Implement support for importing favorites from Edge on Windows 10.

chrome/utility/importer/edge_importer_win.cc

Index: chrome/utility/importer/edge_importer_win.cc
diff --git a/chrome/utility/importer/edge_importer_win.cc b/chrome/utility/importer/edge_importer_win.cc
new file mode 100644
index 0000000000000000000000000000000000000000..c4ffdffe3b8ee462714f76f5ff7b7404a57e05da
--- /dev/null
+++ b/chrome/utility/importer/edge_importer_win.cc
@@ -0,0 +1,555 @@
+// Copyright 2015 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/utility/importer/edge_importer_win.h"
+
+#define JET_UNICODE
+#include <esent.h>
+#undef JET_UNICODE
+#include <Shlobj.h>
+
+#include <algorithm>
+#include <map>
+#include <string>
+#include <tuple>
+#include <vector>
+
+#include "base/files/file_enumerator.h"
+#include "base/files/file_path.h"
+#include "base/files/file_util.h"
+#include "base/memory/ref_counted.h"
+#include "base/memory/scoped_ptr.h"
+#include "base/strings/string16.h"
+#include "base/strings/string_util.h"
+#include "base/time/time.h"
+#include "base/win/windows_version.h"
+#include "chrome/common/importer/edge_importer_utils_win.h"
+#include "chrome/common/importer/imported_bookmark_entry.h"
+#include "chrome/common/importer/importer_bridge.h"
+#include "chrome/grit/generated_resources.h"
+#include "chrome/utility/importer/favicon_reencode.h"
+#include "ui/base/l10n/l10n_util.h"
+#include "url/gurl.h"
+
+namespace {
+
+// Toolbar favorites are placed under this special folder name.
+const base::char16 kFavoritesBarTitle[] = L"_Favorites_Bar_";
+const base::char16 kSpartanDatabaseFile[] = L"spartan.edb";
+
+class EdgeErrorObject {
+ public:
+  EdgeErrorObject() : last_error_(JET_errSuccess) {}
+
+  base::string16 GetErrorMessage() const {
+    WCHAR error_message[kErrorMessagesize] = {};
+    JET_API_PTR err = last_error_;
+    JET_ERR result = JetGetSystemParameter(
+        JET_instanceNil, JET_sesidNil, JET_paramErrorToString, &err,
+        error_message, sizeof(error_message));
+    if (result != JET_errSuccess)
+      return L"";
+
+    return error_message;
+  }
+
+ protected:
+  JET_ERR last_error() const { return last_error_; }
+
+  // This function returns true if the passed error parameter is equal
+  // to JET_errSuccess
+  bool SetLastError(JET_ERR error) {
+    last_error_ = error;
+    return error == JET_errSuccess;
+  }
+
+ private:
+  static const size_t kErrorMessagesize = 1024;

[Ilya Sherman, 2015/12/01 07:31:18]

nit: Please move this to the top of the file, as with the string constants.

[Ilya Sherman, 2015/12/01 07:31:18]

Since you mentioned that this is an arbitrarily chosen size, let's make sure to
document that the choice is arbitrary.  It might help future code spelunkers =)

[Ilya Sherman, 2015/12/01 07:31:18]

nit: s/size/Size

[forshaw, 2015/12/01 11:12:12]

Acknowledged.

[forshaw, 2015/12/01 11:12:13]

Acknowledged.

[forshaw, 2015/12/01 11:12:13]

Acknowledged.

+
+  JET_ERR last_error_;
+
+  DISALLOW_COPY_AND_ASSIGN(EdgeErrorObject);
+};
+
+class EdgeDatabaseTableEnumerator : public EdgeErrorObject {
+ public:
+  EdgeDatabaseTableEnumerator(const base::string16& table_name,
+                              JET_SESID session_id,
+                              JET_TABLEID table_id)
+      : table_id_(table_id), table_name_(table_name), session_id_(session_id) {}
+
+  ~EdgeDatabaseTableEnumerator() {
+    if (table_id_ != JET_tableidNil) {
+      JetCloseTable(session_id_, table_id_);
+    }

[Ilya Sherman, 2015/12/01 07:31:18]

nit: You can now drop the curly braces.

[forshaw, 2015/12/01 11:12:13]

Acknowledged.

