Bug 1032511 - URLSearchParam should decode any non-unicode input correctly, r=hsivonen

dom/base/URLSearchParams.cpp

@@ -5,6 +5,7 @@
 
 #include "URLSearchParams.h"
 #include "mozilla/dom/URLSearchParamsBinding.h"
+#include "mozilla/dom/EncodingUtils.h"
 
 namespace mozilla {
 namespace dom {
@@ -98,30 +99,31 @@ URLSearchParams::ParseInput(const nsACString& aInput,
       name.Assign(string);
     }
 
-    nsAutoCString decodedName;
+    nsAutoString decodedName;
     DecodeString(name, decodedName);
 
-    nsAutoCString decodedValue;
+    nsAutoString decodedValue;
     DecodeString(value, decodedValue);
 
-    AppendInternal(NS_ConvertUTF8toUTF16(decodedName),
-                   NS_ConvertUTF8toUTF16(decodedValue));
+    AppendInternal(decodedName, decodedValue);
   }
 
   NotifyObservers(aObserver);
 }
 
 void
-URLSearchParams::DecodeString(const nsACString& aInput, nsACString& aOutput)
+URLSearchParams::DecodeString(const nsACString& aInput, nsAString& aOutput)
 {
   nsACString::const_iterator start, end;
   aInput.BeginReading(start);
   aInput.EndReading(end);
 
+  nsCString unescaped;
+
   while (start != end) {
     // replace '+' with U+0020
     if (*start == '+') {
-      aOutput.Append(' ');
+      unescaped.Append(' ');
       ++start;
       continue;
     }
@@ -148,20 +150,62 @@ URLSearchParams::DecodeString(const nsACString& aInput, nsACString& aOutput)
 
       if (first != end && second != end &&
           ASCII_HEX_DIGIT(*first) && ASCII_HEX_DIGIT(*second)) {
-        aOutput.Append(HEX_DIGIT(first) * 16 + HEX_DIGIT(second));
+        unescaped.Append(HEX_DIGIT(first) * 16 + HEX_DIGIT(second));
         start = ++second;
         continue;
 
       } else {
-        aOutput.Append('%');
+        unescaped.Append('%');
         ++start;
         continue;
       }
     }
 
-    aOutput.Append(*start);
+    unescaped.Append(*start);
     ++start;
   }
+
+  ConvertString(unescaped, aOutput);
+}
+
+void
+URLSearchParams::ConvertString(const nsACString& aInput, nsAString& aOutput)
+{
+  aOutput.Truncate();
+
+  if (!mDecoder) {
+    mDecoder = EncodingUtils::DecoderForEncoding("UTF-8");
+    if (!mDecoder) {
+      MOZ_ASSERT(mDecoder, "Failed to create a decoder.");
+      return;
+    }
+  }
+
+  nsACString::const_iterator iter;
+  aInput.BeginReading(iter);
+
+  int32_t inputLength = aInput.Length();
+  int32_t outputLength = 0;
+
+  nsresult rv = mDecoder->GetMaxLength(iter.get(), inputLength,
+                                       &outputLength);
+  if (NS_WARN_IF(NS_FAILED(rv))) {
+    return;
+  }
+
+  const mozilla::fallible_t fallible = mozilla::fallible_t();
+  nsAutoArrayPtr<char16_t> buf(new (fallible) char16_t[outputLength + 1]);
+  if (!buf) {
+    return;
+  }
+
+  rv = mDecoder->Convert(iter.get(), &inputLength, buf, &outputLength);
+  if (NS_SUCCEEDED(rv)) {
+    buf[outputLength] = 0;
+    if (!aOutput.Assign(buf, outputLength, mozilla::fallible_t())) {
+      aOutput.Truncate();
+    }
+  }
 }
 
 /* static */ PLDHashOperator

dom/base/URLSearchParams.h

@@ -13,6 +13,7 @@
 #include "nsClassHashtable.h"
 #include "nsHashKeys.h"
 #include "nsISupports.h"
+#include "nsIUnicodeDecoder.h"
 
 namespace mozilla {
 namespace dom {
@@ -83,7 +84,8 @@ private:
 
   void DeleteAll();
 
-  void DecodeString(const nsACString& aInput, nsACString& aOutput);
+  void DecodeString(const nsACString& aInput, nsAString& aOutput);
+  void ConvertString(const nsACString& aInput, nsAString& aOutput);
 
   void NotifyObservers(URLSearchParamsObserver* aExceptObserver);
 
@@ -98,6 +100,7 @@ private:
   nsClassHashtable<nsStringHashKey, nsTArray<nsString>> mSearchParams;
 
   nsTArray<nsRefPtr<URLSearchParamsObserver>> mObservers;
+  nsCOMPtr<nsIUnicodeDecoder> mDecoder;
 };
 
 } // namespace dom

dom/base/test/mochitest.ini

@@ -74,6 +74,7 @@ skip-if = (buildapp == 'b2g' && toolkit != 'gonk') #Bug 931116, b2g desktop spec
 [test_url_malformedHost.html]
 [test_urlExceptions.html]
 [test_urlSearchParams.html]
+[test_urlSearchParams_utf8.html]
 [test_urlutils_stringify.html]
 [test_window_constructor.html]
 [test_window_cross_origin_props.html]

dom/base/test/test_urlSearchParams_utf8.html

@@ -0,0 +1,40 @@
+
+<!DOCTYPE HTML>
+<html>
+<!--
+https://bugzilla.mozilla.org/show_bug.cgi?id=1032511
+-->
+<head>
+  <meta charset="utf-8">
+  <title>Test for Bug 1032511</title>
+  <script type="application/javascript" src="/tests/SimpleTest/SimpleTest.js"></script>
+  <link rel="stylesheet" type="text/css" href="/tests/SimpleTest/test.css"/>
+</head>
+<body>
+<a target="_blank" href="https://bugzilla.mozilla.org/show_bug.cgi?id=1032511">Mozilla Bug 1032511</a>
+<p id="display"></p>
+<div id="content" style="display: none">
+  <iframe name="x" id="x"></iframe>
+  <iframe name="y" id="y"></iframe>
+</div>
+<pre id="test">
+</pre>
+<a href="http://www.example.net?a=b&c=d" id="anchor">foobar</a>
+<area href="http://www.example.net?a=b&c=d" id="area">foobar</area>
+<script type="application/javascript">
+
+  /** Test for Bug 1032511 **/
+  var a = new URLSearchParams("%e2");
+  ok(a, "a exists");
+  is(a.toString(), '=', "The value should be here.");
+
+  a = new URLSearchParams("a%e2");
+  // This is a known decoder bug that fails to emit a REPLACEMENT CHARACTER.
+  is(a.toString(), 'a=', "The value should be here.");
+
+  a = new URLSearchParams("a%e2b");
+  is(a.toString(), 'a%EF%BF%BDb=', "The value should be here.");
+
+</script>
+</body>
+</html>