OLD | NEW |
| (Empty) |
1 // Copyright (c) 2012 The Chromium Authors. All rights reserved. | |
2 // Use of this source code is governed by a BSD-style license that can be | |
3 // found in the LICENSE file. | |
4 | |
5 #include <algorithm> | |
6 #include <fstream> | |
7 | |
8 #include "base/auto_reset.h" | |
9 #include "base/files/file_path.h" | |
10 #include "base/files/file_util.h" | |
11 #include "base/files/scoped_temp_dir.h" | |
12 #include "base/path_service.h" | |
13 #include "base/run_loop.h" | |
14 #include "base/strings/string16.h" | |
15 #include "base/strings/string_util.h" | |
16 #include "base/strings/utf_string_conversions.h" | |
17 #include "chrome/browser/autocomplete/scored_history_match_builder_impl.h" | |
18 #include "chrome/browser/bookmarks/bookmark_model_factory.h" | |
19 #include "chrome/browser/history/history_backend.h" | |
20 #include "chrome/browser/history/history_service.h" | |
21 #include "chrome/browser/history/history_service_factory.h" | |
22 #include "chrome/browser/history/in_memory_url_index.h" | |
23 #include "chrome/browser/history/url_index_private_data.h" | |
24 #include "chrome/common/chrome_paths.h" | |
25 #include "chrome/test/base/history_index_restore_observer.h" | |
26 #include "chrome/test/base/testing_profile.h" | |
27 #include "components/bookmarks/test/bookmark_test_helpers.h" | |
28 #include "components/history/core/browser/history_database.h" | |
29 #include "components/history/core/browser/in_memory_url_index_types.h" | |
30 #include "content/public/test/test_browser_thread_bundle.h" | |
31 #include "sql/transaction.h" | |
32 #include "testing/gtest/include/gtest/gtest.h" | |
33 | |
34 using base::ASCIIToUTF16; | |
35 | |
36 // The test version of the history url database table ('url') is contained in | |
37 // a database file created from a text file('url_history_provider_test.db.txt'). | |
38 // The only difference between this table and a live 'urls' table from a | |
39 // profile is that the last_visit_time column in the test table contains a | |
40 // number specifying the number of days relative to 'today' to which the | |
41 // absolute time should be set during the test setup stage. | |
42 // | |
43 // The format of the test database text file is of a SQLite .dump file. | |
44 // Note that only lines whose first character is an upper-case letter are | |
45 // processed when creating the test database. | |
46 | |
47 namespace history { | |
48 namespace { | |
49 const size_t kMaxMatches = 3; | |
50 const char kTestLanguages[] = "en,ja,hi,zh"; | |
51 } // namespace | |
52 | |
53 // ----------------------------------------------------------------------------- | |
54 | |
55 // Observer class so the unit tests can wait while the cache is being saved. | |
56 class CacheFileSaverObserver : public InMemoryURLIndex::SaveCacheObserver { | |
57 public: | |
58 explicit CacheFileSaverObserver(const base::Closure& task); | |
59 | |
60 bool succeeded() { return succeeded_; } | |
61 | |
62 private: | |
63 // SaveCacheObserver implementation. | |
64 void OnCacheSaveFinished(bool succeeded) override; | |
65 | |
66 base::Closure task_; | |
67 bool succeeded_; | |
68 | |
69 DISALLOW_COPY_AND_ASSIGN(CacheFileSaverObserver); | |
70 }; | |
71 | |
72 CacheFileSaverObserver::CacheFileSaverObserver(const base::Closure& task) | |
73 : task_(task), | |
74 succeeded_(false) { | |
75 } | |
76 | |
77 void CacheFileSaverObserver::OnCacheSaveFinished(bool succeeded) { | |
78 succeeded_ = succeeded; | |
79 task_.Run(); | |
80 } | |
81 | |
82 // ----------------------------------------------------------------------------- | |
83 | |
84 class InMemoryURLIndexTest : public testing::Test { | |
85 public: | |
86 InMemoryURLIndexTest(); | |
87 | |
88 protected: | |
89 // Test setup. | |
90 void SetUp() override; | |
91 void TearDown() override; | |
92 | |
93 // Allows the database containing the test data to be customized by | |
94 // subclasses. | |
95 virtual base::FilePath::StringType TestDBName() const; | |
96 | |
97 // Allows the test to control when the InMemoryURLIndex is initialized. | |
98 virtual bool InitializeInMemoryURLIndexInSetUp() const; | |
99 | |
100 // Initialize the InMemoryURLIndex for the tests. | |
101 void InitializeInMemoryURLIndex(); | |
102 | |
103 // Validates that the given |term| is contained in |cache| and that it is | |
104 // marked as in-use. | |
105 void CheckTerm(const URLIndexPrivateData::SearchTermCacheMap& cache, | |
106 base::string16 term) const; | |
107 | |
108 // Pass-through function to simplify our friendship with HistoryService. | |
109 sql::Connection& GetDB(); | |
110 | |
111 // Pass-through functions to simplify our friendship with InMemoryURLIndex. | |
112 URLIndexPrivateData* GetPrivateData() const; | |
113 base::CancelableTaskTracker* GetPrivateDataTracker() const; | |
114 void ClearPrivateData(); | |
115 void set_history_dir(const base::FilePath& dir_path); | |
116 bool GetCacheFilePath(base::FilePath* file_path) const; | |
117 void PostRestoreFromCacheFileTask(); | |
118 void PostSaveToCacheFileTask(); | |
119 const std::set<std::string>& scheme_whitelist(); | |
120 | |
121 | |
122 // Pass-through functions to simplify our friendship with URLIndexPrivateData. | |
123 bool UpdateURL(const URLRow& row); | |
124 bool DeleteURL(const GURL& url); | |
125 | |
126 // Data verification helper functions. | |
127 void ExpectPrivateDataNotEmpty(const URLIndexPrivateData& data); | |
128 void ExpectPrivateDataEmpty(const URLIndexPrivateData& data); | |
129 void ExpectPrivateDataEqual(const URLIndexPrivateData& expected, | |
130 const URLIndexPrivateData& actual); | |
131 | |
132 ScoredHistoryMatchBuilderImpl builder_; | |
133 content::TestBrowserThreadBundle thread_bundle_; | |
134 scoped_ptr<InMemoryURLIndex> url_index_; | |
135 TestingProfile profile_; | |
136 HistoryService* history_service_; | |
137 HistoryDatabase* history_database_; | |
138 }; | |
139 | |
140 InMemoryURLIndexTest::InMemoryURLIndexTest() | |
141 : builder_(ScoredHistoryMatchBuilderImpl::IsBookmarkedCallback()), | |
142 history_service_(nullptr), | |
143 history_database_(nullptr) { | |
144 } | |
145 | |
146 sql::Connection& InMemoryURLIndexTest::GetDB() { | |
147 return history_database_->GetDB(); | |
148 } | |
149 | |
150 URLIndexPrivateData* InMemoryURLIndexTest::GetPrivateData() const { | |
151 DCHECK(url_index_->private_data()); | |
152 return url_index_->private_data(); | |
153 } | |
154 | |
155 base::CancelableTaskTracker* InMemoryURLIndexTest::GetPrivateDataTracker() | |
156 const { | |
157 DCHECK(url_index_->private_data_tracker()); | |
158 return url_index_->private_data_tracker(); | |
159 } | |
160 | |
161 void InMemoryURLIndexTest::ClearPrivateData() { | |
162 return url_index_->ClearPrivateData(); | |
163 } | |
164 | |
165 void InMemoryURLIndexTest::set_history_dir(const base::FilePath& dir_path) { | |
166 return url_index_->set_history_dir(dir_path); | |
167 } | |
168 | |
169 bool InMemoryURLIndexTest::GetCacheFilePath(base::FilePath* file_path) const { | |
170 DCHECK(file_path); | |
171 return url_index_->GetCacheFilePath(file_path); | |
172 } | |
173 | |
174 void InMemoryURLIndexTest::PostRestoreFromCacheFileTask() { | |
175 url_index_->PostRestoreFromCacheFileTask(); | |
