Make mb_ereg_replace() pattern argument a string

UPGRADING

@@ -94,6 +94,9 @@ PHP 8.0 UPGRADE NOTES
 
   . The 'e' modifier for mb_ereg_replace() has been removed.
     mb_ereg_replace_callback() should be used instead.
+  . A non-string pattern argument to mb_ereg_replace() will now be interpreted
+    as a string instead of an ASCII codepoint. The previous behavior may be
+    restored with an explicit call to chr().
 
 - SPL:
   . SplFileObject::fgetss() has been removed.

ext/mbstring/php_mbregex.c

@@ -967,8 +967,6 @@ PHP_FUNCTION(mb_eregi)
 /* {{{ _php_mb_regex_ereg_replace_exec */
 static void _php_mb_regex_ereg_replace_exec(INTERNAL_FUNCTION_PARAMETERS, OnigOptionType options, int is_callable)
 {
-	zval *arg_pattern_zval;
-
 	char *arg_pattern;
 	size_t arg_pattern_len;
 
@@ -991,7 +989,6 @@ static void _php_mb_regex_ereg_replace_exec(INTERNAL_FUNCTION_PARAMETERS, OnigOp
 	OnigUChar *pos;
 	OnigUChar *string_lim;
 	char *description = NULL;
-	char pat_buf[6];
 
 	const mbfl_encoding *enc;
 
@@ -1010,16 +1007,16 @@ static void _php_mb_regex_ereg_replace_exec(INTERNAL_FUNCTION_PARAMETERS, OnigOp
 		size_t option_str_len = 0;
 
 		if (!is_callable) {
-			if (zend_parse_parameters(ZEND_NUM_ARGS(), "zss|s",
+			if (zend_parse_parameters(ZEND_NUM_ARGS(), "sss|s",
-						&arg_pattern_zval,
+						&arg_pattern, &arg_pattern_len,
 						&replace, &replace_len,
 						&string, &string_len,
 						&option_str, &option_str_len) == FAILURE) {
 				RETURN_FALSE;
 			}
 		} else {
-			if (zend_parse_parameters(ZEND_NUM_ARGS(), "zfs|s",
+			if (zend_parse_parameters(ZEND_NUM_ARGS(), "sfs|s",
-						&arg_pattern_zval,
+						&arg_pattern, &arg_pattern_len,
 						&arg_replace_fci, &arg_replace_fci_cache,
 						&string, &string_len,
 						&option_str, &option_str_len) == FAILURE) {
@@ -1046,26 +1043,7 @@ static void _php_mb_regex_ereg_replace_exec(INTERNAL_FUNCTION_PARAMETERS, OnigOp
 		php_error_docref(NULL, E_WARNING, "The 'e' option is no longer supported, use mb_ereg_replace_callback instead");
 		RETURN_FALSE;
 	}
-	if (Z_TYPE_P(arg_pattern_zval) == IS_STRING) {
-		arg_pattern = Z_STRVAL_P(arg_pattern_zval);
-		arg_pattern_len = Z_STRLEN_P(arg_pattern_zval);
-	} else {
-		php_error_docref(NULL, E_DEPRECATED,
-			"Non-string patterns will be interpreted as strings in the future. "
-			"Use an explicit chr() call to preserve the current behavior");
 
-		/* FIXME: this code is not multibyte aware! */
-		convert_to_long_ex(arg_pattern_zval);
-		pat_buf[0] = (char)Z_LVAL_P(arg_pattern_zval);
-		pat_buf[1] = '\0';
-		pat_buf[2] = '\0';
-		pat_buf[3] = '\0';
-		pat_buf[4] = '\0';
-		pat_buf[5] = '\0';
-
-		arg_pattern = pat_buf;
-		arg_pattern_len = 1;
-	}
 	/* create regex pattern buffer */
 	re = php_mbregex_compile_pattern(arg_pattern, arg_pattern_len, options, MBREX(current_mbctype), syntax);
 	if (re == NULL) {

