LibURL: Use UTF-8 for percent encoding URL fragments

(cherry picked from commit c10cb8ac8d6c897c8fb184d5f0c2a09b8f699b1d)
nico · Nov 15, 2024 · 0cf4dd8 · 0cf4dd8
1 parent 0206d2d
commit 0cf4dd8
Show file tree

Hide file tree

Showing 2 changed files with 39 additions and 10 deletions.
diff --git a/Tests/LibURL/TestURL.cpp b/Tests/LibURL/TestURL.cpp
@@ -339,11 +339,38 @@ TEST_CASE(unicode)
 
 TEST_CASE(query_with_non_ascii)
 {
-    URL::URL url { "http://example.com/?utf8=✓"sv };
-    EXPECT(url.is_valid());
-    EXPECT_EQ(url.serialize_path(), "/"sv);
-    EXPECT_EQ(url.query(), "utf8=%E2%9C%93");
-    EXPECT(!url.fragment().has_value());
+    {
+        URL::URL url = URL::Parser::basic_parse("http://example.com/?utf8=✓"sv);
+        EXPECT(url.is_valid());
+        EXPECT_EQ(url.serialize_path(), "/"sv);
+        EXPECT_EQ(url.query(), "utf8=%E2%9C%93");
+        EXPECT(!url.fragment().has_value());
+    }
+    {
+        URL::URL url = URL::Parser::basic_parse("http://example.com/?shift_jis=✓"sv, {}, nullptr, {}, "shift_jis"sv);
+        EXPECT(url.is_valid());
+        EXPECT_EQ(url.serialize_path(), "/"sv);
+        EXPECT_EQ(url.query(), "shift_jis=%26%2310003%3B");
+        EXPECT(!url.fragment().has_value());
+    }
+}
+
+TEST_CASE(fragment_with_non_ascii)
+{
+    {
+        URL::URL url = URL::Parser::basic_parse("http://example.com/#✓"sv);
+        EXPECT(url.is_valid());
+        EXPECT_EQ(url.serialize_path(), "/"sv);
+        EXPECT(!url.query().has_value());
+        EXPECT_EQ(url.fragment(), "%E2%9C%93");
+    }
+    {
+        URL::URL url = URL::Parser::basic_parse("http://example.com/#✓"sv, {}, nullptr, {}, "shift_jis"sv);
+        EXPECT(url.is_valid());
+        EXPECT_EQ(url.serialize_path(), "/"sv);
+        EXPECT(!url.query().has_value());
+        EXPECT_EQ(url.fragment(), "%E2%9C%93");
+    }
 }
 
 TEST_CASE(complete_file_url_with_base)

diff --git a/Userland/Libraries/LibURL/Parser.cpp b/Userland/Libraries/LibURL/Parser.cpp
@@ -1688,10 +1688,12 @@ URL Parser::basic_parse(StringView raw_input, Optional<URL> const& base_url, URL
             break;
         // -> query state, https://url.spec.whatwg.org/#query-state
         case State::Query:
-            // FIXME: 1. If encoding is not UTF-8 and one of the following is true:
-            //           * url is not special
-            //           * url’s scheme is "ws" or "wss"
-            //        then set encoding to UTF-8.
+            // 1. If encoding is not UTF-8 and one of the following is true:
+            //     * url is not special
+            //     * url’s scheme is "ws" or "wss"
+            //  then set encoding to UTF-8.
+            if (!url->is_special() || url->m_data->scheme == "ws" || url->m_data->scheme == "wss")
+                encoder = TextCodec::encoder_for("utf-8"sv);
 
             // 2. If one of the following is true:
             //    * state override is not given and c is U+0023 (#)
@@ -1746,7 +1748,7 @@ URL Parser::basic_parse(StringView raw_input, Optional<URL> const& base_url, URL
                 // NOTE: The percent-encode is done on EOF on the entire buffer.
                 buffer.append_code_point(code_point);
             } else {
-                url->m_data->fragment = percent_encode_after_encoding(*encoder, buffer.string_view(), PercentEncodeSet::Fragment);
+                url->m_data->fragment = percent_encode_after_encoding(*TextCodec::encoder_for("utf-8"sv), buffer.string_view(), PercentEncodeSet::Fragment);
                 buffer.clear();
             }
             break;