Moved ImportantFileWriter to chrome/common in preparation of moving JsonPrefStore to chrome/common.

BUG=None.
TEST=Unit-tests moved and modified.

Review URL: http://codereview.chromium.org/2128014

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@47818 0039d316-1c4b-4281-b951-d872f2087c98

3 files changed, 377 insertions, 0 deletions

diff --git a/chrome/common/important_file_writer.cc b/chrome/common/important_file_writer.cc
new file mode 100644
index 0000000..3466958
--- /dev/null
+++ b/chrome/common/important_file_writer.cc
@@ -0,0 +1,150 @@
+// Copyright (c) 2010 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/common/important_file_writer.h"
+
+#include <stdio.h>
+
+#include <ostream>
+#include <string>
+
+#include "base/file_path.h"
+#include "base/file_util.h"
+#include "base/logging.h"
+#include "base/message_loop_proxy.h"
+#include "base/string_util.h"
+#include "base/task.h"
+#include "base/thread.h"
+#include "base/time.h"
+
+using base::TimeDelta;
+
+namespace {
+
+const int kDefaultCommitIntervalMs = 10000;
+
+class WriteToDiskTask : public Task {
+ public:
+  WriteToDiskTask(const FilePath& path, const std::string& data)
+      : path_(path),
+        data_(data) {
+  }
+
+  virtual void Run() {
+    // Write the data to a temp file then rename to avoid data loss if we crash
+    // while writing the file. Ensure that the temp file is on the same volume
+    // as target file, so it can be moved in one step, and that the temp file
+    // is securely created.
+    FilePath tmp_file_path;
+    FILE* tmp_file = file_util::CreateAndOpenTemporaryFileInDir(
+        path_.DirName(), &tmp_file_path);
+    if (!tmp_file) {
+      LogFailure("could not create temporary file");
+      return;
+    }
+
+    size_t bytes_written = fwrite(data_.data(), 1, data_.length(), tmp_file);
+    if (!file_util::CloseFile(tmp_file)) {
+      file_util::Delete(tmp_file_path, false);
+      LogFailure("failed to close temporary file");
+      return;
+    }
+    if (bytes_written < data_.length()) {
+      file_util::Delete(tmp_file_path, false);
+      LogFailure("error writing, bytes_written=" + UintToString(
+          static_cast<unsigned int>(bytes_written)));
+      return;
+    }
+
+    if (file_util::ReplaceFile(tmp_file_path, path_)) {
+      LogSuccess();
+      return;
+    }
+
+    file_util::Delete(tmp_file_path, false);
+    LogFailure("could not rename temporary file");
+  }
+
+ private:
+  void LogSuccess() {
+    LOG(INFO) << "successfully saved " << path_.value();
+  }
+
+  void LogFailure(const std::string& message) {
+    LOG(WARNING) << "failed to write " << path_.value()
+                 << ": " << message;
+  }
+
+  const FilePath path_;
+  const std::string data_;
+
+  DISALLOW_COPY_AND_ASSIGN(WriteToDiskTask);
+};
+
+}  // namespace
+
+ImportantFileWriter::ImportantFileWriter(
+    const FilePath& path, base::MessageLoopProxy* file_message_loop_proxy)
+        : path_(path),
+          file_message_loop_proxy_(file_message_loop_proxy),
+          serializer_(NULL),
+          commit_interval_(TimeDelta::FromMilliseconds(
+              kDefaultCommitIntervalMs)) {
+  DCHECK(CalledOnValidThread());
+  DCHECK(file_message_loop_proxy_.get());
+}
+
+ImportantFileWriter::~ImportantFileWriter() {
+  // We're usually a member variable of some other object, which also tends
+  // to be our serializer. It may not be safe to call back to the parent object
+  // being destructed.
+  DCHECK(!HasPendingWrite());
+}
+
+bool ImportantFileWriter::HasPendingWrite() const {
+  DCHECK(CalledOnValidThread());
+  return timer_.IsRunning();
+}
+
+void ImportantFileWriter::WriteNow(const std::string& data) {
+  DCHECK(CalledOnValidThread());
+
+  if (HasPendingWrite())
+    timer_.Stop();
+
+  // TODO(sanjeevr): Add a DCHECK for the return value of PostTask.
+  // (Some tests fail if we add the DCHECK and they need to be fixed first).
+  file_message_loop_proxy_->PostTask(FROM_HERE,
+                                     new WriteToDiskTask(path_, data));
+}
+
+void ImportantFileWriter::ScheduleWrite(DataSerializer* serializer) {
+  DCHECK(CalledOnValidThread());
+
+  DCHECK(serializer);
+  serializer_ = serializer;
+
+  if (!MessageLoop::current()) {
+    // Happens in unit tests.