176 } | |
177 | |
178 void InMemoryURLIndexTest::PostSaveToCacheFileTask() { | |
179 url_index_->PostSaveToCacheFileTask(); | |
180 } | |
181 | |
182 const std::set<std::string>& InMemoryURLIndexTest::scheme_whitelist() { | |
183 return url_index_->scheme_whitelist(); | |
184 } | |
185 | |
186 bool InMemoryURLIndexTest::UpdateURL(const URLRow& row) { | |
187 return GetPrivateData()->UpdateURL(history_service_, | |
188 row, | |
189 url_index_->languages_, | |
190 url_index_->scheme_whitelist_, | |
191 GetPrivateDataTracker()); | |
192 } | |
193 | |
194 bool InMemoryURLIndexTest::DeleteURL(const GURL& url) { | |
195 return GetPrivateData()->DeleteURL(url); | |
196 } | |
197 | |
198 void InMemoryURLIndexTest::SetUp() { | |
199 // We cannot access the database until the backend has been loaded. | |
200 ASSERT_TRUE(profile_.CreateHistoryService(true, false)); | |
201 profile_.CreateBookmarkModel(true); | |
202 bookmarks::test::WaitForBookmarkModelToLoad( | |
203 BookmarkModelFactory::GetForProfile(&profile_)); | |
204 profile_.BlockUntilHistoryProcessesPendingRequests(); | |
205 profile_.BlockUntilHistoryIndexIsRefreshed(); | |
206 history_service_ = HistoryServiceFactory::GetForProfile( | |
207 &profile_, ServiceAccessType::EXPLICIT_ACCESS); | |
208 ASSERT_TRUE(history_service_); | |
209 HistoryBackend* backend = history_service_->history_backend_.get(); | |
210 history_database_ = backend->db(); | |
211 | |
212 // Create and populate a working copy of the URL history database. | |
213 base::FilePath history_proto_path; | |
214 PathService::Get(chrome::DIR_TEST_DATA, &history_proto_path); | |
215 history_proto_path = history_proto_path.Append( | |
216 FILE_PATH_LITERAL("History")); | |
217 history_proto_path = history_proto_path.Append(TestDBName()); | |
218 EXPECT_TRUE(base::PathExists(history_proto_path)); | |
219 | |
220 std::ifstream proto_file(history_proto_path.value().c_str()); | |
221 static const size_t kCommandBufferMaxSize = 2048; | |
222 char sql_cmd_line[kCommandBufferMaxSize]; | |
223 | |
224 sql::Connection& db(GetDB()); | |
225 ASSERT_TRUE(db.is_open()); | |
226 { | |
227 sql::Transaction transaction(&db); | |
228 transaction.Begin(); | |
229 while (!proto_file.eof()) { | |
230 proto_file.getline(sql_cmd_line, kCommandBufferMaxSize); | |
231 if (!proto_file.eof()) { | |
232 // We only process lines which begin with a upper-case letter. | |
233 // TODO(mrossetti): Can iswupper() be used here? | |
234 if (sql_cmd_line[0] >= 'A' && sql_cmd_line[0] <= 'Z') { | |
235 std::string sql_cmd(sql_cmd_line); | |
236 sql::Statement sql_stmt(db.GetUniqueStatement(sql_cmd_line)); | |
237 EXPECT_TRUE(sql_stmt.Run()); | |
238 } | |
239 } | |
240 } | |
241 transaction.Commit(); | |
242 } | |
243 | |
244 // Update the last_visit_time table column in the "urls" table | |
245 // such that it represents a time relative to 'now'. | |
246 sql::Statement statement(db.GetUniqueStatement( | |
247 "SELECT" HISTORY_URL_ROW_FIELDS "FROM urls;")); | |
248 ASSERT_TRUE(statement.is_valid()); | |
249 base::Time time_right_now = base::Time::NowFromSystemTime(); | |
250 base::TimeDelta day_delta = base::TimeDelta::FromDays(1); | |
251 { | |
252 sql::Transaction transaction(&db); | |
253 transaction.Begin(); | |
254 while (statement.Step()) { | |
255 URLRow row; | |
256 history_database_->FillURLRow(statement, &row); | |
257 base::Time last_visit = time_right_now; | |
258 for (int64 i = row.last_visit().ToInternalValue(); i > 0; --i) | |
259 last_visit -= day_delta; | |
260 row.set_last_visit(last_visit); | |
261 history_database_->UpdateURLRow(row.id(), row); | |
262 } | |
263 transaction.Commit(); | |
264 } | |
265 | |
266 // Update the visit_time table column in the "visits" table | |
267 // such that it represents a time relative to 'now'. | |
268 statement.Assign(db.GetUniqueStatement( | |
269 "SELECT" HISTORY_VISIT_ROW_FIELDS "FROM visits;")); | |
270 ASSERT_TRUE(statement.is_valid()); | |
271 { | |
272 sql::Transaction transaction(&db); | |
273 transaction.Begin(); | |
274 while (statement.Step()) { | |
275 VisitRow row; | |
276 history_database_->FillVisitRow(statement, &row); | |
277 base::Time last_visit = time_right_now; | |
278 for (int64 i = row.visit_time.ToInternalValue(); i > 0; --i) | |
279 last_visit -= day_delta; | |
280 row.visit_time = last_visit; | |
281 history_database_->UpdateVisitRow(row); | |
282 } | |
283 transaction.Commit(); | |
284 } | |
285 | |
286 if (InitializeInMemoryURLIndexInSetUp()) | |
287 InitializeInMemoryURLIndex(); | |
288 } | |
289 | |
290 void InMemoryURLIndexTest::TearDown() { | |
291 // Ensure that the InMemoryURLIndex no longer observer HistoryService before | |
292 // it is destroyed in order to prevent HistoryService calling dead observer. | |
293 if (url_index_) | |
294 url_index_->ShutDown(); | |
295 } | |
296 | |
297 base::FilePath::StringType InMemoryURLIndexTest::TestDBName() const { | |
298 return FILE_PATH_LITERAL("url_history_provider_test.db.txt"); | |
299 } | |
300 | |
301 bool InMemoryURLIndexTest::InitializeInMemoryURLIndexInSetUp() const { | |
302 return true; | |
303 } | |
304 | |
305 void InMemoryURLIndexTest::InitializeInMemoryURLIndex() { | |
306 DCHECK(!url_index_); | |
307 url_index_.reset( | |
308 new InMemoryURLIndex(history_service_, base::FilePath(), kTestLanguages)); | |
309 url_index_->Init(); | |
310 url_index_->RebuildFromHistory(history_database_); | |
311 } | |
312 | |
313 void InMemoryURLIndexTest::CheckTerm( | |
314 const URLIndexPrivateData::SearchTermCacheMap& cache, | |
315 base::string16 term) const { | |
316 URLIndexPrivateData::SearchTermCacheMap::const_iterator cache_iter( | |
317 cache.find(term)); | |
318 ASSERT_TRUE(cache.end() != cache_iter) | |
319 << "Cache does not contain '" << term << "' but should."; | |
320 URLIndexPrivateData::SearchTermCacheItem cache_item = cache_iter->second; | |
321 EXPECT_TRUE(cache_item.used_) | |
322 << "Cache item '" << term << "' should be marked as being in use."; | |
323 } | |
324 | |
325 void InMemoryURLIndexTest::ExpectPrivateDataNotEmpty( | |
326 const URLIndexPrivateData& data) { | |
327 EXPECT_FALSE(data.word_list_.empty()); | |
328 // available_words_ will be empty since we have freshly built the | |
329 // data set for these tests. | |
330 EXPECT_TRUE(data.available_words_.empty()); | |
331 EXPECT_FALSE(data.word_map_.empty()); | |
332 EXPECT_FALSE(data.char_word_map_.empty()); | |
333 EXPECT_FALSE(data.word_id_history_map_.empty()); | |
334 EXPECT_FALSE(data.history_id_word_map_.empty()); | |
335 EXPECT_FALSE(data.history_info_map_.empty()); | |
336 } | |
337 | |
338 void InMemoryURLIndexTest::ExpectPrivateDataEmpty( | |
339 const URLIndexPrivateData& data) { | |
340 EXPECT_TRUE(data.word_list_.empty()); | |
341 EXPECT_TRUE(data.available_words_.empty()); | |
342 EXPECT_TRUE(data.word_map_.empty()); | |
343 EXPECT_TRUE(data.char_word_map_.empty()); | |
344 EXPECT_TRUE(data.word_id_history_map_.empty()); | |
345 EXPECT_TRUE(data.history_id_word_map_.empty()); | |
346 EXPECT_TRUE(data.history_info_map_.empty()); | |
347 } | |
348 | |
349 // Helper function which compares two maps for equivalence. The maps' values | |
350 // are associative containers and their contents are compared as well. | |
