Use shared, immutable array for return value of mb_list_encodings

This will allow us to easily check in other mbstring functions if the
list of all supported encodings, returned by mb_list_encodings, is
passed in as input to another function.

Co-authored-by: Ilija Tovilo <[email protected]>

Author: alexdowad

Zend/Optimizer/zend_func_infos.h

@@ -227,7 +227,7 @@ static const func_info_t func_infos[] = {
 	F1("mb_strtoupper", MAY_BE_STRING),
 	F1("mb_strtolower", MAY_BE_STRING),
 	F1("mb_detect_encoding", MAY_BE_STRING|MAY_BE_FALSE),
-	F1("mb_list_encodings", MAY_BE_ARRAY|MAY_BE_ARRAY_KEY_LONG|MAY_BE_ARRAY_OF_STRING),
+	FN("mb_list_encodings", MAY_BE_ARRAY|MAY_BE_ARRAY_KEY_LONG|MAY_BE_ARRAY_OF_STRING),
 	F1("mb_encoding_aliases", MAY_BE_ARRAY|MAY_BE_ARRAY_KEY_LONG|MAY_BE_ARRAY_OF_STRING),
 	F1("mb_encode_mimeheader", MAY_BE_STRING),
 	F1("mb_decode_mimeheader", MAY_BE_STRING),

ext/mbstring/mbstring.c

@@ -1016,6 +1016,7 @@ ZEND_TSRMLS_CACHE_UPDATE();
 	mbstring_globals->internal_encoding_set = 0;
 	mbstring_globals->http_output_set = 0;
 	mbstring_globals->http_input_set = 0;
+	mbstring_globals->all_encodings_list = NULL;
 }
 /* }}} */
 
@@ -1156,6 +1157,13 @@ PHP_RSHUTDOWN_FUNCTION(mbstring)
 	MBSTRG(outconv_enabled) = false;
 	MBSTRG(outconv_state) = 0;
 
+	if (MBSTRG(all_encodings_list)) {
+		GC_DELREF(MBSTRG(all_encodings_list));
+		zend_hash_destroy(MBSTRG(all_encodings_list));
+		efree(MBSTRG(all_encodings_list));
+		MBSTRG(all_encodings_list) = NULL;
+	}
+
 #ifdef HAVE_MBREGEX
 	PHP_RSHUTDOWN(mb_regex) (INIT_FUNC_ARGS_PASSTHRU);
 #endif
@@ -3205,10 +3213,22 @@ PHP_FUNCTION(mb_list_encodings)
 {
 	ZEND_PARSE_PARAMETERS_NONE();
 
-	array_init(return_value);
-	for (const mbfl_encoding **encodings = mbfl_get_supported_encodings(); *encodings; encodings++) {
-		add_next_index_string(return_value, (*encodings)->name);
+	if (MBSTRG(all_encodings_list) == NULL) {
+		/* Initialize shared array of supported encoding names
+		 * This is done so that we can check if `mb_list_encodings()` is being
+		 * passed to other mbstring functions using a cheap pointer equality check */
+		HashTable *array = emalloc(sizeof(HashTable));
+		zend_hash_init(array, 80, NULL, zval_ptr_dtor_str, false);
+		for (const mbfl_encoding **encodings = mbfl_get_supported_encodings(); *encodings; encodings++) {
+			zval tmp;
+			ZVAL_STRING(&tmp, (*encodings)->name);
+			zend_hash_next_index_insert(array, &tmp);
+		}
+		MBSTRG(all_encodings_list) = array;
 	}
+
+	GC_ADDREF(MBSTRG(all_encodings_list));
+	RETURN_ARR(MBSTRG(all_encodings_list));
 }
 /* }}} */
 

ext/mbstring/mbstring.h

@@ -88,6 +88,7 @@ ZEND_BEGIN_MODULE_GLOBALS(mbstring)
 	size_t current_detect_order_list_size;
 	enum mbfl_no_encoding *default_detect_order_list;
 	size_t default_detect_order_list_size;
+	HashTable *all_encodings_list;
 	int filter_illegal_mode;
 	uint32_t filter_illegal_substchar;
 	int current_filter_illegal_mode;

ext/mbstring/mbstring.stub.php

@@ -140,7 +140,6 @@ function mb_detect_encoding(string $string, array|string|null $encodings = null,
 
 /**
  * @return array<int, string>
- * @refcount 1
  */
 function mb_list_encodings(): array {}