Moving vertical boundaries after C-x 3

Moving vertical boundaries after C-x 3

[Richard Stallman]

To move a vertical boundary left or right, you need to find the
place where the mouse pointer appears as a sideways two-headed arrow.
I see there is no tooltip for that.  Could someone make a tooltip for it?

Moving vertical boundaries after C-x 3

[Richard Stallman]

[I sent this message a week ago but did not get a response.]

To move a vertical boundary left or right, you need to find the
place where the mouse pointer appears as a sideways two-headed arrow.
I see there is no tooltip for that.  Could someone make a tooltip for it?

Moving vertical boundaries after C-x 3

[Nick Roberts]

Richard Stallman writes:
 > [I sent this message a week ago but did not get a response.]
 > 
 > To move a vertical boundary left or right, you need to find the
 > place where the mouse pointer appears as a sideways two-headed arrow.
 > I see there is no tooltip for that.  Could someone make a tooltip for it?

Done.  This was pretty self evident from the cursor hape anyway.  How about
adding:

(global-set-key [vertical-line mouse-3] 'mouse-delete-other-windows)
(global-set-key [vertical-line mouse-2] 'mouse-delete-window)

and adjusting the help string accordingly?


-- 
Nick                                           http://www.inet.net.nz/~nickrob

RE: Moving vertical boundaries after C-x 3

[Drew Adams]

    How about adding:
    (global-set-key [vertical-line mouse-3] 'mouse-delete-other-windows)
    (global-set-key [vertical-line mouse-2] 'mouse-delete-window)

Why?

Re: Moving vertical boundaries after C-x 3

[Eli Zaretskii]

> From: "Drew Adams" <drew.adams@oracle.com>
> Date: Thu, 13 Apr 2006 08:37:02 -0700
> 
>     How about adding:
>     (global-set-key [vertical-line mouse-3] 'mouse-delete-other-windows)
>     (global-set-key [vertical-line mouse-2] 'mouse-delete-window)
> 
> Why?

For consistency with the mode line, I guess.

Re: Moving vertical boundaries after C-x 3

[Stefan Monnier]

>> How about adding:
>> (global-set-key [vertical-line mouse-3] 'mouse-delete-other-windows)
>> (global-set-key [vertical-line mouse-2] 'mouse-delete-window)
>> 
>> Why?

> For consistency with the mode line, I guess.
     ^^
     in

AFAICT, the mode-line's bindings are reversed: mouse-2 runs
mouse-delete-other-windows and mouse-3 runs mouse-delete-window.


        Stefan

Re: Moving vertical boundaries after C-x 3

[Richard Stallman]

      How about
    adding:

    (global-set-key [vertical-line mouse-3] 'mouse-delete-other-windows)
    (global-set-key [vertical-line mouse-2] 'mouse-delete-window)

    and adjusting the help string accordingly?

Why suggest those?

Re: Moving vertical boundaries after C-x 3

[Nick Roberts]

Richard Stallman writes:
 >       How about
 >     adding:
 > 
 >     (global-set-key [vertical-line mouse-3] 'mouse-delete-other-windows)
 >     (global-set-key [vertical-line mouse-2] 'mouse-delete-window)
 > 
 >     and adjusting the help string accordingly?
 > 
 > Why suggest those?

For consistency with the mode line.  Its no big deal, I doubt that many will
use it but I see no inconvenience in having it.

To avoid any further ambiguity I attach the proposed patch below.  I note that
[vertical-line mouse-1] is already bound to mouse-select-window.  I have
changed the help string in xdisp.c to match that for the mode-line.  I set it
using make_string but I can't see that done anywhere else, so I wonder if
there's a better way.

I also note that even [vertical-line C-mouse-2] has a binding
(mouse-split-window-vertically).  This function works properly when the window
extends from top to bottom but once it has already been split vertically, it
splits the lower window in the wrong place.

-- 
Nick                                           http://www.inet.net.nz/~nickrob


*** mouse.el	31 Mar 2006 11:03:14 +1200	1.297
--- mouse.el	14 Apr 2006 08:23:25 +1200	
***************
*** 2404,2409 ****
--- 2404,2411 ----
  (global-set-key [header-line mouse-1] 'mouse-select-window)
  (global-set-key [mode-line mouse-2] 'mouse-delete-other-windows)
  (global-set-key [mode-line mouse-3] 'mouse-delete-window)
+ (global-set-key [vertical-line mouse-2] 'mouse-delete-other-windows)
+ (global-set-key [vertical-line mouse-3] 'mouse-delete-window)
  (global-set-key [mode-line C-mouse-2] 'mouse-split-window-horizontally)
  (global-set-key [vertical-scroll-bar C-mouse-2] 'mouse-split-window-vertically)
  (global-set-key [vertical-line C-mouse-2] 'mouse-split-window-vertically)


