bug#64658: 30.0.50; NotifyGrab and NotifyUngrab should be filtered with XInput2

bug#64658: 30.0.50; NotifyGrab and NotifyUngrab should be filtered with XInput2

[左 智杰]

In i3wm + xfce4 setup, when multiple frames opened, switch between frames
using only keyboard, and leave the mouse as:

+-----------------+  +-----------------+
|                 |  |                 |
|                 |  |                 |
|                 |  |                 |
|                 |  |                 |
|                 |  |                 |
|                 |  |                 |
|     FRAME 1     |  |     FRAME 2     |
|    *FOCUSED*    |  |   *UNFOCUSED*   |
|                 |  |          ^      |
|                 |  |          |      |
|                 |  |        MOUSE    |
|                 |  |         HERE    |
|                 |  |                 |
+-----------------+  +-----------------+

This will make the ~emacsclient -e '(selected-frame)'~ returns wrong frame, due to
XInput2 sends a ~NotifyGrab~ with the unfocused frame, making Emacs selects it.

We've make a small but ugly changes to fix this issue:

#+begin_src
diff --git a/src/xterm.c b/src/xterm.c
index 9ecead03b0..d4836afeeb 100644
--- a/src/xterm.c
+++ b/src/xterm.c
@@ -13193,6 +13193,11 @@ xi_focus_handle_for_device (struct x_display_info *dpyinfo,
   if (!device)
     return;
 
+  /* Same as x_detect_focus_change, issue persists in XInput2. */
+  if (event->evtype == XI_FocusIn || event->evtype == XI_FocusOut)
+    if (event->mode == XINotifyGrab || event->mode == XINotifyUngrab)
+      return;
+
   switch (event->evtype)
     {
     case XI_FocusIn:
#+end_src

Wonder if we're correct?

Yours sincerely,
Zigit Zo

---

In GNU Emacs 30.0.50 (build 2, x86_64-pc-linux-gnu, GTK+ Version
 3.24.38, cairo version 1.17.8) of 2023-07-16 built on space
Repository revision: 2df086d121e8ebee872236d0436205144047c5c9
Repository branch: master
Windowing system distributor 'The X.Org Foundation', version 11.0.12101008
System Description: Arch Linux

Configured using:
 'configure --prefix=/home/byte/Workspace/emacs/opt'

Configured features:
ACL CAIRO DBUS FREETYPE GIF GLIB GMP GNUTLS GPM GSETTINGS HARFBUZZ JPEG
JSON LCMS2 LIBSYSTEMD LIBXML2 MODULES NOTIFY INOTIFY PDUMPER PNG RSVG
SECCOMP SOUND SQLITE3 THREADS TIFF TOOLKIT_SCROLL_BARS TREE_SITTER WEBP
X11 XDBE XIM XINPUT2 XPM GTK3 ZLIB

Important settings:
  value of $LANG: zh_CN.UTF-8
  value of $XMODIFIERS: @im=fcitx
  locale-coding-system: utf-8-unix

Major mode: C/*l

Minor modes in effect:
  server-mode: t
  bug-reference-prog-mode: t
  text-scale-mode: t
  tooltip-mode: t
  global-eldoc-mode: t
  show-paren-mode: t
  electric-indent-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
  blink-cursor-mode: t
  line-number-mode: t
  indent-tabs-mode: t
  transient-mark-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  auto-compression-mode: t
  abbrev-mode: t

Load-path shadows:
None found.

Features:
(shadow sort mail-extr emacsbug message mailcap yank-media puny dired
dired-loaddefs rfc822 mml mml-sec password-cache epa derived epg rfc6068
epg-config gnus-util text-property-search time-date subr-x mm-decode
mm-bodies mm-encode mail-parse rfc2231 mailabbrev gmm-utils mailheader
sendmail rfc2047 rfc2045 ietf-drums mm-util mail-prsvr mail-utils pp
server vc-git diff-mode easy-mmode vc vc-dispatcher bug-reference
byte-opt gv bytecomp byte-compile cc-mode cc-fonts cc-guess cc-menus
cc-cmds cc-styles cc-align cc-engine cc-vars cc-defs cl-loaddefs cl-lib
format-spec face-remap china-util rmc iso-transl tooltip cconv eldoc
paren electric uniquify ediff-hook vc-hooks lisp-float-type elisp-mode
mwheel term/x-win x-win term/common-win x-dnd tool-bar dnd fontset image
regexp-opt fringe tabulated-list replace newcomment text-mode lisp-mode
prog-mode register page tab-bar menu-bar rfn-eshadow isearch easymenu
timer select scroll-bar mouse jit-lock font-lock syntax font-core
term/tty-colors frame minibuffer nadvice seq simple cl-generic
indonesian philippine cham georgian utf-8-lang misc-lang vietnamese
tibetan thai tai-viet lao korean japanese eucjp-ms cp51932 hebrew greek
romanian slovak czech european ethiopic indian cyrillic chinese
composite emoji-zwj charscript charprop case-table epa-hook
jka-cmpr-hook help abbrev obarray oclosure cl-preloaded button loaddefs
theme-loaddefs faces cus-face macroexp files window text-properties
overlay sha1 md5 base64 format env code-pages mule custom widget keymap
hashtable-print-readable backquote threads dbusbind inotify lcms2
dynamic-setting system-font-setting font-render-setting cairo gtk
x-toolkit xinput2 x multi-tty move-toolbar make-network-process emacs)