351 template<typename T> | |
352 void ExpectMapOfContainersIdentical(const T& expected, const T& actual) { | |
353 ASSERT_EQ(expected.size(), actual.size()); | |
354 for (typename T::const_iterator expected_iter = expected.begin(); | |
355 expected_iter != expected.end(); ++expected_iter) { | |
356 typename T::const_iterator actual_iter = actual.find(expected_iter->first); | |
357 ASSERT_TRUE(actual.end() != actual_iter); | |
358 typename T::mapped_type const& expected_values(expected_iter->second); | |
359 typename T::mapped_type const& actual_values(actual_iter->second); | |
360 ASSERT_EQ(expected_values.size(), actual_values.size()); | |
361 for (typename T::mapped_type::const_iterator set_iter = | |
362 expected_values.begin(); set_iter != expected_values.end(); ++set_iter) | |
363 EXPECT_EQ(actual_values.count(*set_iter), | |
364 expected_values.count(*set_iter)); | |
365 } | |
366 } | |
367 | |
368 void InMemoryURLIndexTest::ExpectPrivateDataEqual( | |
369 const URLIndexPrivateData& expected, | |
370 const URLIndexPrivateData& actual) { | |
371 EXPECT_EQ(expected.word_list_.size(), actual.word_list_.size()); | |
372 EXPECT_EQ(expected.word_map_.size(), actual.word_map_.size()); | |
373 EXPECT_EQ(expected.char_word_map_.size(), actual.char_word_map_.size()); | |
374 EXPECT_EQ(expected.word_id_history_map_.size(), | |
375 actual.word_id_history_map_.size()); | |
376 EXPECT_EQ(expected.history_id_word_map_.size(), | |
377 actual.history_id_word_map_.size()); | |
378 EXPECT_EQ(expected.history_info_map_.size(), actual.history_info_map_.size()); | |
379 EXPECT_EQ(expected.word_starts_map_.size(), actual.word_starts_map_.size()); | |
380 // WordList must be index-by-index equal. | |
381 size_t count = expected.word_list_.size(); | |
382 for (size_t i = 0; i < count; ++i) | |
383 EXPECT_EQ(expected.word_list_[i], actual.word_list_[i]); | |
384 | |
385 ExpectMapOfContainersIdentical(expected.char_word_map_, | |
386 actual.char_word_map_); | |
387 ExpectMapOfContainersIdentical(expected.word_id_history_map_, | |
388 actual.word_id_history_map_); | |
389 ExpectMapOfContainersIdentical(expected.history_id_word_map_, | |
390 actual.history_id_word_map_); | |
391 | |
392 for (HistoryInfoMap::const_iterator expected_info = | |
393 expected.history_info_map_.begin(); | |
394 expected_info != expected.history_info_map_.end(); ++expected_info) { | |
395 HistoryInfoMap::const_iterator actual_info = | |
396 actual.history_info_map_.find(expected_info->first); | |
397 // NOTE(yfriedman): ASSERT_NE can't be used due to incompatibility between | |
398 // gtest and STLPort in the Android build. See | |
399 // http://code.google.com/p/googletest/issues/detail?id=359 | |
400 ASSERT_TRUE(actual_info != actual.history_info_map_.end()); | |
401 const URLRow& expected_row(expected_info->second.url_row); | |
402 const URLRow& actual_row(actual_info->second.url_row); | |
403 EXPECT_EQ(expected_row.visit_count(), actual_row.visit_count()); | |
404 EXPECT_EQ(expected_row.typed_count(), actual_row.typed_count()); | |
405 EXPECT_EQ(expected_row.last_visit(), actual_row.last_visit()); | |
406 EXPECT_EQ(expected_row.url(), actual_row.url()); | |
407 const VisitInfoVector& expected_visits(expected_info->second.visits); | |
408 const VisitInfoVector& actual_visits(actual_info->second.visits); | |
409 EXPECT_EQ(expected_visits.size(), actual_visits.size()); | |
410 for (size_t i = 0; | |
411 i < std::min(expected_visits.size(), actual_visits.size()); ++i) { | |
412 EXPECT_EQ(expected_visits[i].first, actual_visits[i].first); | |
413 EXPECT_EQ(expected_visits[i].second, actual_visits[i].second); | |
414 } | |
415 } | |
416 | |
417 for (WordStartsMap::const_iterator expected_starts = | |
418 expected.word_starts_map_.begin(); | |
419 expected_starts != expected.word_starts_map_.end(); | |
420 ++expected_starts) { | |
421 WordStartsMap::const_iterator actual_starts = | |
422 actual.word_starts_map_.find(expected_starts->first); | |
423 // NOTE(yfriedman): ASSERT_NE can't be used due to incompatibility between | |
424 // gtest and STLPort in the Android build. See | |
425 // http://code.google.com/p/googletest/issues/detail?id=359 | |
426 ASSERT_TRUE(actual_starts != actual.word_starts_map_.end()); | |
427 const RowWordStarts& expected_word_starts(expected_starts->second); | |
428 const RowWordStarts& actual_word_starts(actual_starts->second); | |
429 EXPECT_EQ(expected_word_starts.url_word_starts_.size(), | |
430 actual_word_starts.url_word_starts_.size()); | |
431 EXPECT_TRUE(std::equal(expected_word_starts.url_word_starts_.begin(), | |
432 expected_word_starts.url_word_starts_.end(), | |
433 actual_word_starts.url_word_starts_.begin())); | |
434 EXPECT_EQ(expected_word_starts.title_word_starts_.size(), | |
435 actual_word_starts.title_word_starts_.size()); | |
436 EXPECT_TRUE(std::equal(expected_word_starts.title_word_starts_.begin(), | |
437 expected_word_starts.title_word_starts_.end(), | |
438 actual_word_starts.title_word_starts_.begin())); | |
439 } | |
440 } | |
441 | |
442 //------------------------------------------------------------------------------ | |
443 | |
444 class LimitedInMemoryURLIndexTest : public InMemoryURLIndexTest { | |
445 protected: | |
446 base::FilePath::StringType TestDBName() const override; | |
447 bool InitializeInMemoryURLIndexInSetUp() const override; | |
448 }; | |
449 | |
450 base::FilePath::StringType LimitedInMemoryURLIndexTest::TestDBName() const { | |
451 return FILE_PATH_LITERAL("url_history_provider_test_limited.db.txt"); | |
452 } | |
453 | |
454 bool LimitedInMemoryURLIndexTest::InitializeInMemoryURLIndexInSetUp() const { | |
455 return false; | |
456 } | |
457 | |
458 TEST_F(LimitedInMemoryURLIndexTest, Initialization) { | |
459 // Verify that the database contains the expected number of items, which | |
460 // is the pre-filtered count, i.e. all of the items. | |
461 sql::Statement statement(GetDB().GetUniqueStatement("SELECT * FROM urls;")); | |
462 ASSERT_TRUE(statement.is_valid()); | |
463 uint64 row_count = 0; | |
464 while (statement.Step()) ++row_count; | |
465 EXPECT_EQ(1U, row_count); | |
466 | |
467 InitializeInMemoryURLIndex(); | |
468 URLIndexPrivateData& private_data(*GetPrivateData()); | |
469 | |
470 // history_info_map_ should have the same number of items as were filtered. | |
471 EXPECT_EQ(1U, private_data.history_info_map_.size()); | |
472 EXPECT_EQ(35U, private_data.char_word_map_.size()); | |
473 EXPECT_EQ(17U, private_data.word_map_.size()); | |
474 } | |
475 | |
476 #if defined(OS_WIN) | |
477 // Flaky on windows trybots: http://crbug.com/351500 | |
478 #define MAYBE_Retrieval DISABLED_Retrieval | |
479 #else | |
480 #define MAYBE_Retrieval Retrieval | |
481 #endif | |
482 TEST_F(InMemoryURLIndexTest, MAYBE_Retrieval) { | |
483 // See if a very specific term gives a single result. | |
484 ScoredHistoryMatches matches = url_index_->HistoryItemsForTerms( | |
485 ASCIIToUTF16("DrudgeReport"), base::string16::npos, kMaxMatches, | |
486 builder_); | |
487 ASSERT_EQ(1U, matches.size()); | |
488 | |
489 // Verify that we got back the result we expected. | |
490 EXPECT_EQ(5, matches[0].url_info.id()); | |
491 EXPECT_EQ("http://drudgereport.com/", matches[0].url_info.url().spec()); | |
