Fixed Incompatabilities with Minigrace

[IsaacOscar]

I've made several changes to increase the ammount of code that works in both kernan and minigrace, specifically:

• I have made the parser more forgiving:
  • Blocks can now have patterns that arn't paranthesised, when their is no abiguity, e.g. {Foo.Bar -> hello } is shorthand for { _ : Foo.Bar -> hello }
  • \e in string literals is now the ANSI escape character.
  • semicolons can now be used as statement seperators (and not just terminators) e.g "foo; bar".
    statements can end via means other than a newline or semicolon e.g. in "class{ foo }", "foo" is ended by the "}".
  • new lines are now allowed before expressions (e.g. foo(\nbar), where \n is a newline character).
  • binary operators may now have spaces arround them
  • I deleted checks that lines are correctly indented (because it was rejecting lots of code minigrace accepts), i.e. it is now more permisive (as opposed to more restrictive) then minigrace.
  • I have added "matches" methods to patterns (including types, strings and numbers) that returns a Boolean (like minigrace), they still have Kernan's match method though.
• Objects now have a default "hash" method that uses object-identity.
• Strings now have *, replace(_) with(_), and split(_) methods.
• Booleans now support || and && when the RHS is a block.
• Numbers niow have a truncated method.
• I have significantly modifies sequences, for simplicity, kernan now only provides one kind (a "Sequence") that is not lazy. I also implemented ==, indexOf, do(_) separatedBy(_), map(_), ++(_), size, sizeIfUnknown, first, last, indices, hash, at(_) and contains(_). Note that because I'm lazy, I didn't make map lazy (as it is in minigrace).
• I fixed a bug where matching against type-intersections behaved the same as variants.

Incompatabilities I have not fixed (that can't simply be fixed by using a custom dialect)

• Many more missing methods....
• support for ...
• Support for is abstract and is required
• Using outer as an expression (and not just a receiver)
• Using outer.outer.
• Having closing square brackets right next to a closing curley, e.g. {[ "hello" ]}, one must write {[ "hello" ] },
• It is not an error if a match-case has multiple cases match
• The confidential isMe and myIdentityHash methods.

[IsaacOscar]

I've now made it so that "\n" is the ASCII line feed character (as it is in every other language...) and not the Unicode line-seperator character.