url: prevent pathname setter from erasing path of path-only URLs

This change prevents the pathname setter from erasing the path of path-only URLs as that would make them cannot-be-a-base URLs. The changes in all files except `src/node_url.cc` have been done by running: ```console git node wpt url ``` Fixes: #39059 Signed-off-by: Darshan Sen <darshan.sen@postman.com> PR-URL: #39060 Reviewed-By: Anna Henningsen <anna@addaleax.net> Reviewed-By: Khaidi Chu <i@2333.moe> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Tiancheng "Timothy" Gu <timothygu99@gmail.com>
nodejs · Jul 18, 2021 · 00cac65 · 00cac65
1 parent fdf625b
commit 00cac65
Show file tree

Hide file tree

Showing 5 changed files with 331 additions and 3 deletions.
diff --git a/src/node_url.cc b/src/node_url.cc
@@ -1477,6 +1477,9 @@ void URL::Parse(const char* input,
           if (ch != '/') {
             continue;
           }
+        } else if (has_state_override && !(url->flags & URL_FLAGS_HAS_HOST)) {
+          url->flags |= URL_FLAGS_HAS_PATH;
+          url->path.emplace_back("");
         }
         break;
       case kPath:

diff --git a/test/fixtures/wpt/README.md b/test/fixtures/wpt/README.md
@@ -22,7 +22,7 @@ Last update:
 - interfaces: https://github.com/web-platform-tests/wpt/tree/fcb671ed8b/interfaces
 - resources: https://github.com/web-platform-tests/wpt/tree/972ca5b669/resources
 - streams: https://github.com/web-platform-tests/wpt/tree/8f60d94439/streams
-- url: https://github.com/web-platform-tests/wpt/tree/1fcb39223d/url
+- url: https://github.com/web-platform-tests/wpt/tree/77d54aa9e0/url
 
 [Web Platform Tests]: https://github.com/web-platform-tests/wpt
 [`git node wpt`]: https://github.com/nodejs/node-core-utils/blob/main/docs/git-node.md#git-node-wpt
diff --git a/test/fixtures/wpt/url/resources/setters_tests.json b/test/fixtures/wpt/url/resources/setters_tests.json
@@ -990,6 +990,26 @@
                 "hostname": "test",
                 "port": "12"
             }
+        },
+        {
+            "comment": "Leading / is not stripped",
+            "href": "http://example.com/",
+            "new_value": "///bad.com",
+            "expected": {
+                "href": "http://example.com/",
+                "host": "example.com",
+                "hostname": "example.com"
+            }
+        },
+        {
+            "comment": "Leading / is not stripped",
+            "href": "sc://example.com/",
+            "new_value": "///bad.com",
+            "expected": {
+                "href": "sc:///",
+                "host": "",
+                "hostname": ""
+            }
         }
     ],
     "hostname": [
@@ -1345,6 +1365,26 @@
                 "hostname": "",
                 "pathname": "//p"
             }
+        },
+        {
+            "comment": "Leading / is not stripped",
+            "href": "http://example.com/",
+            "new_value": "///bad.com",
+            "expected": {
+                "href": "http://example.com/",
+                "host": "example.com",
+                "hostname": "example.com"
+            }
+        },
+        {
+            "comment": "Leading / is not stripped",
+            "href": "sc://example.com/",
+            "new_value": "///bad.com",
+            "expected": {
+                "href": "sc:///",
+                "host": "",
+                "hostname": ""
+            }
         }
     ],
     "port": [
@@ -1571,6 +1611,51 @@
                 "pathname": "me@example.net"
             }
         },
+        {
+            "comment": "Special URLs cannot have their paths erased",
+            "href": "file:///some/path",
+            "new_value": "",
+            "expected": {
+                "href": "file:///",
+                "pathname": "/"
+            }
+        },
+        {
+            "comment": "Non-special URLs can have their paths erased",
+            "href": "foo://somehost/some/path",
+            "new_value": "",
+            "expected": {
+                "href": "foo://somehost",
+                "pathname": ""
+            }
+        },
+        {
+            "comment": "Non-special URLs with an empty host can have their paths erased",
+            "href": "foo:///some/path",
+            "new_value": "",
+            "expected": {
+                "href": "foo://",
+                "pathname": ""
+            }
+        },
+        {
+            "comment": "Path-only URLs cannot have their paths erased",
+            "href": "foo:/some/path",
+            "new_value": "",
+            "expected": {
+                "href": "foo:/",
+                "pathname": "/"
+            }
+        },
+        {
+            "comment": "Path-only URLs always have an initial slash",
+            "href": "foo:/some/path",
+            "new_value": "test",
+            "expected": {
+                "href": "foo:/test",
+                "pathname": "/test"
+            }
+        },
         {
             "href": "unix:/run/foo.socket?timeout=10",
             "new_value": "/var/log/../run/bar.socket",
@@ -1667,6 +1752,24 @@
                 "pathname": "/%23"
             }
         },
+        {
+            "comment": "? doesn't mess up encoding",
+            "href": "http://example.net",
+            "new_value": "/?é",
+            "expected": {
+                "href": "http://example.net/%3F%C3%A9",
+                "pathname": "/%3F%C3%A9"
+            }
+        },
+        {
+            "comment": "# doesn't mess up encoding",
+            "href": "http://example.net",
+            "new_value": "/#é",
+            "expected": {
+                "href": "http://example.net/%23%C3%A9",
+                "pathname": "/%23%C3%A9"
+            }
+        },
         {
             "comment": "File URLs and (back)slashes",
             "href": "file://monkey/",