Implementation of a Matching strategy for URLs in the Declarative WebRequest API.

chrome/browser/extensions/api/declarative/url_matcher_unittest.cc

+// Copyright (c) 2012 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "chrome/browser/extensions/api/declarative/url_matcher.h"
+
+#include "base/string_util.h"
+#include "testing/gtest/include/gtest/gtest.h"
+
+namespace extensions {
+
+TEST(URLMatcherConditionTest, Constructor) {
+  SubstringPattern pattern("example.com", 1);
+  URLMatcherCondition m1(URLMatcherCondition::HOST_SUFFIX, &pattern);
+  EXPECT_EQ(URLMatcherCondition::HOST_SUFFIX, m1.criterion());
+  EXPECT_EQ(&pattern, m1.substring_pattern());
+}
+
+TEST(URLMatcherConditionTest, IsFullUrlCondition) {
+  SubstringPattern pattern("example.com", 1);
+  EXPECT_FALSE(URLMatcherCondition(URLMatcherCondition::HOST_SUFFIX,
+      &pattern).IsFullUrlCondition());
+
+  EXPECT_TRUE(URLMatcherCondition(URLMatcherCondition::HOST_CONTAINS,
+      &pattern).IsFullUrlCondition());
+  EXPECT_TRUE(URLMatcherCondition(URLMatcherCondition::PATH_CONTAINS,
+      &pattern).IsFullUrlCondition());
+  EXPECT_TRUE(URLMatcherCondition(URLMatcherCondition::QUERY_CONTAINS,
+      &pattern).IsFullUrlCondition());
+
+  EXPECT_TRUE(URLMatcherCondition(URLMatcherCondition::URL_PREFIX,
+      &pattern).IsFullUrlCondition());
+  EXPECT_TRUE(URLMatcherCondition(URLMatcherCondition::URL_SUFFIX,
+      &pattern).IsFullUrlCondition());
+  EXPECT_TRUE(URLMatcherCondition(URLMatcherCondition::URL_CONTAINS,
+      &pattern).IsFullUrlCondition());
+  EXPECT_TRUE(URLMatcherCondition(URLMatcherCondition::URL_EQUALS,
+      &pattern).IsFullUrlCondition());
+}
+
+TEST(URLMatcherConditionTest, SubstringMatching) {
+  // TODO(battre): Re-implement this.
+  /*
+  URLMatcherConditionFactory condition_factory;
+
+  // Check that substring_pattern is filled.
+  URLMatcherCondition m1(URLMatcherCondition::HOST_SUFFIX, "example.com");
+  EXPECT_EQ(condition_factory.CreateHostSuffixPattern("example.com"),
+            m1.substring_pattern());
+
+  // Check that patterns are recycled.
+  URLMatcherCondition m2(URLMatcherCondition::HOST_SUFFIX, "example.com");
+  m2.BuildSubstringPattern(&condition_factory);
+  EXPECT_EQ(m1.substring_pattern(), m1.substring_pattern());
+
+  // Check that SubstringPattern::ID are handled correctly.
+  GURL url("http://www.example.com/index.html");
+
+  std::set<SubstringPattern::ID> not_matching;
+  not_matching.insert(m1.substring_pattern().id() + 1);
+  EXPECT_FALSE(m1.IsMatch(not_matching, url));
+
+  std::set<SubstringPattern::ID> matching;
+  matching.insert(m1.substring_pattern().id());
+  EXPECT_TRUE(m1.IsMatch(matching, url));
+
+  // Check that URLs are tested correctly for containment.
+  URLMatcherCondition m3(URLMatcherCondition::HOST_CONTAINS, "example");
+  m3.BuildSubstringPattern(&condition_factory);
+  matching.insert(m3.substring_pattern().id());
+  EXPECT_TRUE(m3.IsFullUrlCondition());
+  EXPECT_FALSE(m3.IsMatch(matching, GURL("http://www.foo.com/example")));
+  EXPECT_TRUE(m3.IsMatch(matching, GURL("http://www.example.com/foo")));
+  */
+}
+
+
+namespace {
+
+bool Matches(scoped_ptr<URLMatcherCondition> condition, std::string text) {
+  return text.find(condition->substring_pattern()->pattern()) !=
+      std::string::npos;
+}
+
+}  // namespace
+
+TEST(URLMatcherConditionFactoryTest, GURLCharacterSet) {
+  // GURL guarantees that neither domain, nor path, nor query may contain
+  // non ASCII-7 characters. We test this here, because a change to this
+  // guarantee breaks this implementation horribly.