*** xdisp.c	13 Apr 2006 13:08:36 +1200	1.1086
--- xdisp.c	14 Apr 2006 08:41:08 +1200	
***************
*** 22360,22366 ****
    if (part == ON_VERTICAL_BORDER)
      {
        cursor = FRAME_X_OUTPUT (f)->horizontal_drag_cursor;
!       help_echo_string = make_string ("drag-mouse-1: resize", 20);
      }
    else if (part == ON_LEFT_FRINGE || part == ON_RIGHT_FRINGE
  	   || part == ON_SCROLL_BAR)
--- 22360,22366 ----
    if (part == ON_VERTICAL_BORDER)
      {
        cursor = FRAME_X_OUTPUT (f)->horizontal_drag_cursor;
!       help_echo_string = make_string ("mouse-1: select (drag to resize), mouse-2: delete others, mouse-3: delete this", 78);
      }
    else if (part == ON_LEFT_FRINGE || part == ON_RIGHT_FRINGE
  	   || part == ON_SCROLL_BAR)

Re: Moving vertical boundaries after C-x 3

[Stefan Monnier]

> I set it using make_string but I can't see that done anywhere else, so
> I wonder if there's a better way.

Elsewhere, they use `build_string'.


        Stefan

Click event bug? (was Re: Moving vertical...)

[Nick Roberts]

 > I also note that even [vertical-line C-mouse-2] has a binding
 > (mouse-split-window-vertically).  This function works properly when the
 > window extends from top to bottom but once it has already been split
 > vertically, it splits the lower window in the wrong place.

I think this is because the click event on the vertical line returns the column
number measured from the top of the frame instead of the top of the window.


-- 
Nick                                           http://www.inet.net.nz/~nickrob

Click event bug? (was Re: Moving vertical...)

[Nick Roberts]

Nick Roberts writes:
 >  > I also note that even [vertical-line C-mouse-2] has a binding
 >  > (mouse-split-window-vertically).  This function works properly when the
 >  > window extends from top to bottom but once it has already been split
 >  > vertically, it splits the lower window in the wrong place.
 > 
 > I think this is because the click event on the vertical line returns the
 > column number measured from the top of the frame instead of the top of the
 > window.

I mean row number of course.

-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Moving vertical boundaries after C-x 3

[Richard Stallman]

When you click on a mode line to delete "the window", it is clear
which window that is.  The mode line belongs to a window.

When you click on the vertical line, which window gets deleted?  Is it
the one whose scroll bar you're under?  That would be the logical
answer.  But I think it will not be very clear to users, and therefore
I think that it is better not to make this change.

Re: Moving vertical boundaries after C-x 3

[Nick Roberts]

Richard Stallman writes:
 > When you click on a mode line to delete "the window", it is clear
 > which window that is.  The mode line belongs to a window.
 > 
 > When you click on the vertical line, which window gets deleted?

The window on the left.

 > Is it the one whose scroll bar you're under?

Have you tried this?  Using the default toolkit there is no vertical line when
the scroll bar is present.  This is true both with the fringe (where there are
narrow margins) and without it.

 > That would be the logical answer.  But I think it will not be very clear to
 > users, and therefore I think that it is better not to make this change.

The same logic can be applied to mouse-split-window-vertically (currently on
[vertical-line C-mouse-2]).  Which window gets split?  You only have to try
it once to find out, it's not rocket science.

Anyway it's a bit tricky selecting the vertical line, but this is true for
resizing.  I think it would be sensible to add these bindings and make the
vertical line a few pixels wider.

-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Click event bug? (was Re: Moving vertical...)

[Richard Stallman]

    I think this is because the click event on the vertical line returns the column
    number measured from the top of the frame instead of the top of the window.

Could you fix it?

Re: Moving vertical boundaries after C-x 3

[Richard Stallman]

    Have you tried this?  Using the default toolkit there is no
    vertical line when the scroll bar is present.  This is true both
    with the fringe (where there are narrow margins) and without it.

Isn't there still a short vertical line, which one can use to drag the
scroll bar left or right?  That is what this discussion started about.
That is where this matters for most users.

I don't see much benefit in making it possible to delete the
window with a click on that small vertical line.

