Fcitx5 5.1.15: fcitx5-configtool fails with QDBusError("org.freedesktop.DBus.Error.NoReply")

[HeaoYe]

Summary

Hi,

I updated Fcitx5 to 5.1.15 today on Arch Linux, but it stopped working properly.

When I run:

fcitx5 -r -D

it seems to start normally:

I2025-10-26 12:10:02.830343 instance.cpp:1416] Override Enabled Addons: {}
I2025-10-26 12:10:02.832660 instance.cpp:1417] Override Disabled Addons: {}
...
I2025-10-26 12:10:05.893371 addonmanager.cpp:204] Loaded addon notifications

However:

• I cannot switch input methods at all.
• Running fcitx5-configtool gives this error:

fcitx5-config: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.")

When I try to terminate fcitx5 -r -D using Ctrl-C, it prints a backtrace but never exits:

^C=========================
Fcitx 5.1.15 -- Get Signal No.: 2
...
/usr/lib/libFcitx5Core.so.7(_ZN5fcitx8Instance4execEv+0x5e) [0x7f293e1fc99e]
fcitx5(+0xe4be) [0x563ad01cb4be]
/usr/lib/libc.so.6(+0x27675) [0x7f293d827675]

It also cannot be killed with a normal kill [PID].
Only pkill -9 fcitx5 works.

Finally, I downgraded the following packages, which were updated along with fcitx5 today:

• fcitx5 → 5.1.15 → 5.1.14
• fcitx5-qt → 5.1.11-1 → 5.1.10-4
• fcitx5-chinese-addons → 5.1.10-1 → 5.1.9-1
• fcitx5-configtool → 5.1.11-1 → 5.1.10-1
  After that, everything works fine again, including switching input methods and opening the config tool.

Steps to Reproduce

sudo pacman -S fcitx5-im
fcitx5 -r -D

Cannot terminate fcitx5 by Ctrl-C and kill [PID].
Cannot use input method.

---

fcitx5-configtool

fcitx5-configtool will show nothing in its window and print an error msg:

fcitx5-config: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.")

Expected Behavior

• Input methods can be switched.
• fcitx5-configtool can modify Fcitx5 configurations.
• fcitx5-diagnose completes without blocking.

Output of fcitx5-diagnose command

When I run fcitx5-diagnose with version 5.1.15, it hangs after outputting:

# Fcitx State:
...
5.  DBus interface:

    Using `dbus-send` to check dbus.

So the following is the output when using version 5.1.14:

System Info:

1. uname -a:

   Linux 6.17.5-arch1-1 #1 SMP PREEMPT_DYNAMIC Thu, 23 Oct 2025 18:49:03 +0000 x86_64 GNU/Linux
   

2. lsb_release:

   lsb_release not found.

3. /etc/lsb-release:

   /etc/lsb-release not found.

4. /etc/os-release:

   NAME="Arch Linux"
   PRETTY_NAME="Arch Linux"
   ID=arch
   BUILD_ID=rolling
   ANSI_COLOR="38;2;23;147;209"
   HOME_URL="https://archlinux.org/"
   DOCUMENTATION_URL="https://wiki.archlinux.org/"
   SUPPORT_URL="https://bbs.archlinux.org/"
   BUG_REPORT_URL="https://gitlab.archlinux.org/groups/archlinux/-/issues"
   PRIVACY_POLICY_URL="https://terms.archlinux.org/docs/privacy-policy/"
   LOGO=archlinux-logo
   

5. Desktop Environment:

   Cannot determine desktop environment.

