Note that this is a somewhat heavily modified custom version of librtmp.
I modified all the platform specific code that we were using for the
OBS1 to make it platform-independent.

I don't really like the code in this library, but it works well enough,
so I can't really fault anyone for it.  It's just very..  unclean.  Even
for a C library, quite unclean.  Some parts are also a little less safe
than I'd prefer as well.

The serializer code is meant to be used as a means of reading/writing
data from any arbitrary type of input/output.

The array output serializer makes it so we can stream data to a dynamic
array on the fly.

Don't clear any data if it's a dummy texture.

 - Add dummy GL texture support to allow libobs texture references to be
   created for GL without

 - Add a texture_getobj function to allow the retrieval of the
   context-specific object, such as the D3D texture pointer, or the
   OpenGL texture object handle.

 - Also cleaned up the export stuff.  I realized it was all totally
   superfluous.  Kind of a dumb moment, but nice to clean it up
   regardless.

 - Make it so that encoders can be assigned to outputs.  If an encoder
   is destroyed, it will automatically remove itself from that output.
   I specifically didn't want to do reference counting because it leaves
   too much potential for unchecked references and it just felt like it
   would be more trouble than it's worth.

 - Add a 'flags' value to the output definition structure.  This lets
   the output specify if it uses video/audio, and whether the output is
   meant to be used with OBS encoders or not.

 - Remove boilerplate code for outputs.  This makes it easier to program
   outputs.  The boilerplate code involved before was mostly just
   involving connecting to the audio/video data streams directly in each
   output plugin.

   Instead of doing that, simply add plugin callback functions for
   receiving video/audio (either encoded or non-encoded, whichever it's
   set to use), and then call obs_output_begin_data_capture and
   obs_output_end_data_capture to automatically handle setting up
   connections to raw or encoded video/audio streams for the plugin.

 - Remove 'active' function from output callbacks, as it's no longer
   really needed now that the libobs output context automatically knows
   when the output is active or not.

 - Make it so that an encoder cannot be destroyed until all data
   connections to the encoder have been removed.

 - Change the 'start' and 'stop' functions in the encoder interface to
   just an 'initialize' callback, which initializes the encoder.

 - Make it so that the encoder must be initialized first before the data
   stream can be started.  The reason why initialization was separated
   from starting the encoder stream was because we need to be able to
   check that the settings used with the encoder *can* be used first.

   This problem was especially annoying if you had both video/audio
   encoding.  Before, you'd have to check the return value from
   obs_encoder_start, and if that second encoder fails, then you
   basically had to stop the first encoder again, making for
   unnecessary boilerplate code whenever starting up two encoders.

This fixes some problems with multiple swap chains (e.g. flickering
between the render views)

Also add support for capturing displays that are not the main display as
preparation for moving the code from test-plugins into a proper plugin
in the future

I neglected to make it initialize the actual control values with the
current settings

 - Add a properties window for sources so that you can now actually edit
   the settings for sources.  Also, display the source by itself in the
   window (Note: not working on mac, and possibly not working on linux).

   When changing the settings for a source, it will call
   obs_source_update on that source when you have modified any values
   automatically.

 - Add a properties 'widget', eventually I want to turn this in to a
   regular nice properties view like you'd see in the designer, but
   right now it just uses a form layout in a QScrollArea with regular
   controls to display the properties.  It's clunky but works for the
   time being.

 - Make it so that swap chains and the main graphics subsystem will
   automatically use at least one backbuffer if none was specified

 - Fix bug where displays weren't added to the main display array

 - Make it so that you can get the properties of a source via the actual
   pointer of a source/encoder/output in addition to being able to look
   up properties via identifier.

 - When registering source types, check for required functions (wasn't
   doing it before).  getheight/getwidth should not be optional if it's
   a video source as well.

