pcre: minor refactoring and improvements

ext/pcre/php_pcre.c

@@ -369,10 +369,6 @@ static char *_pcre2_config_str(uint32_t what)
 /* {{{ PHP_MINFO_FUNCTION(pcre) */
 static PHP_MINFO_FUNCTION(pcre)
 {
-#ifdef HAVE_PCRE_JIT_SUPPORT
-	uint32_t flag = 0;
-	char *jit_target = _pcre2_config_str(PCRE2_CONFIG_JITTARGET);
-#endif
 	char *version = _pcre2_config_str(PCRE2_CONFIG_VERSION);
 	char *unicode = _pcre2_config_str(PCRE2_CONFIG_UNICODE_VERSION);
 
@@ -384,6 +380,8 @@ static PHP_MINFO_FUNCTION(pcre)
 	free(unicode);
 
 #ifdef HAVE_PCRE_JIT_SUPPORT
+	uint32_t flag = 0;
+	char *jit_target = _pcre2_config_str(PCRE2_CONFIG_JITTARGET);
 	if (!pcre2_config(PCRE2_CONFIG_JIT, &flag)) {
 		php_info_print_table_row(2, "PCRE JIT Support", flag ? "enabled" : "disabled");
 	} else {
@@ -579,7 +577,11 @@ PHPAPI pcre_cache_entry* pcre_get_compiled_regex_cache_ex(zend_string *regex, bo
 #else
 	uint32_t			 coptions = 0;
 #endif
+#if PCRE2_MAJOR >= 10 && PCRE2_MINOR >= 45
+	uint32_t			 eoptions = PCRE2_EXTRA_NEVER_CALLOUT;
+#else
 	uint32_t			 eoptions = 0;
+#endif
 	PCRE2_UCHAR	         error[128];
 	PCRE2_SIZE           erroffset;
 	int                  errnumber;
@@ -710,7 +712,7 @@ PHPAPI pcre_cache_entry* pcre_get_compiled_regex_cache_ex(zend_string *regex, bo
 			/* PCRE specific options */
 			case 'A':	coptions |= PCRE2_ANCHORED;		break;
 			case 'D':	coptions |= PCRE2_DOLLAR_ENDONLY;break;
-#ifdef PCRE2_EXTRA_CASELESS_RESTRICT
+#ifdef PCRE2_EXTRA_CASELESS_RESTRICT /* Added in 10.43 (16-February-2024) */
 			case 'r':	eoptions |= PCRE2_EXTRA_CASELESS_RESTRICT; break;
 #endif
 			case 'S':	/* Pass. */					break;
@@ -720,9 +722,7 @@ PHPAPI pcre_cache_entry* pcre_get_compiled_regex_cache_ex(zend_string *regex, bo
 	/* In  PCRE,  by  default, \d, \D, \s, \S, \w, and \W recognize only ASCII
 	   characters, even in UTF-8 mode. However, this can be changed by setting
 	   the PCRE2_UCP option. */
-#ifdef PCRE2_UCP
 						coptions |= PCRE2_UCP;
-#endif
 						/* The \C escape sequence is unsafe in PCRE2_UTF mode */
 						coptions |= PCRE2_NEVER_BACKSLASH_C;
 				break;
@@ -781,10 +781,15 @@ PHPAPI pcre_cache_entry* pcre_get_compiled_regex_cache_ex(zend_string *regex, bo
 			zend_string_release_ex(key, 0);
 		}
 		const char *err_msg = (const char*) error;
-		if (errnumber == PCRE2_ERROR_BACKSLASH_C_CALLER_DISABLED) {
-			err_msg = "using \\C is incompatible with the 'u' modifier";
-		} else {
-			pcre2_get_error_message(errnumber, error, sizeof(error));
+		switch (errnumber) {
+			case PCRE2_ERROR_BACKSLASH_K_IN_LOOKAROUND:
+				err_msg = "\\K is not allowed in lookarounds";
+				break;
+			case PCRE2_ERROR_BACKSLASH_C_CALLER_DISABLED:
+				err_msg = "using \\C is incompatible with the 'u' modifier";
+				break;
+			default:
+				pcre2_get_error_message(errnumber, error, sizeof(error));
 		}
 		php_error_docref(NULL,E_WARNING, "Compilation failed: %s at offset %zu", err_msg, erroffset);
 		pcre_handle_exec_error(PCRE2_ERROR_INTERNAL);

ext/pcre/tests/bug70345.phpt

@@ -19,8 +19,8 @@ preg_match($regex, $subject, $matches);
 var_dump($matches);
 ?>
 --EXPECTF--
-Warning: preg_split(): Compilation failed: \K is not allowed in lookarounds (but see PCRE2_EXTRA_ALLOW_LOOKAROUND_BSK) at offset 9 in %s on line %d
+Warning: preg_split(): Compilation failed: \K is not allowed in lookarounds at offset 9 in %s on line %d
 bool(false)
 
