http://mail.python.org/pipermail/python-dev/2008-June/079988.html

Python 2.6 should stick with PyString_* in its codebase.  The PyBytes_* names
in the spirit of 3.0 are available via a #define only.  See the email thread.

backwards compatibility. Add Py_Refcnt, Py_Type, Py_Size, and
PyVarObject_HEAD_INIT.

(unsigned long vs. unsigned int).

Use Py_UnpackTuple instead of PyArg_ParseTuple where possible.

convince gcc (4.0.x) the variables are never used uninitialized (and raising
a proper exception if they ever are.)

Probably should be backported.

AFMT_S32_BE, AFMT_MPEG.

attributes to oss_audio_device object: 'closed', 'name', and 'mode'.

_EXPORT_INT calls in #ifdef's, to avoid breaking the build on MkLinux
(Linux 2.0).

build (assert(gc->gc.gc_refs != 0) in visit_decref()).

Because OSSAudioError is a global, we must compensate (twice!) for
PyModule_AddObject()'s "helpful" decref of the object it adds.

8 bits is 1 byte, and that isn't about to change any time soon.  (I hope!)

  * it no longer takes ssize, which served no purpose apart from
    scolding you if you got it wrong
  * changed the order of the three remaining required arguments
    to (format, channels, rate) to match the order in which they
    must be set
  * replaced the optional argument 'emulate' with 'strict': if strict
    true, and the audio device does not accept the requested sampling
    parameters, raise OSSAudioError
  * return a tuple (format, channels, rate) reflecting the sampling
    parameters that were actually set

Change the canonical name of ossaudiodev.error to
ossaudiodev.OSSAudioError (keep an alias for backwards compatibility).

Remove 'audio_types' list and 'n_audio_types' (no longer needed now that
setparameters() no longer has an 'ssize' argument to police).

that way it applies to *only* the ioctl() call, and also happens for the
other blocking ioctls (POST, RESET).

  * sync(), because it waits for hardware buffers to flush, which
    can take several seconds depending on cirumstances (according
    to the OSS docs)
  * close(), because it does an implicit sync()

Bug spotted by Joerg Lehmann <joerg@luga.de>.

opening it in non-blocking mode.  Both Guido and David Hammerton have
reported that this fixes their problems with ossaudiodev -- hooray!

it instead of the OS-specific <linux/soundcard.h> or <machine/soundcard.h>.

Mixers devices have an ioctl-only interface, no read/write -- so the
flags passed to open() don't really matter.  Thus, drop the 'mode'
parameter to openmixer() (ie. second arg to newossmixerobject()) and
always open mixers with O_RDWR.

SOUND_DEVICE_NAMES as 'control_labels' and 'control_names'.

what it is.

OSSType to OSSAudioType.

  devices(), stereodevices(), recdevices() ->
  controls(), stereocontrols(), reccontrols()

Based on recommendation of Hannu Savolainen <hannu@opensound.com>:

  The right term to use for things like bass/treble/mic/vol/etc is
  "control".

  "Device" refers to different mixer devices (/dev/mixer0 to /dev/mixerN).

  "Channel" cannot be used because it refers to mono/stereo/multich
  channels. In fact most mixer controls have left/right channels so ...

setrecsrc() -> set_recsrc().

  * channels() -> devices()
  * stereochannels() -> stereodevices()
  * recchannels() -> recdevices()
  * getvol() -> get()
  * setvol() -> set()
This is for (slightly) more consistency with the OSS ioctl names
(READ_DEVMASK, READ_RECMASK, READ_STEREODEVS).

Also make sure the C function names correspond more closely to the
Python method names for mixer methods.

(It added nothing, and served no obvious purpose.)

Export SOUND_MIXER_NRDEVICES constant.

  * add oss_mixer_t and OSSMixerType
  * add newossmixerobject(), oss_mixer_dealloc(), ossopenmixer()
  * add _do_ioctl_1_internal() to support mixer ioctls
  * add mixer methods: oss_mixer_{close,fileno,channels,stereo_channels,
    rec_channels,getvol,setvol,getrecsrc,setrecsrc}()
  * add oss_mixer_methods list
  * add oss_mixer_getattr() (why?!)
  * export SOUND_MIXER_* constants from soundcard.h

_do_ioctl_1() so they take a file descriptor rather than an oss_t
pointer.