Fix #80266: parse_url silently drops port number 0

As of commit 81b2f3e[1], `parse_url()` accepts URLs with a zero port,
but does not report that port, what is wrong in hindsight.

Since the port number is stored as `unsigned short` there is no way to
distinguish between port zero and no port.  For BC reasons, we thus
introduce `parse_url_ex2()` which accepts an output parameter that
allows that distinction, and use the new function to fix the behavior.

The introduction of `parse_url_ex2()` has been suggested by Nikita.

[1] <http://git.php.net/?p=php-src.git;a=commit;h=81b2f3e5d9fcdffd87a4fcd12bd8c708a97091e1>

Closes GH-6399.

6 files changed, 26 insertions, 8 deletions

diff --git a/NEWS b/NEWS
index f6963784d8..4513508271 100644
--- a/NEWS
+++ b/NEWS
@@ -29,6 +29,9 @@ PHP                                                                        NEWS
   . Fixed bug #70461 (disable md5 code when it is not supported in net-snmp).
     (Alexander Bergmann, cmb)
 
+- Standard:
+  . Fixed bug #80266 (parse_url silently drops port number 0). (cmb, Nikita)
+
 29 Oct 2020, PHP 7.3.24
 
 - Core:
diff --git a/ext/standard/tests/url/parse_url_basic_001.phpt b/ext/standard/tests/url/parse_url_basic_001.phpt
index 063fc28832..7ffd001856 100644
--- a/ext/standard/tests/url/parse_url_basic_001.phpt
+++ b/ext/standard/tests/url/parse_url_basic_001.phpt
@@ -859,11 +859,13 @@ echo "Done";
   string(3) "%:x"
 }
 
---> https://example.com:0/: array(3) {
+--> https://example.com:0/: array(4) {
   ["scheme"]=>
   string(5) "https"
   ["host"]=>
   string(11) "example.com"
+  ["port"]=>
+  int(0)
   ["path"]=>
   string(1) "/"
 }
diff --git a/ext/standard/tests/url/parse_url_basic_004.phpt b/ext/standard/tests/url/parse_url_basic_004.phpt
index 042daefeda..533d304ee7 100644
--- a/ext/standard/tests/url/parse_url_basic_004.phpt
+++ b/ext/standard/tests/url/parse_url_basic_004.phpt
@@ -112,7 +112,7 @@ echo "Done";
 --> /   : NULL
 --> /rest/Users?filter={"id":"123"}   : NULL
 --> %:x   : NULL
---> https://example.com:0/   : NULL
+--> https://example.com:0/   : int(0)
 --> http:///blah.com   : bool(false)
 --> http://:80   : bool(false)
 --> http://user@:80   : bool(false)
diff --git a/ext/standard/tests/url/parse_url_unterminated.phpt b/ext/standard/tests/url/parse_url_unterminated.phpt
index 8af50dbe28..8f46eb3dbe 100644
--- a/ext/standard/tests/url/parse_url_unterminated.phpt
+++ b/ext/standard/tests/url/parse_url_unterminated.phpt
@@ -861,11 +861,13 @@ echo "Done";
   string(3) "%:x"
 }
 
---> https://example.com:0/: array(3) {
+--> https://example.com:0/: array(4) {
   ["scheme"]=>
   string(5) "https"
   ["host"]=>
   string(11) "example.com"
+  ["port"]=>
+  int(0)
   ["path"]=>
   string(1) "/"
 }
diff --git a/ext/standard/url.c b/ext/standard/url.c
index fde4ff5377..98dc0f786d 100644
--- a/ext/standard/url.c
+++ b/ext/standard/url.c
@@ -102,14 +102,21 @@ static const char *binary_strcspn(const char *s, const char *e, const char *char
 	return e;
 }
 