+  GURL url("http://www.föö.com/föö?föö#föö");
+  EXPECT_TRUE(IsStringASCII(url.host()));
+  EXPECT_TRUE(IsStringASCII(url.path()));
+  EXPECT_TRUE(IsStringASCII(url.query()));
+  EXPECT_FALSE(IsStringASCII(url.ref()));
+}
+
+// Basic tests of SubstringPattern objects
+TEST(URLMatcherConditionFactoryTest, TestSingletonProperty) {
+  // TODO(battre): Re-implement this.
+  /*URLMatcherConditionFactory matcher;
+  SubstringPattern p1 = matcher.CreateHostEqualsCondition("www.google.com");
+  SubstringPattern p2 = matcher.CreateHostEqualsCondition("www.google.com");
+  EXPECT_EQ(p1, p2);
+  SubstringPattern p3 = matcher.CreateHostEqualsCondition("www.google.de");
+  EXPECT_NE(p2.id(), p3.id());
+  EXPECT_NE(p2.pattern(), p3.pattern());*/
+}
+
+TEST(URLMatcherConditionFactoryTest, TestComponentSearches) {
+  GURL gurl("https://www.google.com/webhp?sourceid=chrome-instant&ie=UTF-8"
+      "&ion=1#hl=en&output=search&sclient=psy-ab&q=chrome%20is%20awesome");
+  URLMatcherConditionFactory matcher;
+  std::string url = matcher.CanonlicalizeURLForComponentSearches(gurl);
+
+  // Test host component.
+  EXPECT_TRUE(Matches(matcher.CreateHostPrefixCondition(""), url));
+  EXPECT_TRUE(Matches(matcher.CreateHostPrefixCondition("www.goog"), url));
+  EXPECT_TRUE(
+      Matches(matcher.CreateHostPrefixCondition("www.google.com"), url));
+  EXPECT_TRUE(
+      Matches(matcher.CreateHostPrefixCondition(".www.google.com"), url));
+  EXPECT_FALSE(Matches(matcher.CreateHostPrefixCondition("google.com"), url));
+  EXPECT_FALSE(
+      Matches(matcher.CreateHostPrefixCondition("www.google.com/"), url));
+  EXPECT_FALSE(Matches(matcher.CreateHostPrefixCondition("webhp"), url));
+
+  EXPECT_TRUE(Matches(matcher.CreateHostSuffixCondition(""), url));
+  EXPECT_TRUE(Matches(matcher.CreateHostSuffixCondition("com"), url));
+  EXPECT_TRUE(Matches(matcher.CreateHostSuffixCondition(".com"), url));
+  EXPECT_TRUE(
+      Matches(matcher.CreateHostSuffixCondition("www.google.com"), url));
+  EXPECT_TRUE(
+      Matches(matcher.CreateHostSuffixCondition(".www.google.com"), url));
+  EXPECT_FALSE(Matches(matcher.CreateHostSuffixCondition("www"), url));
+  EXPECT_FALSE(
+      Matches(matcher.CreateHostSuffixCondition("www.google.com/"), url));
+  EXPECT_FALSE(Matches(matcher.CreateHostSuffixCondition("webhp"), url));
+
+  EXPECT_FALSE(Matches(matcher.CreateHostEqualsCondition(""), url));
+  EXPECT_FALSE(Matches(matcher.CreateHostEqualsCondition("www"), url));
+  EXPECT_TRUE(
+      Matches(matcher.CreateHostEqualsCondition("www.google.com"), url));
+  EXPECT_FALSE(
+      Matches(matcher.CreateHostEqualsCondition("www.google.com/"), url));
+
+
+  // Test path component.