+  }
+
+  const base::string16& table_name() { return table_name_; }
+
+  bool Reset() {
+    return SetLastError(JetMove(session_id_, table_id_, JET_MoveFirst, 0));
+  }
+
+  bool Next() {
+    return SetLastError(JetMove(session_id_, table_id_, JET_MoveNext, 0));
+  }
+
+  template <typename T>
+  bool RetrieveColumn(const base::string16& column_name, T* value) {
+    const JET_COLUMNBASE& column_base = GetColumnByName(column_name);
+    if (column_base.cbMax == 0)
+      return false;
+    std::vector<uint8_t> column_data(column_base.cbMax);
+    unsigned long actual_size = 0;
+    JET_ERR err = JetRetrieveColumn(
+        session_id_, table_id_, column_base.columnid, &column_data[0],
+        column_data.size(), &actual_size, 0, nullptr);
+    SetLastError(err);
+    if (err != JET_errSuccess && err != JET_wrnColumnNull) {
+      return false;
+    }
+
+    if (err == JET_errSuccess) {
+      column_data.resize(actual_size);
+      if (!ValidateAndConvertValue(column_base.coltyp, column_data, value))
+        return false;
+    } else {
+      *value = T();
+    }
+
+    return true;
+  }
+
+ private:
+  template <typename T>
+  bool ValidateAndConvertValueGeneric(const JET_COLTYP match_column_type,
+                                      const JET_COLTYP column_type,
+                                      const std::vector<uint8_t>& column_data,
+                                      T* value) {
+    if ((column_type == match_column_type) &&
+        (column_data.size() >= sizeof(T))) {
+      memcpy(value, &column_data[0], sizeof(T));
+      return true;
+    }
+    SetLastError(JET_errInvalidColumnType);
+    return false;
+  }
+
+  bool ValidateAndConvertValue(const JET_COLTYP column_type,
+                               const std::vector<uint8_t>& column_data,
+                               bool* value) {
+    if ((column_type == JET_coltypBit) && (column_data.size() >= 1)) {
+      *value = (column_data[0] & 1) == 1;
+      return true;
+    }
+    SetLastError(JET_errInvalidColumnType);
+    return false;
+  }
+
+  bool ValidateAndConvertValue(const JET_COLTYP column_type,
+                               const std::vector<uint8_t>& column_data,
+                               base::string16* value) {
+    if ((column_type == JET_coltypLongText) &&
+        ((column_data.size() % sizeof(base::char16)) == 0)) {
+      base::string16& value_ref = *value;
+      size_t char_length = column_data.size() / sizeof(base::char16);
+      value_ref.resize(column_data.size() / sizeof(base::char16));

[Ilya Sherman, 2015/12/01 07:31:18]

nit: Reuse char_length here?

[forshaw, 2015/12/01 11:12:12]

Acknowledged.

+      memcpy(&value_ref[0], &column_data[0], column_data.size());
+      // Remove any trailing NUL characters.
+      while (char_length > 0) {
+        if (value_ref[char_length - 1])
+          break;
+        char_length--;
+      }
+      value_ref.resize(char_length);
+      return true;
+    }
+    SetLastError(JET_errInvalidColumnType);
+    return false;
+  }
+
+  bool ValidateAndConvertValue(const JET_COLTYP column_type,
+                               const std::vector<uint8_t>& column_data,
+                               GUID* value) {
+    return ValidateAndConvertValueGeneric(JET_coltypGUID, column_type,
+                                          column_data, value);
+  }
+
+  bool ValidateAndConvertValue(const JET_COLTYP column_type,
+                               const std::vector<uint8_t>& column_data,
+                               int32_t* value) {
+    return ValidateAndConvertValueGeneric(JET_coltypLong, column_type,
+                                          column_data, value);
+  }
+
+  bool ValidateAndConvertValue(const JET_COLTYP column_type,
+                               const std::vector<uint8_t>& column_data,
+                               int64_t* value) {
+    return ValidateAndConvertValueGeneric(JET_coltypLongLong, column_type,
+                                          column_data, value);
+  }
+
+  bool ValidateAndConvertValue(const JET_COLTYP column_type,
+                               const std::vector<uint8_t>& column_data,
+                               FILETIME* value) {
+    return ValidateAndConvertValueGeneric(JET_coltypLongLong, column_type,
+                                          column_data, value);
+  }
+
+  bool ValidateAndConvertValue(const JET_COLTYP column_type,
+                               const std::vector<uint8_t>& column_data,
+                               uint32_t* value) {
+    return ValidateAndConvertValueGeneric(JET_coltypUnsignedLong, column_type,
+                                          column_data, value);
+  }
+
+  const JET_COLUMNBASE& GetColumnByName(const base::string16& column_name) {
+    auto found_col = columns_by_name_.find(column_name);
+    if (found_col == columns_by_name_.end()) {
+      JET_COLUMNBASE column_base = {};
+      column_base.cbStruct = sizeof(JET_COLUMNBASE);
+      if (!SetLastError(JetGetTableColumnInfo(
+              session_id_, table_id_, column_name.c_str(), &column_base,
+              sizeof(column_base), JET_ColInfoBase))) {
+        // 0 means indicates an invalid column.

