Segment delimiters and path trimming.

Author: o0Ignition0o

src/lib.rs

@@ -1397,8 +1397,12 @@ impl Url {
                 }
                 parser.parse_cannot_be_a_base_path(parser::Input::new(path));
             } else {
+                let path_start = parser.serialization.len();
                 let mut has_host = true; // FIXME
                 parser.parse_path_start(scheme_type, &mut has_host, parser::Input::new(path));
+                if scheme_type.is_file() {
+                    parser::trim_path(&mut parser.serialization, path_start);
+                }
             }
         });
         self.restore_after_path(old_after_path_pos, &after_path);

src/parser.rs

@@ -798,8 +798,24 @@ impl<'a> Parser<'a> {
                 self.serialization.push_str(before_query);
                 // FIXME spec says just "remove last entry", not the "pop" algorithm
                 self.pop_path(scheme_type, base_url.path_start as usize);
-                let remaining =
-                    self.parse_path(scheme_type, &mut true, base_url.path_start as usize, input);
+                // A special url always has a path.
+                // A path always starts with '/'
+                if self.serialization.len() == base_url.path_start as usize {
+                    if SchemeType::from(base_url.scheme()).is_special() || !input.is_empty() {
+                        self.serialization.push('/');
+                    }
+                }
+                let remaining = match input.split_first() {
+                    (Some('/'), remaining) => self.parse_path(
+                        scheme_type,
+                        &mut true,
+                        base_url.path_start as usize,
+                        remaining,
+                    ),
+                    _ => {
+                        self.parse_path(scheme_type, &mut true, base_url.path_start as usize, input)
+                    }
+                };
                 self.with_query_and_fragment(
                     scheme_type,
                     base_url.scheme_end,
@@ -1427,12 +1443,14 @@ impl<'a> Parser<'a> {
 
 // Trim path start forward slashes when no authority is present
 // https://github.com/whatwg/url/issues/232
-fn trim_path(serialization: &mut String, path_start: usize) {
+pub fn trim_path(serialization: &mut String, path_start: usize) {
     let path = serialization.split_off(path_start);
     if path.starts_with("/") {
         let mut trimmed_path = "/".to_string();
         trimmed_path.push_str(path.trim_start_matches("/"));
         serialization.push_str(&trimmed_path);
+    } else {
+        serialization.push_str(&path);
     }
 }
 

src/quirks.rs

@@ -236,10 +236,10 @@ pub fn set_pathname(url: &mut Url, new_pathname: &str) {
     if url.cannot_be_a_base() {
         return;
     }
-    if !SchemeType::from(url.scheme()).is_special()
-            || Some('/') == new_pathname.chars().nth(0)
-            // \\ is a segment delimiter for 'special' URLs"
-            || Some('\\') == new_pathname.chars().nth(0)
+    if Some('/') == new_pathname.chars().nth(0)
+        || SchemeType::from(url.scheme()).is_special()
+        // \ is a segment delimiter for 'special' URLs"
+        && Some('\\') == new_pathname.chars().nth(0)
     {
         url.set_path(new_pathname)
     } else {