+  EXPECT_TRUE(Matches(matcher.CreatePathPrefixCondition(""), url));
+  EXPECT_TRUE(Matches(matcher.CreatePathPrefixCondition("/web"), url));
+  EXPECT_TRUE(Matches(matcher.CreatePathPrefixCondition("/webhp"), url));
+  EXPECT_FALSE(Matches(matcher.CreatePathPrefixCondition("webhp"), url));
+  EXPECT_FALSE(Matches(matcher.CreatePathPrefixCondition("/webhp?"), url));
+
+  EXPECT_TRUE(Matches(matcher.CreatePathSuffixCondition(""), url));
+  EXPECT_TRUE(Matches(matcher.CreatePathSuffixCondition("webhp"), url));
+  EXPECT_TRUE(Matches(matcher.CreatePathSuffixCondition("/webhp"), url));
+  EXPECT_FALSE(Matches(matcher.CreatePathSuffixCondition("/web"), url));
+  EXPECT_FALSE(Matches(matcher.CreatePathSuffixCondition("/webhp?"), url));
+
+  EXPECT_TRUE(Matches(matcher.CreatePathEqualsCondition("/webhp"), url));
+  EXPECT_FALSE(Matches(matcher.CreatePathEqualsCondition("webhp"), url));
+  EXPECT_FALSE(Matches(matcher.CreatePathEqualsCondition("/webhp?"), url));
+  EXPECT_FALSE(
+      Matches(matcher.CreatePathEqualsCondition("www.google.com"), url));
+
+
+  // Test query component.
+  EXPECT_TRUE(Matches(matcher.CreateQueryPrefixCondition(""), url));
+  EXPECT_TRUE(Matches(matcher.CreateQueryPrefixCondition("?sourceid"), url));
+  EXPECT_FALSE(Matches(matcher.CreatePathPrefixCondition("sourceid"), url));
+
+  EXPECT_TRUE(Matches(matcher.CreateQuerySuffixCondition(""), url));
+  EXPECT_TRUE(Matches(matcher.CreateQuerySuffixCondition("ion=1"), url));
+  EXPECT_FALSE(Matches(matcher.CreatePathPrefixCondition("?sourceid"), url));
+  EXPECT_FALSE(Matches(matcher.CreateQuerySuffixCondition("www"), url));
+
+  EXPECT_TRUE(Matches(matcher.CreateQueryEqualsCondition(
+      "?sourceid=chrome-instant&ie=UTF-8&ion=1"), url));
+  EXPECT_FALSE(Matches(matcher.CreateQueryEqualsCondition(
+        "sourceid=chrome-instant&ie=UTF-8&ion="), url));
+  EXPECT_FALSE(
+      Matches(matcher.CreateQueryEqualsCondition("www.google.com"), url));
+
+
+  // Test adjacent components
+  EXPECT_TRUE(Matches(matcher.CreateHostSuffixPathPrefixCondition(
+      "google.com", "/webhp"), url));
+  EXPECT_TRUE(Matches(matcher.CreateHostSuffixPathPrefixCondition(
+        "", "/webhp"), url));
+  EXPECT_TRUE(Matches(matcher.CreateHostSuffixPathPrefixCondition(
+        "google.com", ""), url));
+  EXPECT_FALSE(Matches(matcher.CreateHostSuffixPathPrefixCondition(
+        "www", ""), url));
+}
+
+TEST(URLMatcherConditionFactoryTest, TestFullSearches) {
+  GURL gurl("https://www.google.com/webhp?sourceid=chrome-instant&ie=UTF-8"
+      "&ion=1#hl=en&output=search&sclient=psy-ab&q=chrome%20is%20awesome");
+  URLMatcherConditionFactory matcher;
+  std::string url = matcher.CanonlicalizeURLForFullSearches(gurl);
+
+  EXPECT_TRUE(Matches(matcher.CreateURLPrefixCondition(""), url));
+  EXPECT_TRUE(Matches(matcher.CreateURLPrefixCondition("www.goog"), url));
+  EXPECT_TRUE(Matches(matcher.CreateURLPrefixCondition("www.google.com"), url));
+  EXPECT_TRUE(
+      Matches(matcher.CreateURLPrefixCondition(".www.google.com"), url));
+  EXPECT_TRUE(
+      Matches(matcher.CreateURLPrefixCondition("www.google.com/"), url));
+  EXPECT_FALSE(Matches(matcher.CreateURLPrefixCondition("webhp"), url));
+
+  EXPECT_TRUE(Matches(matcher.CreateURLSuffixCondition(""), url));
+  EXPECT_TRUE(Matches(matcher.CreateURLSuffixCondition("ion=1"), url));
+  EXPECT_FALSE(Matches(matcher.CreateURLSuffixCondition("www"), url));
