Re: [yoshimi-user] MIDI control

  • From: Kristian Amlie <kristian@xxxxxxxxxx>
  • To: yoshimi-user@xxxxxxxxxxxxxxxxxxxxx
  • Date: Sat, 05 Apr 2014 17:50:11 +0200

On 02/04/14 21:36, Will Godfrey wrote:

Right then, back to the serious stuff.

There is a new branch called new_midi for this experimental work.

I've started by creating a new function in jackEngine called convergedMidi

The idea is that this makes a nice detached point where (eventually) two
ringbuffers can be fed, each with only what it needs. This works perfectly for
jack midi, but current alsa midi segfaults. A few printf statements prove that
it gets there OK and with the right data, but it's when it actually makes the
ringbuffer call when it dies.

The reason is quite obvious after a little thought. Without the jack engine
started there *is* no ringbuffer! My first thought was to find a way to set up
the ringbuffer independently of which midi source was set, but then I had
another thought.

What if someone wanted to compile and run yoshimi in a purely alsa
environment? There would be no jack headers to link to. Is this a realistic
situation? If so, we need to use a different ringbuffer. I know the boost
library has one but whether it is as efficient and simple as the jack one I
wouldn't have a clue :(

If noone objects on the list I would take that as the guide. Personally
I have no problem requiring jack to compile Yoshimi; I couldn't imagine
using Yoshimi without it.

--
Kristian



Other related posts: