Re: [AD] Further MIDI fix

[Eric Botcazou <ebotcazou@xxxxxxxxxx>]

> However, there are still some residual issues if a controller was in use
> (eg. sustain) since these are not cancelled at the end of one song and can
> affect the next one.
>
> The proposed fix is to the reset_controllers method in midi.c:
>
>    if (midi_driver->raw_midi) {
>       midi_driver->raw_midi(0xB0+channel);
>       midi_driver->raw_midi(10);
>       midi_driver->raw_midi(midi_channel[channel].pan);
> +      /* All Controllers Off */
> +      midi_driver->raw_midi(0xB0+channel);
> +      midi_driver->raw_midi(121);
> +      midi_driver->raw_midi(0);
>    }
>
> Any opinions?

Sounds reasonable.  Do you have a specific testcase for which it makes a 
difference?

-- 
Eric Botcazou