GNU bug report logs - #17633
24.3.91; mouse-drag-line: Symbol's function definition is void: mouse--remap-link-click-p

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 17633 in the body.
You can then email your comments to 17633 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Christopher Schmidt <ch <at> ristopher.com>
To: bug-gnu-emacs <at> gnu.org
Subject: 24.3.91; mouse-drag-line: Symbol's function definition is void:
 mouse--remap-link-click-p
Date: Thu, 29 May 2014 13:45:09 -0400 (EDT)

    emacs -q
    # Drag the mode line with the mouse.

    mouse-drag-line: Symbol's function definition is void: mouse--remap-link-click-p

This is a regression.

    In GNU Emacs 24.3.91.1 (x86_64-unknown-linux-gnu, GTK+ Version 2.24.10)
    Repository revision: 117171 eliz <at> gnu.org-20140529164749-nz9x3hi1ipkv9cfk
    Windowing system distributor `The X.Org Foundation', version 11.0.11204000

Message #8 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: bug-gnu-emacs <at> gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is void:
 mouse--remap-link-click-p
Date: Thu, 29 May 2014 14:48:12 -0400

>     emacs -q
>     # Drag the mode line with the mouse.
>     mouse-drag-line: Symbol's function definition is void: mouse--remap-link-click-p

Indeed, we have a call to that function, while I removed it a year ago.
Not sure how I managed not to notice that call.  I do wonder tho: do you
have a recipe to reproduce the problem?
I tried to hit the bug, but wasn't able to (even after reading the code
and trying to infer the circumstances that would lead to it).

This is a piece of code that we should probably revisit, so having such
a recipe would help me understand how the code works.


        Stefan

Message #11 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Christopher Schmidt <ch <at> ristopher.com>
To: bug-gnu-emacs <at> gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is void:
 mouse--remap-link-click-p
Date: Thu, 29 May 2014 15:06:56 -0400 (EDT)

Stefan Monnier <monnier <at> iro.umontreal.ca> writes:
> I do wonder tho: do you have a recipe to reproduce the problem?  I
> tried to hit the bug, but wasn't able to (even after reading the code
> and trying to infer the circumstances that would lead to it).

>>     emacs -q
>>     # Drag the mode line with the mouse.

That is my recipe.

FWIW there is another major bug:

    emacs -q
    # Click the mode line's buffer identification string ("*GNU Emacs*")

    button-activate: Symbol's function definition is void: nil

This bug only occurs when point is on a character with a keymap.  emacs
seems to use the keymap of the character at point rather than the one of
the character that was actually clicked at.

Message #14 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: bug-gnu-emacs <at> gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is void:
 mouse--remap-link-click-p
Date: Thu, 29 May 2014 16:32:34 -0400

>>> emacs -q
>>> # Drag the mode line with the mouse.
> That is my recipe.

Hmm.... when I try that, I get a frame with a single window, so I can
try to drag the mode line, but it won't budge.  And I don't get
any error.
So, I tried to add a C-x 2.  I can then move the mode line, but then
again, I don't get any error.

Oh, wait, right, I see it now: I had used "emacs -Q" rather than "emacs -q".
Thanks.

> FWIW there is another major bug:

>     emacs -q
>     # Click the mode line's buffer identification string ("*GNU Emacs*")
>     button-activate: Symbol's function definition is void: nil
> This bug only occurs when point is on a character with a keymap.  emacs
> seems to use the keymap of the character at point rather than the one of
> the character that was actually clicked at.

Can you make it a separate bug-report?


        Stefan

Message #17 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Christopher Schmidt <ch <at> ristopher.com>
To: bug-gnu-emacs <at> gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is void:
 mouse--remap-link-click-p
Date: Thu, 29 May 2014 16:47:05 -0400 (EDT)

Stefan Monnier <monnier <at> iro.umontreal.ca> writes:
>> FWIW there is another major bug:
>
>>     emacs -q
>>     # Click the mode line's buffer identification string ("*GNU Emacs*")
>>     button-activate: Symbol's function definition is void: nil
>> This bug only occurs when point is on a character with a keymap.
>> emacs seems to use the keymap of the character at point rather than
>> the one of the character that was actually clicked at.
>
> Can you make it a separate bug-report?

AFAICS both issues are manifestations of the same bug because both of
them only come up if point is on a character with a dedicated keymap.

Message #20 received at 17633 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: 17633 <at> debbugs.gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is void:
 mouse--remap-link-click-p
Date: Sat, 31 May 2014 08:34:28 -0400

> AFAICS both issues are manifestations of the same bug because both of
> them only come up if point is on a character with a dedicated keymap.

