Re: PATCH: focus follows mouse in C

Re: PATCH: focus follows mouse in C

[Richard Stallman]

    +  if (x_autoselect_window_p)
    +    {
    +      int area;
    +      Lisp_Object window;
    +
    +      window = window_from_coordinates (frame, XINT (event->x), XINT (event->y), &area, 0);
    +
    +      if (WINDOW_LIVE_P (window) && !MINI_WINDOW_P (XWINDOW(window)))
    +	Fselect_window(window);
    +    }
    +

It looks like this selects the window on every motion of the mouse.
That may be what you don't like.  Maybe it should select the window
only when the mouse moves into a different window.

Also, it seems to be completely unwilling to select a minibuffer window.
When the minibuffer is active, it should be selectable like any other window.

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

Re: PATCH: focus follows mouse in C

[Pavel Janík]

   From: Richard Stallman <rms@gnu.org>
   Date: Thu, 7 Feb 2002 07:56:45 -0700 (MST)

   >     +  if (x_autoselect_window_p)
   >     +    {
   >     +      int area;
   >     +      Lisp_Object window;
   >     +
   >     +      window = window_from_coordinates (frame, XINT (event->x), XINT (event->y), &area, 0);
   >     +
   >     +      if (WINDOW_LIVE_P (window) && !MINI_WINDOW_P (XWINDOW(window)))
   >     +	Fselect_window(window);
   >     +    }
   >     +
   > 
   > It looks like this selects the window on every motion of the mouse.
   > That may be what you don't like.  Maybe it should select the window
   > only when the mouse moves into a different window.

For the user, it has this behaviour. Yes, it is ugly to call
Fselect_window for each window and wait for it to execute

  if (EQ (window, selected_window))
    return window;

It should be directly here. Fixed.

   > Also, it seems to be completely unwilling to select a minibuffer window.
   > When the minibuffer is active, it should be selectable like any other window.

Nice idea, here it is:

  if (x_autoselect_window_p)
    {
      int area;
      Lisp_Object window;

      window = window_from_coordinates (frame, XINT (event->x), XINT (event->y), &area, 0);

      /* Window will be selected only when it is not selected now.
	 Minibuffer window will be selected iff it is active.  */
      if (!EQ (window, selected_window)
	  && (!MINI_WINDOW_P (XWINDOW(window))
	      || (EQ (window, minibuf_window) && minibuf_level > 0)))
	Fselect_window (window);
    }

Maybe we can also use `area' here to allow user to customize this
behaviour. E.g. mouse over mode-line/fringe/scroll-bar does/doesn't select
window or something similar. Maybe a customizable list
x-autoselect-window-parts with full value '(text mode-line vertical-border
header-line left-fringe right-fringe)? If something will be missing here,
mouse over it can not select the window? Well, maybe I'm crazy, but the
implementation of this will surely wait for feedback from real users of
this feature.

What I do not like when x-autoselect-window is t, is connected with my
working habits. I think it can be *really* useful to people who use
focus-follows-mouse in their window managers and are used to that
behaviour.
-- 
Pavel Janík

Ask three people which VM is best, and you will get four answers ;)
                  -- Hubert Mantel about VM in 2.4.x

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

Re: PATCH: focus follows mouse in C

[Richard Stallman]

       > It looks like this selects the window on every motion of the mouse.
       > That may be what you don't like.  Maybe it should select the window
       > only when the mouse moves into a different window.

    For the user, it has this behaviour.

If I type C-x o so that some other window (which the mouse is not in)
is selected, and then move the mouse slightly, does that reselect the
window that the mouse is in?

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

Re: PATCH: focus follows mouse in C

[Pavel Janík]

   From: Richard Stallman <rms@gnu.org>
   Date: Fri, 8 Feb 2002 16:24:23 -0700 (MST)

   >     For the user, it has this behaviour.
   > 
   > If I type C-x o so that some other window (which the mouse is not in)
   > is selected, and then move the mouse slightly, does that reselect the
   > window that the mouse is in?

Of course, because "focus follows mouse".
-- 
Pavel Janík

Take care to branch the right way on equality.
                  --  The Elements of Programming Style (Kernighan & Plaugher)

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

Re: PATCH: focus follows mouse in C

[Eli Zaretskii]

> From: Pavel@Janik.cz (Pavel =?iso-8859-2?q?Jan=EDk?=)
> Date: Sat, 09 Feb 2002 08:29:26 +0100
> 
>    From: Richard Stallman <rms@gnu.org>
>    Date: Fri, 8 Feb 2002 16:24:23 -0700 (MST)
> 
>    >     For the user, it has this behaviour.
>    > 
>    > If I type C-x o so that some other window (which the mouse is not in)
>    > is selected, and then move the mouse slightly, does that reselect the
>    > window that the mouse is in?
> 
> Of course, because "focus follows mouse".

