Refactoring the private API

[danielrbrowne]

This PR:

Aims to improve the readability of the private API of the SDK, by simplifying it and bringing it more inline with Swift best practices. This introduces no major breaking changes to the public API, but can be seen as a precursor for the same kind of improvements being made to the public API for a future 10.0.0 release.

• Renames some types in the private API (I.e. By removing the redundant Pusher prefix and/or using more descriptive nouns)
• Consolidates the private cryptography functionality into the Crypto type
• Moves some protocol definitions to their own source files (for consistency)
• Ensures the Constants.API enum is not exposed publicly
  • This was a mistake when introduced in the 9.0.0 release, and it was not meant to be part of the public API. Technically this is a breaking change, however there is no end-user utility to exposing this enum so I consider this appropriate for a 9.x release.
  • A non-breaking API change adding a defaultHost to PusherHost was made to allow for this change.
• Moves the private WebSocket URL constructor method to an extension method for the URL Foundation type
  • Previously, this was implemented as a global private method
• Removes redundant public and open access level specifiers in Tests
  • There is no need to expose any functionality outside of the Tests target, as it is self-contained

[codecov-io]

Codecov Report

Merging #339 (f6eb619) into master (18570cb) will decrease coverage by 0.03%.
The diff coverage is 69.03%.

@@            Coverage Diff             @@
##           master     #339      +/-   ##
==========================================
- Coverage   89.46%   89.42%   -0.04%     
==========================================
  Files          50       50              
  Lines        3550     3538      -12     
==========================================
- Hits         3176     3164      -12     
  Misses        374      374              

Impacted Files	Coverage Δ
Sources/Helpers/Logger.swift	53.84% <ø> (ø)
Sources/Models/ChannelsProtocolCloseCode.swift	83.33% <ø> (ø)
Sources/Models/PusherClientOptions.swift	100.00% <ø> (ø)
Sources/Models/PusherPresenceChannel.swift	82.81% <0.00%> (ø)
Sources/ObjC/PusherClientOptions+ObjectiveC.swift	0.00% <ø> (ø)
Sources/ObjC/PusherHost+ObjectiveC.swift	0.00% <0.00%> (ø)
Sources/Helpers/EventParser.swift	73.91% <33.33%> (ø)
Sources/Models/PusherChannel.swift	92.75% <33.33%> (ø)
Sources/Services/PusherConnection.swift	79.22% <45.28%> (ø)
...xtensions/PusherConnection+WebsocketDelegate.swift	63.28% <50.00%> (ø)
... and 21 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 18570cb...f6eb619. Read the comment docs.

[TomKemp]

Looks good 👍 Left some comments that don't require changes.