 - Add an RAII OBSObj wrapper to obs.hpp for non-reference-counted
   libobs pointers

 - Add an RAII OBSSignal wrapper to obs.hpp for libobs signals to
   automatically disconnect them on destruction

 - Move the "scale and center" calculation in window-basic-main.cpp to
   its own function and in its own source file

 - Add an 'update' callback to WASAPI audio sources

The ones that were being used returned the previous value rather than
the new value

Microsoft's garbage compiler just doesn't even..  read the names of
enums.  It sees an enum and goes "durr, that's an int" without even
properly evaluating it.  Just total garbage, as per usual.

If integers are used, then store as integers.  If doubles are used, then
store as doubles.  This way precision issues are prevented.

Removed some warnings associated with it, was the wrong variable type.

You have to use qt5_use_modules otherwise it just won't always work.  I
don't know how it was working before with cmake, but apparently this
does the trick to ensure that all the necessary libraries and includes
are added.  ..Might require cmake 3.8.9 though.  ..I do hope that
doesn't end up being a problem.

Also, rename atomic functions to be consistent with the rest of the
platform/threading functions, and move atomic functions to threading*
files rather than platform* files

Automatically query defaults for sources/outputs/encoders and set them
before calling create

* Not entirely sure if this will fix it, johnnypatterson claimed it
  would, but this checked out on my mac so at the very least I don't
  think it should hurt anything

GCC warned of this, but strangely not clang.

 - Implement OBS encoder interface.  It was previously incomplete, but
   now is reaching some level of completion, though probably should
   still be considered preliminary.

   I had originally implemented it so that encoders only have a 'reset'
   function to reset their parameters, but I felt that having both a
   'start' and 'stop' function would be useful.

   Encoders are now assigned to a specific video/audio media output each
   rather than implicitely assigned to the main obs video/audio
   contexts.  This allows separate encoder contexts that aren't
   necessarily assigned to the main video/audio context (which is useful
   for things such as recording specific sources).  Will probably have
   to do this for regular obs outputs as well.

   When creating an encoder, you must now explicitely state whether that
   encoder is an audio or video encoder.

   Audio and video can optionally be automatically converted depending
   on what the encoder specifies.

   When something 'attaches' to an encoder, the first attachment starts
   the encoder, and the encoder automatically attaches to the media
   output context associated with it.  Subsequent attachments won't have
   the same effect, they will just start receiving the same encoder data
   when the next keyframe plays (along with SEI if any).  When detaching
   from the encoder, the last detachment will fully stop the encoder and
   detach the encoder from the media output context associated with the
   encoder.

   SEI must actually be exported separately; because new encoder
   attachments may not always be at the beginning of the stream, the
   first keyframe they get must have that SEI data in it.  If the
   encoder has SEI data, it needs only add one small function to simply
   query that SEI data, and then that data will be handled automatically
   by libobs for all subsequent encoder attachments.

 - Implement x264 encoder plugin, move x264 files to separate plugin to
   separate necessary dependencies.

 - Change video/audio frame output structures to not use const
   qualifiers to prevent issues with non-const function usage elsewhere.
   This was an issue when writing the x264 encoder, as the x264 encoder
   expects non-const frame data.

   Change stagesurf_map to return a non-const data type to prevent this
   as well.

 - Change full range parameter of video scaler to be an enum rather than
   boolean

device_id would end up being null, which would mess with the functions.

if (data->output->flags & AVFMT_RAWPICTURE)

If this was true, the 'ret' variable would be used without
initialization.

linux input plugins for desktop and audio capture

Otherwise deadlock

Make sure it locks the write mutex before freeing the packets, and put
the detach code in the main thread loop rather than off in a separate
function for clarity

The event functions got renamed in obs.

We use the default format to avoid any resampling by pulse audio,
since obs will probably resample anyways if necessary.

pulse audio provides it's own convenience function to
iterate the mainloop, no need to use a custom one.
pulseaudio provides timing information for the stream, we use
that instead of system time.

Instead of copying the data from pulse to a local buffer and
then push it to obs, the pointer from pulse is now used and
pa_stream_drop() called afterwards.