, I think, is to fix the menu code to avoid showing duplicates
in such a case.


Meanwhile, the bindings for C-tab really are duplicates.
I will get rid of the extra ones.

Re: minibuffer and current-local-map

[Jason Rumney]

Richard Stallman wrote:
> I see the duplicate in the normal value of
> minibuffer-local-completion-map.  The parent of that map is
> minibuffer-local-map, and both of them have menu bar submaps.
>   

The duplication seems to be related to this comment in menu-bar.el

 (dolist (map (list minibuffer-local-map
            ;; This shouldn't be necessary, but there's a funny
            ;; bug in keymap.c that I don't understand yet.  -stef
            minibuffer-local-completion-map))
    (define-key map [menu-bar minibuf]
      (cons "Minibuf" (make-sparse-keymap "Minibuf"))))

Re: minibuffer and current-local-map

[Richard Stallman]

    > I see the duplicate in the normal value of
    > minibuffer-local-completion-map.  The parent of that map is
    > minibuffer-local-map, and both of them have menu bar submaps.
    >   

    The duplication seems to be related to this comment in menu-bar.el

     (dolist (map (list minibuffer-local-map
		;; This shouldn't be necessary, but there's a funny
		;; bug in keymap.c that I don't understand yet.  -stef
		minibuffer-local-completion-map))
	(define-key map [menu-bar minibuf]
	  (cons "Minibuf" (make-sparse-keymap "Minibuf"))))

There might be a bug in creating the submaps some other way,
which this works around.  But I think that is irrelevant here,
because the submap data looks correct to me.  If fixing the other 
bug makes it possible to create the same submaps in another way
that ought to be equivalent, that won't get rid of the bug in 
looking at them.

Re: minibuffer and current-local-map

[Glenn Morris]

I no longer have the original message to reply to. I'm replying to:

http://lists.gnu.org/archive/html/emacs-devel/2007-11/msg00602.html

  From:   Richard Stallman
  Subject:        Re: minibuffer and current-local-map
  Date:           Fri, 09 Nov 2007 17:00:03 -0500

The issue is:

emacs -Q
M-x
C-mouse-3 in the minibuffer brings up a menu with two "Minibuff" entries.

rms said:

  The solution, I think, is to fix the menu code to avoid showing
  duplicates in such a case.


Here is a simple patch that just takes the first of any duplicate
(same name) menus.


*** mouse.el.~1.315.2.4.~	2007-12-06 20:27:17.000000000 -0800
--- mouse.el	2008-01-06 11:11:51.000000000 -0800
***************
*** 173,183 ****
  	 ;; default to the edit menu.
  	 (newmap (if ancestor
  		     (make-sparse-keymap (concat mode-name " Mode"))
! 		   menu-bar-edit-menu)))
      (if ancestor
  	;; Make our menu inherit from the desired keymap which we want
  	;; to display as the menu now.
! 	(set-keymap-parent newmap ancestor))
      (popup-menu newmap event prefix)))
  
  
--- 173,190 ----
  	 ;; default to the edit menu.
  	 (newmap (if ancestor
  		     (make-sparse-keymap (concat mode-name " Mode"))
! 		   menu-bar-edit-menu))
!          uniq)
      (if ancestor
  	;; Make our menu inherit from the desired keymap which we want
  	;; to display as the menu now.
! 	(set-keymap-parent newmap
!                            (progn
!                              (dolist (e ancestor)
!                                (unless (and (listp e)
!                                             (assoc (car e) uniq))
!                                  (setq uniq (append uniq (list e)))))
!                              uniq)))
      (popup-menu newmap event prefix)))

Re: minibuffer and current-local-map

[Richard Stallman]

If it works, please install it, and thanks.

Re: minibuffer and current-local-map

[Stephen Berman]

On Mon, 07 Jan 2008 06:31:17 -0500 Richard Stallman <rms@gnu.org> wrote:

> If it works, please install it, and thanks.

It solves the problem I reported.  Thanks, Glenn.

Steve Berman