Add implicit trailing dot domain matching support to URLPattern.

extensions/common/url_pattern.cc

Index: extensions/common/url_pattern.cc
diff --git a/extensions/common/url_pattern.cc b/extensions/common/url_pattern.cc
index 91ba65bec02583f4666ea4a68e76c96fe3ef66e6..c71a473825728bad5ad29b9d3ae02c55779c2257 100644
--- a/extensions/common/url_pattern.cc
+++ b/extensions/common/url_pattern.cc
@@ -18,6 +18,7 @@
 #include "content/public/common/url_constants.h"
 #include "extensions/common/constants.h"
 #include "net/base/registry_controlled_domains/registry_controlled_domain.h"
+#include "net/base/url_util.h"
 #include "url/gurl.h"
 #include "url/url_util.h"
 
@@ -113,6 +114,10 @@ std::string StripTrailingWildcard(const std::string& path) {
   return wildcard_index == path_last ? path.substr(0, path_last) : path;
 }
 
+bool EndsWithDot(base::StringPiece str) {
+  return base::EndsWith(str, ".", base::CompareCase::SENSITIVE);
+}
+
 }  // namespace
 
 // static
@@ -404,8 +409,10 @@ bool URLPattern::MatchesHost(const std::string& host) const {
 }
 
 bool URLPattern::MatchesHost(const GURL& test) const {
+  const std::string test_host(CanonicalizeHostForMatching(test));

[Devlin, 2016/10/28 19:17:27]

This code is called under some pretty performance-sensitive times (e.g. web
request matching, etc).  We need to do a pretty big overhaul to make it all
faster, but I wonder if we can at least avoid a string copy here.  Could we
instead use a StringPiece test_host and StringPiece pattern_host that both have
the trailing dot (if any) removed, and thus avoid the copies?

+
   // If the hosts are exactly equal, we have a match.
-  if (test.host() == host_)
+  if (test_host == host_)
     return true;
 
   // If we're matching subdomains, and we have no host in the match pattern,
@@ -424,14 +431,14 @@ bool URLPattern::MatchesHost(const GURL& test) const {
     return false;
 
   // Check if the test host is a subdomain of our host.
-  if (test.host().length() <= (host_.length() + 1))
+  if (test_host.length() <= (host_.length() + 1))
     return false;
 
-  if (test.host().compare(test.host().length() - host_.length(),
-                          host_.length(), host_) != 0)
+  if (test_host.compare(test_host.length() - host_.length(), host_.length(),
+                        host_) != 0)
     return false;
 
-  return test.host()[test.host().length() - host_.length() - 1] == '.';
+  return test_host[test_host.length() - host_.length() - 1] == '.';
 }
 
 bool URLPattern::ImpliesAllHosts() const {
@@ -607,6 +614,23 @@ std::vector<URLPattern> URLPattern::ConvertToExplicitSchemes() const {
   return result;
 }
 
+std::string URLPattern::CanonicalizeHostForMatching(const GURL& url) const {
+  std::string url_host = url.host();
+  if (url_host.empty() || url.HostIsIPAddress())
+    return url_host;
+
+  const bool pattern_host_ends_with_dot = EndsWithDot(host_);
+  const bool url_host_ends_with_dot = EndsWithDot(url_host);
+  if (pattern_host_ends_with_dot != url_host_ends_with_dot) {
+    if (url_host_ends_with_dot)
+      url_host = net::TrimEndingDot(url_host);
+    else
+      url_host += ".";
+  }
+
+  return url_host;
+}
+
 // static
 const char* URLPattern::GetParseResultString(
     URLPattern::ParseResult parse_result) {