That sounds like a misfeature to me: it's very easy to make small
mouse movements just by tapping on the table or on the keyboard.  Some
users might become annoyed enough to not use the feature, just because
of this.

How about if window reselection will only be done if the mouse is in a
window different from the one it was the last time, at least as a user
option?

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

Re: PATCH: focus follows mouse in C

[Pavel Janík]

   From: "Eli Zaretskii" <eliz@is.elta.co.il>
   Date: Sat, 09 Feb 2002 12:01:46 +0200

   > > Of course, because "focus follows mouse".
   > 
   > That sounds like a misfeature to me: it's very easy to make small
   > mouse movements just by tapping on the table or on the keyboard.  Some
   > users might become annoyed enough to not use the feature, just because
   > of this.
   > 
   > How about if window reselection will only be done if the mouse is in a
   > window different from the one it was the last time, at least as a user
   > option?

Yes, now I finally understand your and RMS' point. I agree and I think that
it should default to behaviour described by you. I think that when I have
two windows and point is in the bottom one and I do C-x o, small mouse
movement (still in bottom window) should not (by default, but user should
be able to change it) select the bottom window again. This should be it
(module cus-start.el etc. changes):

--- xterm.c.~1.703.~	Sun Jan 27 17:15:53 2002
+++ xterm.c	Sat Feb  9 12:45:39 2002
@@ -253,6 +253,15 @@
 
 static int any_help_event_p;
 
+/* Non-zero means autoselect window with the mouse cursor.  */
+
+int x_autoselect_window_p;
+
+/* Non-zero means always autoselect window even if user switched to
+   the different window.  */
+
+int x_autoselect_always_p;
+
 /* Non-zero means draw block and hollow cursor as wide as the glyph
    under it.  For example, if a block cursor is over a tab, it will be
    drawn as wide as that tab on the display.  */
@@ -6605,6 +6614,26 @@
   last_mouse_motion_event = *event;
   XSETFRAME (last_mouse_motion_frame, frame);
 
+  if (x_autoselect_window_p)
+    {
+      int area;
+      Lisp_Object window;
+      static Lisp_Object last_window;
+
+      window = window_from_coordinates (frame, XINT (event->x), XINT (event->y), &area, 0);
+
+      /* Window will be selected only when it is not selected now and
+	 last mouse movement event was not in it.  Minubuffer window
+	 will be selected iff it is active.  */
+      if ( (x_autoselect_always_p || !EQ (window, last_window))
+	  && !EQ (window, selected_window)
+	  && (!MINI_WINDOW_P (XWINDOW (window))
+	      || (EQ (window, minibuf_window) && minibuf_level > 0)))
+	Fselect_window (window);
+
+      last_window=window;
+    }
+
   if (event->window != FRAME_X_WINDOW (frame))
     {
       frame->mouse_moved = 1;
@@ -15002,6 +15031,15 @@
   previous_help_echo = Qnil;
   staticpro (&previous_help_echo);
   help_echo_pos = -1;
+
+  DEFVAR_BOOL ("x-autoselect-window", &x_autoselect_window_p,
+    doc: /* *Non-nil means autoselect window with mouse pointer.  */);
+  x_autoselect_window_p = 0;
+
+  DEFVAR_BOOL ("x-autoselect-always", &x_autoselect_always_p,
+    doc: /* *Non-nil means always autoselect window with mouse pointer
+even if user switched to different window.  */);
+  x_autoselect_always_p = 0;
 
   DEFVAR_BOOL ("x-stretch-cursor", &x_stretch_cursor_p,
     doc: /* *Non-nil means draw block cursor as wide as the glyph under it.

-- 
Pavel Janík

panic("IRQ, you lose...");
                  -- 2.2.16 arch/mips/sgi/kernel/indy_int.c

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

Re: PATCH: focus follows mouse in C

[Eli Zaretskii]

> From: Pavel@Janik.cz (Pavel =?iso-8859-2?q?Jan=EDk?=)
> Date: Sat, 09 Feb 2002 12:53:12 +0100
> 
> --- xterm.c.~1.703.~	Sun Jan 27 17:15:53 2002
> +++ xterm.c	Sat Feb  9 12:45:39 2002
> @@ -253,6 +253,15 @@

Ahem.. why is this only in xterm.c?  Is it just because this is a
prototype?  I mean, we surely want this for any version that supports
a mouse, including the Windows, the Mac, and the DOS ports, right?

And what about the tty version with xt-mouse or an equivalent?

