Enhancement: better input validation

src/Helpers/Utils.php

@@ -129,9 +129,22 @@ public static function removeBackslashes($data)
         return $data;
     }
 
+    /**
+     * Decode strings recursively to prevent double (or more) encoded strings
+     *
+     * @unreleased
+     */
+    public static function recursiveUrlDecode(string $data): string
+    {
+        $decoded = urldecode($data);
+
+        return $decoded === $data ? $data : self::recursiveUrlDecode($decoded);
+    }
+
     /**
      * The regular expression attempts to capture the basic structure of all data types that can be serialized by PHP.
      *
+     * @unreleased Decode the string and remove any character not allowed in a serialized string
      * @since 3.19.3 Support all types of serialized data instead of only objects and arrays
      * @since 3.17.2
      */
@@ -141,9 +154,17 @@ public static function containsSerializedDataRegex($data): bool
             return false;
         }
 
+        $data = self::recursiveUrlDecode($data);
+
+        /**
+         * This regular expression removes any special character that is not:
+         * a Letter (a-zA-Z), number (0-9), or any of the characters {}, :, ;, ", ', ., [, ], (, ), ,
+         */
+        $data = preg_replace('/[^a-zA-Z0-9:{};"\'.\[\](),]/', '', $data);
+
         $pattern = '/
-        (a:\d+:\{.*\}) |         # Matches arrays (e.g: a:2:{i:0;s:5:"hello";i:1;i:42;})
-        (O:\d+:"[^"]+":\{.*\}) | # Matches objects (e.g: O:8:"stdClass":1:{s:4:"name";s:5:"James";})
+        (a:\d+:\{.*}) |         # Matches arrays (e.g: a:2:{i:0;s:5:"hello";i:1;i:42;})
+        (O:\d+:"[^"]+":\{.*}) | # Matches objects (e.g: O:8:"stdClass":1:{s:4:"name";s:5:"James";})
         (s:\d+:"[^"]*";) |       # Matches strings (e.g: s:5:"hello";)
         (i:\d+;) |               # Matches integers (e.g: i:42;)
         (b:[01];) |              # Matches booleans (e.g: b:1; or b:0;)

tests/Unit/Helpers/UtilsTest.php

@@ -82,15 +82,14 @@ public function testMaybeSafeUnserialize($data, bool $expected)
     }
 
     /**
+     * @unreleased Test encoded strings and strings with special characters
      * @since 3.19.3 Test all types of serialized data
      * @since 3.17.2
      */
     public function serializedDataProvider(): array
     {
         return [
             [serialize('bar'), true],
-            ['\\' . serialize('backslash-bypass'), true],
-            ['\\\\' . serialize('double-backslash-bypass'), true],
             ['foo', false],
             [serialize('qux'), true],
             ['bar', false],
@@ -103,6 +102,50 @@ public function serializedDataProvider(): array
             ['Lorem ipsum b:1; dolor sit amet', true], // boolean
             ['Lorem ipsum d:3.14; dolor sit amet', true], // float
             ['Lorem ipsum N; dolor sit amet', true], // NULL
+            // Strings with special characters (e.g: emojis, spaces, control characters) that are not part of a predefined set of safe characters for serialized data structures (used to try to bypass the validations)
+            [
+                // emojis bypass sample
+                'O😼:8:"stdClass":1:{s😼:4:"name";s😼:5:"James";}',
+                true,
+            ],
+            [
+                // spaces bypass sample
+                'O :8:"stdClass":1:{s :4:"name";s :5:"James";}',
+                true,
+            ],
+            // Bypass with simple methods
+            [
+                // backslash
+                '\\' . serialize('backslash-bypass'),
+                true,
+            ],
+            [
+                // double-backslash
+                '\\\\' . serialize('double-backslash-bypass'),
+                true,
+            ],
+            // Bypass with encoding string method - URL-encoded
+            [
+                // Single encode for O:8:"stdClass":1:{s:4:"name";s:5:"James";}
+                'O%3A8%3A%22stdClass%22%3A1%3A%7Bs%3A4%3A%22name%22%3Bs%3A5%3A%22James%22%3B%7D',
+                true,
+            ],
+            [
+                // Double encode for O:8:"stdClass":1:{s:4:"name";s:5:"James";}
+                'O%253A8%253A%2522stdClass%2522%253A1%253A%257Bs%253A4%253A%2522name%2522%253Bs%253A5%253A%2522James%2522%253B%257D',
+                true,
+            ],
+            // Samples using multiple obfuscation techniques together
+            [
+                // Single URL-encoded for O😼:8:"stdClass":1:{s😼:4:"name";s😼:5:"James";}
+                'O%F0%9F%98%BC%3A8%3A%22stdClass%22%3A1%3A%7Bs%F0%9F%98%BC%3A4%3A%22name%22%3Bs%F0%9F%98%BC%3A5%3A%22James%22%3B%7D',
+                true,
+            ],
+            [
+                // Double URL-encoded for O😼:8:"stdClass":1:{s😼:4:"name";s😼:5:"James";}
+                'O%25F0%259F%2598%25BC%253A8%253A%2522stdClass%2522%253A1%253A%257Bs%25F0%259F%2598%25BC%253A4%253A%2522name%2522%253Bs%25F0%259F%2598%25BC%253A5%253A%2522James%2522%253B%257D',
+                true,
+            ],
         ];
     }
 }