* Remove unused variable

This variable is unused since the first commit to this file
in the commit history, so it's probably safe to remove it.

* Lex identifier as label only if it's at line start

* Stop matching identifiers that begin with a digit

This is so we don't match numbers as labels with the new labels rule.

* Add label tests

* Fix existing tests

* Java: fix lexing of 'record' soft keyword

Refactor the Java lexer to treat `record` as a soft keyword.

Previously, the lexer assumed record is a reserved word, even though
it is a soft keyword which can be used as a variable name.

This refactor lexes record as a keyword only if it appears at the
beggining of the line, with some potential other keywords like public
and private preceding it.

* Remove repetition in capture group

* Update test output

Detect function with return types of more than a single word length
(like `unsigned int` or `long long`).

* Add square, angle and single underscore to punct; simplify string regex

* Add dollar as operator in LOAD DDL; add pipe as op for resused edge names

* Add to test file to test recent lexer fixes

* Added PR# to new test cases