Merge 152067 - Let url filter test the scheme in urlContains/Equals/Prefix/Suffix criteria

BUG=140238

Review URL: https://chromiumcodereview.appspot.com/10823313

TBR=battre@chromium.org
Review URL: https://chromiumcodereview.appspot.com/10859003

git-svn-id: svn://svn.chromium.org/chrome/branches/1229/src@152447 0039d316-1c4b-4281-b951-d872f2087c98

5 files changed, 67 insertions, 35 deletions

diff --git a/chrome/common/extensions/api/events.json b/chrome/common/extensions/api/events.json
index da3369413..4b505cf 100644
--- a/chrome/common/extensions/api/events.json
+++ b/chrome/common/extensions/api/events.json
@@ -256,22 +256,22 @@
           },
           "urlContains": {
             "type": "string",
-            "description": "Matches if the URL contains a specified string.",
+            "description": "Matches if the URL (without fragment identifier) contains a specified string. Port numbers are stripped from the URL if they match the default port number.",
             "optional": true
           },
           "urlEquals": {
             "type": "string",
-            "description": "Matches if the URL is equal to a specified string.",
+            "description": "Matches if the URL (without fragment identifier) is equal to a specified string. Port numbers are stripped from the URL if they match the default port number.",
             "optional": true
           },
           "urlPrefix": {
             "type": "string",
-            "description": "Matches if the URL starts with a specified string.",
+            "description": "Matches if the URL (without fragment identifier) starts with a specified string. Port numbers are stripped from the URL if they match the default port number.",
             "optional": true
           },
           "urlSuffix": {
             "type": "string",
-            "description": "Matches if the URL ends with a specified string.",
+            "description": "Matches if the URL (without fragment identifier) ends with a specified string. Port numbers are stripped from the URL if they match the default port number.",
             "optional": true
           },
           "schemes": {
diff --git a/chrome/common/extensions/docs/apps/events.html b/chrome/common/extensions/docs/apps/events.html
index e4edf3d..320b141 100644
--- a/chrome/common/extensions/docs/apps/events.html
+++ b/chrome/common/extensions/docs/apps/events.html
@@ -1265,7 +1265,7 @@ refers to an array of rules including filled optional parameters.
                 </div>
               </em>
           </dt>
-          <dd>Matches if the URL contains a specified string.</dd>
+          <dd>Matches if the URL (without fragment identifier) contains a specified string. Port numbers are stripped from the URL if they match the default port number.</dd>
           <!-- OBJECT PROPERTIES -->
           <!-- OBJECT METHODS -->
           <!-- OBJECT EVENT FIELDS -->
@@ -1289,7 +1289,7 @@ refers to an array of rules including filled optional parameters.
                 </div>
               </em>
           </dt>
-          <dd>Matches if the URL is equal to a specified string.</dd>
+          <dd>Matches if the URL (without fragment identifier) is equal to a specified string. Port numbers are stripped from the URL if they match the default port number.</dd>
           <!-- OBJECT PROPERTIES -->
           <!-- OBJECT METHODS -->
           <!-- OBJECT EVENT FIELDS -->
@@ -1313,7 +1313,7 @@ refers to an array of rules including filled optional parameters.
                 </div>
               </em>
           </dt>
-          <dd>Matches if the URL starts with a specified string.</dd>
+          <dd>Matches if the URL (without fragment identifier) starts with a specified string. Port numbers are stripped from the URL if they match the default port number.</dd>
           <!-- OBJECT PROPERTIES -->
           <!-- OBJECT METHODS -->
           <!-- OBJECT EVENT FIELDS -->
@@ -1337,7 +1337,7 @@ refers to an array of rules including filled optional parameters.
                 </div>
               </em>
           </dt>
-          <dd>Matches if the URL ends with a specified string.</dd>
+          <dd>Matches if the URL (without fragment identifier) ends with a specified string. Port numbers are stripped from the URL if they match the default port number.</dd>
           <!-- OBJECT PROPERTIES -->
           <!-- OBJECT METHODS -->
           <!-- OBJECT EVENT FIELDS -->
diff --git a/chrome/common/extensions/docs/extensions/events.html b/chrome/common/extensions/docs/extensions/events.html
index 7874258..88ae41c 100644
--- a/chrome/common/extensions/docs/extensions/events.html
+++ b/chrome/common/extensions/docs/extensions/events.html
@@ -1395,7 +1395,7 @@ refers to an array of rules including filled optional parameters.
                 </div>
               </em>
           </dt>
-          <dd>Matches if the URL contains a specified string.</dd>
+          <dd>Matches if the URL (without fragment identifier) contains a specified string. Port numbers are stripped from the URL if they match the default port number.</dd>
           <!-- OBJECT PROPERTIES -->
           <!-- OBJECT METHODS -->
           <!-- OBJECT EVENT FIELDS -->
@@ -1419,7 +1419,7 @@ refers to an array of rules including filled optional parameters.
                 </div>
               </em>
           </dt>
-          <dd>Matches if the URL is equal to a specified string.</dd>
+          <dd>Matches if the URL (without fragment identifier) is equal to a specified string. Port numbers are stripped from the URL if they match the default port number.</dd>
           <!-- OBJECT PROPERTIES -->
           <!-- OBJECT METHODS -->
           <!-- OBJECT EVENT FIELDS -->
@@ -1443,7 +1443,7 @@ refers to an array of rules including filled optional parameters.
                 </div>
               </em>
           </dt>
-          <dd>Matches if the URL starts with a specified string.</dd>
+          <dd>Matches if the URL (without fragment identifier) starts with a specified string. Port numbers are stripped from the URL if they match the default port number.</dd>
           <!-- OBJECT PROPERTIES -->
           <!-- OBJECT METHODS -->
           <!-- OBJECT EVENT FIELDS -->
@@ -1467,7 +1467,7 @@ refers to an array of rules including filled optional parameters.
                 </div>
               </em>
           </dt>
-          <dd>Matches if the URL ends with a specified string.</dd>
+          <dd>Matches if the URL (without fragment identifier) ends with a specified string. Port numbers are stripped from the URL if they match the default port number.</dd>
           <!-- OBJECT PROPERTIES -->
           <!-- OBJECT METHODS -->
           <!-- OBJECT EVENT FIELDS -->
diff --git a/chrome/common/extensions/matcher/url_matcher.cc b/chrome/common/extensions/matcher/url_matcher.cc
index 722efd9..0eb5ab5 100644
--- a/chrome/common/extensions/matcher/url_matcher.cc
+++ b/chrome/common/extensions/matcher/url_matcher.cc
@@ -8,7 +8,9 @@
 #include <iterator>
 