492 EXPECT_EQ(ASCIIToUTF16("DRUDGE REPORT 2010"), matches[0].url_info.title()); | |
493 EXPECT_TRUE(matches[0].can_inline); | |
494 | |
495 // Make sure a trailing space prevents inline-ability but still results | |
496 // in the expected result. | |
497 matches = url_index_->HistoryItemsForTerms(ASCIIToUTF16("DrudgeReport "), | |
498 base::string16::npos, kMaxMatches, | |
499 builder_); | |
500 ASSERT_EQ(1U, matches.size()); | |
501 EXPECT_EQ(5, matches[0].url_info.id()); | |
502 EXPECT_EQ("http://drudgereport.com/", matches[0].url_info.url().spec()); | |
503 EXPECT_EQ(ASCIIToUTF16("DRUDGE REPORT 2010"), matches[0].url_info.title()); | |
504 EXPECT_FALSE(matches[0].can_inline); | |
505 | |
506 // Search which should result in multiple results. | |
507 matches = url_index_->HistoryItemsForTerms( | |
508 ASCIIToUTF16("drudge"), base::string16::npos, kMaxMatches, builder_); | |
509 ASSERT_EQ(2U, matches.size()); | |
510 // The results should be in descending score order. | |
511 EXPECT_GE(matches[0].raw_score, matches[1].raw_score); | |
512 | |
513 // Search which should result in nearly perfect result. | |
514 matches = url_index_->HistoryItemsForTerms( | |
515 ASCIIToUTF16("Nearly Perfect Result"), base::string16::npos, kMaxMatches, | |
516 builder_); | |
517 ASSERT_EQ(1U, matches.size()); | |
518 // The results should have a very high score. | |
519 EXPECT_GT(matches[0].raw_score, 900); | |
520 EXPECT_EQ(32, matches[0].url_info.id()); | |
521 EXPECT_EQ("https://nearlyperfectresult.com/", | |
522 matches[0].url_info.url().spec()); // Note: URL gets lowercased. | |
523 EXPECT_EQ(ASCIIToUTF16("Practically Perfect Search Result"), | |
524 matches[0].url_info.title()); | |
525 EXPECT_FALSE(matches[0].can_inline); | |
526 | |
527 // Search which should result in very poor result. | |
528 matches = url_index_->HistoryItemsForTerms( | |
529 ASCIIToUTF16("qui c"), base::string16::npos, kMaxMatches, builder_); | |
530 ASSERT_EQ(1U, matches.size()); | |
531 // The results should have a poor score. | |
532 EXPECT_LT(matches[0].raw_score, 500); | |
533 EXPECT_EQ(33, matches[0].url_info.id()); | |
534 EXPECT_EQ("http://quiteuselesssearchresultxyz.com/", | |
535 matches[0].url_info.url().spec()); // Note: URL gets lowercased. | |
536 EXPECT_EQ(ASCIIToUTF16("Practically Useless Search Result"), | |
537 matches[0].url_info.title()); | |
538 EXPECT_FALSE(matches[0].can_inline); | |
539 | |
540 // Search which will match at the end of an URL with encoded characters. | |
541 matches = url_index_->HistoryItemsForTerms( | |
542 ASCIIToUTF16("Mice"), base::string16::npos, kMaxMatches, builder_); | |
543 ASSERT_EQ(1U, matches.size()); | |
544 EXPECT_EQ(30, matches[0].url_info.id()); | |
545 EXPECT_FALSE(matches[0].can_inline); | |
546 | |
547 // Check that URLs are not escaped an escape time. | |
548 matches = url_index_->HistoryItemsForTerms(ASCIIToUTF16("1% wikipedia"), | |
549 base::string16::npos, kMaxMatches, | |
550 builder_); | |
551 ASSERT_EQ(1U, matches.size()); | |
552 EXPECT_EQ(35, matches[0].url_info.id()); | |
553 EXPECT_EQ("http://en.wikipedia.org/wiki/1%25_rule_(Internet_culture)", | |
554 matches[0].url_info.url().spec()); | |
555 | |
556 // Verify that a single term can appear multiple times in the URL and as long | |
557 // as one starts the URL it is still inlined. | |
558 matches = url_index_->HistoryItemsForTerms( | |
559 ASCIIToUTF16("fubar"), base::string16::npos, kMaxMatches, builder_); | |
560 ASSERT_EQ(1U, matches.size()); | |
561 EXPECT_EQ(34, matches[0].url_info.id()); | |
562 EXPECT_EQ("http://fubarfubarandfubar.com/", matches[0].url_info.url().spec()); | |
563 EXPECT_EQ(ASCIIToUTF16("Situation Normal -- FUBARED"), | |
564 matches[0].url_info.title()); | |
565 EXPECT_TRUE(matches[0].can_inline); | |
566 } | |
567 | |
568 TEST_F(InMemoryURLIndexTest, CursorPositionRetrieval) { | |
569 // See if a very specific term with no cursor gives an empty result. | |
570 ScoredHistoryMatches matches = url_index_->HistoryItemsForTerms( | |
571 ASCIIToUTF16("DrudReport"), base::string16::npos, kMaxMatches, builder_); | |
572 ASSERT_EQ(0U, matches.size()); | |
573 | |
574 // The same test with the cursor at the end should give an empty result. | |
575 matches = url_index_->HistoryItemsForTerms(ASCIIToUTF16("DrudReport"), 10u, | |
576 kMaxMatches, builder_); | |
577 ASSERT_EQ(0U, matches.size()); | |
578 | |
579 // If the cursor is between Drud and Report, we should find the desired | |
580 // result. | |
581 matches = url_index_->HistoryItemsForTerms(ASCIIToUTF16("DrudReport"), 4u, | |
582 kMaxMatches, builder_); | |
583 ASSERT_EQ(1U, matches.size()); | |
584 EXPECT_EQ("http://drudgereport.com/", matches[0].url_info.url().spec()); | |
585 EXPECT_EQ(ASCIIToUTF16("DRUDGE REPORT 2010"), matches[0].url_info.title()); | |
586 | |
587 // Now check multi-word inputs. No cursor should fail to find a | |
588 // result on this input. | |
589 matches = url_index_->HistoryItemsForTerms(ASCIIToUTF16("MORTGAGERATE DROPS"), | |
590 base::string16::npos, kMaxMatches, | |
591 builder_); | |
592 ASSERT_EQ(0U, matches.size()); | |
593 | |
594 // Ditto with cursor at end. | |
595 matches = url_index_->HistoryItemsForTerms(ASCIIToUTF16("MORTGAGERATE DROPS"), | |
596 18u, kMaxMatches, builder_); | |
597 ASSERT_EQ(0U, matches.size()); | |
598 | |
599 // If the cursor is between MORTAGE And RATE, we should find the | |
600 // desired result. | |
601 matches = url_index_->HistoryItemsForTerms(ASCIIToUTF16("MORTGAGERATE DROPS"), | |
602 8u, kMaxMatches, builder_); | |
603 ASSERT_EQ(1U, matches.size()); | |
604 EXPECT_EQ("http://www.reuters.com/article/idUSN0839880620100708", | |
605 matches[0].url_info.url().spec()); | |
606 EXPECT_EQ(ASCIIToUTF16( | |
607 "UPDATE 1-US 30-yr mortgage rate drops to new record low | Reuters"), | |
608 matches[0].url_info.title()); | |
609 } | |
610 | |
611 TEST_F(InMemoryURLIndexTest, URLPrefixMatching) { | |
612 // "drudgere" - found, can inline | |
613 ScoredHistoryMatches matches = url_index_->HistoryItemsForTerms( | |
614 ASCIIToUTF16("drudgere"), base::string16::npos, kMaxMatches, builder_); | |
615 ASSERT_EQ(1U, matches.size()); | |
616 EXPECT_TRUE(matches[0].can_inline); | |
617 | |
618 // "drudgere" - found, can inline | |
619 matches = url_index_->HistoryItemsForTerms( | |
620 ASCIIToUTF16("drudgere"), base::string16::npos, kMaxMatches, builder_); | |
621 ASSERT_EQ(1U, matches.size()); | |
622 EXPECT_TRUE(matches[0].can_inline); | |
623 | |
624 // "www.atdmt" - not found | |
625 matches = url_index_->HistoryItemsForTerms( | |
626 ASCIIToUTF16("www.atdmt"), base::string16::npos, kMaxMatches, builder_); | |
627 EXPECT_EQ(0U, matches.size()); | |
628 | |
629 // "atdmt" - found, cannot inline | |
630 matches = url_index_->HistoryItemsForTerms( | |
631 ASCIIToUTF16("atdmt"), base::string16::npos, kMaxMatches, builder_); | |
632 ASSERT_EQ(1U, matches.size()); | |
633 EXPECT_FALSE(matches[0].can_inline); | |
634 | |
635 // "view.atdmt" - found, can inline | |
636 matches = url_index_->HistoryItemsForTerms( | |
637 ASCIIToUTF16("view.atdmt"), base::string16::npos, kMaxMatches, builder_); | |
638 ASSERT_EQ(1U, matches.size()); | |
639 EXPECT_TRUE(matches[0].can_inline); | |
640 | |
