Fixed #78385: Distinguish absent/empty query/fragment

http://example.com/foo   => query = null, fragment = null
http://example.com/foo?  => query = "",   fragment = null
http://example.com/foo#  => query = null, fragment = ""
http://example.com/foo?# => query = "",   fragment = ""

Closes GH-5078.

8 files changed, 102 insertions, 34 deletions

diff --git a/NEWS b/NEWS
index d46a30a379..2d0556cbab 100644
--- a/NEWS
+++ b/NEWS
@@ -92,6 +92,8 @@ PHP                                                                        NEWS
     (peter279k)
   . Fixed bug #76859 (stream_get_line skips data if used with data-generating
     filter). (kkopachev)
+  . Fixed bug #78385 (parse_url() does not include 'query' when question mark
+    is the last char). (Islam Israfilov)
 
 - tidy:
   . Removed the unused $use_include_path parameter from tidy_repair_string().
diff --git a/UPGRADING b/UPGRADING
index 65c8ac2720..1d616a53fd 100644
--- a/UPGRADING
+++ b/UPGRADING
@@ -301,6 +301,14 @@ PHP 8.0 UPGRADE NOTES
   . FILTER_SANITIZE_MAGIC_QUOTES has been removed.
   . Calling implode() with parameters in a reverse order ($pieces, $glue) is no
     longer supported.
+  . parse_url() will now distinguish absent and empty queries and fragments:
+
+        http://example.com/foo   => query = null, fragment = null
+        http://example.com/foo?  => query = "",   fragment = null
+        http://example.com/foo#  => query = null, fragment = ""
+        http://example.com/foo?# => query = "",   fragment = ""
+
+    Previously all cases resulted in query and fragment being null.
 
 - tidy:
   . The $use_include_path parameter, which was not used internally, has been
diff --git a/ext/standard/tests/strings/url_t.phpt b/ext/standard/tests/strings/url_t.phpt
index 79ff3bc4a8..1ff7baff90 100644
--- a/ext/standard/tests/strings/url_t.phpt
+++ b/ext/standard/tests/strings/url_t.phpt
@@ -213,30 +213,36 @@ $sample_urls = array (
   string(10) "/index.php"
 }
 
---> www.php.net/?: array(1) {
+--> www.php.net/?: array(2) {
   ["path"]=>
   string(12) "www.php.net/"
+  ["query"]=>
+  string(0) ""
 }
 
---> www.php.net:80/?: array(3) {
+--> www.php.net:80/?: array(4) {
   ["host"]=>
   string(11) "www.php.net"
   ["port"]=>
   int(80)
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
 }
 
---> http://www.php.net/?: array(3) {
+--> http://www.php.net/?: array(4) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
   string(11) "www.php.net"
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
 }
 
---> http://www.php.net:80/?: array(4) {
+--> http://www.php.net:80/?: array(5) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
@@ -245,6 +251,8 @@ $sample_urls = array (
   int(80)
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
 }
 
 --> http://www.php.net:80/index.php: array(4) {
@@ -359,7 +367,7 @@ $sample_urls = array (
   string(10) "/index.php"
 }
 
---> http://www.php.net:80/index.php?: array(4) {
+--> http://www.php.net:80/index.php?: array(5) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
@@ -368,6 +376,8 @@ $sample_urls = array (
   int(80)
   ["path"]=>
   string(10) "/index.php"
+  ["query"]=>
+  string(0) ""
 }
 
 --> http://www.php.net:80/#foo: array(5) {
@@ -383,7 +393,7 @@ $sample_urls = array (
   string(3) "foo"
 }
 
---> http://www.php.net:80/?#: array(4) {
+--> http://www.php.net:80/?#: array(6) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
@@ -392,6 +402,10 @@ $sample_urls = array (
   int(80)
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
+  ["fragment"]=>
+  string(0) ""
 }
 
 --> http://www.php.net:80/?test=1: array(5) {
diff --git a/ext/standard/tests/url/parse_url_basic_001.phpt b/ext/standard/tests/url/parse_url_basic_001.phpt
index d59f080fbf..3977a3fc2e 100644
--- a/ext/standard/tests/url/parse_url_basic_001.phpt
+++ b/ext/standard/tests/url/parse_url_basic_001.phpt
@@ -144,30 +144,36 @@ echo "Done";
   string(10) "/index.php"
 }
 
---> www.php.net/?: array(1) {
+--> www.php.net/?: array(2) {
   ["path"]=>
   string(12) "www.php.net/"
+  ["query"]=>
+  string(0) ""
 }
 
---> www.php.net:80/?: array(3) {
+--> www.php.net:80/?: array(4) {
   ["host"]=>
   string(11) "www.php.net"
   ["port"]=>
   int(80)
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
 }
 
---> http://www.php.net/?: array(3) {
+--> http://www.php.net/?: array(4) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
   string(11) "www.php.net"
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
 }
 