6. XDG SESSION TYPE:

   XDG_SESSION_TYPE='wayland'
   

7. Bash Version:

   BASH_VERSION='5.3.3(1)-release'
   

Environment:

1. DISPLAY:

   DISPLAY=':1'
   
   
   WAYLAND_DISPLAY='wayland-1'
   

2. Keyboard Layout:

   1. setxkbmap:

      WARNING: Running setxkbmap against an Xwayland server
      xkb_keymap {
      	xkb_keycodes  { include "evdev+aliases(qwerty)"	};
      	xkb_types     { include "complete"	};
      	xkb_compat    { include "complete"	};
      	xkb_symbols   { include "pc+us+inet(evdev)"	};
      	xkb_geometry  { include "pc(pc105)"	};
      };
      

   2. xprop:

      _XKB_RULES_NAMES(STRING) = "evdev", "pc105", "us", "", ""
      

3. Locale:

   1. All locales:

      C
      C.utf8
      en_US.utf8
      POSIX
      zh_CN.utf8
      

   2. Current locale:

      LANG=en_US.UTF-8
      LC_CTYPE="en_US.UTF-8"
      LC_NUMERIC="en_US.UTF-8"
      LC_TIME="en_US.UTF-8"
      LC_COLLATE="en_US.UTF-8"
      LC_MONETARY="en_US.UTF-8"
      LC_MESSAGES="en_US.UTF-8"
      LC_PAPER="en_US.UTF-8"
      LC_NAME="en_US.UTF-8"
      LC_ADDRESS="en_US.UTF-8"
      LC_TELEPHONE="en_US.UTF-8"
      LC_MEASUREMENT="en_US.UTF-8"
      LC_IDENTIFICATION="en_US.UTF-8"
      LC_ALL=
      

4. Directories:

   1. Home:

      /home/xxx
      

   2. ${XDG_CONFIG_HOME}:

      Environment variable XDG_CONFIG_HOME is not set.

      Current value of XDG_CONFIG_HOME is ~/.config (/home/xxx/.config).

   3. Fcitx5 Settings Directory:

      Current fcitx5 settings directory is ~/.config/fcitx5 (/home/xxx/.config/fcitx5).

5. Current user:

   The script is run as xxx (1000).

Fcitx State:

1. executable:

   Found fcitx5 at /usr/bin/fcitx5.

2. version:

   Fcitx version: 5.1.14

3. process:

   Found 1 fcitx5 process:

      3747 fcitx5
   

4. fcitx5-remote:

   fcitx5-remote works properly.

5. DBus interface:

   Using dbus-send to check dbus.

   Owner of DBus name org.fcitx.Fcitx5 is :1.50.

   PID of DBus name org.fcitx.Fcitx5 owner is 3747.

   Debug information from dbus:

      Group [wayland:] has 4 InputContext(s)
     IC [b993e8f2b02745359b21138deef613d9] program:google-chrome frontend:wayland_v2 cap:100000072 focus:0
     IC [a8d2e357e22e4e59a670d53cfd8ea496] program:kitty frontend:wayland_v2 cap:100000072 focus:1
     IC [49d7dde9b012483faaa6b16c32dbdb1a] program:kitty frontend:wayland_v2 cap:100000072 focus:0
     IC [9d5b3015fd2644da91e7415f8809916f] program: frontend:wayland_v2 cap:100000072 focus:0
   Group [x11::1] has 0 InputContext(s)
   Input Context without group
   

Frontends setup:

The environment variable checked by this script only shows the environment under current shell. It is still possible that you did not set the environment to the whole graphic desktop session. You may inspect the actual environment variable of a certain process by using xargs -0 -L1 /proc/$PID/environ for a certain process that you find not working.

Xim:

1. ${XMODIFIERS}:

   Environment variable XMODIFIERS is set to "@im=fcitx" correctly.
   Xim Server Name from Environment variable is fcitx.

2. XIM_SERVERS on root window:

   Xim server name is the same with that set in the environment variable.

Qt:

1. qt4 - ${QT4_IM_MODULE}:

   Environment variable QT_IM_MODULE is set to "fcitx" correctly.

   fcitx5-qt4-immodule-probing not found.

2. qt5 - ${QT_IM_MODULE}:

   Environment variable QT_IM_MODULE is set to "fcitx" correctly.

