flatten Keyword into TokenType

flatten Keyword into TokenType

This change makes Keyword extend TokenType so that
1) the TokenType of a KeywordToken is the keyword itself
2) TokenType.KEYWORD can be removed

This simplifies the issue of determining the precedence
when handling "as" and "is" which we consider to be keywords
and the spec considers to be operators.
It also addresses the situation where analyzer considers
these two keywords to have the type TokenType.KEYWORD
while fasta considers them to have their own unique
token type.

* change Keyword to extend TokenType
* make TokenType.AS be the same object as Keyword.AS
* make TokenType.IS be the same object as Keyword.IS
* rename Keyword.syntax --> Keyword.lexeme
* replace token.type == TokenType.KEYWORD with token.type.isKeyword
* remove TokenType.KEYWORD
* remove unnecessary Keyword.info method

This has been split into two separate CLs
so that existing references can be updated
prior to removal of the deprecated API.
See

* https://codereview.chromium.org/2842213003/
* https://codereview.chromium.org/2848493003/

[danrubel, 2017-04-25 17:26:07]

danrubel@google.com changed reviewers:
+ ahe@google.com, paulberry@google.com

[danrubel, 2017-04-27 21:51:58]

Description was changed from

==========
flatten Keyword into TokenType

This change makes Keyword extend TokenType so that
1) the TokenType of a KeywordToken is the keyword itself
2) TokenType.KEYWORD can be removed

This simplifies the issue of determining the precedence
when handling "as" and "is" which we consider to be keywords
and the spec considers to be operators.
It also addresses the situation where analyzer considers
these two keywords to have the type TokenType.KEYWORD
while fasta considers them to have their own unique
token type.

* change Keyword to extend TokenType
* make TokenType.AS be the same object as Keyword.AS
* make TokenType.IS be the same object as Keyword.IS
* rename Keyword.syntax --> Keyword.lexeme
* replace token.type == TokenType.KEYWORD with token.type.isKeyword
* remove TokenType.KEYWORD
* remove unnecessary Keyword.info method
==========

to

==========
flatten Keyword into TokenType

This change makes Keyword extend TokenType so that
1) the TokenType of a KeywordToken is the keyword itself
2) TokenType.KEYWORD can be removed

This simplifies the issue of determining the precedence
when handling "as" and "is" which we consider to be keywords
and the spec considers to be operators.
It also addresses the situation where analyzer considers
these two keywords to have the type TokenType.KEYWORD
while fasta considers them to have their own unique
token type.

* change Keyword to extend TokenType
* make TokenType.AS be the same object as Keyword.AS
* make TokenType.IS be the same object as Keyword.IS
* rename Keyword.syntax --> Keyword.lexeme
* replace token.type == TokenType.KEYWORD with token.type.isKeyword
* remove TokenType.KEYWORD
* remove unnecessary Keyword.info method

This has been split into two separate CLs
so that existing references can be updated
prior to removal of the deprecated API.
See

* https://codereview.chromium.org/2842213003/
* https://codereview.chromium.org/2848493003/
==========