[MiNT] Daily builds (again)

["Helmut Karlowski" <helmut.karlowski@ish.de>]

Lars Schmidbauer wrote:

> The bug is still there (infinite repeat of the last pressed key or no key
> is accepted anymore), but something is new:
> If the error appears, message pid 0 (MiNT): kbd_repeat: reset kbd_lock
> is written on screen (or console).

This is really interesting. It should not appear as kbd_lock is only set in bioskeys().

The message can only appear when bioskeys does not return (in case of
crash), or it is periodically called by the Eiffel-driver, which would
be strange.

> In case of infinite repeat, it's possible to press another key which is
> repeated then for a short time, then system freezes.
> If no key is accepted anymore, pressing any key writes the same message
> on screen again, mouse still works.

This one: pid 0 (MiNT): kbd_repeat: reset kbd_lock ? Or is it ikbd_scan?

> So it's definetly a MiNT/Eiffel problem and not XaAES related. Hope you
> can fix this...

I'd vote for Eiffel (I never thought its XaAES).

I've added a cancel-repeat in bioskeys(), if we are *really* lucky this helps.

I'll commit this tonight.

-Helmut