+    DoScheduledWrite();
+    return;
+  }
+
+  if (!timer_.IsRunning()) {
+    timer_.Start(commit_interval_, this,
+                 &ImportantFileWriter::DoScheduledWrite);
+  }
+}
+
+void ImportantFileWriter::DoScheduledWrite() {
+  DCHECK(serializer_);
+  std::string data;
+  if (serializer_->SerializeData(&data)) {
+    WriteNow(data);
+  } else {
+    LOG(WARNING) << "failed to serialize data to be saved in "
+                 << path_.value();
+  }
+  serializer_ = NULL;
+}
diff --git a/chrome/common/important_file_writer.h b/chrome/common/important_file_writer.h
new file mode 100644
index 0000000..d4f57a0
--- /dev/null
+++ b/chrome/common/important_file_writer.h
@@ -0,0 +1,112 @@
+// Copyright (c) 2009 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.
+
+#ifndef CHROME_COMMON_IMPORTANT_FILE_WRITER_H_
+#define CHROME_COMMON_IMPORTANT_FILE_WRITER_H_
+
+#include <string>
+
+#include "base/basictypes.h"
+#include "base/file_path.h"
+#include "base/non_thread_safe.h"
+#include "base/ref_counted.h"
+#include "base/time.h"
+#include "base/timer.h"
+
+namespace base {
+class MessageLoopProxy;
+class Thread;
+}
+
+// Helper to ensure that a file won't be corrupted by the write (for example on
+// application crash). Consider a naive way to save an important file F:
+//
+// 1. Open F for writing, truncating it.
+// 2. Write new data to F.
+//
+// It's good when it works, but it gets very bad if step 2. doesn't complete.
+// It can be caused by a crash, a computer hang, or a weird I/O error. And you
+// end up with a broken file.
+//
+// To be safe, we don't start with writing directly to F. Instead, we write to
+// to a temporary file. Only after that write is successful, we rename the
+// temporary file to target filename.
+//
+// If you want to know more about this approach and ext3/ext4 fsync issues, see
+// http://valhenson.livejournal.com/37921.html
+class ImportantFileWriter : public NonThreadSafe {
+ public:
+  // Used by ScheduleSave to lazily provide the data to be saved. Allows us
+  // to also batch data serializations.
+  class DataSerializer {
+   public:
+    virtual ~DataSerializer() {}
+
+    // Should put serialized string in |data| and return true on successful
+    // serialization. Will be called on the same thread on which
+    // ImportantFileWriter has been created.
+    virtual bool SerializeData(std::string* data) = 0;
+  };
+
+  // Initialize the writer.
+  // |path| is the name of file to write.
+  // |file_message_loop_proxy| is the MessageLoopProxy for a thread on which
+  // file I/O can be done.
+  // All non-const methods, ctor and dtor must be called on the same thread.
+  ImportantFileWriter(const FilePath& path,
+                      base::MessageLoopProxy* file_message_loop_proxy);
+
+  // You have to ensure that there are no pending writes at the moment
+  // of destruction.
+  ~ImportantFileWriter();
+
+  FilePath path() const { return path_; }
+
+  // Returns true if there is a scheduled write pending which has not yet
+  // been started.
+  bool HasPendingWrite() const;
+
+  // Save |data| to target filename. Does not block. If there is a pending write
+  // scheduled by ScheduleWrite, it is cancelled.
+  void WriteNow(const std::string& data);
+
+  // Schedule a save to target filename. Data will be serialized and saved
+  // to disk after the commit interval. If another ScheduleWrite is issued
+  // before that, only one serialization and write to disk will happen, and
+  // the most recent |serializer| will be used. This operation does not block.
+  // |serializer| should remain valid through the lifetime of
+  // ImportantFileWriter.
+  void ScheduleWrite(DataSerializer* serializer);
+
+  // Serialize data pending to be saved and execute write on backend thread.
+  void DoScheduledWrite();
+
+  base::TimeDelta commit_interval() const {
+    return commit_interval_;
+  }
+
+  void set_commit_interval(const base::TimeDelta& interval) {
+    commit_interval_ = interval;
+  }
+
+ private:
+  // Path being written to.
+  const FilePath path_;
+
+  // MessageLoopProxy for the thread on which file I/O can be done.
+  scoped_refptr<base::MessageLoopProxy> file_message_loop_proxy_;
+
+  // Timer used to schedule commit after ScheduleWrite.
