Audio calls can cause memory corruption

[tux3]

No description provided.

[dubslow]

How well do you have this tracked down? I can throw a full blown debugging run at it...

[tux3]

Barely. It's pretty hard to track down.

[msteinbeck]

The new Audiobuffer based on a non locking circular ringbuffer will be ready soon.

[tux3]

I just found out that I'm completely retarded and my QTimers and other call objects are happilly living in the wrong thread.
This could explain the corruption and/or pulseaudio crashes since this is not even supposed to work and explicitely forbidden by Qt.

[tux3]

Actually, I might just have fixed the memory corruption with 6ba8be3

I've had an audio call last 10 minutes (and still going) without crashing on Windows.

If nobody can reproduce anymore, I'll close this bug.

[msteinbeck]

This commit fixes some bugs I had some problems with for the ringbuffer, very nice!

[tux3]

It seems to be fixed. Finally.