641 // "view.atdmt" - found, can inline | |
642 matches = url_index_->HistoryItemsForTerms( | |
643 ASCIIToUTF16("view.atdmt"), base::string16::npos, kMaxMatches, builder_); | |
644 ASSERT_EQ(1U, matches.size()); | |
645 EXPECT_TRUE(matches[0].can_inline); | |
646 | |
647 // "cnn.com" - found, can inline | |
648 matches = url_index_->HistoryItemsForTerms( | |
649 ASCIIToUTF16("cnn.com"), base::string16::npos, kMaxMatches, builder_); | |
650 ASSERT_EQ(2U, matches.size()); | |
651 // One match should be inline-able, the other not. | |
652 EXPECT_TRUE(matches[0].can_inline != matches[1].can_inline); | |
653 | |
654 // "www.cnn.com" - found, can inline | |
655 matches = url_index_->HistoryItemsForTerms( | |
656 ASCIIToUTF16("www.cnn.com"), base::string16::npos, kMaxMatches, builder_); | |
657 ASSERT_EQ(1U, matches.size()); | |
658 EXPECT_TRUE(matches[0].can_inline); | |
659 | |
660 // "ww.cnn.com" - found because we allow mid-term matches in hostnames | |
661 matches = url_index_->HistoryItemsForTerms( | |
662 ASCIIToUTF16("ww.cnn.com"), base::string16::npos, kMaxMatches, builder_); | |
663 ASSERT_EQ(1U, matches.size()); | |
664 | |
665 // "www.cnn.com" - found, can inline | |
666 matches = url_index_->HistoryItemsForTerms( | |
667 ASCIIToUTF16("www.cnn.com"), base::string16::npos, kMaxMatches, builder_); | |
668 ASSERT_EQ(1U, matches.size()); | |
669 EXPECT_TRUE(matches[0].can_inline); | |
670 | |
671 // "tp://www.cnn.com" - not found because we don't allow tp as a mid-term | |
672 // match | |
673 matches = url_index_->HistoryItemsForTerms(ASCIIToUTF16("tp://www.cnn.com"), | |
674 base::string16::npos, kMaxMatches, | |
675 builder_); | |
676 ASSERT_EQ(0U, matches.size()); | |
677 } | |
678 | |
679 TEST_F(InMemoryURLIndexTest, ProperStringMatching) { | |
680 // Search for the following with the expected results: | |
681 // "atdmt view" - found | |
682 // "atdmt.view" - not found | |
683 // "view.atdmt" - found | |
684 ScoredHistoryMatches matches = url_index_->HistoryItemsForTerms( | |
685 ASCIIToUTF16("atdmt view"), base::string16::npos, kMaxMatches, builder_); | |
686 ASSERT_EQ(1U, matches.size()); | |
687 matches = url_index_->HistoryItemsForTerms( | |
688 ASCIIToUTF16("atdmt.view"), base::string16::npos, kMaxMatches, builder_); | |
689 ASSERT_EQ(0U, matches.size()); | |
690 matches = url_index_->HistoryItemsForTerms( | |
691 ASCIIToUTF16("view.atdmt"), base::string16::npos, kMaxMatches, builder_); | |
692 ASSERT_EQ(1U, matches.size()); | |
693 } | |
694 | |
695 TEST_F(InMemoryURLIndexTest, HugeResultSet) { | |
696 // Create a huge set of qualifying history items. | |
697 for (URLID row_id = 5000; row_id < 6000; ++row_id) { | |
698 URLRow new_row(GURL("http://www.brokeandaloneinmanitoba.com/"), row_id); | |
699 new_row.set_last_visit(base::Time::Now()); | |
700 EXPECT_TRUE(UpdateURL(new_row)); | |
701 } | |
702 | |
703 ScoredHistoryMatches matches = url_index_->HistoryItemsForTerms( | |
704 ASCIIToUTF16("b"), base::string16::npos, kMaxMatches, builder_); | |
705 URLIndexPrivateData& private_data(*GetPrivateData()); | |
706 ASSERT_EQ(kMaxMatches, matches.size()); | |
707 // There are 7 matches already in the database. | |
708 ASSERT_EQ(1008U, private_data.pre_filter_item_count_); | |
709 ASSERT_EQ(500U, private_data.post_filter_item_count_); | |
710 ASSERT_EQ(kMaxMatches, private_data.post_scoring_item_count_); | |
711 } | |
712 | |
713 #if defined(OS_WIN) | |
714 // Flaky on windows trybots: http://crbug.com/351500 | |
715 #define MAYBE_TitleSearch DISABLED_TitleSearch | |
716 #else | |
717 #define MAYBE_TitleSearch TitleSearch | |
718 #endif | |
719 TEST_F(InMemoryURLIndexTest, MAYBE_TitleSearch) { | |
720 // Signal if someone has changed the test DB. | |
721 EXPECT_EQ(29U, GetPrivateData()->history_info_map_.size()); | |
722 | |
723 // Ensure title is being searched. | |
724 ScoredHistoryMatches matches = url_index_->HistoryItemsForTerms( | |
725 ASCIIToUTF16("MORTGAGE RATE DROPS"), base::string16::npos, kMaxMatches, | |
726 builder_); | |
727 ASSERT_EQ(1U, matches.size()); | |
728 | |
729 // Verify that we got back the result we expected. | |
730 EXPECT_EQ(1, matches[0].url_info.id()); | |
731 EXPECT_EQ("http://www.reuters.com/article/idUSN0839880620100708", | |
732 matches[0].url_info.url().spec()); | |
733 EXPECT_EQ(ASCIIToUTF16( | |
734 "UPDATE 1-US 30-yr mortgage rate drops to new record low | Reuters"), | |
735 matches[0].url_info.title()); | |
736 } | |
737 | |
738 TEST_F(InMemoryURLIndexTest, TitleChange) { | |
739 // Verify current title terms retrieves desired item. | |
740 base::string16 original_terms = | |
741 ASCIIToUTF16("lebronomics could high taxes influence"); | |
742 ScoredHistoryMatches matches = url_index_->HistoryItemsForTerms( | |
743 original_terms, base::string16::npos, kMaxMatches, builder_); | |
744 ASSERT_EQ(1U, matches.size()); | |
745 | |
746 // Verify that we got back the result we expected. | |
747 const URLID expected_id = 3; | |
748 EXPECT_EQ(expected_id, matches[0].url_info.id()); | |
749 EXPECT_EQ("http://www.businessandmedia.org/articles/2010/20100708120415.aspx", | |
750 matches[0].url_info.url().spec()); | |
751 EXPECT_EQ(ASCIIToUTF16( | |
752 "LeBronomics: Could High Taxes Influence James' Team Decision?"), | |
753 matches[0].url_info.title()); | |
754 URLRow old_row(matches[0].url_info); | |
755 | |
756 // Verify new title terms retrieves nothing. | |
757 base::string16 new_terms = ASCIIToUTF16("does eat oats little lambs ivy"); | |
758 matches = url_index_->HistoryItemsForTerms(new_terms, base::string16::npos, | |
759 kMaxMatches, builder_); | |
760 ASSERT_EQ(0U, matches.size()); | |
761 | |
762 // Update the row. | |
763 old_row.set_title(ASCIIToUTF16("Does eat oats and little lambs eat ivy")); | |
764 EXPECT_TRUE(UpdateURL(old_row)); | |
765 | |
766 // Verify we get the row using the new terms but not the original terms. | |
767 matches = url_index_->HistoryItemsForTerms(new_terms, base::string16::npos, | |
768 kMaxMatches, builder_); | |
769 ASSERT_EQ(1U, matches.size()); | |
770 EXPECT_EQ(expected_id, matches[0].url_info.id()); | |
771 matches = url_index_->HistoryItemsForTerms( | |
772 original_terms, base::string16::npos, kMaxMatches, builder_); | |
773 ASSERT_EQ(0U, matches.size()); | |
774 } | |
775 | |
776 TEST_F(InMemoryURLIndexTest, NonUniqueTermCharacterSets) { | |
777 // The presence of duplicate characters should succeed. Exercise by cycling | |
778 // through a string with several duplicate characters. | |
779 ScoredHistoryMatches matches = url_index_->HistoryItemsForTerms( | |
780 ASCIIToUTF16("ABRA"), base::string16::npos, kMaxMatches, builder_); | |
781 ASSERT_EQ(1U, matches.size()); | |
782 EXPECT_EQ(28, matches[0].url_info.id()); | |
783 EXPECT_EQ("http://www.ddj.com/windows/184416623", | |
784 matches[0].url_info.url().spec()); | |
785 | |
786 matches = url_index_->HistoryItemsForTerms( | |
787 ASCIIToUTF16("ABRACAD"), base::string16::npos, kMaxMatches, builder_); | |
788 ASSERT_EQ(1U, matches.size()); | |
789 EXPECT_EQ(28, matches[0].url_info.id()); | |
790 | |
791 matches = url_index_->HistoryItemsForTerms( | |
792 ASCIIToUTF16("ABRACADABRA"), base::string16::npos, kMaxMatches, builder_); | |
793 ASSERT_EQ(1U, matches.size()); | |
