Overhaul the MySQL lexer (#1527)

* Overhaul the MySQL lexer

Fixes #975, #1063, #1453

Changes include:

Documentation
-------------

* Note in the lexer docstring that Oracle MySQL is the target syntax.
  MariaDB syntax is not a target (though there is significant overlap).

Unit tests
----------

* Add 140 unit tests for MySQL.

Literals
--------

* Hexadecimal/binary/date/time/timestamp literals are supported.
* Integer mantissas are supported for scientific notation.
* In-string escapes are now tokenized properly.
* Support the "unknown" constant.

Comments
--------

* Optimizer hints are now supported, and keywords are
  recognized and tokenized as preprocessor instructions.
* Remove nested multi-line comment support, which is no
  longer supported in MySQL.

Variables
---------

* Support the '@' prefix for variable names.
* Lift restrictions on characters in unquoted variable names.
  (MySQL does not impose a restriction on lead characters.)
* Support single/double/backtick-quoted variable names, including escapes.
* Support the '@@' prefix for system variable names.
* Support '?' as a variable so people can demonstrate prepared statements.

Keywords
--------

* Keyword / data type / function are now in a separate, auto-updating file.
* Support 25 additional data types (including spatial and JSON types).
* Support 460 additional MySQL keywords.
* Support 372 MySQL functions.
  Explicit function support resolves a bug that causes non-function
  items to be treated as functions simply because they have a trailing
  opening parenthesis.
* Support exceptions for the 'SET' keyword, which is both a datatype and
  a keyword depending on context.

Schema object names
-------------------

* Support Unicode in MySQL schema object names.
* Support parsing of backtick-quoted schema object name escapes.
  (Escapes do not produce a distinct token type at this time.)

Operators
---------

* Remove non-operator characters from the list of operators.
* Remove non-punctuation characters from the list of punctuation.

* Cleanup items based on feedback

* Remove an unnecessary optional newline lookahead for single-line comments

1 files changed, 1 insertions, 1 deletions

diff --git a/AUTHORS b/AUTHORS
index 9b0ad009..5058c612 100644
--- a/AUTHORS
+++ b/AUTHORS
@@ -138,7 +138,7 @@ Other contributors, listed alphabetically, are:
 * Stephen McKamey -- Duel/JBST lexer
 * Brian McKenna -- F# lexer
 * Charles McLaughlin -- Puppet lexer
-* Kurt McKee -- Tera Term macro lexer, PostgreSQL updates
+* Kurt McKee -- Tera Term macro lexer, PostgreSQL updates, MySQL overhaul
 * Lukas Meuser -- BBCode formatter, Lua lexer
 * Cat Miller -- Pig lexer
 * Paul Miller -- LiveScript lexer