- Add WASAPI audio capture for windows, input and output

 - Check for null pointer in os_dlopen

 - Add exception-safe 'WinHandle' and 'CoTaskMemPtr' helper classes that
   will automatically call CloseHandle on handles and call CoTaskMemFree
   on certain types of memory returned from windows functions

 - Changed the wide <-> MBS/UTF8 conversion functions so that you use
   buffers (like these functions are *supposed* to behave), and changed
   the ones that allocate to a different naming scheme to be safe

Somehow this code didn't get included with the last commit.

Also, don't have it repeat trying to reconnect if no devices are found

 - Split input and output audio captures so that they're different
   sources.  This allows easier handling and enumeration of audio
   devices without having to do some sort of string processing.

   This way the user interface code can handle this a bit more easily,
   and so that it doesn't confuse users either.  This should be done for
   all audio capture sources for all operating systems.  You don't have
   to duplicate any code, you just need to create input/output wrapper
   functions to designate the audio as input or output before creation.

 - Make it detect soundflower and wavtap devices as mac "output" devices
   (even though they're actually input) for the mac output capture, and
   make it so that users can select a default output capture and
   automatically use soundflower or wavtap.

   I'm not entirely happy about having to do this, but because mac is
   designed this way, this is really the only way to handle it that
   makes it easier for users and UI code to deal with.

   Note that soundflower and wavtap are still also designated as input
   devices, so will still show up in input device enumeration.

 - Remove pragma messages because they were kind polluting the other
   compiler messages and just getting in the way.  In the future we can
   just do a grep for TODO to find them.

 - Redo list property again, this time using a safer internal array,
   rather than requiring sketchy array inputs.  Having functions handle
   everything behind the scenes is much safer.

 - Remove the reference counter debug log code, as it was included
   unintentionally in a commit.

Categories added an unnecessary complexity to making properties, and
would very likely almost never be used in most cases, and were more of a
display feature.  The main issue is that it made property data more
complex to work with, and I just didn't feel comfortable with that.

Also, added a function to allow you to retrieve a porperty just by its
name.

As usual, microsoft treats all enums as integers, rather than actually
even checking the enum type.  Worthless compiler.  Just complete
garbage.

When a source/output/etc has a property of a 'list' type, there was no
way to get the names associated with its values.  That, and it only
supported lists of either text, or enums (0..[value] only).

Now, you can associate translated names with those values, and use
integer, float, or string values.  Put it all in to one function as well
to simplify its usage.

I plan on using this to help get enumerations from devices/etc for
certain types of sources.  For example, if I get the properties of an
audio source, I'd like to have a list of available devices with it as
well.

- Signals and dynamic callbacks now require declarations to be made
  before being used.  What this does is allows us to get information
  about the functions dynamically which can be relayed to the user and
  plugins for future extended usage (this should have big implications
  later for scripting in particular, hopefully).

- Reduced the number of types calldata uses from "everything I could
  think of" to simply integer, float, bool, pointer/object, string.
  Integer data is now stored as long long.  Floats are now stored as
  doubles (check em).

- Use a more consistent naming scheme for lexer error/warning macros.

- Fixed a rather nasty bug where switching to an existing scene would
  cause it to increment sourceSceneRefs, which would mean that it would
  never end up never properly removing the source when the user clicks
  removed (stayed in limbo, obs_source_remove never got called)

If the default device changes, set the reconnect interval to 200
milliseconds so it pretty much immediately tries to reinitialize the
audio with the newly selected default device.  Otherwise, use 2000
millisecond intervals, and assume disconnection.

Also, reduced FFmpeg logging to just regular FFmpeg information rather
than everything FFmpeg logs.

Happened because I compiled on windows and it only compiled
windows-specific code.

See, it can sometimes be a bit confusing.  These functions should
definitely not fail under normal circumstances, and these errors may
affect the user and/or application in some way.

LOG_ERROR should be used in places where though recoverable (or at least
something that can be handled safely), was unexpected, and may affect
the user/application.

LOG_WARNING should be used in places where it's not entirely unexpected,
is recoverable, and doesn't really affect the user/application.

I can't believe I wasn't doing this.  This is why file output was
getting corrupted.  Audio and video send in data from separate threads.
I should be embarassed for not having considered that.

Key lesson:  Increase threading paranoia levels.  Apparently my
threading paranoid levels are lackluster.

I'm not entirely sure what's going on with my FFmpeg code, but it's
definitely not generating the proper footers for MP4 files, despite the
fact that the footer function succeeds.  Going to use AVIs for the time
being still.

Also, make it use 'veryfast' preset.  Still testing this, might have to
revise this later.

The mutex must be unlocked regardless of what the return value is or it
will be locked forever if the return value isn't 0.

After a mac just boots up, it often takes about 700 milliseconds for
audio devices to work on first use, so it would often have issues with
the 700ms audio buffering time, and audio data would get cut off.  Just
increasing the buffering a little bit fixes the issue.

Forgot to add a '!'.

Actually, if initializing failed at all, it would never properly
uninitialize because the 'initialized' variable was never set until the
very end.  Instead, set the "initialized" flag from the beginning to
ensure initialization.

If coreaudio_start failed, it wouldn't free the audio data properly.
Fixed that issue.

These address structures are very confusing and I wish apple designed
better system APIs.

Last commit was on mac so I was unable to incude the VS project files
additions for threading-windows.c.

Was pointing to a bad filename for the new threading stuff

 - Add CoreAudio device input capture for mac audio capturing.  The code
   should cover just about everything for capturing mac input device
   audio.  Because of the way mac audio is designed, users may have no
   choice but to obtain the open source soundflower software to capture
   their mac's desktop audio.  It may be necessary for us to distribute
   it with the program as well.

 - Hide event backend

 - Use win32 events for windows

 - Allow timed waits for events

 - Fix a few warnings

FFmpeg test output wasn't make any attempt to sync data before.  Should
be much more accurate now.

Also, added a restart message to audio settings if base audio settings
are changed.

the signals for scenes could have potentially conflicted with default
source signals.  "remove" should be used for source removal, for
example.  Changed the scene signals to "item-add" and "item-remove" for
its items.

Resetting audio while libobs is active is a real pain.  I think I'm just
going to do audio resetting later, or maybe just require restart
regardless just because having to shut down audio streams/lines while
there's sources currently active requires recreating all the audio
lines for each audio source.  Very painful.

Video fortunately is no big deal, so at least there's that.

Split off activate to activate and show callbacks, and split off
deactivate to deactivate and hide callbacks.  Sources didn't previously
have a means to know whether it was actually being displayed in the main
view or just happened to be visible somewhere.  Now, for things like
transition sources, they have a means of knowing when they have actually
been "activated" so they can initiate their sequence.

A source is now only considered "active" when it's being displayed by
the main view.  When a source is shown in the main view, the activate
callback/signal is triggered.  When it's no longer being displayed by
the main view, deactivate callback/signal is triggered.

When a source is just generally visible to see by any view, the show
callback/signal is triggered.  If it's no longer visible by any views,
then the hide callback/signal is triggered.

Presentation volume will now only be active when a source is active in
the main view rather than also in auxilary views.

Also fix a potential bug where parents wouldn't properly increment or
decrement all the activation references of a child source when a child
was added or removed.

Implement a few audio options in to the user interface as well as a few
inline audio functions in audio-io.h.

Make it so ffmpeg plugin automatically converts to the desired format.

Use regular interleaved float internally for audio instead of planar
float.