   Using fcitx5-qt5-immodule-probing to check the actual im module to be used under current environment:

   QT_QPA_PLATFORM=wayland
   QT_IM_MODULE=fcitx
   IM_MODULE_CLASSNAME=fcitx::QFcitxPlatformInputContext
   

3. qt6 - ${QT_IM_MODULE}:

   Environment variable QT_IM_MODULE is set to "fcitx" correctly.

   Using fcitx5-qt6-immodule-probing to check the actual im module to be used under current environment:

   QT_QPA_PLATFORM=wayland
   QT_IM_MODULE=fcitx
   IM_MODULE_CLASSNAME=fcitx::QFcitxPlatformInputContext
   

4. Qt IM module files:

   Found fcitx5 im module for qt: /usr/lib/qt/plugins/platforminputcontexts/libfcitx5platforminputcontextplugin.so.
   Found fcitx5 qt6 module: /usr/lib/fcitx5/qt6/libfcitx-quickphrase-editor5.so.
   Found fcitx5 im module for qt6: /usr/lib/qt6/plugins/platforminputcontexts/libfcitx5platforminputcontextplugin.so.
   Found unknown fcitx qt module: /usr/lib/qt6/plugins/plasma/kcms/systemsettings/kcm_fcitx5.so.

   Following error may not be accurate because guessing Qt version from path depends on how your distribution packages Qt. It is not a critical error if you do not use any Qt application with certain version of Qt or you are using text-input support by Qt under Wayland.
   Cannot find fcitx5 input method module for Qt4.

Gtk:

1. gtk - ${GTK_IM_MODULE}:

   Environment variable GTK_IM_MODULE is set to "fcitx" correctly.

   Using fcitx5-gtk2-immodule-probing to check the actual im module to be used under current environment:

   Using fcitx5-gtk3-immodule-probing to check the actual im module to be used under current environment:

   GTK_IM_MODULE=fcitx
   

   Using fcitx5-gtk4-immodule-probing to check the actual im module to be used under current environment:

   GTK_IM_MODULE=fcitx
   

2. gtk-query-immodules:

   1. gtk 2:

      Cannot find gtk-query-immodules for gtk 2

      Cannot find fcitx5 im module for gtk 2.

   2. gtk 3:

      Found gtk-query-immodules for gtk 3.24.51 at /usr/bin/gtk-query-immodules-3.0.
      Version Line:

      # Created by /usr/bin/gtk-query-immodules-3.0 from gtk+-3.24.51
      

      Found fcitx5 im modules for gtk 3.24.51.

      "/usr/lib/gtk-3.0/3.0.0/immodules/im-fcitx5.so" 
      "fcitx" "Fcitx5 (Flexible Input Method Framework5)" "fcitx5" "/usr/locale" "ja:ko:zh:*" 
      "fcitx5" "Fcitx5 (Flexible Input Method Framework5)" "fcitx5" "/usr/locale" "ja:ko:zh:*" 
      

3. Gtk IM module cache:

   1. gtk 2:

      Cannot find immodules cache for gtk 2

      Cannot find fcitx5 im module for gtk 2 in cache.

   2. gtk 3:

      Found immodules cache for gtk 3.24.51 at /usr/lib/gtk-3.0/3.0.0/immodules.cache.
      Version Line:

      # Created by /usr/bin/gtk-query-immodules-3.0 from gtk+-3.24.51
      

      Found fcitx5 im modules for gtk 3.24.51.

      "/usr/lib/gtk-3.0/3.0.0/immodules/im-fcitx5.so" 
      "fcitx" "Fcitx5 (Flexible Input Method Framework5)" "fcitx5" "/usr/locale" "ja:ko:zh:*" 
      "fcitx5" "Fcitx5 (Flexible Input Method Framework5)" "fcitx5" "/usr/locale" "ja:ko:zh:*" 
      

4. Gtk IM module files:

   1. gtk 2:

      All found Gtk 2 immodule files exist.

   2. gtk 3:

      All found Gtk 3 immodule files exist.

   3. gtk 4:

      All found Gtk 4 immodule files exist.

Configuration:

Fcitx Addons:

1. Addon Config Dir:

   Found fcitx5 addon config directory: /usr/share/fcitx5/addon.