+
+  EXPECT_TRUE(Matches(matcher.CreateURLContainsCondition(""), url));
+  EXPECT_TRUE(Matches(matcher.CreateURLContainsCondition("www.goog"), url));
+  EXPECT_TRUE(Matches(matcher.CreateURLContainsCondition(".www.goog"), url));
+  EXPECT_TRUE(Matches(matcher.CreateURLContainsCondition("webhp"), url));
+  EXPECT_TRUE(Matches(matcher.CreateURLContainsCondition("?"), url));
+  EXPECT_TRUE(Matches(matcher.CreateURLContainsCondition("sourceid"), url));
+  EXPECT_TRUE(Matches(matcher.CreateURLContainsCondition("ion=1"), url));
+  EXPECT_FALSE(Matches(matcher.CreateURLContainsCondition("foobar"), url));
+  EXPECT_FALSE(Matches(matcher.CreateURLContainsCondition("search"), url));
+
+  EXPECT_TRUE(Matches(matcher.CreateURLEqualsCondition(
+      "www.google.com/webhp?sourceid=chrome-instant&ie=UTF-8&ion=1"), url));
+  EXPECT_FALSE(
+      Matches(matcher.CreateURLEqualsCondition("www.google.com"), url));
+}
+
+/*
+// TODO(battre): Re-implement this.
+TEST(URLMatcherConditionSetTest, Constructors) {
+  URLMatcherConditionFactory condition_factory;
+  URLMatcherCondition m1(URLMatcherCondition::HOST_SUFFIX, "example.com");
+  m1.BuildSubstringPattern(&condition_factory);
+  URLMatcherCondition m2(URLMatcherCondition::PATH_CONTAINS, "foo");
+  m2.BuildSubstringPattern(&condition_factory);
+
+  std::vector<URLMatcherCondition> conditions;
+  conditions.push_back(m1);
+  conditions.push_back(m2);
+
+  URLMatcherConditionSet condition_set(1, conditions);
+  EXPECT_EQ(1, condition_set.id());
+  EXPECT_EQ(2u, condition_set.conditions().size());
+
+  std::vector<URLMatcherCondition> other_conditions;
+  other_conditions.push_back(m1);
+  URLMatcherConditionSet condition_set2(2, other_conditions);
+  condition_set2 = condition_set;
+  EXPECT_EQ(1, condition_set2.id());
+  EXPECT_EQ(2u, condition_set2.conditions().size());
+
+  URLMatcherConditionSet condition_set3(condition_set);
+  EXPECT_EQ(1, condition_set2.id());
+  EXPECT_EQ(2u, condition_set2.conditions().size());
+}
+
+TEST(URLMatcherConditionSetTest, Matching) {
+  GURL url1("http://www.example.com/foo?bar=1");
+  GURL url2("http://foo.example.com/index.html");
+
+  URLMatcherConditionFactory condition_factory;
+  URLMatcherCondition m1(URLMatcherCondition::HOST_SUFFIX, "example.com");
+  m1.BuildSubstringPattern(&condition_factory);
+  URLMatcherCondition m2(URLMatcherCondition::PATH_CONTAINS, "foo");
+  m2.BuildSubstringPattern(&condition_factory);
+
+  std::vector<URLMatcherCondition> conditions;
+  conditions.push_back(m1);
+  conditions.push_back(m2);
+
+  URLMatcherConditionSet condition_set(1, conditions);
+  EXPECT_EQ(1, condition_set.id());
+  EXPECT_EQ(2u, condition_set.conditions().size());
+
+  std::set<SubstringPattern::ID> matching_substring_patterns;
+  matching_substring_patterns.insert(m1.substring_pattern().id());
+  EXPECT_FALSE(condition_set.IsMatch(matching_substring_patterns, url1));
+
+  matching_substring_patterns.insert(m2.substring_pattern().id());
+  EXPECT_TRUE(condition_set.IsMatch(matching_substring_patterns, url1));
+  EXPECT_FALSE(condition_set.IsMatch(matching_substring_patterns, url2));
+}
+*/
+
+TEST(URLMatcherTest, FullTest) {
+  GURL url1("http://www.example.com/foo?bar=1");
+  GURL url2("http://foo.example.com/index.html");
+
+  URLMatcher matcher;
+  URLMatcherConditionFactory* factory = matcher.condition_factory();
+
+  // First insert.