 #include "base/logging.h"
+#include "content/public/common/url_constants.h"
 #include "googleurl/src/gurl.h"
+#include "googleurl/src/url_canon.h"
 
 namespace extensions {
 
@@ -88,19 +90,23 @@ namespace extensions {
 // Case 2: url_{prefix,suffix,equals,contains} searches.
 // =====================================================
 //
-// Step 1: as above
+// Step 1: as above, except that
+// - the scheme is not removed
+// - the port is not removed if it is specified and does not match the default
+//   port for the given scheme.
 //
 // Step 2:
 // Translate URL to String and add the following position markers:
 // - BU = Beginning of URL
 // - EU = End of URL
-// Furthermore, the hostname is canonicalized to start with a ".".
 //
-// -> www.example.com/index.html?search=foo becomes
-// BU .www.example.com/index.html?search=foo EU
+// -> http://www.example.com:8080/index.html?search=foo#first_match becomes
+// BU http://www.example.com:8080/index.html?search=foo EU
+// -> http://www.example.com:80/index.html?search=foo#first_match becomes
+// BU http://www.example.com/index.html?search=foo EU
 //
-// url_prefix(prefix) = BU add_missing_dot_prefix(prefix)
-// -> url_prefix("www.example") = BU .www.example
+// url_prefix(prefix) = BU prefix
+// -> url_prefix("http://www.example") = BU http://www.example
 //
 // url_contains(substring) = substring
 // -> url_contains("index") = index
@@ -113,7 +119,7 @@ namespace extensions {
 // by a combination of a url_contains() query followed by an explicit test:
 //
 // host_contains(str) = url_contains(str) followed by test whether str occurs
-//   in host comonent of original URL.
+//   in host component of original URL.
 // -> host_contains("example.co") = example.co
 //    followed by gurl.host().find("example.co");
 //
@@ -160,7 +166,7 @@ bool URLMatcherCondition::operator<(const URLMatcherCondition& rhs) const {
 