---> http://www.php.net:80/?: array(4) {
+--> http://www.php.net:80/?: array(5) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
@@ -176,6 +182,8 @@ echo "Done";
   int(80)
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
 }
 
 --> http://www.php.net:80/index.php: array(4) {
@@ -290,7 +298,7 @@ echo "Done";
   string(10) "/index.php"
 }
 
---> http://www.php.net:80/index.php?: array(4) {
+--> http://www.php.net:80/index.php?: array(5) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
@@ -299,6 +307,8 @@ echo "Done";
   int(80)
   ["path"]=>
   string(10) "/index.php"
+  ["query"]=>
+  string(0) ""
 }
 
 --> http://www.php.net:80/#foo: array(5) {
@@ -314,7 +324,7 @@ echo "Done";
   string(3) "foo"
 }
 
---> http://www.php.net:80/?#: array(4) {
+--> http://www.php.net:80/?#: array(6) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
@@ -323,6 +333,10 @@ echo "Done";
   int(80)
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
+  ["fragment"]=>
+  string(0) ""
 }
 
 --> http://www.php.net:80/?test=1: array(5) {
@@ -731,11 +745,13 @@ echo "Done";
   string(4) "/:80"
 }
 
---> http://x:?: array(2) {
+--> http://x:?: array(3) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
   string(1) "x"
+  ["query"]=>
+  string(0) ""
 }
 
 --> x:blah.com: array(2) {
@@ -754,18 +770,22 @@ echo "Done";
 
 --> x://::abc/?: bool(false)
 
---> http://::?: array(2) {
+--> http://::?: array(3) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
   string(1) ":"
+  ["query"]=>
+  string(0) ""
 }
 
---> http://::#: array(2) {
+--> http://::#: array(3) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
   string(1) ":"
+  ["fragment"]=>
+  string(0) ""
 }
 
 --> x://::6.5: array(3) {
diff --git a/ext/standard/tests/url/parse_url_basic_008.phpt b/ext/standard/tests/url/parse_url_basic_008.phpt
index 761eb61179..142da3d561 100644
--- a/ext/standard/tests/url/parse_url_basic_008.phpt
+++ b/ext/standard/tests/url/parse_url_basic_008.phpt
@@ -38,10 +38,10 @@ echo "Done";
 --> http://www.php.net:80   : NULL
 --> http://www.php.net:80/   : NULL
 --> http://www.php.net/index.php   : NULL
---> www.php.net/?   : NULL
---> www.php.net:80/?   : NULL
---> http://www.php.net/?   : NULL
---> http://www.php.net:80/?   : NULL
+--> www.php.net/?   : string(0) ""
+--> www.php.net:80/?   : string(0) ""
+--> http://www.php.net/?   : string(0) ""
+--> http://www.php.net:80/?   : string(0) ""
 --> http://www.php.net:80/index.php   : NULL
 --> http://www.php.net:80/foo/bar/index.php   : NULL
 --> http://www.php.net:80/this/is/a/very/deep/directory/structure/and/file.php   : NULL
@@ -52,9 +52,9 @@ echo "Done";
 --> http://www.php.net:80/this/../a/../deep/directory/   : NULL
 --> http://www.php.net:80/this/is/a/very/deep/directory/../file.php   : NULL
 --> http://www.php.net:80/index.php   : NULL
---> http://www.php.net:80/index.php?   : NULL
+--> http://www.php.net:80/index.php?   : string(0) ""
 --> http://www.php.net:80/#foo   : NULL
---> http://www.php.net:80/?#   : NULL
+--> http://www.php.net:80/?#   : string(0) ""
 --> http://www.php.net:80/?test=1   : string(6) "test=1"
 --> http://www.php.net/?test=1&   : string(7) "test=1&"
 --> http://www.php.net:80/?&   : string(1) "&"
@@ -91,11 +91,11 @@ echo "Done";
 --> gg:9130731   : NULL
 --> http://user:@pass@host/path?argument?value#etc   : string(14) "argument?value"
 --> http://10.10.10.10/:80   : NULL
---> http://x:?   : NULL
+--> http://x:?   : string(0) ""
 --> x:blah.com   : NULL
 --> x:/blah.com   : NULL
 --> x://::abc/?   : bool(false)
---> http://::?   : NULL
+--> http://::?   : string(0) ""
 --> http://::#   : NULL
 --> x://::6.5   : NULL
 --> http://?:/   : bool(false)
diff --git a/ext/standard/tests/url/parse_url_basic_009.phpt b/ext/standard/tests/url/parse_url_basic_009.phpt
index f08787761d..3634046de2 100644
--- a/ext/standard/tests/url/parse_url_basic_009.phpt
+++ b/ext/standard/tests/url/parse_url_basic_009.phpt
@@ -54,7 +54,7 @@ echo "Done";
 --> http://www.php.net:80/index.php   : NULL
 --> http://www.php.net:80/index.php?   : NULL
 --> http://www.php.net:80/#foo   : string(3) "foo"
---> http://www.php.net:80/?#   : NULL
+--> http://www.php.net:80/?#   : string(0) ""
 --> http://www.php.net:80/?test=1   : NULL
 --> http://www.php.net/?test=1&   : NULL
 --> http://www.php.net:80/?&   : NULL
@@ -96,7 +96,7 @@ echo "Done";
 --> x:/blah.com   : NULL
 --> x://::abc/?   : bool(false)
 --> http://::?   : NULL
---> http://::#   : NULL
+--> http://::#   : string(0) ""
 --> x://::6.5   : NULL
 --> http://?:/   : bool(false)
 --> http://@?:/   : bool(false)
diff --git a/ext/standard/tests/url/parse_url_unterminated.phpt b/ext/standard/tests/url/parse_url_unterminated.phpt
index 94975ee889..7a0cba8ada 100644
--- a/ext/standard/tests/url/parse_url_unterminated.phpt
+++ b/ext/standard/tests/url/parse_url_unterminated.phpt
@@ -146,30 +146,36 @@ echo "Done";
   string(10) "/index.php"
 }
 
