Add dollar-quoting support to MySQL multi-query parser

Author: JanTvrdik

src/MySqlMultiQueryParser.php

@@ -52,6 +52,7 @@ private function getQueryPattern(string $delimiter): string
 							|   \" (*PRUNE)                                           (?: \\\\.    | [^\"]           )*+ \"
 							|   \` (*PRUNE)                                           (?: [^\`]++  | \`\`            )*+ \`
 							|   /\\* (*PRUNE)                                         (?: [^*]++   | \\*(?!/)        )*+ \\*/
+							|   (\\$(?:[a-zA-Z_\\x80-\\xFF][\\w\\x80-\\xFF]*+)?\\$) (*PRUNE) (?: [^$]++   | (?!\\g{-1})\\$ )*+ \\g{-1}
 							|   --[^\\n]*+(?:\\n|\\z)
 							|   \\#[^\\n]*+(?:\\n|\\z)
 							|   (?!$delimiterPattern) .

tests/cases/MySqlMultiQueryParserTest.phpt

@@ -29,7 +29,7 @@ class MySqlMultiQueryParserTest extends MultiQueryParserTestCase
 
 	protected function getExpectedFileQueryCount(): int
 	{
-		return 60;
+		return 61;
 	}
 
 
@@ -242,6 +242,21 @@ class MySqlMultiQueryParserTest extends MultiQueryParserTestCase
 
 			// Escaped backticks (doubled) inside backtick identifiers
 			["SELECT `col``name` FROM t;", ["SELECT `col``name` FROM t"]],
+
+			// Dollar-quoted strings in JavaScript stored programs
+			[
+				'CREATE FUNCTION gcd(a INT, b INT) RETURNS INT NO SQL LANGUAGE JAVASCRIPT AS $mle$ let x = a; let y = b; $mle$;',
+				['CREATE FUNCTION gcd(a INT, b INT) RETURNS INT NO SQL LANGUAGE JAVASCRIPT AS $mle$ let x = a; let y = b; $mle$'],
+			],
+			[
+				'CREATE FUNCTION js_add(a INT, b INT) RETURNS INT LANGUAGE JAVASCRIPT AS $$ return a + b; $$;',
+				['CREATE FUNCTION js_add(a INT, b INT) RETURNS INT LANGUAGE JAVASCRIPT AS $$ return a + b; $$'],
+			],
+			// Nested dollar-quoted strings with different tags
+			[
+				'CREATE FUNCTION nested() RETURNS INT LANGUAGE JAVASCRIPT AS $mle$ let s = $inner$;$inner$; return 1; $mle$;',
+				['CREATE FUNCTION nested() RETURNS INT LANGUAGE JAVASCRIPT AS $mle$ let s = $inner$;$inner$; return 1; $mle$'],
+			],
 		];
 	}
 
@@ -297,6 +312,11 @@ class MySqlMultiQueryParserTest extends MultiQueryParserTestCase
 				["SELECT `col;na", "me` FROM t;"],
 				["SELECT `col;name` FROM t"],
 			],
+			// Dollar-quoted string spanning chunks
+			[
+				['CREATE FUNCTION f() RETURNS INT LANGUAGE JAVASCRIPT AS $mle$ let x = 1;', ' return x; $mle$;'],
+				['CREATE FUNCTION f() RETURNS INT LANGUAGE JAVASCRIPT AS $mle$ let x = 1; return x; $mle$'],
+			],
 		];
 	}
 }

tests/data/mysql.sql

@@ -225,6 +225,21 @@ INSERT INTO contents (id, type, thread_id, replied_at) VALUES (1, 'thread', NULL
 INSERT INTO contents (id, type, thread_id, replied_at) VALUES (2, 'comment', 1, '2020-01-01 12:00:00');
 INSERT INTO contents (id, type, thread_id, replied_at) VALUES (3, 'comment', 1, '2020-01-02 12:00:00');
 
+CREATE FUNCTION gcd(a INT, b INT)
+	RETURNS INT
+	NO SQL
+	LANGUAGE JAVASCRIPT AS
+$mle$
+	let x = Math.abs(a);
+	let y = Math.abs(b);
+	while (y) {
+		var t = y;
+		y = x % y;
+		x = t;
+	}
+	return x;
+$mle$;
+
 # Hash comment with semicolons; should be ignored; entirely
 SELECT `backtick;identifier` FROM authors WHERE name = 'test';
 # Another hash comment