Fix bug #81598: Use C.UTF-8 as LC_CTYPE locale by default

Unfortunately, libedit is locale based and does not accept UTF-8
input when the C locale is used. This patch switches the default
locale to C.UTF-8 instead (if it is available). This makes libedit
work and I believe it shouldn't affect behavior of single-byte
locale-dependent functions that PHP otherwise uses.

Closes GH-7635.

NEWS

@@ -15,6 +15,10 @@ PHP                                                                        NEWS
     (devnexen)
   . Introduced MYSQLI_IS_MARIADB. (devnexen)
 
+- Readline:
+  . Fixed bug #81598 (Cannot input unicode characters in PHP 8 interactive
+    shell). (Nikita)
+
 - Reflection:
   . Fixed bug #81681 (ReflectionEnum throwing exceptions). (cmb)
 

Zend/tests/lc_ctype_inheritance.phpt

@@ -16,8 +16,8 @@ var_dump(setlocale(LC_CTYPE, "de_DE", "de-DE") !== false);
 var_dump(bin2hex(strtoupper("\xe4")));
 var_dump(preg_match('/\w/', "\xe4"));
 ?>
---EXPECT--
-string(1) "C"
+--EXPECTF--
+string(%d) "C%r(\.UTF-8)?%r"
 string(2) "e4"
 int(0)
 bool(true)

Zend/zend_operators.c

@@ -2662,6 +2662,15 @@ ZEND_API void zend_update_current_locale(void) /* {{{ */
 }
 /* }}} */
 
+ZEND_API void zend_reset_lc_ctype_locale(void)
+{
+	/* Use the C.UTF-8 locale so that readline can process UTF-8 input, while not interfering
+	 * with single-byte locale-dependent functions used by PHP. */
+	if (!setlocale(LC_CTYPE, "C.UTF-8")) {
+		setlocale(LC_CTYPE, "C");
+	}
+}
+
 static zend_always_inline void zend_str_tolower_impl(char *dest, const char *str, size_t length) /* {{{ */ {
 	unsigned char *p = (unsigned char*)str;
 	unsigned char *q = (unsigned char*)dest;

Zend/zend_operators.h

@@ -472,6 +472,8 @@ ZEND_API zend_long ZEND_FASTCALL zend_atol(const char *str, size_t str_len);
 
 ZEND_API void zend_update_current_locale(void);
 
+ZEND_API void zend_reset_lc_ctype_locale(void);
+
 /* The offset in bytes between the value and type fields of a zval */
 #define ZVAL_OFFSETOF_TYPE	\
 	(offsetof(zval, u1.type_info) - offsetof(zval, value))

ext/snmp/snmp.c

@@ -1999,7 +1999,7 @@ PHP_MINIT_FUNCTION(snmp)
 
 	init_snmp("snmpapp");
 	/* net-snmp corrupts the CTYPE locale during initialization. */
-	setlocale(LC_CTYPE, "C");
+	zend_reset_lc_ctype_locale();
 
 #ifdef NETSNMP_DS_LIB_DONT_PERSIST_STATE
 	/* Prevent update of the snmpapp.conf file */

ext/standard/basic_functions.c

@@ -528,6 +528,7 @@ PHP_RSHUTDOWN_FUNCTION(basic) /* {{{ */
 	 * to the value in startup environment */
 	if (BG(locale_changed)) {
 		setlocale(LC_ALL, "C");
+		zend_reset_lc_ctype_locale();
 		zend_update_current_locale();
 		if (BG(ctype_string)) {
 			zend_string_release_ex(BG(ctype_string), 0);

main/main.c

@@ -2087,6 +2087,7 @@ int php_module_startup(sapi_module_struct *sf, zend_module_entry *additional_mod
 	zuf.getenv_function = sapi_getenv;
 	zuf.resolve_path_function = php_resolve_path_for_zend;
 	zend_startup(&zuf);
+	zend_reset_lc_ctype_locale();
 	zend_update_current_locale();
 
 	zend_observer_startup();