> +  DEFVAR_BOOL ("x-autoselect-window", &x_autoselect_window_p,
> +    doc: /* *Non-nil means autoselect window with mouse pointer.  */);
> +  x_autoselect_window_p = 0;
> +
> +  DEFVAR_BOOL ("x-autoselect-always", &x_autoselect_always_p,
> +    doc: /* *Non-nil means always autoselect window with mouse pointer
> +even if user switched to different window.  */);
> +  x_autoselect_always_p = 0;

Isn't it better to have a single option which can be nil, t, or
something non-nil and not t?

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

Re: PATCH: focus follows mouse in C

[Pavel Janík]

   From: "Eli Zaretskii" <eliz@is.elta.co.il>
   Date: Sat, 09 Feb 2002 16:52:29 +0200

   > > --- xterm.c.~1.703.~	Sun Jan 27 17:15:53 2002
   > > +++ xterm.c	Sat Feb  9 12:45:39 2002
   > > @@ -253,6 +253,15 @@
   > 
   > Ahem.. why is this only in xterm.c?  Is it just because this is a
   > prototype?  I mean, we surely want this for any version that supports
   > a mouse, including the Windows, the Mac, and the DOS ports, right?

Of course. I know that, Eli. I sent only changes to xterm.c. I will not
prepare changes for all ports because I do not have such systems and do not
want to do changes in something I can not test. I will send changes to
respective maintainers who will check and test them. But I will do that all
when we settle on the right solution...

   > And what about the tty version with xt-mouse or an equivalent?

The same as above applies.

   > > +  DEFVAR_BOOL ("x-autoselect-window", &x_autoselect_window_p,
   > > +    doc: /* *Non-nil means autoselect window with mouse pointer.  */);
   > > +  x_autoselect_window_p = 0;
   > > +
   > > +  DEFVAR_BOOL ("x-autoselect-always", &x_autoselect_always_p,
   > > +    doc: /* *Non-nil means always autoselect window with mouse pointer
   > > +even if user switched to different window.  */);
   > > +  x_autoselect_always_p = 0;
   > 
   > Isn't it better to have a single option which can be nil, t, or
   > something non-nil and not t?

Perhaps. Any other opinions?
-- 
Pavel Janík

No matter how hard you try, you can't make a racehorse out of a pig.
You can, however, make a faster pig.
                  -- An unknown author in GNU Emacs about Emacs's byte-opt

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

Re: PATCH: focus follows mouse in C

[Pavel Janík]

   From: "Eli Zaretskii" <eliz@is.elta.co.il>
   Date: Sat, 09 Feb 2002 16:52:29 +0200

   > And what about the tty version with xt-mouse or an equivalent?

It is not possible at all in xterm, because xterm can not sent mouse
movement events, only button presses/releases (see ctlseqs.PS from
X documentation, page 5). The same is stated in the xterm manual page:

       Xterm allows character-based applications to receive mouse
       events (currently button-press  and  release  events,  and
       button-motion  events) as keyboard control sequences.  See
       Xterm Control Sequences for details.

On the other hand, it is doable on a virtual console (at least on my system
which is based on Linux kernel) because t-mouse.el (resp. mev program from
gpm suite) can report mouse movements (current t-mouse.el which is part of
gpm, ignores mouse movements, because the cmdline argument "-e-move" is
used for mev).
-- 
Pavel Janík

Use data arrays to avoid repetitive control sequences.
                  --  The Elements of Programming Style (Kernighan & Plaugher)

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

Re: PATCH: focus follows mouse in C

[Al Petrofsky]

> From: Pavel@Janik.cz (Pavel =?iso-8859-2?q?Jan=EDk?=)
>    From: "Eli Zaretskii" <eliz@is.elta.co.il>
> 
>    > And what about the tty version with xt-mouse or an equivalent?
> 
> It is not possible at all in xterm, because xterm can not sent mouse
> movement events, only button presses/releases (see ctlseqs.PS from
> X documentation, page 5). The same is stated in the xterm manual page:
> 
>        Xterm allows character-based applications to receive mouse
>        events (currently button-press  and  release  events,  and
>        button-motion  events) as keyboard control sequences.  See
>        Xterm Control Sequences for details.

In x11r6.5, xterm cannot send any motion events, but in XFree86 3.3
and later, xterm can send both button-motion events (i.e. motion
events while a button is pressed) and regular pointer-motion events.
Some of the documentation is still a little behind.

> On the other hand, it is doable on a virtual console (at least on my system
> which is based on Linux kernel) because t-mouse.el (resp. mev program from
> gpm suite) can report mouse movements (current t-mouse.el which is part of
> gpm, ignores mouse movements, because the cmdline argument "-e-move" is
> used for mev).

The gpm design has always seemed fundamentally broken to me, because
it sends mouse events over a completely separate channel (a unix
domain socket) rather than through the tty.  This makes it impossible
to rlogin to a remote host and have applications on the remote host
receive the mouse events.

