Escape backslash in special URL path components

Closes servo#468
ulf5 · Nov 6, 2018 · 20bebc2 · 20bebc2
1 parent a07eac0
commit 20bebc2
Show file tree

Hide file tree

Showing 3 changed files with 39 additions and 3 deletions.
diff --git a/percent_encoding/lib.rs b/percent_encoding/lib.rs
@@ -140,6 +140,11 @@ define_encode_set! {
     /// space, double quote ("), hash (#), inequality qualifiers (<), (>), backtick (`),
     /// question mark (?), and curly brackets ({), (}), percent sign (%), forward slash (/) are
     /// encoded.
+    ///
+    /// # Note
+    ///
+    /// For [special URLs](https://url.spec.whatwg.org/#is-special), the backslash (\) character should
+    /// additionally be escaped, but that is *not* included in this encode set.
     pub PATH_SEGMENT_ENCODE_SET = [DEFAULT_ENCODE_SET] | {'%', '/'}
 }
 

diff --git a/src/parser.rs b/src/parser.rs
@@ -19,9 +19,24 @@ use host::{Host, HostInternal};
 use percent_encoding::{
     utf8_percent_encode, percent_encode,
     SIMPLE_ENCODE_SET, DEFAULT_ENCODE_SET, USERINFO_ENCODE_SET, QUERY_ENCODE_SET,
-    PATH_SEGMENT_ENCODE_SET
+    PATH_SEGMENT_ENCODE_SET, EncodeSet
 };
 
+// The backslash (\) character is treated as a path separator in special URLs
+// so it needs to be additionally escaped in that case.
+#[derive(Clone)]
+struct SPECIAL_PATH_SEGMENT_ENCODE_SET;
+
+impl EncodeSet for SPECIAL_PATH_SEGMENT_ENCODE_SET {
+    #[inline]
+    fn contains(&self, byte: u8) -> bool {
+        match byte {
+            b'\\' => true,
+            _ => PATH_SEGMENT_ENCODE_SET.contains(byte)
+        }
+    }
+}
+
 pub type ParseResult<T> = Result<T, ParseError>;
 
 macro_rules! simple_enum_error {
@@ -1011,8 +1026,13 @@ impl<'a> Parser<'a> {
                     _ => {
                         self.check_url_code_point(c, &input);
                         if self.context == Context::PathSegmentSetter {
-                            self.serialization.extend(utf8_percent_encode(
-                                utf8_c, PATH_SEGMENT_ENCODE_SET));
+                            if scheme_type.is_special() {
+                                self.serialization.extend(utf8_percent_encode(
+                                    utf8_c, SPECIAL_PATH_SEGMENT_ENCODE_SET));
+                            } else {
+                                self.serialization.extend(utf8_percent_encode(
+                                    utf8_c, PATH_SEGMENT_ENCODE_SET));
+                            }
                         } else {
                             self.serialization.extend(utf8_percent_encode(
                                 utf8_c, DEFAULT_ENCODE_SET));

diff --git a/tests/unit.rs b/tests/unit.rs
@@ -109,6 +109,17 @@ fn new_directory_paths() {
     }
 }
 
+#[test]
+fn path_backslash_fun() {
+    let mut special_url = "http://foobar.com".parse::<Url>().unwrap();
+    special_url.path_segments_mut().unwrap().push("foo\\bar");
+    assert_eq!(special_url.as_str(), "http://foobar.com/foo%5Cbar");
+
+    let mut nonspecial_url = "thing://foobar.com".parse::<Url>().unwrap();
+    nonspecial_url.path_segments_mut().unwrap().push("foo\\bar");
+    assert_eq!(nonspecial_url.as_str(), "thing://foobar.com/foo\\bar");
+}
+
 #[test]
 fn from_str() {
     assert!("http://testing.com/this".parse::<Url>().is_ok());