I believe Adam Lindsay said this around Tue, 12 Nov 2002:
You might like to try experimenting with a current CVS snapshot and 0.36-test5. I'm looking at it now, too.
Okay, I gave a look. I grabbed the above tarball and took the file portaudio/pa_mac_core/pa_mac_core.c (dated 2002/10/18), and put it in the place of the existing pa_mac_core.c in pd-0.36-test5. I added the following to LDFLAGS in the main makefile: -framework AudioUnit -framework AudioToolbox (I'm they're needed to support PAv18-patch's new "sample rate conversion, mono-to-stereo conversion, and buffer size adaptation" functionality.)
I deleted the existing pa_mac_core.o, and did a new make. It built, and it runs on my machine. It might work for you *if* the emi2|6 is falling prey to the portaudio non-functionality I described.
Was it Pat Pagano who was also asking about getting full-duplex working? This could be it...
Here's something: the new version of pa_mac_core loses the very nice latency measurement (like "PortAudio on OS X - Latency = 2048 frames, 46 msec"). Simply copy-pasting the bit of code (from line 990 in the old pa_mac_core to line 1318 in the new) didn't give an accurate measurement. So the new code ain't perfect, but it's hopefully a general improvement.
-- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Adam Lindsay +44(0)1524 594 537 atl@comp.lancs.ac.uk http://www.comp.lancs.ac.uk/computing/users/atl/ -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=