-al

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

Re: PATCH: focus follows mouse in C

[Eli Zaretskii]

On Sat, 9 Feb 2002, Al Petrofsky wrote:

> In x11r6.5, xterm cannot send any motion events, but in XFree86 3.3
> and later, xterm can send both button-motion events (i.e. motion
> events while a button is pressed) and regular pointer-motion events.
> Some of the documentation is still a little behind.

Is there _any_ docs on this available?  Anything at all?

I think we'll need changes on the C level to support this feature,
since mouse motion is tracked on the C level in Emacs.

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

Re: PATCH: focus follows mouse in C

[Al Petrofsky]

> From: Eli Zaretskii <eliz@is.elta.co.il>
> On Sat, 9 Feb 2002, Al Petrofsky wrote:
> 
> > In x11r6.5, xterm cannot send any motion events, but in XFree86 3.3
> > and later, xterm can send both button-motion events (i.e. motion
> > events while a button is pressed) and regular pointer-motion events.
> > Some of the documentation is still a little behind.
> 
> Is there _any_ docs on this available?  Anything at all?

Yes, it is documented in xc/doc/specs/xterm/ctlseqs.ms in the XFree86
tree.  The button-event mode is correctly documented.  The any-event
mode is mistakenly described thusly:

  Any-event mode is the same as button-event mode, except that all motion
  events are reported instead of just those that enter a new character cell.

That should be "instead of just those that occur while a button is
held down".

-al

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

Re: PATCH: focus follows mouse in C

[Richard Stallman]

    Yes, now I finally understand your and RMS' point. I agree and I think that
    it should default to behaviour described by you. I think that when I have
    two windows and point is in the bottom one and I do C-x o, small mouse
    movement (still in bottom window) should not (by default, but user should
    be able to change it) select the bottom window again. This should be it
    (module cus-start.el etc. changes):

With this change, do you like the results better?
Is it a convenient feature now?

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

Re: PATCH: focus follows mouse in C

[Pavel Janík]

   From: Richard Stallman <rms@gnu.org>
   Date: Sun, 10 Feb 2002 19:08:59 -0700 (MST)

   >     Yes, now I finally understand your and RMS' point. I agree and I think that
   >     it should default to behaviour described by you. I think that when I have
   >     two windows and point is in the bottom one and I do C-x o, small mouse
   >     movement (still in bottom window) should not (by default, but user should
   >     be able to change it) select the bottom window again. This should be it
   >     (module cus-start.el etc. changes):
   > 
   > With this change, do you like the results better?
   > Is it a convenient feature now?

Yes, it is much better and in fact, I;m now using it. I will install it
today.
-- 
Pavel Janík

printk(KERN_WARNING "Multi-volume CD somehow got mounted.\n");
                  -- 2.2.16 fs/isofs/inode.c

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

Re: PATCH: focus follows mouse in C

[Richard Stallman]

       > With this change, do you like the results better?
       > Is it a convenient feature now?

    Yes, it is much better and in fact, I;m now using it. I will install it
    today.

Thanks.

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

Re: PATCH: focus follows mouse in C

[Pavel Janík]

   From: "Eli Zaretskii" <eliz@is.elta.co.il>
   Date: Sat, 09 Feb 2002 16:52:29 +0200

   > Ahem.. why is this only in xterm.c?  Is it just because this is a
   > prototype?  I mean, we surely want this for any version that supports
   > a mouse, including the Windows, the Mac, and the DOS ports, right?

I have just committed that change and notified Jason and Andrew. Windows
and Mac ports should be fairly easy (they use almost the same
infrastructure). But what about DOS? The only changed thing in xterm.c is
note_mouse_highlight. I do not know how it works under DOS, Eli can you
help here? Is Emacs able to use mouse under DOS? The last time I met DOS,
I was 10 years younger then now ;-)
-- 
Pavel Janík

Don't just echo the code with comments - make every comment count.
                  --  The Elements of Programming Style (Kernighan & Plaugher)

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

Re: PATCH: focus follows mouse in C

[Eli Zaretskii]

> From: Pavel@Janik.cz (Pavel =?iso-8859-2?q?Jan=EDk?=)
> Date: Sat, 16 Feb 2002 11:35:51 +0100
> 
> I have just committed that change and notified Jason and Andrew. Windows
> and Mac ports should be fairly easy (they use almost the same
> infrastructure). But what about DOS? The only changed thing in xterm.c is
> note_mouse_highlight. I do not know how it works under DOS, Eli can you
> help here?

I already committed the changes to msdos.c to support this new
feature.  The comments I sent some time ago are based on playing with
this in the MS-DOS port.

> Is Emacs able to use mouse under DOS?

Yes, of course.

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