Ensure that www.website.com matches website.com

This cl ensures that www.website.com matches website.com
in SupervisedUserUrlFilter.

Bug: 1032811
Change-Id: Ic76ffde2457d029fbb25acaf6f41e048cbca17b5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2092125
Commit-Queue: Yilkal Abe <yilkal@chromium.org>
Reviewed-by: Aga Wronska <agawronska@chromium.org>
Cr-Commit-Position: refs/heads/master@{#753788}

chrome/browser/supervised_user/supervised_user_url_filter.cc

@@ -20,6 +20,7 @@
 #include "base/stl_util.h"
 #include "base/strings/string_number_conversions.h"
 #include "base/strings/string_util.h"
+#include "base/strings/utf_string_conversions.h"
 #include "base/task/post_task.h"
 #include "base/task/thread_pool.h"
 #include "base/task_runner_util.h"
@@ -29,6 +30,7 @@
 #include "chrome/common/chrome_features.h"
 #include "components/policy/core/browser/url_blacklist_manager.h"
 #include "components/policy/core/browser/url_util.h"
+#include "components/url_formatter/url_formatter.h"
 #include "components/url_matcher/url_matcher.h"
 #include "components/variations/service/variations_service.h"
 #include "content/public/browser/browser_thread.h"
@@ -244,8 +246,21 @@ bool SupervisedUserURLFilter::HasFilteredScheme(const GURL& url) {
 bool SupervisedUserURLFilter::HostMatchesPattern(
     const std::string& canonical_host,
     const std::string& pattern) {
-  std::string trimmed_pattern = pattern;
+  // If |canonical_host| starts with |www.| but |pattern| starts with neither
+  // |www.| nor |*.| then trim |www.| part of canonical host.
+  bool is_host_www =
+      base::StartsWith(canonical_host, "www.", base::CompareCase::SENSITIVE);
+  bool patern_accepts =
+      base::StartsWith(pattern, "www.", base::CompareCase::SENSITIVE) ||
+      base::StartsWith(pattern, "*.", base::CompareCase::SENSITIVE);
+
   std::string trimmed_host = canonical_host;
+  if (is_host_www && !patern_accepts) {
+    trimmed_host = base::UTF16ToASCII(
+        url_formatter::StripWWW(base::ASCIIToUTF16(canonical_host)));
+  }
+
+  std::string trimmed_pattern = pattern;
   if (base::EndsWith(pattern, ".*", base::CompareCase::SENSITIVE)) {
     size_t registry_length = GetCanonicalHostRegistryLength(
         trimmed_host, EXCLUDE_UNKNOWN_REGISTRIES, EXCLUDE_PRIVATE_REGISTRIES);

chrome/browser/supervised_user/supervised_user_url_filter.h

@@ -92,10 +92,11 @@ class SupervisedUserURLFilter {
   //   or accounts.google.com).
   // - If the pattern ends with ".*", it matches the host on any known TLD
   //   (e.g. the pattern "google.*" would match google.com or google.co.uk).
-  // See the SupervisedUserURLFilterTest.HostMatchesPattern unit test for more
-  // examples.
-  // Asterisks in other parts of the pattern are not allowed.
-  // |host| and |pattern| are assumed to be normalized to lower-case.
+  // If the |host| starts with "www." but the |pattern| starts with neither
+  // "www." nor "*.", the function strips the "www." part of |host| and tries to
+  // match again. See the SupervisedUserURLFilterTest.HostMatchesPattern unit
+  // test for more examples. Asterisks in other parts of the pattern are not
+  // allowed. |host| and |pattern| are assumed to be normalized to lower-case.
   // This method is public for testing.
   static bool HostMatchesPattern(const std::string& canonical_host,
                                  const std::string& pattern);

chrome/browser/supervised_user/supervised_user_url_filter_unittest.cc

@@ -336,6 +336,8 @@ TEST_F(SupervisedUserURLFilterTest, HasFilteredScheme) {
 }
 
 TEST_F(SupervisedUserURLFilterTest, HostMatchesPattern) {
+  EXPECT_TRUE(SupervisedUserURLFilter::HostMatchesPattern("www.google.com",
+                                                          "google.com"));
   EXPECT_TRUE(
       SupervisedUserURLFilter::HostMatchesPattern("www.google.com",
                                                   "*.google.com"));