Indeed, I installed the patch below which seems to fix them.
Can you confirm it fixes the problems you've seen?


        Stefan


=== modified file 'lisp/ChangeLog'
--- lisp/ChangeLog	2014-05-31 01:01:19 +0000
+++ lisp/ChangeLog	2014-05-31 12:32:03 +0000
@@ -1,5 +1,8 @@
 2014-05-31  Stefan Monnier  <monnier <at> iro.umontreal.ca>
 
+	* mouse.el (mouse-posn-property): Ignore posn-point for mode-line
+	clicks (bug#17633).
+
 	* leim/quail/latin-pre.el ("latin-2-prefix"): Use ",," rather than ", "
 	for the single comma, since ", " is *very* common in normal French text
 	(bug#17643).

=== modified file 'lisp/mouse.el'
--- lisp/mouse.el	2014-05-18 22:53:27 +0000
+++ lisp/mouse.el	2014-05-31 12:30:47 +0000
@@ -658,7 +658,10 @@
 	    (str (posn-string pos)))
 	(or (and str
 		 (get-text-property (cdr str) property (car str)))
-	    (and pt
+            ;; FIXME: mouse clicks on the mode-line come with a position in
+            ;; (nth 5).  Maybe we should change the C code instead so that
+            ;; mouse-clicks don't include a position there!
+	    (and pt (not (memq (posn-area pos) '(mode-line header-line)))
 		 (get-char-property pt property w))))
     (get-char-property pos property)))
 

Message #25 received at 17633-done <at> debbugs.gnu.org (full text, mbox):

From: Christopher Schmidt <ch <at> ristopher.com>
To: 17633-done <at> debbugs.gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is void:
 mouse--remap-link-click-p
Date: Sat, 31 May 2014 08:55:49 -0400 (EDT)

Stefan Monnier <monnier <at> iro.umontreal.ca> writes:
>> AFAICS both issues are manifestations of the same bug because both of
>> them only come up if point is on a character with a dedicated keymap.
>
> Indeed, I installed the patch below which seems to fix them.  Can you
> confirm it fixes the problems you've seen?

The problems are gone now.  Thanks a lot!

Message #28 received at 17633 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: 17633 <at> debbugs.gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is
 void:	mouse--remap-link-click-p
Date: Sat, 31 May 2014 16:00:55 +0300

> From: Stefan Monnier <monnier <at> iro.umontreal.ca>
> Date: Sat, 31 May 2014 08:34:28 -0400
> 
>  2014-05-31  Stefan Monnier  <monnier <at> iro.umontreal.ca>
>  
> +	* mouse.el (mouse-posn-property): Ignore posn-point for mode-line
> +	clicks (bug#17633).
> +
>  	* leim/quail/latin-pre.el ("latin-2-prefix"): Use ",," rather than ", "
>  	for the single comma, since ", " is *very* common in normal French text
>  	(bug#17643).
> 
> === modified file 'lisp/mouse.el'
> --- lisp/mouse.el	2014-05-18 22:53:27 +0000
> +++ lisp/mouse.el	2014-05-31 12:30:47 +0000
> @@ -658,7 +658,10 @@
>  	    (str (posn-string pos)))
>  	(or (and str
>  		 (get-text-property (cdr str) property (car str)))
> -	    (and pt
> +            ;; FIXME: mouse clicks on the mode-line come with a position in
> +            ;; (nth 5).  Maybe we should change the C code instead so that
> +            ;; mouse-clicks don't include a position there!
> +	    (and pt (not (memq (posn-area pos) '(mode-line header-line)))
>  		 (get-char-property pt property w))))
>      (get-char-property pos property)))

Can you explain the problem in more detail?  I'm afraid I don't
understand the problem with mouse clicks on mode line.

Message #31 received at 17633 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 17633 <at> debbugs.gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is
 void:	mouse--remap-link-click-p
Date: Sat, 31 May 2014 10:55:54 -0400

> Can you explain the problem in more detail?  I'm afraid I don't
> understand the problem with mouse clicks on mode line.

I think the real fix (which I intend to install into trunk) is the
one below: including `point' in a mouse-click is simply incorrect since
the current position of point has nothing to do with the mouse event.


