Fix broken utf-8 text for content-on-dino-page

See bug for more information about the problem, and solution.

This was manually tested.

Bug: 910458
Change-Id: I663a1354f12a917c478144530311b31e4bd1657f
Reviewed-on: https://chromium-review.googlesource.com/c/1357520Reviewed-by: Cathy Li <chili@chromium.org>
Commit-Queue: Dan H <harringtond@google.com>
Cr-Commit-Position: refs/heads/master@{#613106}

chrome/renderer/net/available_offline_content_helper.cc

@@ -10,6 +10,7 @@
 #include "base/json/json_writer.h"
 #include "base/metrics/histogram_macros.h"
 #include "base/numerics/safe_conversions.h"
+#include "base/strings/utf_string_conversions.h"
 #include "base/values.h"
 #include "chrome/common/available_offline_content.mojom.h"
 #include "components/error_page/common/net_error_info.h"
@@ -22,6 +23,23 @@ namespace {
 using chrome::mojom::AvailableOfflineContentPtr;
 using chrome::mojom::AvailableContentType;
 
+// Converts a string to base-64 data. This is done for security purposes, to
+// avoid potential XSS. Note that when this value is decoded in javascript, we
+// want to use the atob() function, but that function only handles latin-1
+// characters. Additionally, javascript needs UTF16 strings. So we instead
+// encode to UTF16, and then store that data as base64.
+std::string ConvertToUTF16Base64(const std::string& text) {
+  base::string16 text_utf16 = base::UTF8ToUTF16(text);
+  std::string utf16_bytes;
+  for (base::char16 c : text_utf16) {
+    utf16_bytes.push_back(static_cast<char>(c >> 8));
+    utf16_bytes.push_back(static_cast<char>(c & 0xff));
+  }
+  std::string encoded;
+  base::Base64Encode(utf16_bytes, &encoded);
+  return encoded;
+}
+
 base::Value AvailableContentToValue(const AvailableOfflineContentPtr& content) {
   // All pieces of text content downloaded from the web will be base64 encoded
   // to lessen security risks when this dictionary is passed as a string to
@@ -30,13 +48,13 @@ base::Value AvailableContentToValue(const AvailableOfflineContentPtr& content) {
   base::Value value(base::Value::Type::DICTIONARY);
   value.SetKey("ID", base::Value(content->id));
   value.SetKey("name_space", base::Value(content->name_space));
-  base::Base64Encode(content->title, &base64_encoded);
-  value.SetKey("title_base64", base::Value(base64_encoded));
-  base::Base64Encode(content->snippet, &base64_encoded);
-  value.SetKey("snippet_base64", base::Value(base64_encoded));
+  value.SetKey("title_base64",
+               base::Value(ConvertToUTF16Base64(content->title)));
+  value.SetKey("snippet_base64",
+               base::Value(ConvertToUTF16Base64(content->snippet)));
   value.SetKey("date_modified", base::Value(content->date_modified));
-  base::Base64Encode(content->attribution, &base64_encoded);
-  value.SetKey("attribution_base64", base::Value(base64_encoded));
+  value.SetKey("attribution_base64",
+               base::Value(ConvertToUTF16Base64(content->attribution)));
   value.SetKey("thumbnail_data_uri",
                base::Value(content->thumbnail_data_uri.spec()));
   value.SetKey("content_type",

chrome/renderer/net/net_error_helper_core_unittest.cc

@@ -2696,23 +2696,23 @@ const std::string GetExpectedAvailableContentAsJson() {
   std::string want_json = R"([
     {
       "ID": "ID",
-      "attribution_base64": "YXR0cmlidXRpb24=",
+      "attribution_base64": "AGEAdAB0AHIAaQBiAHUAdABpAG8Abg==",
       "content_type": 0,
       "date_modified": "date_modified",
       "name_space": "name_space",
-      "snippet_base64": "c25pcHBldA==",
+      "snippet_base64": "AHMAbgBpAHAAcABlAHQ=",
       "thumbnail_data_uri": "data:image/png;base64,abc",
-      "title_base64": "dGl0bGU="
+      "title_base64": "AHQAaQB0AGwAZQ=="
     },
     {
       "ID": "ID2",
-      "attribution_base64": "YXR0cmlidXRpb24y",
+      "attribution_base64": "AGEAdAB0AHIAaQBiAHUAdABpAG8AbgAy",
       "content_type": 3,
       "date_modified": "date_modified2",
       "name_space": "name_space2",
-      "snippet_base64": "c25pcHBldDI=",
+      "snippet_base64": "AHMAbgBpAHAAcABlAHQAMg==",
       "thumbnail_data_uri": "data:image/png;base64,abc",
-      "title_base64": "dGl0bGUy"
+      "title_base64": "AHQAaQB0AGwAZQAy"
     }
   ])";
   base::ReplaceChars(want_json, base::kWhitespaceASCII, "", &want_json);

components/neterror/resources/neterror.js

@@ -2,6 +2,17 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
+// Decodes a UTF16 string that is encoded as base64.
+function decodeUTF16Base64ToString(encoded_text) {
+  var data = atob(encoded_text);
+  var result = '';
+  for (var i = 0; i < data.length; i += 2) {
+    result +=
+        String.fromCharCode(data.charCodeAt(i) * 256 + data.charCodeAt(i + 1));
+  }
+  return result;
+}
+
 function toggleHelpBox() {
   var helpBoxOuter = document.getElementById('details');
   helpBoxOuter.classList.toggle(HIDDEN_CLASS);
@@ -286,9 +297,11 @@ function offlineContentAvailable(isShown, suggestions) {
   // plain text.
   for (var index = 0; index < suggestions.length; index++) {
     document.getElementById(`offline-content-suggestion-title-${index}`)
-        .textContent = atob(suggestions[index].title_base64);
+        .textContent =
+        decodeUTF16Base64ToString(suggestions[index].title_base64);
     document.getElementById(`offline-content-suggestion-attribution-${index}`)
-        .textContent = atob(suggestions[index].attribution_base64);
+        .textContent =
+        decodeUTF16Base64ToString(suggestions[index].attribution_base64);
   }
 
   var contentListElement = document.getElementById('offline-content-list');
@@ -324,8 +337,7 @@ function onDocumentLoad() {
   var showSavedCopyButtonVisible =
       loadTimeData.valueExists('showSavedCopyButton') &&
       loadTimeData.getValue('showSavedCopyButton').msg;
-  var downloadButtonVisible =
-      loadTimeData.valueExists('downloadButton') &&
+  var downloadButtonVisible = loadTimeData.valueExists('downloadButton') &&
       loadTimeData.getValue('downloadButton').msg;
 
   // If offline content suggestions will be visible, the usual buttons will not