Changes

Imported audio system from Legacy. Includes the TweetNacl lib for
encrypting audio, the Opus wrapper, and the opus natives. Added .ignore
information to ignore the test src folder.
 (commit: 8a104f3)

Removed AudioManager#moveAudioConnection, moved logic into
#openAudioConnection.
 (commit: 310d574)

Created ConnectionListener system. All connection related statuses are
now represented by ConnectionStatus. Changed the audio keepalive
(Websocket and UDP) to use a single runnable instead of 2 different
threads. Audio KeepAlive now uses a ScheduledThreadPoolExecutor which
should hugely reduce the amount of threads used by audio
 (commit: 8bd97f1)

Added a ratelimit queue to the Main Websocket to protect it from
websocket message send spam.
 (commit: 5f2bd43)

First pass on audio reconnect. Supports mWS resume and individual aWS
reconnect. Need to implement mWS reconnect/invalidate support.
 (commit: a97afac)

Added support to preserve audio connections and AudioManager settings
during full JDA reconnect/invalidate.
 (commit: 0bc967b)

Separated the audio packet sending logic from the packet encoding and
encrypting. Allows for custom senders through IAudioSendFactory and
IAudioSendSystem. Threads are now killed when audio handlers are set to
null. Fixed a CPU leak related to ScheduledThreadPoolExecutor in
AudioWebSocket Added proper naming to many threads.
 (commit: b6de2f9)

Added handling for AudioWebSocket failing to connect. Fixes #110
 (commit: d0e072b)