2. Addon List:

   1. Found 27 enabled addons:

      Simplified and Traditional Chinese Translation 5.1.9
      Classic User Interface 5.1.14
      Clipboard 5.1.14
      Cloud Pinyin 5.1.9
      DBus 5.1.14
      DBus Frontend 5.1.14
      Emoji 5.1.14
      Fcitx4 Frontend 5.1.14
      Full width character 5.1.9
      IBus Frontend 5.1.14
      Input method selector 5.1.14
      Keyboard 5.1.14
      KDE Input Method Panel 5.1.14
      Status Notifier 5.1.14
      Notification 5.1.14
      Pinyin 5.1.9
      Extra Pinyin functionality 5.1.9
      Punctuation 5.1.9
      Quick Phrase 5.1.14
      Spell 5.1.14
      Table 5.1.9
      Unicode 5.1.14
      DBus Virtual Keyboard 5.1.14
      Wayland 5.1.14
      Wayland Input method frontend 5.1.14
      XCB 5.1.14
      X Input Method Frontend 5.1.14
      

   2. Found 0 disabled addons:

3. Addon Libraries:

   All libraries for all addons are found.

4. User Interface:

   Found 3 enabled user interface addons:

   Classic User Interface
   KDE Input Method Panel
   DBus Virtual Keyboard
   

Input Methods:

1. /home/xxx/.config/fcitx5/profile:

   [Groups/0]
   # Group Name
   Name=Default
   # Layout
   Default Layout=us
   # Default Input Method
   DefaultIM=pinyin
   
   [Groups/0/Items/0]
   # Name
   Name=keyboard-us
   # Layout
   Layout=
   
   [Groups/0/Items/1]
   # Name
   Name=pinyin
   # Layout
   Layout=
   
   [GroupOrder]
   0=Default
   

Log:

1. date:

   Sun Oct 26 01:00:31 PM CST 2025
   

2. /home/xxx/.config/fcitx5/crash.log:

   /home/xxx/.config/fcitx5/crash.log not found.

[awa2333]

What is your WM? Is it Hyprland? I found someone said that it only crash on hyprland wm.

[ludovicolea]

What is your WM? Is it Hyprland? I found someone said that it only crash on hyprland wm.

It crashes on my hyprland v0.51.1-6 as well, after updating to fcitx5 v5.1.15-1. Fcitx5 configtool fails to load any configs, and I also cant switch input method.

~ 
❯ fcitx5-configtool
fcitx5-config: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.")

[tokyo4j]

I'm having the same issue in labwc on Arch Linux

$ pacman -Q | grep fcitx
fcitx5 5.1.15-1
fcitx5-configtool 5.1.11-1
fcitx5-gtk 5.1.4-2
fcitx5-mozc 2.32.5981.102.g29e2b95-1
fcitx5-qt 5.1.11-1

EDIT: I'm bisecting the recent commits

[HeaoYe]

What is your WM? Is it Hyprland? I found someone said that it only crash on hyprland wm.

Yes, I'm using Hyprland.

[au-summer]

I am also facing this problem on Hyprland. After downgrading to 5.1.14 the problem disappears.

[tokyo4j]

I confirmed reverting cce4115 fixes the issue.

[HeaoYe]

I confirmed reverting cce4115 fixes the issue.

Do you think this is caused by hyprland/labwc not supporting ext_data_control protocol yet？

[tokyo4j]

I confirmed reverting cce4115 fixes the issue.

Do you think this is caused by hyprland/labwc not supporting ext_data_control protocol yet？

At least labwc supports ext-data-control protocol. I don't know the cause of the issue.

EDIT: Hyprland seems to support the protocol, too: https://github.com/hyprwm/Hyprland/blob/6ea4769b39fccec22ae2acd015511a745d378840/src/protocols/ExtDataDevice.hpp

EDIT2: I patched labwc to drop ext-data-control support and fcitx5 works fine. So strangely, cce4115 which adds ext-data-control protocol support is breaking for compositors who support ext-data-control protocol.

