Allow dashes in music videos on the first pass

When the global volume has been set low, and a sequence of tracks are being played, there are complaints of occasional jumps to full volume.
This is down to a race condition where the volume request can arrive at OMXAudio before it has been initialised.
The fix is simple, don't send the volume change until m_CurrentAudio.started.

Currently, once demuxer has reached EOF, we send the EOF messages to audio/video players
and set their inited/started flags to false. But if started is false we ignore any
PLAYER_DISPLAYTIME coming back from players, which stops the elapsed time from updating.
This can affect the final ~8 seconds of the file.

The fix delays changing these flags until audio/video players have signalled EOS.

[fix] upnp: parentID of the root container should be -1, fixes #14279

GUIBaseContainer: remove unneeded call from CGUIBaseContainer::Render() - we are calling it in CGUIBaseContainer::Process() earlier

[fix] non-media items trumped upnp listings, fixes #14268

 for m3u8 filename - fixes stream quality selection (e.x. for airplay ^^) -
 right way this time

The cache buffer size was being initialized to the write size
of a PA stream, which is not the buffer size of the PA server.
The cache buffer size will now use the PA server buffer size.
(trac bug 13902)

Signed-off-by: Michael Cronenworth <mike@cchtml.com>

PA is calling back functions on stream destruction and may try
to read a free'd stream variable. (trac bug 13903)

Signed-off-by: Michael Cronenworth <mike@cchtml.com>

Upnp Playlist bug fixes for Frodo

dvdplayer: fix abort of ffmpeg streams

Frodo airplayconflicts

[CA/AE] - fixes

- also adapt to some new findings on revers engineering itunes airplay traffic (add sessioncounter to reverse event and fix http request formatting)
- fixes #14191

Conflicts:

	xbmc/network/AirPlayServer.cpp

In non-blocking mode, lowest-level read protocols are
supposed block only for a short amount of time to let
retry_transfer_wrapper() check for interrupts.

Also, checking the interrupt_callback in the receiving thread is
wrong, as interrupt_callback is not guaranteed to be thread-safe
and the job is already done by retry_transfer_wrapper(). The error
code was also incorrect.

Bug reported by Andrey Utkin.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

The old interface could not handle being called from another thread

Not this still doesn't resolve all issues. We are still not
setting up a correct channel map for the output

m_endtime is only initialised when we underrun. Use m_av_clock->OMXAudioBuffer() to protect access to uninitialised variable.

This can cause player to get a decode error from audio_decode component
closing stream when packet doesn't actually fit.

Could be useful for detecting underrun conditions