bug#15999: 24.3.50; Redisplay broken with split windows

bug#15999: 24.3.50; Redisplay broken with split windows

[Stephen Berman]

A recent change -- my guess is this one (I don't have time now to bisect
to confirm this suspicion):

115272: Stefan Monnier 2013-11-28 Refine redisplay optimizations to only
redisplay *some* frames/windows

-- has broken redisplay with split windows.  To reproduce:

0. emacs -Q
1. M-x display-time-mode
3. C-x 2 (or C- x 3)
4. C-x b, so one window displays *scratch* and the other *Messages*.
5. When the displayed time changes, only the current buffer is updated;
   in the other buffer the time display is unchanged, and the scroll bar
   disappears (note my build does not use toolkit scroll bars, I don't
   know if that makes a differences.)  This states of affairs remains
   until the non-updated buffer becomes current.


In GNU Emacs 24.3.50.14 (x86_64-suse-linux-gnu, GTK+ Version 3.10.2)
 of 2013-11-29 on rosalinde
Bzr revision: 115285 schwab@linux-m68k.org-20131129112504-5izsans1kw88maan
Windowing system distributor `The X.Org Foundation', version 11.0.11403901
System Description:	openSUSE 13.1 (Bottle) (x86_64)

Configured using:
 `configure --without-toolkit-scroll-bars CFLAGS=-g3'

Important settings:
  value of $LANG: en_US.UTF-8
  value of $XMODIFIERS: @im=ibus
  locale-coding-system: utf-8-unix
  default enable-multibyte-characters: t

Major mode: Message

Minor modes in effect:
  diff-auto-refine-mode: t
  gnus-message-citation-mode: t
  mml-mode: t
  shell-dirtrack-mode: t
  show-paren-mode: t
  recentf-mode: t
  display-time-mode: t
  electric-pair-mode: t
  electric-indent-mode: t
  tooltip-mode: t
  mouse-wheel-mode: t
  tool-bar-mode: t
  menu-bar-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  auto-compression-mode: t
  temp-buffer-resize-mode: t
  column-number-mode: t
  line-number-mode: t
  auto-fill-function: message-do-auto-fill
  abbrev-mode: t

bug#15999: 24.3.50; Redisplay broken with split windows

[Stefan Monnier]

> 115272: Stefan Monnier 2013-11-28 Refine redisplay optimizations to only
> redisplay *some* frames/windows

Sounds highly likely, indeed.

> 0. emacs -Q
> 1. M-x display-time-mode
> 3. C-x 2 (or C- x 3)
> 4. C-x b, so one window displays *scratch* and the other *Messages*.
> 5. When the displayed time changes, only the current buffer is updated;
>    in the other buffer the time display is unchanged, and the scroll bar
>    disappears (note my build does not use toolkit scroll bars, I don't
>    know if that makes a differences.)  This states of affairs remains
>    until the non-updated buffer becomes current.

- The mode-line not being redrawn in the other window is expected.
  The patch below should fix it.
- But for some reason I couldn't reproduce it (looks like something
  somewhere in my tests is causing the extra redisplay needed).
  Ah, I could reproduce it by building --without-toolkit-scroll-bars, so
  it looks like the presence of toolkit scrollbars somehow plays an
  important role.
- The scroll-bar disappearance is not expected.  If the patch below
  makes this problem disappear, it's only because it hides it, not
  because it fixes it.
  

        Stefan

=== modified file 'lisp/time.el'
--- lisp/time.el	2013-08-09 18:49:36 +0000
+++ lisp/time.el	2013-11-29 17:43:47 +0000
@@ -471,7 +471,7 @@
     ;; This is inside the let binding, but we are not going to document
     ;; what variables are available.
     (run-hooks 'display-time-hook))
-  (force-mode-line-update))
+  (force-mode-line-update 'all))
 
 (defun display-time-file-nonempty-p (file)
   (let ((remote-file-name-inhibit-cache (- display-time-interval 5)))

bug#15999: 24.3.50; Redisplay broken with split windows

[Stefan Monnier]

>    in the other buffer the time display is unchanged, and the scroll bar
>    disappears (note my build does not use toolkit scroll bars, I don't

OK, I figured out why the scrollbar disappeared.  Should all be fixed
now, thank you,


        Stefan

bug#15999: 24.3.50; Redisplay broken with split windows

[Stephen Berman]

On Fri, 29 Nov 2013 13:21:10 -0500 Stefan Monnier <monnier@iro.umontreal.ca> wrote:

>>    in the other buffer the time display is unchanged, and the scroll bar
>>    disappears (note my build does not use toolkit scroll bars, I don't
>
> OK, I figured out why the scrollbar disappeared.  Should all be fixed
> now, thank you,

Confirmed; thanks.

Steve Berman

bug#15999: 24.3.50; Redisplay broken with split windows

[Stephen Berman]

On Fri, 29 Nov 2013 21:02:13 +0100 Stephen Berman <stephen.berman@gmx.net> wrote:

> On Fri, 29 Nov 2013 13:21:10 -0500 Stefan Monnier <monnier@iro.umontreal.ca> wrote:
>
>>>    in the other buffer the time display is unchanged, and the scroll bar
>>>    disappears (note my build does not use toolkit scroll bars, I don't
>>
>> OK, I figured out why the scrollbar disappeared.  Should all be fixed
>> now, thank you,
>
> Confirmed; thanks.

Whoops, spoke too soon -- I only looked at the scroll bar.  The time in
the mode line of the non-selected windows is still not being updated.

Steve Berman

bug#15999: 24.3.50; Redisplay broken with split windows

[Stephen Berman]

On Fri, 29 Nov 2013 21:06:05 +0100 Stephen Berman <stephen.berman@gmx.net> wrote:

> On Fri, 29 Nov 2013 21:02:13 +0100 Stephen Berman <stephen.berman@gmx.net> wrote:
>
>> On Fri, 29 Nov 2013 13:21:10 -0500 Stefan Monnier <monnier@iro.umontreal.ca> wrote:
>>
>>>>    in the other buffer the time display is unchanged, and the scroll bar
>>>>    disappears (note my build does not use toolkit scroll bars, I don't
>>>
>>> OK, I figured out why the scrollbar disappeared.  Should all be fixed
>>> now, thank you,
>>
>> Confirmed; thanks.
>
> Whoops, spoke too soon -- I only looked at the scroll bar.  The time in
> the mode line of the non-selected windows is still not being updated.

It's fixed now, due to r115299 I presume; thanks.

Steve Berman

bug#15999: 24.3.50; Redisplay broken with split windows

[Stefan Monnier]

> It's fixed now, due to r115299 I presume; thanks.

Indeed that should have fixed it, thank you,


        Stefan