Drop tokens.

[DiscreteTom]

Currently we store token in ASTNode.token and in Lexer.errors. But for most cases, these tokens are not used, which will waste a lot of memory.

For the ASTNode, a better way is to define some transformer to transform a token into a terminator ASTNode then we can drop the token.

For Lexer.errors, we can define a callback to receive those error tokens.

[DiscreteTom]

But with re-lex, lexer callbacks might not accurate.

[DiscreteTom]

// add token transformer
parserBuilder.lexer(lexer, {
	t1: token => nodeData
})

[DiscreteTom]

For lexer.error, maybe we can batch process them in parser.commit? This is to prevent the inconsistency caused by re-lex.