---> www.php.net/?: array(1) {
+--> www.php.net/?: array(2) {
   ["path"]=>
   string(12) "www.php.net/"
+  ["query"]=>
+  string(0) ""
 }
 
---> www.php.net:80/?: array(3) {
+--> www.php.net:80/?: array(4) {
   ["host"]=>
   string(11) "www.php.net"
   ["port"]=>
   int(80)
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
 }
 
---> http://www.php.net/?: array(3) {
+--> http://www.php.net/?: array(4) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
   string(11) "www.php.net"
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
 }
 
---> http://www.php.net:80/?: array(4) {
+--> http://www.php.net:80/?: array(5) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
@@ -178,6 +184,8 @@ echo "Done";
   int(80)
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
 }
 
 --> http://www.php.net:80/index.php: array(4) {
@@ -292,7 +300,7 @@ echo "Done";
   string(10) "/index.php"
 }
 
---> http://www.php.net:80/index.php?: array(4) {
+--> http://www.php.net:80/index.php?: array(5) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
@@ -301,6 +309,8 @@ echo "Done";
   int(80)
   ["path"]=>
   string(10) "/index.php"
+  ["query"]=>
+  string(0) ""
 }
 
 --> http://www.php.net:80/#foo: array(5) {
@@ -316,7 +326,7 @@ echo "Done";
   string(3) "foo"
 }
 
---> http://www.php.net:80/?#: array(4) {
+--> http://www.php.net:80/?#: array(6) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
@@ -325,6 +335,10 @@ echo "Done";
   int(80)
   ["path"]=>
   string(1) "/"
+  ["query"]=>
+  string(0) ""
+  ["fragment"]=>
+  string(0) ""
 }
 
 --> http://www.php.net:80/?test=1: array(5) {
@@ -733,11 +747,13 @@ echo "Done";
   string(4) "/:80"
 }
 
---> http://x:?: array(2) {
+--> http://x:?: array(3) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
   string(1) "x"
+  ["query"]=>
+  string(0) ""
 }
 
 --> x:blah.com: array(2) {
@@ -756,18 +772,22 @@ echo "Done";
 
 --> x://::abc/?: bool(false)
 
---> http://::?: array(2) {
+--> http://::?: array(3) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
   string(1) ":"
+  ["query"]=>
+  string(0) ""
 }
 
---> http://::#: array(2) {
+--> http://::#: array(3) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
   string(1) ":"
+  ["fragment"]=>
+  string(0) ""
 }
 
 --> x://::6.5: array(3) {
diff --git a/ext/standard/url.c b/ext/standard/url.c
index c2d9340e1c..c73818f08e 100644
--- a/ext/standard/url.c
+++ b/ext/standard/url.c
@@ -291,6 +291,8 @@ PHPAPI php_url *php_url_parse_ex(char const *str, size_t length)
 		if (p < e) {
 			ret->fragment = zend_string_init(p, (e - p), 0);
 			php_replace_controlchars_ex(ZSTR_VAL(ret->fragment), ZSTR_LEN(ret->fragment));
+		} else {
+			ret->fragment = ZSTR_EMPTY_ALLOC();
 		}
 		e = p-1;
 	}
@@ -301,6 +303,8 @@ PHPAPI php_url *php_url_parse_ex(char const *str, size_t length)
 		if (p < e) {
 			ret->query = zend_string_init(p, (e - p), 0);
 			php_replace_controlchars_ex(ZSTR_VAL(ret->query), ZSTR_LEN(ret->query));
+		} else {
+			ret->query = ZSTR_EMPTY_ALLOC();
 		}
 		e = p-1;
 	}