[openbox] Key Chain Broken by Multiple Keyboards

[Mikael Magnusson]

On 6/22/23, admin at nosuck.org <admin at nosuck.org> wrote:
> Hello, all.  I recently started using two keyboards simultaneously, one for
> my left hand and another for my right.  My hands are much more comfortable
> now, and I recommend split keyboards to you, too, if you happen to have two
> hands.
>
> However, many Openbox binding chains, such as “W-a p” below, suddenly fail
> to execute:
>
>> <keybind key="W-a">
>> 	<keybind key="x"><action name="Execute"><command>setsid ~/audio
>> stop</command></action></keybind>
>> 	<keybind key="p"><action name="Execute"><command>setsid ~/audio
>> next</command></action></keybind>
>> </keybind>
>
> The “W-a x” binding executes without issue.
>
> This seems to happen because “p” is on the right-hand keyboard, while “W-”
> (Super_L), “a”, and “x” are all on the left-hand keyboard.  Is this a known
> issue?  Can you suggest any workarounds, please?  I miss my binding chains.
>
> They were awfully convenient.
>
> Also, chains that start with, say, “W-h” instead of “W-a” (with “h” on the
> right-hand keyboard) are never even initiated.  A non-chained “W-h” binding,
>
> meanwhile, executes without issue--so the problem seems specific to chains.
>
> Thank you.

I looked at the keyboard code and I can't see any obvious reason why
chains would be different in this regard, we do two things when
entering a chain: change an internal variable to say where in the
binding tree we are currently, and bind the keys inside the chain and
unbind the keys outside. I'm not sure if there is some funny implicit
business going on due to rebinding keys while a key is held down which
gives us a passive grab... That said, Openbox shouldn't even know it's
a different keyboard.

What does 'xinput list' show for you? In particular, are both keyboard
slave devices listed under the same master device?


-- 
Mikael Magnusson