794 EXPECT_EQ(28, matches[0].url_info.id()); | |
795 | |
796 matches = url_index_->HistoryItemsForTerms( | |
797 ASCIIToUTF16("ABRACADABR"), base::string16::npos, kMaxMatches, builder_); | |
798 ASSERT_EQ(1U, matches.size()); | |
799 EXPECT_EQ(28, matches[0].url_info.id()); | |
800 | |
801 matches = url_index_->HistoryItemsForTerms( | |
802 ASCIIToUTF16("ABRACA"), base::string16::npos, kMaxMatches, builder_); | |
803 ASSERT_EQ(1U, matches.size()); | |
804 EXPECT_EQ(28, matches[0].url_info.id()); | |
805 } | |
806 | |
807 TEST_F(InMemoryURLIndexTest, TypedCharacterCaching) { | |
808 // Verify that match results for previously typed characters are retained | |
809 // (in the term_char_word_set_cache_) and reused, if possible, in future | |
810 // autocompletes. | |
811 | |
812 URLIndexPrivateData::SearchTermCacheMap& cache( | |
813 GetPrivateData()->search_term_cache_); | |
814 | |
815 // The cache should be empty at this point. | |
816 EXPECT_EQ(0U, cache.size()); | |
817 | |
818 // Now simulate typing search terms into the omnibox and check the state of | |
819 // the cache as each item is 'typed'. | |
820 | |
821 // Simulate typing "r" giving "r" in the simulated omnibox. The results for | |
822 // 'r' will be not cached because it is only 1 character long. | |
823 url_index_->HistoryItemsForTerms(ASCIIToUTF16("r"), base::string16::npos, | |
824 kMaxMatches, builder_); | |
825 EXPECT_EQ(0U, cache.size()); | |
826 | |
827 // Simulate typing "re" giving "r re" in the simulated omnibox. | |
828 // 're' should be cached at this point but not 'r' as it is a single | |
829 // character. | |
830 url_index_->HistoryItemsForTerms(ASCIIToUTF16("r re"), base::string16::npos, | |
831 kMaxMatches, builder_); | |
832 ASSERT_EQ(1U, cache.size()); | |
833 CheckTerm(cache, ASCIIToUTF16("re")); | |
834 | |
835 // Simulate typing "reco" giving "r re reco" in the simulated omnibox. | |
836 // 're' and 'reco' should be cached at this point but not 'r' as it is a | |
837 // single character. | |
838 url_index_->HistoryItemsForTerms(ASCIIToUTF16("r re reco"), | |
839 base::string16::npos, kMaxMatches, builder_); | |
840 ASSERT_EQ(2U, cache.size()); | |
841 CheckTerm(cache, ASCIIToUTF16("re")); | |
842 CheckTerm(cache, ASCIIToUTF16("reco")); | |
843 | |
844 // Simulate typing "mort". | |
845 // Since we now have only one search term, the cached results for 're' and | |
846 // 'reco' should be purged, giving us only 1 item in the cache (for 'mort'). | |
847 url_index_->HistoryItemsForTerms(ASCIIToUTF16("mort"), base::string16::npos, | |
848 kMaxMatches, builder_); | |
849 ASSERT_EQ(1U, cache.size()); | |
850 CheckTerm(cache, ASCIIToUTF16("mort")); | |
851 | |
852 // Simulate typing "reco" giving "mort reco" in the simulated omnibox. | |
853 url_index_->HistoryItemsForTerms(ASCIIToUTF16("mort reco"), | |
854 base::string16::npos, kMaxMatches, builder_); | |
855 ASSERT_EQ(2U, cache.size()); | |
856 CheckTerm(cache, ASCIIToUTF16("mort")); | |
857 CheckTerm(cache, ASCIIToUTF16("reco")); | |
858 | |
859 // Simulate a <DELETE> by removing the 'reco' and adding back the 'rec'. | |
860 url_index_->HistoryItemsForTerms(ASCIIToUTF16("mort rec"), | |
861 base::string16::npos, kMaxMatches, builder_); | |
862 ASSERT_EQ(2U, cache.size()); | |
863 CheckTerm(cache, ASCIIToUTF16("mort")); | |
864 CheckTerm(cache, ASCIIToUTF16("rec")); | |
865 } | |
866 | |
867 TEST_F(InMemoryURLIndexTest, AddNewRows) { | |
868 // Verify that the row we're going to add does not already exist. | |
869 URLID new_row_id = 87654321; | |
870 // Newly created URLRows get a last_visit time of 'right now' so it should | |
871 // qualify as a quick result candidate. | |
872 EXPECT_TRUE(url_index_->HistoryItemsForTerms(ASCIIToUTF16("brokeandalone"), | |
873 base::string16::npos, | |
874 kMaxMatches, builder_).empty()); | |
875 | |
876 // Add a new row. | |
877 URLRow new_row(GURL("http://www.brokeandaloneinmanitoba.com/"), new_row_id++); | |
878 new_row.set_last_visit(base::Time::Now()); | |
879 EXPECT_TRUE(UpdateURL(new_row)); | |
880 | |
881 // Verify that we can retrieve it. | |
882 EXPECT_EQ(1U, url_index_->HistoryItemsForTerms(ASCIIToUTF16("brokeandalone"), | |
883 base::string16::npos, | |
884 kMaxMatches, builder_).size()); | |
885 | |
886 // Add it again just to be sure that is harmless and that it does not update | |
887 // the index. | |
888 EXPECT_FALSE(UpdateURL(new_row)); | |
889 EXPECT_EQ(1U, url_index_->HistoryItemsForTerms(ASCIIToUTF16("brokeandalone"), | |
890 base::string16::npos, | |
891 kMaxMatches, builder_).size()); | |
892 | |
893 // Make up an URL that does not qualify and try to add it. | |
894 URLRow unqualified_row(GURL("http://www.brokeandaloneinmanitoba.com/"), | |
895 new_row_id++); | |
896 EXPECT_FALSE(UpdateURL(new_row)); | |
897 } | |
898 | |
899 TEST_F(InMemoryURLIndexTest, DeleteRows) { | |
900 ScoredHistoryMatches matches = url_index_->HistoryItemsForTerms( | |
901 ASCIIToUTF16("DrudgeReport"), base::string16::npos, kMaxMatches, | |
902 builder_); | |
903 ASSERT_EQ(1U, matches.size()); | |
904 | |
905 // Delete the URL then search again. | |
906 EXPECT_TRUE(DeleteURL(matches[0].url_info.url())); | |
907 EXPECT_TRUE(url_index_->HistoryItemsForTerms(ASCIIToUTF16("DrudgeReport"), | |
908 base::string16::npos, | |
909 kMaxMatches, builder_).empty()); | |
910 | |
911 // Make up an URL that does not exist in the database and delete it. | |
912 GURL url("http://www.hokeypokey.com/putyourrightfootin.html"); | |
913 EXPECT_FALSE(DeleteURL(url)); | |
914 } | |
915 | |
916 TEST_F(InMemoryURLIndexTest, ExpireRow) { | |
917 ScoredHistoryMatches matches = url_index_->HistoryItemsForTerms( | |
918 ASCIIToUTF16("DrudgeReport"), base::string16::npos, kMaxMatches, | |
919 builder_); | |
920 ASSERT_EQ(1U, matches.size()); | |
921 | |
922 // Determine the row id for the result, remember that id, broadcast a | |
923 // delete notification, then ensure that the row has been deleted. | |
924 URLRows deleted_rows; | |
925 deleted_rows.push_back(matches[0].url_info); | |
926 url_index_->OnURLsDeleted(nullptr, false, false, deleted_rows, | |
927 std::set<GURL>()); | |
928 EXPECT_TRUE(url_index_->HistoryItemsForTerms(ASCIIToUTF16("DrudgeReport"), | |
929 base::string16::npos, | |
930 kMaxMatches, builder_).empty()); | |
931 } | |
932 | |
933 TEST_F(InMemoryURLIndexTest, WhitelistedURLs) { | |
934 struct TestData { | |
935 const std::string url_spec; | |
936 const bool expected_is_whitelisted; | |
937 } data[] = { | |
938 // URLs with whitelisted schemes. | |
939 { "about:histograms", true }, | |
940 { "chrome://settings", true }, | |
941 { "file://localhost/Users/joeschmoe/sekrets", true }, | |
942 { "ftp://public.mycompany.com/myfile.txt", true }, | |
943 { "http://www.google.com/translate", true }, | |
944 { "https://www.gmail.com/", true }, | |
945 { "mailto:support@google.com", true }, | |
946 // URLs with unacceptable schemes. | |
947 { "aaa://www.dummyhost.com;frammy", false }, | |
948 { "aaas://www.dummyhost.com;frammy", false }, | |
949 { "acap://suzie@somebody.com", false }, | |
950 { "cap://cal.example.com/Company/Holidays", false }, | |
951 { "cid:foo4*foo1@bar.net", false }, | |
952 { "crid://example.com/foobar", false }, | |
953 { "data:image/png;base64,iVBORw0KGgoAAAANSUhE=", false }, | |
