FYI: make it work across all GUI app with fcitx5 dbus

[QiangF]

I find it could be possible to monitor if we are in an input field of any GUI app via fcitx5

(when (bound-and-true-p exwm-enable)
  (defvar exwm-inside-input-field nil)
  (defun exwm-input-field-entry-handler (&rest args)
    (setq exwm-inside-input-field t))
  (defun exwm-input-field-exit-handler ()
    (setq exwm-inside-input-field nil))

  (dbus-register-signal
   :session fcitx-service
   nil                                  ; PATH: Wildcard, listen on all object paths
   fcitx-ic-interface "CurrentIM"
   #'exwm-input-field-entry-handler)

  (dbus-register-signal
   :session fcitx-service
   nil
   fcitx-ic-interface "NotifyFocusOut"
   #'exwm-input-field-exit-handler))

See the whole package here https://github.com/QiangF/imbot

[sinic]

Oh, cool, thanks. So the package you linked pretty much supersedes my module?
I will keep this issue open for visibility, then, and maybe archive this project once I get around to try out yours.

[QiangF]

I believe the two signals are there. I think you can just use the default english input method if you don't need other input method.

[sinic]

Finally got around to try this out properly (using dbus-register-monitor instead of dbus-register-signal, as with the latter I wasn't receiving any signals directly addressed to Fcitx).

It seems to work fine, but I'm concerned about what exactly to do in the handlers: EXWM's grab and release functions take a X window ID, which is available to the callbacks in my GTK IM module (even though I still suspect a bug lurking there), but not to the callbacks in your solution. In my experiment, I simply used the ID of the currently focused window, but I don't think there's any guarantee that the signal was caused by the focused window.

I'll stick to my IM module for now, but still, thanks for bringing this to my attention.