 bool URLMatcherCondition::IsFullURLCondition() const {
   // For these criteria the SubstringMatcher needs to be executed on the
-  // GURL that is canonlizaliced with
+  // GURL that is canonicalized with
   // URLMatcherConditionFactory::CanonicalizeURLForFullSearches.
   switch (criterion_) {
     case HOST_CONTAINS:
@@ -314,14 +320,26 @@ URLMatcherConditionFactory::CreateHostEqualsPathPrefixCondition(
 
 std::string URLMatcherConditionFactory::CanonicalizeURLForFullSearches(
     const GURL& url) {
-  return kBeginningOfURL + CanonicalizeHostname(url.host()) + url.path() +
-      (url.has_query() ? "?" + url.query() : "") + kEndOfURL;
+  GURL::Replacements replacements;
+  replacements.ClearPassword();
+  replacements.ClearUsername();
+  replacements.ClearRef();
+  // Clear port if it is implicit from scheme.
+  if (url.has_port()) {
+    const std::string& port = url.scheme();
+    if (url_canon::DefaultPortForScheme(port.c_str(), port.size()) ==
+            url.EffectiveIntPort()) {
+      replacements.ClearPort();
+    }
+  }
+  return kBeginningOfURL + url.ReplaceComponents(replacements).spec() +
+      kEndOfURL;
 }
 
 URLMatcherCondition URLMatcherConditionFactory::CreateURLPrefixCondition(
     const std::string& prefix) {
   return CreateCondition(URLMatcherCondition::URL_PREFIX,
-      kBeginningOfURL + CanonicalizeHostname(prefix));
+      kBeginningOfURL + prefix);
 }
 
 URLMatcherCondition URLMatcherConditionFactory::CreateURLSuffixCondition(
@@ -337,7 +355,7 @@ URLMatcherCondition URLMatcherConditionFactory::CreateURLContainsCondition(
 URLMatcherCondition URLMatcherConditionFactory::CreateURLEqualsCondition(
     const std::string& str) {
   return CreateCondition(URLMatcherCondition::URL_EQUALS,
-      kBeginningOfURL + CanonicalizeHostname(str) + kEndOfURL);
+      kBeginningOfURL + str + kEndOfURL);
 }
 
 void URLMatcherConditionFactory::ForgetUnusedPatterns(
diff --git a/chrome/common/extensions/matcher/url_matcher_unittest.cc b/chrome/common/extensions/matcher/url_matcher_unittest.cc
index add8530..acf414f 100644
--- a/chrome/common/extensions/matcher/url_matcher_unittest.cc
+++ b/chrome/common/extensions/matcher/url_matcher_unittest.cc
@@ -230,7 +230,7 @@ TEST(URLMatcherConditionFactoryTest, TestSingletonProperty) {
 }
 
 TEST(URLMatcherConditionFactoryTest, TestComponentSearches) {
-  GURL gurl("https://www.google.com/webhp?sourceid=chrome-instant&ie=UTF-8"
+  GURL gurl("https://www.google.com:1234/webhp?sourceid=chrome-instant&ie=UTF-8"
       "&ion=1#hl=en&output=search&sclient=psy-ab&q=chrome%20is%20awesome");
   URLMatcherConditionFactory factory;
   std::string url = factory.CanonicalizeURLForComponentSearches(gurl);
@@ -326,18 +326,21 @@ TEST(URLMatcherConditionFactoryTest, TestComponentSearches) {
 }
 