954 { "dict://dict.org/d:shortcake:", false }, | |
955 { "dns://192.168.1.1/ftp.example.org?type=A", false }, | |
956 { "fax:+358.555.1234567", false }, | |
957 { "geo:13.4125,103.8667", false }, | |
958 { "go:Mercedes%20Benz", false }, | |
959 { "gopher://farnsworth.ca:666/gopher", false }, | |
960 { "h323:farmer-john;sixpence", false }, | |
961 { "iax:johnQ@example.com/12022561414", false }, | |
962 { "icap://icap.net/service?mode=translate&lang=french", false }, | |
963 { "im:fred@example.com", false }, | |
964 { "imap://michael@minbari.org/users.*", false }, | |
965 { "info:ddc/22/eng//004.678", false }, | |
966 { "ipp://example.com/printer/fox", false }, | |
967 { "iris:dreg1//example.com/local/myhosts", false }, | |
968 { "iris.beep:dreg1//example.com/local/myhosts", false }, | |
969 { "iris.lws:dreg1//example.com/local/myhosts", false }, | |
970 { "iris.xpc:dreg1//example.com/local/myhosts", false }, | |
971 { "iris.xpcs:dreg1//example.com/local/myhosts", false }, | |
972 { "ldap://ldap.itd.umich.edu/o=University%20of%20Michigan,c=US", false }, | |
973 { "mid:foo4%25foo1@bar.net", false }, | |
974 { "modem:+3585551234567;type=v32b?7e1;type=v110", false }, | |
975 { "msrp://atlanta.example.com:7654/jshA7weztas;tcp", false }, | |
976 { "msrps://atlanta.example.com:7654/jshA7weztas;tcp", false }, | |
977 { "news:colorectal.info.banned", false }, | |
978 { "nfs://server/d/e/f", false }, | |
979 { "nntp://www.example.com:6543/info.comp.lies/1234", false }, | |
980 { "pop://rg;AUTH=+APOP@mail.mycompany.com:8110", false }, | |
981 { "pres:fred@example.com", false }, | |
982 { "prospero://host.dom//pros/name", false }, | |
983 { "rsync://syler@lost.com/Source", false }, | |
984 { "rtsp://media.example.com:554/twister/audiotrack", false }, | |
985 { "service:acap://some.where.net;authentication=KERBEROSV4", false }, | |
986 { "shttp://www.terces.com/secret", false }, | |
987 { "sieve://example.com//script", false }, | |
988 { "sip:+1-212-555-1212:1234@gateway.com;user=phone", false }, | |
989 { "sips:+1-212-555-1212:1234@gateway.com;user=phone", false }, | |
990 { "sms:+15105551212?body=hello%20there", false }, | |
991 { "snmp://tester5@example.com:8161/bridge1;800002b804616263", false }, | |
992 { "soap.beep://stockquoteserver.example.com/StockQuote", false }, | |
993 { "soap.beeps://stockquoteserver.example.com/StockQuote", false }, | |
994 { "tag:blogger.com,1999:blog-555", false }, | |
995 { "tel:+358-555-1234567;postd=pp22", false }, | |
996 { "telnet://mayor_margie:one2rule4All@www.mycity.com:6789/", false }, | |
997 { "tftp://example.com/mystartupfile", false }, | |
998 { "tip://123.123.123.123/?urn:xopen:xid", false }, | |
999 { "tv:nbc.com", false }, | |
1000 { "urn:foo:A123,456", false }, | |
1001 { "vemmi://zeus.mctel.fr/demo", false }, | |
1002 { "wais://www.mydomain.net:8765/mydatabase", false }, | |
1003 { "xmpp:node@example.com", false }, | |
1004 { "xmpp://guest@example.com", false }, | |
1005 }; | |
1006 | |
1007 const std::set<std::string>& whitelist(scheme_whitelist()); | |
1008 for (size_t i = 0; i < arraysize(data); ++i) { | |
1009 GURL url(data[i].url_spec); | |
1010 EXPECT_EQ(data[i].expected_is_whitelisted, | |
1011 URLIndexPrivateData::URLSchemeIsWhitelisted(url, whitelist)); | |
1012 } | |
1013 } | |
1014 | |
1015 TEST_F(InMemoryURLIndexTest, ReadVisitsFromHistory) { | |
1016 const HistoryInfoMap& history_info_map = GetPrivateData()->history_info_map_; | |
1017 | |
1018 // Check (for URL with id 1) that the number of visits and their | |
1019 // transition types are what we expect. We don't bother checking | |
1020 // the timestamps because it's too much trouble. (The timestamps go | |
1021 // through a transformation in InMemoryURLIndexTest::SetUp(). We | |
1022 // assume that if the count and transitions show up with the right | |
1023 // information, we're getting the right information from the history | |
1024 // database file.) | |
1025 HistoryInfoMap::const_iterator entry = history_info_map.find(1); | |
1026 ASSERT_TRUE(entry != history_info_map.end()); | |
1027 { | |
1028 const VisitInfoVector& visits = entry->second.visits; | |
1029 EXPECT_EQ(3u, visits.size()); | |
1030 EXPECT_EQ(0u, visits[0].second); | |
1031 EXPECT_EQ(1u, visits[1].second); | |
1032 EXPECT_EQ(0u, visits[2].second); | |
1033 } | |
1034 | |
1035 // Ditto but for URL with id 35. | |
1036 entry = history_info_map.find(35); | |
1037 ASSERT_TRUE(entry != history_info_map.end()); | |
1038 { | |
1039 const VisitInfoVector& visits = entry->second.visits; | |
1040 EXPECT_EQ(2u, visits.size()); | |
1041 EXPECT_EQ(1u, visits[0].second); | |
1042 EXPECT_EQ(1u, visits[1].second); | |
1043 } | |
1044 | |
1045 // The URL with id 32 has many visits listed in the database, but we | |
1046 // should only read the most recent 10 (which are all transition type 0). | |
1047 entry = history_info_map.find(32); | |
1048 ASSERT_TRUE(entry != history_info_map.end()); | |
1049 { | |
1050 const VisitInfoVector& visits = entry->second.visits; | |
1051 EXPECT_EQ(10u, visits.size()); | |
1052 for (size_t i = 0; i < visits.size(); ++i) | |
1053 EXPECT_EQ(0u, visits[i].second); | |
1054 } | |
1055 } | |
1056 | |
1057 TEST_F(InMemoryURLIndexTest, CacheSaveRestore) { | |
1058 base::ScopedTempDir temp_directory; | |
1059 ASSERT_TRUE(temp_directory.CreateUniqueTempDir()); | |
1060 set_history_dir(temp_directory.path()); | |
1061 | |
1062 URLIndexPrivateData& private_data(*GetPrivateData()); | |
1063 | |
1064 // Ensure that there is really something there to be saved. | |
1065 EXPECT_FALSE(private_data.word_list_.empty()); | |
1066 // available_words_ will already be empty since we have freshly built the | |
1067 // data set for this test. | |
1068 EXPECT_TRUE(private_data.available_words_.empty()); | |
1069 EXPECT_FALSE(private_data.word_map_.empty()); | |
1070 EXPECT_FALSE(private_data.char_word_map_.empty()); | |
1071 EXPECT_FALSE(private_data.word_id_history_map_.empty()); | |
1072 EXPECT_FALSE(private_data.history_id_word_map_.empty()); | |
1073 EXPECT_FALSE(private_data.history_info_map_.empty()); | |
1074 EXPECT_FALSE(private_data.word_starts_map_.empty()); | |
1075 | |
1076 // Make sure the data we have was built from history. (Version 0 | |
1077 // means rebuilt from history.) | |
1078 EXPECT_EQ(0, private_data.restored_cache_version_); | |
1079 | |
1080 // Capture the current private data for later comparison to restored data. | |
1081 scoped_refptr<URLIndexPrivateData> old_data(private_data.Duplicate()); | |
1082 const base::Time rebuild_time = private_data.last_time_rebuilt_from_history_; | |
1083 | |
1084 { | |
1085 // Save then restore our private data. | |
1086 base::RunLoop run_loop; | |
1087 CacheFileSaverObserver save_observer(run_loop.QuitClosure()); | |
1088 url_index_->set_save_cache_observer(&save_observer); | |
1089 PostSaveToCacheFileTask(); | |
1090 run_loop.Run(); | |
1091 EXPECT_TRUE(save_observer.succeeded()); | |
1092 } | |
1093 | |
1094 // Clear and then prove it's clear before restoring. | |
1095 ClearPrivateData(); | |
1096 EXPECT_TRUE(private_data.word_list_.empty()); | |
1097 EXPECT_TRUE(private_data.available_words_.empty()); | |
1098 EXPECT_TRUE(private_data.word_map_.empty()); | |