EDIT3: Patching labwc to drop wlr-data-control support also fixed the issue. So the problem is that fcitx5 misbehaves if the compositor supports both of wlr-data-control and ext-data-control.

[Pix3lexe]

Same issue here on Hyprland. After downgrading to 5.1.14 it works again.

[ymfqa]

Same bug on Hyprland,change to 5.1.14 can work

[HeaoYe]

It seems that when both ext_data_control and wlr_data_control are available. The globalCreated() signal callback first creates ZwlrDataControlManagerV1.

fcitx5/src/modules/clipboard/waylandclipboard.cpp

Lines 284 to 293 in cce4115

	} else if (interface ==
	wayland::ZwlrDataControlManagerV1::interface) {
	if (ptr != wlr_manager_) {
	deviceMap_.clear();
	wlr_manager_ =
	display_
	->getGlobal<wayland::ZwlrDataControlManagerV1>();
	}
	refreshSeat();
	} else if (interface == wayland::WlSeat::interface) {

And then put ZwlrDataControlDeviceV1 into deviceMap_ in function refreshSeat().

fcitx5/src/modules/clipboard/waylandclipboard.cpp

Lines 334 to 340 in cce4115

	} else if (wlr_manager_) {
	auto *device = wlr_manager_->getDataDevice(seat.get());
	deviceMap_.emplace(
	seat.get(),
	std::make_unique<DataDevice<wayland::ZwlrDataControlDeviceV1>>(
	this, device));
	}

Later, the same callback is triggered again, calling deviceMap_.clear() and creating an ExtDataControlManagerV1.

fcitx5/src/modules/clipboard/waylandclipboard.cpp

Lines 277 to 284 in cce4115

	if (interface == wayland::ExtDataControlManagerV1::interface) {
	if (ptr != ext_manager_) {
	deviceMap_.clear();
	ext_manager_ =
	display_->getGlobal<wayland::ExtDataControlManagerV1>();
	}
	refreshSeat();
	} else if (interface ==

However, since deviceMap_ already contains ZwlrDataControlDeviceV1 objects, clearing will destroy the corresponding Wayland objects, but their event loop callbacks might not be properly unregistered. This leads to invalid object references and results in runtime errors being printed in my terminal：

I2025-10-26 22:15:16.279421 addonmanager.cpp:204] Loaded addon notificationitem
I2025-10-26 22:15:16.279903 addonmanager.cpp:204] Loaded addon notifications
❌ not a valid new object id (4278190085), message data_offer(n)**
❌ E2025-10-26 22:15:16.296288 waylandeventreader.cpp:129] Wayland connection got error: 22**
I2025-10-26 22:15:16.296355 waylandmodule.cpp:365] Connection removed 
I2025-10-26 22:15:16.298600 addonmanager.cpp:306] Unloading addon notifications

This likely indicates a use-after-free issue due to event callbacks being triggered after the associated objects have been destroyed.

I think that’s why this issue disappears when either ext-data-control or wlr-data-control support is removed.

---

fcitx5/src/modules/clipboard/waylandclipboard.cpp

Lines 273 to 274 in cce4115

	display_->requestGlobals<wayland::ExtDataControlManagerV1>();
	display_->requestGlobals<wayland::ZwlrDataControlManagerV1>();

Maybe pre-detecting the supported protocol and requesting only one of them could serve as a temporary solution.

if (ext_data_control_supported)
    display_->requestGlobals<wayland::ExtDataControlManagerV1>();
else:
    display_->requestGlobals<wayland::ZwlrDataControlManagerV1>();

[wengxt]

I would still think the bug is on compositor side that they managed both similar with the same, code thus when a client binds to both, it fails to function normally.

It's tricky to bind to only one of them since the global registry may send globals in any order.

[svicervlad]

Maybe related basecamp/omarchy#2870

[wengxt]

I think I understand for the freeze thing, the invalid object not so sure.
The signal of thread termination was sent before thread eventloop starts.