 TEST(URLMatcherConditionFactoryTest, TestFullSearches) {
-  GURL gurl("https://www.google.com/webhp?sourceid=chrome-instant&ie=UTF-8"
+  // The Port 443 is stripped because it is the default port for https.
+  GURL gurl("https://www.google.com:443/webhp?sourceid=chrome-instant&ie=UTF-8"
       "&ion=1#hl=en&output=search&sclient=psy-ab&q=chrome%20is%20awesome");
   URLMatcherConditionFactory factory;
   std::string url = factory.CanonicalizeURLForFullSearches(gurl);
 
   EXPECT_TRUE(Matches(factory.CreateURLPrefixCondition(""), url));
-  EXPECT_TRUE(Matches(factory.CreateURLPrefixCondition("www.goog"), url));
-  EXPECT_TRUE(Matches(factory.CreateURLPrefixCondition("www.google.com"), url));
-  EXPECT_TRUE(
-      Matches(factory.CreateURLPrefixCondition(".www.google.com"), url));
-  EXPECT_TRUE(
-      Matches(factory.CreateURLPrefixCondition("www.google.com/"), url));
+  EXPECT_TRUE(Matches(factory.CreateURLPrefixCondition(
+      "https://www.goog"), url));
+  EXPECT_TRUE(Matches(factory.CreateURLPrefixCondition(
+      "https://www.google.com"), url));
+  EXPECT_TRUE(Matches(factory.CreateURLPrefixCondition(
+      "https://www.google.com/webhp?"), url));
+  EXPECT_FALSE(Matches(factory.CreateURLPrefixCondition(
+      "http://www.google.com"), url));
   EXPECT_FALSE(Matches(factory.CreateURLPrefixCondition("webhp"), url));
 
   EXPECT_TRUE(Matches(factory.CreateURLSuffixCondition(""), url));
@@ -346,18 +349,29 @@ TEST(URLMatcherConditionFactoryTest, TestFullSearches) {
 
   EXPECT_TRUE(Matches(factory.CreateURLContainsCondition(""), url));
   EXPECT_TRUE(Matches(factory.CreateURLContainsCondition("www.goog"), url));
-  EXPECT_TRUE(Matches(factory.CreateURLContainsCondition(".www.goog"), url));
   EXPECT_TRUE(Matches(factory.CreateURLContainsCondition("webhp"), url));
   EXPECT_TRUE(Matches(factory.CreateURLContainsCondition("?"), url));
   EXPECT_TRUE(Matches(factory.CreateURLContainsCondition("sourceid"), url));
   EXPECT_TRUE(Matches(factory.CreateURLContainsCondition("ion=1"), url));
+  EXPECT_FALSE(Matches(factory.CreateURLContainsCondition(".www.goog"), url));
   EXPECT_FALSE(Matches(factory.CreateURLContainsCondition("foobar"), url));
   EXPECT_FALSE(Matches(factory.CreateURLContainsCondition("search"), url));
+  EXPECT_FALSE(Matches(factory.CreateURLContainsCondition(":443"), url));
 
   EXPECT_TRUE(Matches(factory.CreateURLEqualsCondition(
-      "www.google.com/webhp?sourceid=chrome-instant&ie=UTF-8&ion=1"), url));
+      "https://www.google.com/webhp?sourceid=chrome-instant&ie=UTF-8&ion=1"),
+      url));
   EXPECT_FALSE(
-      Matches(factory.CreateURLEqualsCondition("www.google.com"), url));
+      Matches(factory.CreateURLEqualsCondition("https://www.google.com"), url));
+
+  // Same as above but this time with a non-standard port.
+  gurl = GURL("https://www.google.com:1234/webhp?sourceid=chrome-instant&"
+      "ie=UTF-8&ion=1#hl=en&output=search&sclient=psy-ab&q=chrome%20is%20"
+      "awesome");
+  url = factory.CanonicalizeURLForFullSearches(gurl);
+  EXPECT_TRUE(Matches(factory.CreateURLPrefixCondition(
+      "https://www.google.com:1234/webhp?"), url));
+  EXPECT_TRUE(Matches(factory.CreateURLContainsCondition(":1234"), url));
 }