1099 EXPECT_TRUE(private_data.char_word_map_.empty()); | |
1100 EXPECT_TRUE(private_data.word_id_history_map_.empty()); | |
1101 EXPECT_TRUE(private_data.history_id_word_map_.empty()); | |
1102 EXPECT_TRUE(private_data.history_info_map_.empty()); | |
1103 EXPECT_TRUE(private_data.word_starts_map_.empty()); | |
1104 | |
1105 { | |
1106 base::RunLoop run_loop; | |
1107 HistoryIndexRestoreObserver restore_observer(run_loop.QuitClosure()); | |
1108 url_index_->set_restore_cache_observer(&restore_observer); | |
1109 PostRestoreFromCacheFileTask(); | |
1110 run_loop.Run(); | |
1111 EXPECT_TRUE(restore_observer.succeeded()); | |
1112 } | |
1113 | |
1114 URLIndexPrivateData& new_data(*GetPrivateData()); | |
1115 | |
1116 // Make sure the data we have was reloaded from cache. (Version 0 | |
1117 // means rebuilt from history; anything else means restored from | |
1118 // a cache version.) Also, the rebuild time should not have changed. | |
1119 EXPECT_GT(new_data.restored_cache_version_, 0); | |
1120 EXPECT_EQ(rebuild_time, new_data.last_time_rebuilt_from_history_); | |
1121 | |
1122 // Compare the captured and restored for equality. | |
1123 ExpectPrivateDataEqual(*old_data.get(), new_data); | |
1124 } | |
1125 | |
1126 #if defined(OS_WIN) | |
1127 // http://crbug.com/351500 | |
1128 #define MAYBE_RebuildFromHistoryIfCacheOld DISABLED_RebuildFromHistoryIfCacheOld | |
1129 #else | |
1130 #define MAYBE_RebuildFromHistoryIfCacheOld RebuildFromHistoryIfCacheOld | |
1131 #endif | |
1132 TEST_F(InMemoryURLIndexTest, MAYBE_RebuildFromHistoryIfCacheOld) { | |
1133 base::ScopedTempDir temp_directory; | |
1134 ASSERT_TRUE(temp_directory.CreateUniqueTempDir()); | |
1135 set_history_dir(temp_directory.path()); | |
1136 | |
1137 URLIndexPrivateData& private_data(*GetPrivateData()); | |
1138 | |
1139 // Ensure that there is really something there to be saved. | |
1140 EXPECT_FALSE(private_data.word_list_.empty()); | |
1141 // available_words_ will already be empty since we have freshly built the | |
1142 // data set for this test. | |
1143 EXPECT_TRUE(private_data.available_words_.empty()); | |
1144 EXPECT_FALSE(private_data.word_map_.empty()); | |
1145 EXPECT_FALSE(private_data.char_word_map_.empty()); | |
1146 EXPECT_FALSE(private_data.word_id_history_map_.empty()); | |
1147 EXPECT_FALSE(private_data.history_id_word_map_.empty()); | |
1148 EXPECT_FALSE(private_data.history_info_map_.empty()); | |
1149 EXPECT_FALSE(private_data.word_starts_map_.empty()); | |
1150 | |
1151 // Make sure the data we have was built from history. (Version 0 | |
1152 // means rebuilt from history.) | |
1153 EXPECT_EQ(0, private_data.restored_cache_version_); | |
1154 | |
1155 // Overwrite the build time so that we'll think the data is too old | |
1156 // and rebuild the cache from history. | |
1157 const base::Time fake_rebuild_time = | |
1158 private_data.last_time_rebuilt_from_history_ - | |
1159 base::TimeDelta::FromDays(30); | |
1160 private_data.last_time_rebuilt_from_history_ = fake_rebuild_time; | |
1161 | |
1162 // Capture the current private data for later comparison to restored data. | |
1163 scoped_refptr<URLIndexPrivateData> old_data(private_data.Duplicate()); | |
1164 | |
1165 { | |
1166 // Save then restore our private data. | |
1167 base::RunLoop run_loop; | |
1168 CacheFileSaverObserver save_observer(run_loop.QuitClosure()); | |
1169 url_index_->set_save_cache_observer(&save_observer); | |
1170 PostSaveToCacheFileTask(); | |
1171 run_loop.Run(); | |
1172 EXPECT_TRUE(save_observer.succeeded()); | |
1173 } | |
1174 | |
1175 // Clear and then prove it's clear before restoring. | |
1176 ClearPrivateData(); | |
1177 EXPECT_TRUE(private_data.word_list_.empty()); | |
1178 EXPECT_TRUE(private_data.available_words_.empty()); | |
1179 EXPECT_TRUE(private_data.word_map_.empty()); | |
1180 EXPECT_TRUE(private_data.char_word_map_.empty()); | |
1181 EXPECT_TRUE(private_data.word_id_history_map_.empty()); | |
1182 EXPECT_TRUE(private_data.history_id_word_map_.empty()); | |
1183 EXPECT_TRUE(private_data.history_info_map_.empty()); | |
1184 EXPECT_TRUE(private_data.word_starts_map_.empty()); | |
1185 | |
1186 { | |
1187 base::RunLoop run_loop; | |
1188 HistoryIndexRestoreObserver restore_observer(run_loop.QuitClosure()); | |
1189 url_index_->set_restore_cache_observer(&restore_observer); | |
1190 PostRestoreFromCacheFileTask(); | |
1191 run_loop.Run(); | |
1192 EXPECT_TRUE(restore_observer.succeeded()); | |
1193 } | |
1194 | |
1195 URLIndexPrivateData& new_data(*GetPrivateData()); | |
1196 | |
1197 // Make sure the data we have was rebuilt from history. (Version 0 | |
1198 // means rebuilt from history; anything else means restored from | |
1199 // a cache version.) | |
1200 EXPECT_EQ(0, new_data.restored_cache_version_); | |
1201 EXPECT_NE(fake_rebuild_time, new_data.last_time_rebuilt_from_history_); | |
1202 | |
1203 // Compare the captured and restored for equality. | |
1204 ExpectPrivateDataEqual(*old_data.get(), new_data); | |
1205 } | |
1206 | |
1207 class InMemoryURLIndexCacheTest : public testing::Test { | |
1208 public: | |
1209 InMemoryURLIndexCacheTest() {} | |
1210 | |
1211 protected: | |
1212 void SetUp() override; | |
1213 void TearDown() override; | |
1214 | |
1215 // Pass-through functions to simplify our friendship with InMemoryURLIndex. | |
1216 void set_history_dir(const base::FilePath& dir_path); | |
1217 bool GetCacheFilePath(base::FilePath* file_path) const; | |
1218 | |
1219 base::ScopedTempDir temp_dir_; | |
1220 scoped_ptr<InMemoryURLIndex> url_index_; | |
1221 }; | |
1222 | |
1223 void InMemoryURLIndexCacheTest::SetUp() { | |
1224 ASSERT_TRUE(temp_dir_.CreateUniqueTempDir()); | |
1225 base::FilePath path(temp_dir_.path()); | |
1226 url_index_.reset(new InMemoryURLIndex(nullptr, path, kTestLanguages)); | |
1227 } | |
1228 | |
1229 void InMemoryURLIndexCacheTest::TearDown() { | |
1230 if (url_index_) | |
1231 url_index_->ShutDown(); | |
1232 } | |
1233 | |
1234 void InMemoryURLIndexCacheTest::set_history_dir( | |
1235 const base::FilePath& dir_path) { | |
1236 return url_index_->set_history_dir(dir_path); | |
1237 } | |
1238 | |
1239 bool InMemoryURLIndexCacheTest::GetCacheFilePath( | |
1240 base::FilePath* file_path) const { | |
1241 DCHECK(file_path); | |
1242 return url_index_->GetCacheFilePath(file_path); | |
1243 } | |
1244 | |
1245 TEST_F(InMemoryURLIndexCacheTest, CacheFilePath) { | |
1246 base::FilePath expectedPath = | |
1247 temp_dir_.path().Append(FILE_PATH_LITERAL("History Provider Cache")); | |
1248 std::vector<base::FilePath::StringType> expected_parts; | |
1249 expectedPath.GetComponents(&expected_parts); | |
1250 base::FilePath full_file_path; | |
1251 ASSERT_TRUE(GetCacheFilePath(&full_file_path)); | |
1252 std::vector<base::FilePath::StringType> actual_parts; | |
1253 full_file_path.GetComponents(&actual_parts); | |
1254 ASSERT_EQ(expected_parts.size(), actual_parts.size()); | |
1255 size_t count = expected_parts.size(); | |
1256 for (size_t i = 0; i < count; ++i) | |
1257 EXPECT_EQ(expected_parts[i], actual_parts[i]); | |
1258 // Must clear the history_dir_ to satisfy the dtor's DCHECK. | |
1259 set_history_dir(base::FilePath()); | |
1260 } | |
1261 | |
1262 } // namespace history | |
OLD | NEW |