[Fix] parse_url: don’t treat colon in path as port

ext/standard/tests/url/gh12703.phpt

@@ -0,0 +1,43 @@
+--TEST--
+GH-12703 parse_url() return false on absolute path containing ':' and no query string.
+--FILE--
+<?php
+var_dump(
+    parse_url('/page:1'),
+    parse_url('/page:1', \PHP_URL_SCHEME),
+
+    // Working example for context.
+    parse_url('/page:1?foo=bar'),
+
+    // Examples from similar GH issue (7890)
+    parse_url('//www.examle.com/foo:65535/'),
+    parse_url('//www.examle.com:8080/foo:65535/'),
+);
+
+?>
+--EXPECT--
+array(1) {
+  ["path"]=>
+  string(7) "/page:1"
+}
+NULL
+array(2) {
+  ["path"]=>
+  string(7) "/page:1"
+  ["query"]=>
+  string(7) "foo=bar"
+}
+array(2) {
+  ["host"]=>
+  string(14) "www.examle.com"
+  ["path"]=>
+  string(11) "/foo:65535/"
+}
+array(3) {
+  ["host"]=>
+  string(14) "www.examle.com"
+  ["port"]=>
+  int(8080)
+  ["path"]=>
+  string(11) "/foo:65535/"
+}

ext/standard/tests/url/url_utf8.phpt

@@ -0,0 +1,10 @@
+--TEST--
+Uri: hostnames should be preserved in Unicode form
+--FILE--
+<?php
+
+$parsed = parse_url('http://ουτοπία.δπθ.gr/');
+var_dump($parsed['host']);
+?>
+--EXPECT--
+string(24) "ουτοπία.δπθ.gr"

ext/standard/url.c

@@ -18,6 +18,12 @@
 #include <string.h>
 #include <ctype.h>
 #include <sys/types.h>
+#if defined(__APPLE__)
+#include <wchar.h>
+#endif
+#if defined(__APPLE__)
+#include <wctype.h>
+#endif
 
 #include "php.h"
 
@@ -58,6 +64,23 @@ static void parse_url_free_uri(void *uri)
 
 static void php_replace_controlchars(char *str, size_t len)
 {
+	#if defined(__APPLE__) 
+	{
+		ZEND_ASSERT(str != NULL);
+		wchar_t wbuf[len];
+		memset(wbuf, 0, sizeof(wbuf));
+		size_t wlen = mbstowcs(wbuf, str, len);
+
+		for (size_t i = 0; i < wlen; i++) {
+			if (iswcntrl(wbuf[i])) {
+				wbuf[i] = L'_';
+			}
+		}
+
+		wcstombs(str, wbuf, len);
+		return;
+	}
+	#endif
 	unsigned char *s = (unsigned char *)str;
 	unsigned char *e = (unsigned char *)str + len;
 
@@ -180,8 +203,11 @@ PHPAPI php_url *php_url_parse_ex2(char const *str, size_t length, bool *has_port
 		parse_port:
 		p = e + 1;
 		pp = p;
+
+		const char *first_slash = memchr(s, '/', ue - s);
 
 		while (pp < ue && pp - p < 6 && isdigit(*pp)) {
+			if (first_slash && pp >= first_slash) break;
 			pp++;
 		}
 
@@ -925,4 +951,4 @@ PHP_FUNCTION(get_headers)
 PHP_MINIT_FUNCTION(url)
 {
 	return php_uri_handler_register(&parse_url_uri_handler);
-}
+}