[Portaudio] Solaris support?

[Ross Bencina]

"Dominic Mazzoni" <dominic@minorninth.com> wrote:
> It seems to me like it might make sense to have a pa_unix directory
> which contains all of the code common to implementations which
> use read() and write() on a file descriptor, i.e. the traditional
> unix audio I/O model.  Then pa_unix_oss and pa_solaris would
> implement the device queries differently, but share the same code
> for the main loop

Hi Dominic,

Sounds like a good idea. I'm doing the same thing on windows with the
multi-api support. There's a common .c file which contains code for memory
allocation, timing (for cpu load measurement) and Pa_Sleep().

Another thing I've done is broken the internal port audio stream structure
up into a number of smaller structs (fptr interface, cpuload data, etc)
which each implementation will aggregate allong with their implementation
specific data. This design will easily accomodate a common PaUnixStreamData
struct and supporting functions shared by oss and solaris implementations.

I'll post more about the architecture when I get things into a CVSable
state.

Best wishes,

Ross.