Correction: It is not the “quiet” and “splash” parameters that are the (sole) cause of the display rotation issue, but rather the DBUS events.
In my last article, I wrote that the “quiet” and ‘splash’ parameters were the problem. I need to correct myself here. It wasn’t these parameters that caused the issue where display rotation occasionally failed to work. No, the DBus event message “tabletMode changed” was either not sent or did not arrive. In any case, relying on these events can be problematic. As a workaround, the option in the KDE default configuration that limits auto-rotation to only when “tablet mode” is active was disabled. Since this mechanism listens for this event, that option was disabled. Now, auto-rotation always works regardless of tablet mode. Additionally, in the new version, I have provided a listener written in Python that uses the evdev library and checks whether the hardware keyboard has been disabled in order to provide the tablet launcher.