Memory information:
((conses 16 88211 12696) (symbols 48 8441 0) (strings 32 25666 1087)
 (string-bytes 1 914293) (vectors 16 16380)
 (vector-slots 8 280038 15484) (floats 8 37 44)
 (intervals 56 4640 108) (buffers 984 13))

bug#64658: 30.0.50; NotifyGrab and NotifyUngrab should be filtered with XInput2

[Po Lu via Bug reports for GNU Emacs, the Swiss army knife of text editors]

左 智杰 <hardirq@outlook.com> writes:

> In i3wm + xfce4 setup, when multiple frames opened, switch between frames
> using only keyboard, and leave the mouse as:
>
> +-----------------+  +-----------------+
> |                 |  |                 |
> |                 |  |                 |
> |                 |  |                 |
> |                 |  |                 |
> |                 |  |                 |
> |                 |  |                 |
> |     FRAME 1     |  |     FRAME 2     |
> |    *FOCUSED*    |  |   *UNFOCUSED*   |
> |                 |  |          ^      |
> |                 |  |          |      |
> |                 |  |        MOUSE    |
> |                 |  |         HERE    |
> |                 |  |                 |
> +-----------------+  +-----------------+
>
> This will make the ~emacsclient -e '(selected-frame)'~ returns wrong frame, due to
> XInput2 sends a ~NotifyGrab~ with the unfocused frame, making Emacs selects it.
>
> We've make a small but ugly changes to fix this issue:
>
> #+begin_src
> diff --git a/src/xterm.c b/src/xterm.c
> index 9ecead03b0..d4836afeeb 100644
> --- a/src/xterm.c
> +++ b/src/xterm.c
> @@ -13193,6 +13193,11 @@ xi_focus_handle_for_device (struct x_display_info *dpyinfo,
>    if (!device)
>      return;
>  
> +  /* Same as x_detect_focus_change, issue persists in XInput2. */
> +  if (event->evtype == XI_FocusIn || event->evtype == XI_FocusOut)
> +    if (event->mode == XINotifyGrab || event->mode == XINotifyUngrab)
> +      return;
> +
>    switch (event->evtype)
>      {
>      case XI_FocusIn:
> #+end_src
>
> Wonder if we're correct?

I don't want this hack in the input extension code, because it is much
more sensitive to inconsistencies in the keyboard focus state.

What is the value of event->send_event for these unwanted focus events?

bug#64658: 30.0.50; NotifyGrab and NotifyUngrab should be filtered with XInput2

[Zigit Zo]

They're both 0, so the ~mode~ member is chosen :( For convenient, posted
all members when FocusIn event triggered, the last one is the cause of
~selected-frame~ changes.

#+begin_src
Thread 1 "emacs" hit Breakpoint 5, xi_handle_focus_change (dpyinfo=0x558ec6140c00) at ../../src/xterm.c:13068
13068	{
#2  0x0000558ec403e631 in event_handler_gdk (gxev=0x7ffd3ebb6080, ev=<optimized out>, data=<optimized out>)
     at ../../src/xterm.c:17671
$282 = (struct frame *) 0x558ec5dd5da0
"emacs - FRAME 1 TITLE"
(gdb) p *(XIFocusInEvent*) ((XEvent*) gxev)->xcookie.data
$284 = {
   type = 35,
   serial = 265494,
   send_event = 0,
   display = 0x558ec5ecb250,
   extension = 131,
   evtype = 9,
   time = 170569789,
   deviceid = 3,
   sourceid = 3,
   detail = 3,
   root = 483,
   event = 79691928,
   child = 0,
   root_x = 2769,
   root_y = 993,
   event_x = 2751,
   event_y = 879,
   mode = 3,
   focus = 0,
   same_screen = 1,
   buttons = {
     mask_len = 4,
     mask = 0x558ec9481958 ""
   },
   mods = {
     base = 64,
     latched = 0,
     locked = 0,
     effective = 64
   },
   group = {
     base = 0,
     latched = 0,
     locked = 0,
     effective = 0
   }
}