+  scoped_ptr<URLMatcherConditionSet::Conditions> conditions1(
+      new URLMatcherConditionSet::Conditions());
+  conditions1->push_back(
+      factory->CreateHostSuffixCondition("example.com").release());
+  conditions1->push_back(
+      factory->CreatePathContainsCondition("foo").release());
+
+  const int kConditionSetId1 = 1;
+  scoped_ptr<ScopedVector<const URLMatcherConditionSet> > insert1(
+      new ScopedVector<const URLMatcherConditionSet>());
+  insert1->push_back(
+      new URLMatcherConditionSet(kConditionSetId1, conditions1.Pass()));
+  matcher.AddConditionSets(insert1.Pass());
+  EXPECT_EQ(1u, matcher.MatchUrl(url1).size());
+  EXPECT_EQ(0u, matcher.MatchUrl(url2).size());
+
+  // Second insert.
+  scoped_ptr<URLMatcherConditionSet::Conditions> conditions2(
+        new URLMatcherConditionSet::Conditions());
+  conditions2->push_back(
+      factory->CreateHostSuffixCondition("example.com").release());
+
+  const int kConditionSetId2 = 2;
+  scoped_ptr<ScopedVector<const URLMatcherConditionSet> > insert2(
+      new ScopedVector<const URLMatcherConditionSet>());
+  insert2->push_back(
+      new URLMatcherConditionSet(kConditionSetId2, conditions2.Pass()));
+  matcher.AddConditionSets(insert2.Pass());
+  EXPECT_EQ(2u, matcher.MatchUrl(url1).size());
+  EXPECT_EQ(1u, matcher.MatchUrl(url2).size());
+
+  // This should be the cached singleton.
+  int patternId1 = factory->CreateHostSuffixCondition(
+      "example.com")->substring_pattern()->id();
+
+  // Removal of last insert.
+  std::vector<URLMatcherConditionSet::ID> remove2;
+  remove2.push_back(kConditionSetId2);
+  matcher.RemoveConditionSets(remove2);
+  EXPECT_EQ(1u, matcher.MatchUrl(url1).size());
+  EXPECT_EQ(0u, matcher.MatchUrl(url2).size());
+
+  // Removal of first insert.
+  std::vector<URLMatcherConditionSet::ID> remove1;
+  remove1.push_back(kConditionSetId1);
+  matcher.RemoveConditionSets(remove1);
+  EXPECT_EQ(0u, matcher.MatchUrl(url1).size());
+  EXPECT_EQ(0u, matcher.MatchUrl(url2).size());
+
+  // The cached singleton in matcher.condition_factory_ should be destroyed to
+  // free memory.
+  int patternId2 = factory->CreateHostSuffixCondition(
+      "example.com")->substring_pattern()->id();
+  // If patternId1 and patternId2 are different that indicates that
+  // matcher.condition_factory_ does not leak memory.
+  EXPECT_NE(patternId1, patternId2);
+}
+
+}  // namespace extensions