When seeking, the seek does not immediately get applied to the data
source. Instead it is delayed. This results in a situation where the
sound can be seeked, but a small window will exist where querying the
seek point will not be consistent with the previous seek request.

Public issue https://github.com/mackron/miniaudio/issues/750

Public issue https://github.com/mackron/miniaudio/issues/732

https://github.com/mackron/miniaudio/issues/733

If the state miniaudio side of the device does not match the actual
state of the backend side of it, such as when the device is stopped but
the backend doesn't post a notification, attempting to stop the device
might result in a deadlock.

This is a just a quick workaround hack for the moment while a more
robust solution is figured out.

https://github.com/mackron/miniaudio/issues/717

resolves: https://github.com/mackron/miniaudio/issues/723Signed-off-by: Oldes <oldes.huhuman@gmail.com>

The SSE2 code paths for mono expansion introduced in Version 0.11.15
mixed up the parameters of `_mm_shuffle_ps()`, which in turn caused
adjacent PCM frames to be swapped in the channel-expanded output.

This is just a small project to experiment with a few API ideas. This
is not a replacement for miniaudio or anything so don't panic.

This applies only to duplex devices.

This unifies, as much as possible, the ScriptProcessorNode path with
the AudioWorklets path to avoid some code duplication, and to also make
the two paths more similar to each other to ease in maintenance.

  * Duplex mode now only creates a single AudioContext and AudioWorklet
  * Devices should now automatically start in response to a gesture

Public issue https://github.com/mackron/miniaudio/issues/597