BUG#35140271: Regex split hanging in cursor.execute(..., multi=True) for complex queries.

oscpache · oscpache · commit 82039316f513 · 2023-06-01T12:17:21.000-06:00
Performance degradation was induced while fixing `BUG#34655520:
Wrong MySQLCursor.statement values in the results of
cursor.execute(..., multi=True)`.

With this patch, the performance issue is fixed.

Change-Id: Ib8a35fb768ab5ecce8779de72bc720ffea47f7d8
diff --git a/CHANGES.txt b/CHANGES.txt
@@ -19,6 +19,7 @@ v8.1.0
 - BUG#35338384: PIP installs incompatible Connector/Python packages
 - BUG#35318413: Fix charset mapping for MySQL 8.1.0
 - BUG#35212199: Check for identifier quotes in the database name
+- BUG#35140271: Regex split hanging in cursor.execute(..., multi=True) for complex queries
 - BUG#29115406: CONTRIBUTION - FIX RECV COMPRESS BUG
 
 v8.0.33
diff --git a/lib/mysql/connector/cursor.py b/lib/mysql/connector/cursor.py
@@ -99,7 +99,9 @@
     """,
     re.X,
 )
-RE_SQL_SPLIT_STMTS = re.compile(b""";(?=(?:[^"'`]*["'`].*["'`])*[^"'`]*$)""")
+RE_SQL_SPLIT_STMTS = re.compile(
+    b""";(?=(?:[^"'`]*(?:"[^"]*"|'[^']*'|`[^`]*`))*[^"'`]*$)"""
+)
 RE_SQL_FIND_PARAM = re.compile(b"""%s(?=(?:[^"'`]*["'`][^"'`]*["'`])*[^"'`]*$)""")
 RE_SQL_PYTHON_REPLACE_PARAM = re.compile(r"%\(.*?\)s")
 RE_SQL_PYTHON_CAPTURE_PARAM_NAME = re.compile(r"%\((.*?)\)s")
diff --git a/tests/cext/test_cext_cursor.py b/tests/cext/test_cext_cursor.py
@@ -627,7 +627,8 @@ def test_execute_multi(self):
 
         cur.close()
 
-        # test statement property
+        # Bug#34655520: Wrong MySQLCursor.statement values in the results of
+        # cursor.execute(..., multi=True)
         operations = [
             'select 1; SELECT "`";',
             "SELECT '\"'; SELECT 2; select '```';",
diff --git a/tests/test_cursor.py b/tests/test_cursor.py
@@ -630,6 +630,9 @@ def test_execute(self):
 
     def test_execute_multi(self):
         # Tests when execute(..., multi=True)
+
+        # Bug#34655520: Wrong MySQLCursor.statement values in the results of
+        # cursor.execute(..., multi=True)
         operations = [
             'select 1; SELECT "`";',
             "SELECT '\"'; SELECT 2; select '```';",
@@ -647,6 +650,31 @@ def test_execute_multi(self):
                     self.assertEqual(exp, res_cur.statement)
                     _ = res_cur.fetchall()
 
+        # Bug#35140271: Regex split hanging in cursor.execute(..., multi=True)
+        # for complex queries
+        operations.append("SELECT `';`; select 3; SELECT '`;`;`';")
+        control.append(["SELECT `';`", "select 3", "SELECT '`;`;`'"])
+        for operation, exps in zip(operations, control):
+            executed_list = cursor.RE_SQL_SPLIT_STMTS.split(operation.encode())
+            self.assertEqual(len(executed_list), len(exps) + 1)
+
+        # the bug reports that execution hangs with the following statement;
+        # let's run it and check it does not hang.
+        RE_SQL_SPLIT_STMTS_OLD = re.compile(
+            b""";(?=(?:[^"'`]*["'`][^"'`]*["'`])*[^"'`]*$)"""
+        )
+        operation = """
+        create database test;\n create table
+        test.test(pk INT PRIMARY KEY AUTO_INCREMENT, t CHAR(6));\n INSERT INTO
+        test.test(t) VALUES ('h'), ('e'), ('l'), ('l'), ('o'), (' '), ('w'), ('o'),
+        ('r'), ('l'), ('d'), ('!'), (' '), ('h'), ('e'), ('l'), ('l'), ('o'), (' '),
+        ('w'), ('o'), ('r'), ('l'), ('d'), ('!'); \n -- insert db's elements
+        """
+        self.assertListEqual(
+            cursor.RE_SQL_SPLIT_STMTS.split(operation.encode()),
+            RE_SQL_SPLIT_STMTS_OLD.split(operation.encode()),
+        )
+
     def test_executemany(self):
         """MySQLCursor object executemany()-method"""
         self.check_method(self.cur, "executemany")
