Right. Well, it compiles and runs just fine on my machine when I replace portaudio/pa_mac_core/pa_mac_core.c with the new one and recompile. I'm just using my normal internal sound driver, though. Best I can say is "it compiles, and it doesn't seem to break anything obvious."
that's good news.
These are the changes since the last pa_mac_core.c synched with Pd:
<snip>
the thing that sort of concerns me is that portaudio seems to need patching to support every new device release for OSX. this is not good, and pretty much defeats a lot of what CoreAudio is meant to do. plus the AudioConverter routine adds another buffer to the I/O chain which might mean at least another 64 samples of latency on both ends. pd already has a relatively high latency on OSX right now, so this isn't going to help out.
What should I do?
- Check in the new pa_mac_core.c and let curious people compile the
devel_0_36 branch? 2) Point people to the portaudio tarball and let them experiment with the sources themselves? http://portaudio.com/archives/pa_snapshot_v18.tar.gz 3) Seed a new pd/bin/pd binary to people who want to test the updated portaudio stuff with a range of hardware?
CVS is a good place for testing and experimentation. go ahead and commit it and i will compile a version and give it to some people using devices like the 828 and report back.
also, i would be interested in working on a direct CoreAudio implementation for pd. i'm tied up on a project until the second week in April, so i won't even take a look until then. anyone else interested in doing this?
cgc