ext/mbstring/tests/bug72994.phpt

@@ -1,21 +0,0 @@
---TEST--
-Bug #72994 (mbc_to_code() out of bounds read)
---SKIPIF--
-<?php
-if (!extension_loaded('mbstring')) die('skip mbstring extension not available');
-if (!function_exists('mb_ereg_replace')) die('skip mb_ereg_replace() not available');
-?>
---FILE--
-<?php
-$var1 = mb_ereg_replace($var-232338951,NULL,NULL,NULL);
-var_dump($var1);
-?>
-===DONE===
---EXPECTF--
-Notice: Undefined variable: var in %s on line %d
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
-
-Warning: mb_ereg_replace(): mbregex compile err: invalid code point value in %sbug72994.php on line %d
-bool(false)
-===DONE===

ext/mbstring/tests/mb_ereg_replace_variation1.phpt

@@ -96,87 +96,53 @@ foreach($inputs as $input) {
 
 echo "Done";
 ?>
---EXPECTF--
+--EXPECT--
 *** Testing mb_ereg_replace() : usage variations ***
 
 -- Iteration 1 --
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
 string(10) "string_val"
 
 -- Iteration 2 --
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
 string(10) "string_val"
 
 -- Iteration 3 --
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
 string(10) "string_val"
 
 -- Iteration 4 --
-
+string(10) "string_val"
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
-
-Warning: mb_ereg_replace(): mbregex compile err: invalid code point value in %s on line %d
-bool(false)
 
 -- Iteration 5 --
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
 string(10) "string_val"
 
 -- Iteration 6 --
-
+string(10) "string_val"
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
-
-Warning: mb_ereg_replace(): mbregex compile err: invalid code point value in %s on line %d
-bool(false)
 
 -- Iteration 7 --
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
 string(10) "string_val"
 
 -- Iteration 8 --
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
 string(10) "string_val"
 
 -- Iteration 9 --
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
 string(10) "string_val"
 
 -- Iteration 10 --
-
+string(120) "string_valsstring_valtstring_valrstring_valistring_valnstring_valgstring_val_string_valvstring_valastring_vallstring_val"
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
-string(10) "string_val"
 
 -- Iteration 11 --
-
+string(120) "string_valsstring_valtstring_valrstring_valistring_valnstring_valgstring_val_string_valvstring_valastring_vallstring_val"
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
-string(10) "string_val"
 
 -- Iteration 12 --
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
 string(10) "string_val"
 
 -- Iteration 13 --
-
+string(120) "string_valsstring_valtstring_valrstring_valistring_valnstring_valgstring_val_string_valvstring_valastring_vallstring_val"
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
-string(10) "string_val"
 
 -- Iteration 14 --
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
 string(10) "string_val"
 
 -- Iteration 15 --
-
+string(120) "string_valsstring_valtstring_valrstring_valistring_valnstring_valgstring_val_string_valvstring_valastring_vallstring_val"
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
-string(10) "string_val"
 
 -- Iteration 16 --
 string(120) "string_valsstring_valtstring_valrstring_valistring_valnstring_valgstring_val_string_valvstring_valastring_vallstring_val"
@@ -194,17 +160,11 @@ string(10) "string_val"
 string(10) "string_val"
 
 -- Iteration 21 --
-
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
 string(10) "string_val"
 
 -- Iteration 22 --
-
+string(120) "string_valsstring_valtstring_valrstring_valistring_valnstring_valgstring_val_string_valvstring_valastring_vallstring_val"
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
-string(10) "string_val"
 
 -- Iteration 23 --
-
+string(120) "string_valsstring_valtstring_valrstring_valistring_valnstring_valgstring_val_string_valvstring_valastring_vallstring_val"
-Deprecated: mb_ereg_replace(): Non-string patterns will be interpreted as strings in the future. Use an explicit chr() call to preserve the current behavior in %s on line %d
-string(10) "string_val"
 Done