Support dynamic delimiter in mysql parser

[huydo862003]

Summary

• Support dynamic DELIMITER in mysql parser.

DELIMITER @@

CREATE TABLE Users (
  id int
)@@

Issue

#566

Lasting Changes (Technical)

(please list down: code changes/things that have wide-effect; new libraries/functions added that can be used by others; examples below)

• (Added class EmailValidator to validate email address' validity)
• (Added Tenant#is_trial? check)

Checklist

Please check directly on the box once each of these are done

• Documentation (if necessary)
• Tests (integration test/unit test)
• Integration Tests Passed
• Code Review

[NQPhuc]

This work. Another way is you could define a lexer base class :and move the code here to there. Please check the. For more details, check the PostgresSQL Lexer.g4 and .js for an example

[huydo862003]

I have extracted out a base class nhe:

dbml/packages/dbml-core/src/parse/ANTLR/parsers/mysql/MySqlLexer.g4

Lines 39 to 43 in 09d99e5

	options {
	// MUST READ: This is used to detect dynamic DELIMITERs, the original grammar does not have this
	superClass = MySqlLexerBase;
	caseInsensitive = true;
	}

dbml/packages/dbml-core/src/parse/ANTLR/parsers/mysql/MySqlLexerBase.js

Line 3 in 09d99e5

export default class MySqlLexerBase extends antlr4.Lexer {

[NQPhuc]

Can you add more test for the following cases:

1. DELIMITER GO
2. DELIMITER END
3. DELIMITER inside comment (should be ignored) -- DELIMITER ...
4. DELIMITER appear in quoted string

CREATE TABLE Table_5 (
  id INT COMMENT 'just a comment with semi colon ;',
  name varchar default 'default value with semi colon ;'
);

DELIMITER @

CREATE TABLE Table_6 (
  id INT COMMENT 'just a comment with new delimiter @',
  name varchar default 'default value with new delimiter @'
)@

[huydo862003]

I have fixed all the problems and add some contrived testcases in delimiter.in.sql nhe: https://github.com/holistics/dbml/blob/c646dddacc98307a490e8363c197bbefdfe1347d/packages/dbml-core/__tests__/parser/mysql-parse/input/delimiter.in.sql.

I have also added other unrelated tests to make sure changes to the Lexer does not break anything unrelated to delimiters.