Merge branch 'PHP-8.2'

* PHP-8.2:
  Fix memory leaks in ext-tidy

Author: Girgias

Diff for: ext/tidy/tests/parsing_file_too_large.phpt

@@ -0,0 +1,56 @@
+--TEST--
+Trying to parse a file that is too large (over 4GB)
+--EXTENSIONS--
+tidy
+--SKIPIF--
+<?php
+if (PHP_INT_SIZE != 8) die("skip this test is for 64bit platform only");
+if (getenv("SKIP_SLOW_TESTS")) die("skip slow test");
+?>
+--INI--
+memory_limit="5G"
+--FILE--
+<?php
+
+$path = __DIR__ . '/too_large_test.html';
+$file = fopen($path, 'w+');
+
+// Write over 4GB
+const MIN_FILE_SIZE = 4_294_967_295;
+
+var_dump(fseek($file, MIN_FILE_SIZE+10));
+$s = str_repeat("a", 10);
+$bytes_written = fwrite($file, $s);
+if ($bytes_written === false) {
+    echo "Didn't write bytes\n";
+}
+
+$tidy = new tidy;
+try {
+    var_dump($tidy->parseFile($path));
+} catch (\Throwable $e) {
+    echo $e::class, ': ', $e->getMessage(), PHP_EOL;
+}
+
+try {
+    var_dump(tidy_parse_file($path));
+} catch (\Throwable $e) {
+    echo $e::class, ': ', $e->getMessage(), PHP_EOL;
+}
+
+try {
+    $tidy = new tidy($path);
+} catch (\Throwable $e) {
+    echo $e::class, ': ', $e->getMessage(), PHP_EOL;
+}
+?>
+--CLEAN--
+<?php
+$path = __DIR__ . '/too_large_test.html';
+unlink($path);
+?>
+--EXPECT--
+int(0)
+ValueError: Input string is too long
+ValueError: Input string is too long
+ValueError: Input string is too long

Diff for: ext/tidy/tests/parsing_inexistent_file.phpt

@@ -0,0 +1,22 @@
+--TEST--
+Trying to parse a non existent file
+--EXTENSIONS--
+tidy
+--FILE--
+<?php
+
+$tidy = new tidy;
+var_dump($tidy->parseFile("does_not_exist.html"));
+
+var_dump(tidy_parse_file("does_not_exist.html"));
+
+$tidy = new tidy("does_not_exist.html");
+?>
+--EXPECTF--
+Warning: tidy::parseFile(): Cannot load "does_not_exist.html" into memory in %s on line %d
+bool(false)
+
+Warning: tidy_parse_file(): Cannot load "does_not_exist.html" into memory in %s on line %d
+bool(false)
+
+Warning: tidy::__construct(): Cannot load "does_not_exist.html" into memory in %s on line %d

Diff for: ext/tidy/tidy.c

@@ -1069,19 +1069,20 @@ PHP_FUNCTION(tidy_parse_file)
 		Z_PARAM_BOOL(use_include_path)
 	ZEND_PARSE_PARAMETERS_END();
 
-	tidy_instantiate(tidy_ce_doc, return_value);
-	obj = Z_TIDY_P(return_value);
-
 	if (!(contents = php_tidy_file_to_mem(ZSTR_VAL(inputfile), use_include_path))) {
 		php_error_docref(NULL, E_WARNING, "Cannot load \"%s\" into memory%s", ZSTR_VAL(inputfile), (use_include_path) ? " (using include path)" : "");
 		RETURN_FALSE;
 	}
 
 	if (ZEND_SIZE_T_UINT_OVFL(ZSTR_LEN(contents))) {
+		zend_string_release_ex(contents, 0);
 		zend_value_error("Input string is too long");
 		RETURN_THROWS();
 	}
 
+	tidy_instantiate(tidy_ce_doc, return_value);
+	obj = Z_TIDY_P(return_value);
+
 	TIDY_APPLY_CONFIG(obj->ptdoc->doc, options_str, options_ht);
 
 	if (php_tidy_parse_string(obj, ZSTR_VAL(contents), (uint32_t)ZSTR_LEN(contents), enc) == FAILURE) {
@@ -1372,6 +1373,7 @@ PHP_METHOD(tidy, __construct)
 		}
 
 		if (ZEND_SIZE_T_UINT_OVFL(ZSTR_LEN(contents))) {
+			zend_string_release_ex(contents, 0);
 			zend_value_error("Input string is too long");
 			RETURN_THROWS();
 		}
@@ -1410,6 +1412,7 @@ PHP_METHOD(tidy, parseFile)
 	}
 
 	if (ZEND_SIZE_T_UINT_OVFL(ZSTR_LEN(contents))) {
+		zend_string_release_ex(contents, 0);
 		zend_value_error("Input string is too long");
 		RETURN_THROWS();
 	}