It is my mistake that "silently" changing the default behaviour. I apologize and add some comments to show my intention.
Harold L Hunt II wrote: > It might be nice, but the Windows autorepeat caused problems with > systems running in Japanese... so they need the XKB autorepeat > functionality, since the Windows autorepeat needs to be disabled for them. The Windows autorepeat is not crucial to Japanese language handling. Disabling it was a part of the patch to improve mode key states synchronization. Sorry for confusing you, Harold. I disabled it for the two reasons below: 1) Having both Windows and XKB autorepeat breaks the state synchronization. For example, if we press Caps Lock for a while, Windows toggles the state exactly once but XKB repeatedly toggles the state since it receives many KeyPress events generated by Windows autorepeat. 2) Having two schemes for autorepeat looks not so good as Harold said. Adoption of XKB layer is just my choice. Because we are using an X server, it is natural for me to configure it using X Window's manner (i.e. xset). Please consider keymaps; we set keymaps via xmodmap, setxkbmap and/or XF86Config file -- all those are X Window's manner. Anyway, now we seem to find a better solution to get default values from Windows and enable Windows autorepeat if XKB is disabled. It looks that Alexander is implementing the features, is it right? I really appreciate him. Takuma Murakami ([EMAIL PROTECTED])
