Normalize URL paths: convert /.//p, /..//p, and //p to p

servo · Jun 18, 2024 · d4bea43 · d4bea43
1 parent 3d6dbbb
commit d4bea43
Show file tree

Hide file tree

Showing 2 changed files with 26 additions and 3 deletions.
diff --git a/url/src/lib.rs b/url/src/lib.rs
@@ -1723,6 +1723,32 @@ impl Url {
                 );
             }
         });
+
+        // To handle cases like <non-spec:/> set pathname to </..//p>
+        // For instance, //p should be converted to /..//p here
+        // At this point, we would get "non-spec://p" for serialization
+        // and "/..//p" for path
+        if let Some(path_pos) = path.rfind("//") {
+            if let Some(serialization_pos) = self.serialization.rfind("//") {
+                const PATH_INCREMENT: usize = 2; // length of "/."
+
+                if path_pos + PATH_INCREMENT <= path.len()
+                    && serialization_pos + PATH_INCREMENT <= self.serialization.len()
+                {
+                    let rest_path = &path[(path_pos + PATH_INCREMENT)..];
+                    let rest_serialization =
+                        &self.serialization[(serialization_pos + PATH_INCREMENT)..];
+
+                    // rest should be the same
+                    if rest_path == rest_serialization {
+                        self.serialization
+                            .replace_range(serialization_pos.., &format!("/.//{}", rest_path));
+                        self.path_start += PATH_INCREMENT as u32;
+                    }
+                }
+            }
+        }
+
         self.restore_after_path(old_after_path_pos, &after_path);
     }
 

diff --git a/url/tests/expected_failures.txt b/url/tests/expected_failures.txt
@@ -44,7 +44,4 @@
 <file://monkey/> set pathname to <\\\\>
 <file:///unicorn> set pathname to <//\\/>
 <file:///unicorn> set pathname to <//monkey/..//>
-<non-spec:/> set pathname to </.//p>
-<non-spec:/> set pathname to </..//p>
-<non-spec:/> set pathname to <//p>
 <non-spec:/.//> set pathname to <p>