Re: Moving vertical boundaries after C-x 3

[Nick Roberts]

Richard Stallman writes:
 >     Have you tried this?  Using the default toolkit there is no
 >     vertical line when the scroll bar is present.  This is true both
 >     with the fringe (where there are narrow margins) and without it.
 >
 > Isn't there still a short vertical line, which one can use to drag the
 > scroll bar left or right?  That is what this discussion started about.
 > That is where this matters for most users.

ISTR scroll bars weren't part of the original thread.  There is a very short
line on the mode line where one can drag the scroll bar but nothing along the
side of the scroll bar itself.

 > I don't see much benefit in making it possible to delete the
 > window with a click on that small vertical line.

When the scroll bar is not present, the window edge can be dragged at any point
along its edge.  This is for the Lucid toolkit.  Maybe other toolkits behave
differently.  Its not small, just narrow (7 pixel grabbable width?), thats why
I suggested making it wider.  Incidentally, clicking on the vertical border
is easier on a test terminal (using t-mouse-mode) because it is wider (one
character wide).

On a text terminal, the vertical border is displayed with `|' characters.  This
is historical but now that it has reverse contrast (its face inherits from
mode-line-inactive face), I think using SPC looks more visually appealing (at
least on a Linux Console).  Furthermore, the vertical border could presumably
use mode-line/mode-line-inactive face just as for the mode line.  This would
make a difference on 256 colour xterms.  Below is a patch just for the former
task.


-- 
Nick                                           http://www.inet.net.nz/~nickrob


*** dispnew.c	13 Apr 2006 09:12:44 +1200	1.362
--- dispnew.c	16 Apr 2006 10:35:38 +1200	
***************
*** 2727,2733 ****
  	  right_border_glyph
  	    = ((dp && INTEGERP (DISP_BORDER_GLYPH (dp)))
  	       ? spec_glyph_lookup_face (w, XINT (DISP_BORDER_GLYPH (dp)))
! 	       : '|');
  
  	  if (FAST_GLYPH_FACE (right_border_glyph) <= 0)
  	    right_border_glyph
--- 2727,2733 ----
  	  right_border_glyph
  	    = ((dp && INTEGERP (DISP_BORDER_GLYPH (dp)))
  	       ? spec_glyph_lookup_face (w, XINT (DISP_BORDER_GLYPH (dp)))
! 	       : ' ');
  
  	  if (FAST_GLYPH_FACE (right_border_glyph) <= 0)
  	    right_border_glyph

Re: Click event bug? (was Re: Moving vertical...)

[Nick Roberts]

 >     I think this is because the click event on the vertical line returns the
 >     column number measured from the top of the frame instead of the top of
       ^^^^^^
 I meant row number.

 >     the window.
 > 
 > Could you fix it?

I'm hoping Kim will look at it.

-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Moving vertical boundaries after C-x 3

[M Jared Finder]

Richard Stallman wrote:
>     Have you tried this?  Using the default toolkit there is no
>     vertical line when the scroll bar is present.  This is true both
>     with the fringe (where there are narrow margins) and without it.
> 
> Isn't there still a short vertical line, which one can use to drag the
> scroll bar left or right?  That is what this discussion started about.
> That is where this matters for most users.

Not in GTK and Win32 builds.  If you disable the scrollbar, then there 
is a black line you can grab to resize each window, but with a scrollbar 
on either the left or right side of the window, it's not there.

   -- MJF

Re: Click event bug? (was Re: Moving vertical...)

[Kim F. Storm]

Nick Roberts <nickrob@snap.net.nz> writes:

>  >     I think this is because the click event on the vertical line returns the
>  >     column number measured from the top of the frame instead of the top of
>        ^^^^^^
>  I meant row number.
>
>  >     the window.
>  > 
>  > Could you fix it?
>
> I'm hoping Kim will look at it.

Does this patch do what you need?


Index: window.c
===================================================================
RCS file: /cvsroot/emacs/emacs/src/window.c,v
retrieving revision 1.540
diff -c -r1.540 window.c
*** window.c	20 Mar 2006 22:22:25 -0000	1.540
--- window.c	16 Apr 2006 23:22:02 -0000
***************
*** 661,672 ****
  	  || WINDOW_RIGHTMOST_P (w))
  	{
  	  if (!WINDOW_LEFTMOST_P (w) && abs (*x - x0) < grabbable_width)
! 	    return ON_VERTICAL_BORDER;
  	}
        else
  	{
  	  if (abs (*x - x1) < grabbable_width)
! 	    return ON_VERTICAL_BORDER;
  	}
  