-Warning: preg_match(): Compilation failed: \K is not allowed in lookarounds (but see PCRE2_EXTRA_ALLOW_LOOKAROUND_BSK) at offset 12 in %s on line %d
+Warning: preg_match(): Compilation failed: \K is not allowed in lookarounds at offset 12 in %s on line %d
 NULL

ext/pcre/tests/bug75457.phpt

@@ -1,5 +1,11 @@
 --TEST--
 Bug #75457 (heap-use-after-free in php7.0.25)
+--SKIPIF--
+<?php
+if (PCRE_VERSION_MAJOR >= 10 && PCRE_VERSION_MINOR >= 45) {
+	die("skip callout feature is disable in PCRE2 10.45 and above");
+}
+?>
 --FILE--
 <?php
 $pattern = "/(((?(?C)0?=))(?!()0|.(?0)0)())/";

ext/pcre/tests/callout_syntax_in_pattern.phpt

@@ -0,0 +1,19 @@
+--TEST--
+PCRE2 callout feature is not supported.
+--SKIPIF--
+<?php
+if (PCRE_VERSION_MAJOR >= 10 && PCRE_VERSION_MINOR >= 45) {
+	die("skip callout feature is disable in PCRE2 10.45 and above");
+}
+?>
+--FILE--
+<?php
+
+$subject = 'hello';
+$pattern = '/a(?C1)b(?C2)c/';
+
+var_dump(preg_match($pattern, $subject));
+
+?>
+--EXPECT--
+int(0)

ext/pcre/tests/callout_syntax_in_pattern_disabled.phpt

@@ -0,0 +1,20 @@
+--TEST--
+PCRE2 callout feature is not supported and disabled if pcre2 lib permits it.
+--SKIPIF--
+<?php
+if (!(PCRE_VERSION_MAJOR >= 10 && PCRE_VERSION_MINOR >= 45)) {
+	die("skip needs 10.45 or above");
+}
+?>
+--FILE--
+<?php
+
+$subject = 'hello';
+$pattern = '/a(?C1)b(?C2)c/';
+
+var_dump(preg_match($pattern, $subject));
+
+?>
+--EXPECTF--
+Warning: preg_match(): Compilation failed: using callouts is disabled by the application at offset 4 in %s on line %d
+bool(false)

ext/pcre/tests/disabled_pcre_compile_options.phpt

@@ -0,0 +1,25 @@
+--TEST--
+Disabled PCRE2 regex compilation features.
+--FILE--
+<?php
+
+$subject = 'hello';
+$patterns = [
+	'/\u/', // Unknown unicode escape sequence
+    '/\C/u', // \C in unicode
+    '/a(?=b\K)/', // \K in look around
+];
+
+foreach ($patterns as $pattern) {
+     var_dump(preg_match($pattern, $subject));
+}
+?>
+--EXPECTF--
+Warning: preg_match(): Compilation failed: PCRE2 does not support \F, \L, \l, \N{name}, \U, or \u at offset 2 in %s on line %d
+bool(false)
+
+Warning: preg_match(): Compilation failed: using \C is incompatible with the 'u' modifier at offset 2 in %s on line %d
+bool(false)
+
+Warning: preg_match(): Compilation failed: \K is not allowed in lookarounds at offset 8 in %s on line %d
+bool(false)

ext/pcre/tests/preg_grep_error1.phpt

@@ -13,26 +13,11 @@ $values = [
     '/[a-zA-Z]', //Regex without closing delimiter
     '[a-zA-Z]/', //Regex without opening delimiter
     '/[a-zA-Z]/F',
-    [
-        '[a-z]', //Array of Regexes
-        '[A-Z]',
-        '[0-9]',
-    ],
     '/[a-zA-Z]/', //Regex string
 ];
 $array = [123, 'abc', 'test'];
 foreach ($values as $value) {
-    try {
-        var_dump(preg_grep($value, $array));
-    } catch (TypeError $e) {
-        echo $e::class, ': ', $e->getMessage(), "\n";
-    }
-}
-$value = new stdclass(); //Object
-try {
     var_dump(preg_grep($value, $array));
-} catch (TypeError $e) {
-    echo $e::class, ': ', $e->getMessage(), "\n";
 }
 ?>
 --EXPECTF--
@@ -47,11 +32,9 @@ bool(false)
 
 Warning: preg_grep(): Unknown modifier 'F' in %spreg_grep_error1.php on line %d
 bool(false)
-TypeError: preg_grep(): Argument #1 ($pattern) must be of type string, array given
 array(2) {
   [1]=>
   string(3) "abc"
   [2]=>
   string(4) "test"
 }
-TypeError: preg_grep(): Argument #1 ($pattern) must be of type string, stdClass given