        Stefan


=== modified file 'src/keyboard.c'
--- src/keyboard.c	2014-05-30 04:12:08 +0000
+++ src/keyboard.c	2014-05-31 13:42:02 +0000
@@ -5266,9 +5266,7 @@
 				     &object, &dx, &dy, &width, &height);
 	  if (STRINGP (string))
 	    string_info = Fcons (string, make_number (charpos));
-	  textpos = (w == XWINDOW (selected_window)
-		     && current_buffer == XBUFFER (w->contents))
-	    ? PT : marker_position (w->pointm);
+	  textpos = Qnil;
 
 	  xret = wx;
 	  yret = wy;

Message #34 received at 17633 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: 17633 <at> debbugs.gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is
 void:	mouse--remap-link-click-p
Date: Sat, 31 May 2014 19:44:11 +0300

> From: Stefan Monnier <monnier <at> iro.umontreal.ca>
> Cc: 17633 <at> debbugs.gnu.org
> Date: Sat, 31 May 2014 10:55:54 -0400
> 
> > Can you explain the problem in more detail?  I'm afraid I don't
> > understand the problem with mouse clicks on mode line.
> 
> I think the real fix (which I intend to install into trunk) is the
> one below: including `point' in a mouse-click is simply incorrect since
> the current position of point has nothing to do with the mouse event.

But why does it hurt to include the window's point in the click?
Other kinds of click on non-text parts do include some kind of
position, so why is this one a problem?

Btw, your patch won't compile, because 'textpos' is not a Lisp object.

Message #37 received at 17633 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 17633 <at> debbugs.gnu.org
Subject: Re: bug#17633: 24.3.91; mouse-drag-line: Symbol's function definition
 is void:	mouse--remap-link-click-p
Date: Sat, 31 May 2014 16:11:02 -0400

> But why does it hurt to include the window's point in the click?

Because mouse-posn-property then checks the text-property at point.

> Other kinds of click on non-text parts do include some kind of
> position, so why is this one a problem?

Usually those positions are at least somewhat related to the position of
the mouse.  But indeed mouse-posn-property should probably ignore the
posn-point info for the other non-text parts.

> Btw, your patch won't compile, because 'textpos' is not a Lisp object.

Yes, it was just to show the intention.


        Stefan

Message #40 received at 17633 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: 17633 <at> debbugs.gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is
 void:	mouse--remap-link-click-p
Date: Sun, 01 Jun 2014 05:41:34 +0300

> From: Stefan Monnier <monnier <at> iro.umontreal.ca>
> Cc: 17633 <at> debbugs.gnu.org
> Date: Sat, 31 May 2014 16:11:02 -0400
> 
> > But why does it hurt to include the window's point in the click?
> 
> Because mouse-posn-property then checks the text-property at point.

It shouldn't do that.  This sounds like a problem with
mouse-posn-property rather than with make_lispy_position.

Message #43 received at 17633 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 17633 <at> debbugs.gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is
 void:	mouse--remap-link-click-p
Date: Sun, 01 Jun 2014 10:18:57 -0400

>> > But why does it hurt to include the window's point in the click?
>> Because mouse-posn-property then checks the text-property at point.
> It shouldn't do that.  This sounds like a problem with
> mouse-posn-property rather than with make_lispy_position.

In the case of fringe clicks, I agree that these events's
buffer-position info is worthwhile and that mouse-posn-property should
ignore it, but in the case of mode-line clicks, the click is completely
unrelated to the particular position of point at that moment, os
including that info is just asking for trouble,


        Stefan

Message #46 received at 17633 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: 17633 <at> debbugs.gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is
 void:	mouse--remap-link-click-p
Date: Sun, 01 Jun 2014 17:57:47 +0300

> From: Stefan Monnier <monnier <at> iro.umontreal.ca>
> Cc: 17633 <at> debbugs.gnu.org
> Date: Sun, 01 Jun 2014 10:18:57 -0400
> 
> >> > But why does it hurt to include the window's point in the click?
> >> Because mouse-posn-property then checks the text-property at point.
> > It shouldn't do that.  This sounds like a problem with
> > mouse-posn-property rather than with make_lispy_position.
> 
> In the case of fringe clicks, I agree that these events's
> buffer-position info is worthwhile and that mouse-posn-property should
> ignore it, but in the case of mode-line clicks, the click is completely
> unrelated to the particular position of point at that moment, os
> including that info is just asking for trouble,

It's not necessarily unrelated.  Imagine some feature that wants to
change the effect of such a click depending on point.  There's nothing
especially hilarious about that.

Message #49 received at 17633 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 17633 <at> debbugs.gnu.org
Subject: Re: bug#17633: 24.3.91;
 mouse-drag-line: Symbol's function definition is
 void:	mouse--remap-link-click-p
Date: Sun, 01 Jun 2014 17:13:42 -0400

> It's not necessarily unrelated.  Imagine some feature that wants to
> change the effect of such a click depending on point.  There's nothing
> especially hilarious about that.

No, but there's no need for point to be included in the click event.


        Stefan

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.