        if (*x < x0 || *x >= x1)
--- 661,680 ----
  	  || WINDOW_RIGHTMOST_P (w))
  	{
  	  if (!WINDOW_LEFTMOST_P (w) && abs (*x - x0) < grabbable_width)
! 	    {
! 	      *x = max (0, *x - x0);
! 	      *y -= top_y;
! 	      return ON_VERTICAL_BORDER;
! 	    }
  	}
        else
  	{
  	  if (abs (*x - x1) < grabbable_width)
! 	    {
! 	      *x = min (x1, *x) - x0;
! 	      *y -= top_y;
! 	      return ON_VERTICAL_BORDER;
! 	    }
  	}
  
        if (*x < x0 || *x >= x1)
***************
*** 708,714 ****
  	  && !WINDOW_HAS_VERTICAL_SCROLL_BAR (w)
  	  && !WINDOW_RIGHTMOST_P (w)
  	  && (abs (*x - right_x) < grabbable_width))
! 	return ON_VERTICAL_BORDER;
      }
    else
      {
--- 716,726 ----
  	  && !WINDOW_HAS_VERTICAL_SCROLL_BAR (w)
  	  && !WINDOW_RIGHTMOST_P (w)
  	  && (abs (*x - right_x) < grabbable_width))
! 	{
! 	  *x = min (right_x, *x) - left_x;
! 	  *y -= top_y;
! 	  return ON_VERTICAL_BORDER;
! 	}
      }
    else
      {
***************
*** 720,725 ****
--- 732,739 ----
  	{
  	  /* On the border on the right side of the window?  Assume that
  	     this area begins at RIGHT_X minus a canonical char width.  */
+ 	  *x = min (right_x, *x) - left_x;
+ 	  *y -= top_y;
  	  return ON_VERTICAL_BORDER;
  	}
      }

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: Click event bug? (was Re: Moving vertical...)

[Nick Roberts]

 > >  > Could you fix it?
 > >
 > > I'm hoping Kim will look at it.
 > 
 > Does this patch do what you need?

Yes, so I've committed it.

Thanks,

Nick                                           http://www.inet.net.nz/~nickrob

Re: Moving vertical boundaries after C-x 3

[Richard Stallman]

     > Isn't there still a short vertical line, which one can use to drag the
     > scroll bar left or right?  That is what this discussion started about.
     > That is where this matters for most users.

    ISTR scroll bars weren't part of the original thread.  There is a very short
    line on the mode line where one can drag the scroll bar

Yes, that is what I'm talking about.  That is where the arrow appears,
that's where I asked for a tooltip, and that's where these new mouse clicks
would work, if we define them.

     > I don't see much benefit in making it possible to delete the
     > window with a click on that small vertical line.

    When the scroll bar is not present, the window edge can be dragged at any point
    along its edge.

Yes, but since there usually is a scroll bar, I think this won't be
useful for many users.  I don't think it is worth defining bindings
of marginal utility, that will be difficult to use, just for the
short line on the mode line or for the sake of the few who turn off
scroll bars.

Re: Click event bug? (was Re: Moving vertical...)

[Richard Stallman]

     > Could you fix it?

    I'm hoping Kim will look at it.

Could you please keep track of whether someone else looks at it,
and tell me in 2 weeks what the situation is?

Re: Moving vertical boundaries after C-x 3

[Nick Roberts]

 >      > I don't see much benefit in making it possible to delete the
 >      > window with a click on that small vertical line.
 > 
 >     When the scroll bar is not present, the window edge can be dragged at
 >     any point along its edge.
 > 
 > Yes, but since there usually is a scroll bar, I think this won't be
 > useful for many users.  I don't think it is worth defining bindings
 > of marginal utility, that will be difficult to use, just for the
 > short line on the mode line or for the sake of the few who turn off
 > scroll bars.

I don't use scroll bars and I wonder how you know there are only a few of us.
In any case surely the solution must be to make the vertical line visible
alongside the scroll bar.  Thats what XEmacs does.  It would also make other
commands like mouse-split-window-vertically available to the scroll bar user.

I guess you don't like the text terminal patch that I suggested (using the SPC
character instead of `|') but I would rather you said this, so I have a record
of what you think.


-- 
Nick                                           http://www.inet.net.nz/~nickrob

Re: Click event bug? (was Re: Moving vertical...)

[Richard Stallman]

Thanks for fixing this.  Would you please add some comments
explaining your new code?

Re: Moving vertical boundaries after C-x 3

[Richard Stallman]

    I don't use scroll bars and I wonder how you know there are only a few of us.

It is obvious.

    In any case surely the solution must be to make the vertical line visible
    alongside the scroll bar.

What do you mean, "solution"?  There is no problem.

Please drop this subject.  It is wasting too much time
and I can't afford it!