+  base::OneShotTimer<ImportantFileWriter> timer_;
+
+  // Serializer which will provide the data to be saved.
+  DataSerializer* serializer_;
+
+  // Time delta after which scheduled data will be written to disk.
+  base::TimeDelta commit_interval_;
+
+  DISALLOW_COPY_AND_ASSIGN(ImportantFileWriter);
+};
+
+#endif  // CHROME_COMMON_IMPORTANT_FILE_WRITER_H_
diff --git a/chrome/common/important_file_writer_unittest.cc b/chrome/common/important_file_writer_unittest.cc
new file mode 100644
index 0000000..3d44ae0
--- /dev/null
+++ b/chrome/common/important_file_writer_unittest.cc
@@ -0,0 +1,115 @@
+// Copyright (c) 2009 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/common/important_file_writer.h"
+
+#include "base/compiler_specific.h"
+#include "base/file_path.h"
+#include "base/file_util.h"
+#include "base/logging.h"
+#include "base/message_loop.h"
+#include "base/scoped_temp_dir.h"
+#include "base/thread.h"
+#include "base/time.h"
+#include "testing/gtest/include/gtest/gtest.h"
+
+namespace {
+
+std::string GetFileContent(const FilePath& path) {
+  std::string content;
+  if (!file_util::ReadFileToString(path, &content)) {
+    NOTREACHED();
+  }
+  return content;
+}
+
+class DataSerializer : public ImportantFileWriter::DataSerializer {
+ public:
+  explicit DataSerializer(const std::string& data) : data_(data) {
+  }
+
+  virtual bool SerializeData(std::string* output) {
+    output->assign(data_);
+    return true;
+  }
+
+ private:
+  const std::string data_;
+};
+
+}  // namespace
+
+class ImportantFileWriterTest : public testing::Test {
+ public:
+  ImportantFileWriterTest() { }
+  virtual void SetUp() {
+    ASSERT_TRUE(temp_dir_.CreateUniqueTempDir());
+    file_ = temp_dir_.path().AppendASCII("test-file");
+  }
+
+ protected:
+  FilePath file_;
+  MessageLoop loop_;
+
+ private:
+  ScopedTempDir temp_dir_;
+};
+
+TEST_F(ImportantFileWriterTest, Basic) {
+  ImportantFileWriter writer(file_,
+                             base::MessageLoopProxy::CreateForCurrentThread());
+  EXPECT_FALSE(file_util::PathExists(writer.path()));
+  writer.WriteNow("foo");
+  loop_.RunAllPending();
+
+  ASSERT_TRUE(file_util::PathExists(writer.path()));
+  EXPECT_EQ("foo", GetFileContent(writer.path()));
+}
+
+TEST_F(ImportantFileWriterTest, ScheduleWrite) {
+  ImportantFileWriter writer(file_,
+                             base::MessageLoopProxy::CreateForCurrentThread());
+  writer.set_commit_interval(base::TimeDelta::FromMilliseconds(25));
+  EXPECT_FALSE(writer.HasPendingWrite());
+  DataSerializer serializer("foo");
+  writer.ScheduleWrite(&serializer);
+  EXPECT_TRUE(writer.HasPendingWrite());
+  MessageLoop::current()->PostDelayedTask(FROM_HERE,
+                                          new MessageLoop::QuitTask(), 100);
+  MessageLoop::current()->Run();
+  EXPECT_FALSE(writer.HasPendingWrite());
+  ASSERT_TRUE(file_util::PathExists(writer.path()));
+  EXPECT_EQ("foo", GetFileContent(writer.path()));
+}
+
+TEST_F(ImportantFileWriterTest, DoScheduledWrite) {
+  ImportantFileWriter writer(file_,
+                             base::MessageLoopProxy::CreateForCurrentThread());
+  EXPECT_FALSE(writer.HasPendingWrite());
+  DataSerializer serializer("foo");
+  writer.ScheduleWrite(&serializer);
+  EXPECT_TRUE(writer.HasPendingWrite());
+  writer.DoScheduledWrite();
+  MessageLoop::current()->PostDelayedTask(FROM_HERE,
+                                          new MessageLoop::QuitTask(), 100);
+  MessageLoop::current()->Run();
+  EXPECT_FALSE(writer.HasPendingWrite());
+  ASSERT_TRUE(file_util::PathExists(writer.path()));
+  EXPECT_EQ("foo", GetFileContent(writer.path()));
+}
+
+TEST_F(ImportantFileWriterTest, BatchingWrites) {
+  ImportantFileWriter writer(file_,
+                             base::MessageLoopProxy::CreateForCurrentThread());
+  writer.set_commit_interval(base::TimeDelta::FromMilliseconds(25));
+  DataSerializer foo("foo"), bar("bar"), baz("baz");
+  writer.ScheduleWrite(&foo);
+  writer.ScheduleWrite(&bar);
+  writer.ScheduleWrite(&baz);
+  MessageLoop::current()->PostDelayedTask(FROM_HERE,
+                                          new MessageLoop::QuitTask(), 100);
+  MessageLoop::current()->Run();
+  ASSERT_TRUE(file_util::PathExists(writer.path()));
+  EXPECT_EQ("baz", GetFileContent(writer.path()));
+}