ext/pcre/tests/preg_match_all_error1.phpt

@@ -13,29 +13,14 @@ $regex_array = [
     '/[a-zA-Z]', //Regex without closing delimiter
     '[a-zA-Z]/', //Regex without opening delimiter
     '/[a-zA-Z]/F',
-    [
-        '[a-z]', //Array of Regexes
-        '[A-Z]',
-        '[0-9]',
-    ],
     '/[a-zA-Z]/', //Regex string
 ];
 $subject = 'test';
 foreach ($regex_array as $regex_value) {
-    try {
-        var_dump(preg_match_all($regex_value, $subject, $matches1));
-    } catch (TypeError $e) {
-        echo $e::class, ': ', $e->getMessage(), "\n";
-    }
+    var_dump(preg_match_all($regex_value, $subject, $matches1));
     var_dump($matches1);
 }
-$regex_value = new stdclass(); //Object
-try {
-    var_dump(preg_match_all($regex_value, $subject, $matches));
-} catch (TypeError $e) {
-    echo $e::class, ': ', $e->getMessage(), "\n";
-}
-var_dump($matches);
+
 ?>
 --EXPECTF--
 Warning: preg_match_all(): Delimiter must not be alphanumeric, backslash, or NUL byte in %spreg_match_all_error1.php on line %d
@@ -53,8 +38,6 @@ NULL
 Warning: preg_match_all(): Unknown modifier 'F' in %spreg_match_all_error1.php on line %d
 bool(false)
 NULL
-TypeError: preg_match_all(): Argument #1 ($pattern) must be of type string, array given
-NULL
 int(4)
 array(1) {
   [0]=>
@@ -69,5 +52,3 @@ array(1) {
     string(1) "t"
   }
 }
-TypeError: preg_match_all(): Argument #1 ($pattern) must be of type string, stdClass given
-NULL

ext/pcre/tests/preg_replace_error1.phpt

@@ -25,12 +25,6 @@ $subject = 'a';
 foreach ($regex_array as $regex_value) {
     var_dump(preg_replace($regex_value, $replace, $subject));
 }
-$regex_value = new stdclass(); //Object
-try {
-    var_dump(preg_replace($regex_value, $replace, $subject));
-} catch (Error $e) {
-    echo $e::class, ': ', $e->getMessage(), "\n";
-}
 ?>
 --EXPECTF--
 Warning: preg_replace(): Delimiter must not be alphanumeric, backslash, or NUL byte in %spreg_replace_error1.php on line %d
@@ -46,4 +40,3 @@ Warning: preg_replace(): Unknown modifier 'F' in %spreg_replace_error1.php on li
 NULL
 string(1) "a"
 string(1) "1"
-TypeError: preg_replace(): Argument #1 ($pattern) must be of type array|string, stdClass given

ext/pcre/tests/preg_replace_error2.phpt

@@ -9,7 +9,10 @@ Test preg_replace() function : error conditions - wrong arg types
 * Testing how preg_replace reacts to being passed the wrong type of replacement argument
 */
 $regex = '/[a-zA-Z]/';
-$replace = array('this is a string', array('this is', 'a subarray'),);
+$replace = [
+	'this is a string',
+	['this is', 'a subarray'],
+];
 $subject = 'test';
 foreach($replace as $value) {
     try {
@@ -18,14 +21,7 @@ foreach($replace as $value) {
         echo $e::class, ': ', $e->getMessage(), "\n";
     }
 }
-$value = new stdclass(); //Object
-try {
-    var_dump(preg_replace($regex, $value, $subject));
-} catch (Error $e) {
-    echo $e::class, ': ', $e->getMessage(), "\n";
-}
 ?>
 --EXPECT--
 string(64) "this is a stringthis is a stringthis is a stringthis is a string"
 TypeError: preg_replace(): Argument #1 ($pattern) must be of type array when argument #2 ($replacement) is an array, string given
-TypeError: preg_replace(): Argument #2 ($replacement) must be of type array|string, stdClass given

ext/pcre/tests/preg_split_error1.phpt

@@ -13,26 +13,11 @@ $regex_array = [
     '/[a-zA-Z]', //Regex without closing delimiter
     '[a-zA-Z]/', //Regex without opening delimiter
     '/[a-zA-Z]/F',
-    [
-        '[a-z]', //Array of Regexes
-        '[A-Z]',
-        '[0-9]',
-    ],
     '/[a-zA-Z]/', //Regex string
 ];
 $subject = '1 2 a 3 4 b 5 6';
 foreach ($regex_array as $regex_value) {
-    try {
-        var_dump(preg_split($regex_value, $subject));
-    } catch (TypeError $e) {
-        echo $e::class, ': ', $e->getMessage(), "\n";
-    }
-}
-$regex_value = new stdclass(); //Object
-try {
-    var_dump(preg_split($regex_value, $subject));
-} catch (TypeError $e) {
-    echo $e::class, ': ', $e->getMessage(), "\n";
+     var_dump(preg_split($regex_value, $subject));
 }
 ?>
 --EXPECTF--
@@ -47,7 +32,6 @@ bool(false)
 
 Warning: preg_split(): Unknown modifier 'F' in %spreg_split_error1.php on line %d
 bool(false)
-TypeError: preg_split(): Argument #1 ($pattern) must be of type string, array given
 array(3) {
   [0]=>
   string(4) "1 2 "
@@ -56,4 +40,3 @@ array(3) {
   [2]=>
   string(4) " 5 6"
 }
-TypeError: preg_split(): Argument #1 ($pattern) must be of type string, stdClass given