-/* {{{ php_url_parse
- */
 PHPAPI php_url *php_url_parse_ex(char const *str, size_t length)
 {
+	zend_bool has_port;
+	return php_url_parse_ex2(str, length, &has_port);
+}
+
+/* {{{ php_url_parse_ex2
+ */
+PHPAPI php_url *php_url_parse_ex2(char const *str, size_t length, zend_bool *has_port)
+{
 	char port_buf[6];
 	php_url *ret = ecalloc(1, sizeof(php_url));
 	char const *s, *e, *p, *pp, *ue;
 
+	*has_port = 0;
 	s = str;
 	ue = s + length;
 
@@ -199,6 +206,7 @@ PHPAPI php_url *php_url_parse_ex(char const *str, size_t length)
 			port_buf[pp - p] = '\0';
 			port = ZEND_STRTOL(port_buf, &end, 10);
 			if (port >= 0 && port <= 65535 && end != port_buf) {
+				*has_port = 1;
 				ret->port = (unsigned short) port;
 				if (s + 1 < ue && *s == '/' && *(s + 1) == '/') { /* relative-scheme URL */
 				    s += 2;
@@ -264,6 +272,7 @@ parse_host:
 				port_buf[e - p] = '\0';
 				port = ZEND_STRTOL(port_buf, &end, 10);
 				if (port >= 0 && port <= 65535 && end != port_buf) {
+					*has_port = 1;
 					ret->port = (unsigned short)port;
 				} else {
 					php_url_free(ret);
@@ -332,6 +341,7 @@ PHP_FUNCTION(parse_url)
 	php_url *resource;
 	zend_long key = -1;
 	zval tmp;
+	zend_bool has_port;
 
 	ZEND_PARSE_PARAMETERS_START(1, 2)
 		Z_PARAM_STRING(str, str_len)
@@ -339,7 +349,7 @@ PHP_FUNCTION(parse_url)
 		Z_PARAM_LONG(key)
 	ZEND_PARSE_PARAMETERS_END();
 
-	resource = php_url_parse_ex(str, str_len);
+	resource = php_url_parse_ex2(str, str_len, &has_port);
 	if (resource == NULL) {
 		/* @todo Find a method to determine why php_url_parse_ex() failed */
 		RETURN_FALSE;
@@ -354,7 +364,7 @@ PHP_FUNCTION(parse_url)
 				if (resource->host != NULL) RETVAL_STR_COPY(resource->host);
 				break;
 			case PHP_URL_PORT:
-				if (resource->port != 0) RETVAL_LONG(resource->port);
+				if (has_port) RETVAL_LONG(resource->port);
 				break;
 			case PHP_URL_USER:
 				if (resource->user != NULL) RETVAL_STR_COPY(resource->user);
@@ -390,7 +400,7 @@ PHP_FUNCTION(parse_url)
 		ZVAL_STR_COPY(&tmp, resource->host);
 		zend_hash_add_new(Z_ARRVAL_P(return_value), ZSTR_KNOWN(ZEND_STR_HOST), &tmp);
 	}
-	if (resource->port != 0) {
+	if (has_port) {
 		ZVAL_LONG(&tmp, resource->port);
 		zend_hash_add_new(Z_ARRVAL_P(return_value), ZSTR_KNOWN(ZEND_STR_PORT), &tmp);
 	}
diff --git a/ext/standard/url.h b/ext/standard/url.h
index ec925a2bd0..0e187065df 100644
--- a/ext/standard/url.h
+++ b/ext/standard/url.h
@@ -33,6 +33,7 @@ typedef struct php_url {
 PHPAPI void php_url_free(php_url *theurl);
 PHPAPI php_url *php_url_parse(char const *str);
 PHPAPI php_url *php_url_parse_ex(char const *str, size_t length);
+PHPAPI php_url *php_url_parse_ex2(char const *str, size_t length, zend_bool *has_port);
 PHPAPI size_t php_url_decode(char *str, size_t len); /* return value: length of decoded string */
 PHPAPI size_t php_raw_url_decode(char *str, size_t len); /* return value: length of decoded string */
 PHPAPI zend_string *php_url_encode(char const *s, size_t len);