Issue #108

 - Parses outgoing JOINs to try and guess keys before we get the MODE reply.
   - Parsing algorithm based on Quakenet/Hybrid handling of keys.
     - Keys are swallowed from the key-list for EVERY channel that is
       to join, even if it is not needed, so you would need to use
       "JOIN #Foo,#Bar,#Baz Foo,,Baz" to join keyed channels #Foo and #Baz.
 - Key changes by mode +k and -k will be tracked.
 - Ignores attempts to set the key as "*" if we know a "better" key.
   - Side effect: If the key is actually set to "*" we can only learn it if that
     is what we join with, or it gets changed to that from no-key.

Java 8 introduces a sane API for dates; to store datettimes
we should now be using LocalDateTime.

Having two separate events means the ordering isn't guaranteed.
Handle everything for 311 numerics in one handler.

When we moved to events, we stopped passing null in place of clients.
Instead, fake clients should be used. Apparently ProcessKick wasn't
properly updated.

Fixes issue #100

Issue DMDirc/DMDirc#601

Fixes #82.

Just does ircu stuff for now.

Remove all tests that depend on TestParser - they're rubbish and
need to be completely rewritten to be proper unit tests.

Seems a bit odd for each implementation to define a method that
*always* returns a static String[] when we can just pass that
up to the base class.

Just add a utility method to get the callback proxy instead.

Think this is a bit nicer, and is one less thing to have to change
if we switch to events.

All callbacks are now done through getCallback, which offers
type safety (gasp!) and is potentially extensible in the future
(wow!).

This means that nothing can depend on callbacks existing, so
there are two functional changes:

 - the parser will no longer disconnect automatically on failed
   pings; users have to listen for the event and handle it
   themselves;

 - the parser will no longer attempt to automatically use alt
   nicknames or create new nicknames on connection; users have to
   listen for the event and implement the logic themselves

Both of these were really weird anyway, as adding a listener (even
for debugging purposes) disabled the built-in behaviour.

ReadLine now parses tags out of incoming lines (tags are not currently passed through the encoder)

IRCParser.tokeniseLine() now understands that tags may exist at the start of a line and tokenises accordingly.
(In reality, the only time it ever sees them is during testing)

Parser processLine now looks for timestamps from tags from the ReadLine, not by parsing token[0] itself.

Made IRCReader tests a lot less annoying to work with.

Change-Id: I1d81392650d8193503621b2e69fbe66cd0eb736a
Reviewed-on: http://gerrit.dmdirc.com/4058
Reviewed-by: Greg Holmes <greg@dmdirc.com>
Automatic-Compile: DMDirc Build Manager