Thread 1 "emacs" hit Breakpoint 5, xi_handle_focus_change (dpyinfo=0x558ec6140c00) at ../../src/xterm.c:13068
13068	{
#2  0x0000558ec403e631 in event_handler_gdk (gxev=0x7ffd3ebb6080, ev=<optimized out>, data=<optimized out>)
     at ../../src/xterm.c:17671
$289 = (struct frame *) 0x558ec5dd5da0
"emacs - FRAME 1 TITLE"
(gdb) p *(XIFocusInEvent*) ((XEvent*) gxev)->xcookie.data
$291 = {
   type = 35,
   serial = 270527,
   send_event = 0,
   display = 0x558ec5ecb250,
   extension = 131,
   evtype = 9,
   time = 170569915,
   deviceid = 3,
   sourceid = 3,
   detail = 0,
   root = 483,
   event = 79691928,
   child = 0,
   root_x = 2769,
   root_y = 993,
   event_x = 2751,
   event_y = 879,
   mode = 2,
   focus = 0,
   same_screen = 1,
   buttons = {
     mask_len = 4,
     mask = 0x558ecc272658 ""
   },
   mods = {
     base = 64,
     latched = 0,
     locked = 0,
     effective = 64
   },
   group = {
     base = 0,
     latched = 0,
     locked = 0,
     effective = 0
   }
}

Thread 1 "emacs" hit Breakpoint 5, xi_handle_focus_change (dpyinfo=0x558ec6140c00) at ../../src/xterm.c:13068
13068	{
#2  0x0000558ec403e631 in event_handler_gdk (gxev=0x7ffd3ebb6080, ev=<optimized out>, data=<optimized out>)
     at ../../src/xterm.c:17671
$295 = (struct frame *) 0x558ecde5baf0
"emacs - FRAME 2 TITLE UNFOCUSED"
(gdb) p *(XIFocusInEvent*) ((XEvent*) gxev)->xcookie.data
$297 = {
   type = 35,
   serial = 270527,
   send_event = 0,
   display = 0x558ec5ecb250,
   extension = 131,
   evtype = 9,
   time = 170570127,
   deviceid = 3,
   sourceid = 3,
   detail = 5,
   root = 483,
   event = 79692372,
   child = 0,
   root_x = 2769,
   root_y = 993,
   event_x = 459,
   event_y = 879,
   mode = 1,
   focus = 0,
   same_screen = 1,
   buttons = {
     mask_len = 4,
     mask = 0x558ecb779b88 ""
   },
   mods = {
     base = 64,
     latched = 0,
     locked = 0,
     effective = 64
   },
   group = {
     base = 0,
     latched = 0,
     locked = 0,
     effective = 0
   }
}
#+end_sr

Hope this will be useful.

在 2023/7/16 17:09, Po Lu 写道:

> 左 智杰 <hardirq@outlook.com> writes:
>
>> In i3wm + xfce4 setup, when multiple frames opened, switch between frames
>> using only keyboard, and leave the mouse as:
>>
>> +-----------------+  +-----------------+
>> |                 |  |                 |
>> |                 |  |                 |
>> |                 |  |                 |
>> |                 |  |                 |
>> |                 |  |                 |
>> |                 |  |                 |
>> |     FRAME 1     |  |     FRAME 2     |
>> |    *FOCUSED*    |  |   *UNFOCUSED*   |
>> |                 |  |          ^      |
>> |                 |  |          |      |
>> |                 |  |        MOUSE    |
>> |                 |  |         HERE    |
>> |                 |  |                 |
>> +-----------------+  +-----------------+
>>
>> This will make the ~emacsclient -e '(selected-frame)'~ returns wrong frame, due to
>> XInput2 sends a ~NotifyGrab~ with the unfocused frame, making Emacs selects it.
>>
>> We've make a small but ugly changes to fix this issue:
>>
>> #+begin_src
>> diff --git a/src/xterm.c b/src/xterm.c
>> index 9ecead03b0..d4836afeeb 100644
>> --- a/src/xterm.c
>> +++ b/src/xterm.c
>> @@ -13193,6 +13193,11 @@ xi_focus_handle_for_device (struct x_display_info *dpyinfo,
>>     if (!device)
>>       return;
>>   
>> +  /* Same as x_detect_focus_change, issue persists in XInput2. */
>> +  if (event->evtype == XI_FocusIn || event->evtype == XI_FocusOut)
>> +    if (event->mode == XINotifyGrab || event->mode == XINotifyUngrab)
>> +      return;
>> +
>>     switch (event->evtype)
>>       {
>>       case XI_FocusIn:
>> #+end_src
>>
>> Wonder if we're correct?
> I don't want this hack in the input extension code, because it is much
> more sensitive to inconsistencies in the keyboard focus state.
>
> What is the value of event->send_event for these unwanted focus events?

bug#64658: 30.0.50; NotifyGrab and NotifyUngrab should be filtered with XInput2

[Po Lu via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Zigit Zo <hardirq@outlook.com> writes:

> They're both 0, so the ~mode~ member is chosen :( For convenient, posted
> all members when FocusIn event triggered, the last one is the cause of
> ~selected-frame~ changes.

So I gather that a corresponding XI_FocusOut event is sent after the
window manager shortcut completes, and `selected-frame' returns an
unexpected value only for the duration of the grab?

If so, there's no bug in Emacs: the other frame really is selected for
the duration of the shortcut.