fix: handle url search params encoding correctly (#628)

ada-url · Apr 11, 2024 · 4aeea61 · 4aeea61
1 parent c5c6575
commit 4aeea61
Show file tree

Hide file tree

Showing 2 changed files with 29 additions and 4 deletions.
diff --git a/include/ada/character_sets-inl.h b/include/ada/character_sets-inl.h
@@ -462,23 +462,23 @@ constexpr uint8_t WWW_FORM_URLENCODED_PERCENT_ENCODE[32] = {
     // 30     31     32     33     34     35     36     37
     0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00,
     // 38     39     3A     3B     3C     3D     3E     3F
-    0x00 | 0x00 | 0x00 | 0x00 | 0x10 | 0x00 | 0x40 | 0x80,
+    0x00 | 0x00 | 0x04 | 0x08 | 0x10 | 0x20 | 0x40 | 0x80,
     // 40     41     42     43     44     45     46     47
-    0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00,
+    0x01 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00,
     // 48     49     4A     4B     4C     4D     4E     4F
     0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00,
     // 50     51     52     53     54     55     56     57
     0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00,
     // 58     59     5A     5B     5C     5D     5E     5F
-    0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00,
+    0x00 | 0x00 | 0x00 | 0x08 | 0x00 | 0x20 | 0x40 | 0x00,
     // 60     61     62     63     64     65     66     67
     0x01 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00,
     // 68     69     6A     6B     6C     6D     6E     6F
     0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00,
     // 70     71     72     73     74     75     76     77
     0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00 | 0x00,
     // 78     79     7A     7B     7C     7D     7E     7F
-    0x00 | 0x00 | 0x00 | 0x08 | 0x00 | 0x20 | 0x40 | 0x80,
+    0x00 | 0x00 | 0x00 | 0x08 | 0x10 | 0x20 | 0x40 | 0x80,
     // 80     81     82     83     84     85     86     87
     0x01 | 0x02 | 0x04 | 0x08 | 0x10 | 0x20 | 0x40 | 0x80,
     // 88     89     8A     8B     8C     8D     8E     8F

diff --git a/tests/url_search_params.cpp b/tests/url_search_params.cpp
@@ -232,3 +232,28 @@ TEST(url_search_params, test_to_string_encoding) {
             "q1=foo&q2=foo+bar&q3=foo+bar&q4=foo%2Fbar");
   SUCCEED();
 }
+
+// https://github.com/cloudflare/workerd/issues/1777
+TEST(url_search_params, test_character_set) {
+  auto search_params = ada::url_search_params("key=value");
+
+  // - The application/x-www-form-urlencoded percent-encode set is the component
+  // percent-encode set and U+0021 (!), U+0027 (') to U+0029 RIGHT PARENTHESIS,
+  // inclusive, and U+007E (~).
+  // - The component percent-encode set is the userinfo percent-encode set and
+  // U+0024 ($) to U+0026 (&), inclusive, U+002B (+), and U+002C (,).
+  // - The userinfo percent-encode set is the path percent-encode set and U+002F
+  // (/), U+003A (:), U+003B (;), U+003D (=), U+0040 (@), U+005B ([) to U+005E
+  // (^), inclusive, and U+007C (|).
+  std::vector<char> unique_keys = {'/', ':', ';', '=', '@', '[',  ']', '^', '|',
+                                   '$', '&', '+', ',', '!', '\'', ')', '~'};
+  for (auto& unique_key : unique_keys) {
+    auto value = "value" + std::string(1, unique_key);
+    search_params.set("key", value);
+    // Getting should return the same thing.
+    ASSERT_EQ(search_params.get("key").value(), value);
+    // Stringified version should be percent encoded.
+    ASSERT_NE(search_params.to_string(), "key=" + value);
+  }
+  SUCCEED();
+}