[Ilya Sherman, 2015/12/01 07:31:18]

nit: s/means indicates/indicates

[forshaw, 2015/12/01 11:12:13]

Acknowledged.

+        column_base.cbMax = 0;
+      }
+      columns_by_name_[column_name] = column_base;
+      found_col = columns_by_name_.find(column_name);
+    }
+    return found_col->second;
+  }
+
+  std::map<const base::string16, JET_COLUMNBASE> columns_by_name_;
+  JET_TABLEID table_id_;
+  base::string16 table_name_;
+  JET_SESID session_id_;
+
+  DISALLOW_COPY_AND_ASSIGN(EdgeDatabaseTableEnumerator);
+};
+
+class EdgeDatabaseSession : public EdgeErrorObject {
+ public:
+  EdgeDatabaseSession()
+      : db_id_(JET_dbidNil),
+        instance_id_(JET_instanceNil),
+        session_id_(JET_sesidNil) {}
+
+  ~EdgeDatabaseSession() {
+    // We don't need to collect other ID handles, terminating instance
+    // is enough to shut the entire session down.
+    if (instance_id_ != JET_instanceNil) {
+      JetTerm(instance_id_);
+    }

[Ilya Sherman, 2015/12/01 07:31:18]

nit: You can remove the curly braces here.

[forshaw, 2015/12/01 11:12:13]

Acknowledged.

+  }
+
+  bool OpenDatabase(const base::string16& database_file) {
+    if (IsOpen()) {
+      SetLastError(JET_errOneDatabasePerSession);
+      return false;
+    }
+    if (!SetLastError(JetSetSystemParameter(nullptr, JET_sesidNil,
+                                            JET_paramDatabasePageSize,
+                                            kEdgeDatabasePageSize, nullptr)))
+      return false;
+    if (!SetLastError(JetCreateInstance(&instance_id_, L"EdgeDataImporter")))
+      return false;
+    if (!SetLastError(JetSetSystemParameter(&instance_id_, JET_sesidNil,
+                                            JET_paramRecovery, 0, L"Off")))
+      return false;
+    if (!SetLastError(JetInit(&instance_id_)))
+      return false;
+    if (!SetLastError(
+            JetBeginSession(instance_id_, &session_id_, nullptr, nullptr)))
+      return false;
+    if (!SetLastError(JetAttachDatabase2(session_id_, database_file.c_str(), 0,
+                                         JET_bitDbReadOnly)))
+      return false;
+    if (!SetLastError(JetOpenDatabase(session_id_, database_file.c_str(),
+                                      nullptr, &db_id_, JET_bitDbReadOnly)))
+      return false;
+    return true;
+  }
+
+  scoped_ptr<EdgeDatabaseTableEnumerator> OpenTableEnumerator(
+      const base::string16& table_name) {
+    JET_TABLEID table_id;
+
+    if (!IsOpen()) {
+      SetLastError(JET_errDatabaseNotFound);
+      return nullptr;
+    }
+
+    if (!SetLastError(JetOpenTable(session_id_, db_id_, table_name.c_str(),
+                                   nullptr, 0, JET_bitTableReadOnly,
+                                   &table_id)))
+      return nullptr;
+
+    return make_scoped_ptr(
+        new EdgeDatabaseTableEnumerator(table_name, session_id_, table_id));
+  }
+
+ private:
+  // This is the page size of the Edge data. It's unlikely to change.
+  static const JET_API_PTR kEdgeDatabasePageSize = 8192;
+
+  bool IsOpen() { return instance_id_ != JET_instanceNil; }
+
+  JET_DBID db_id_;
+  JET_INSTANCE instance_id_;
+  JET_SESID session_id_;
+
+  DISALLOW_COPY_AND_ASSIGN(EdgeDatabaseSession);
+};
+
+struct EdgeFavoriteEntry {
+  EdgeFavoriteEntry()
+      : is_folder(false),
+        order_number(0),
+        item_id(GUID_NULL),
+        parent_id(GUID_NULL) {}
+
+  base::string16 title;
+  GURL url;
+  base::FilePath favicon_file;
+  bool is_folder;
+  int64_t order_number;
+  base::Time date_updated;
+  GUID item_id;
+  GUID parent_id;
+
+  std::vector<const EdgeFavoriteEntry*> children;
+
+  ImportedBookmarkEntry ToBookmarkEntry(
+      bool in_toolbar,
+      const std::vector<base::string16>& path) const {
+    ImportedBookmarkEntry entry;
+    entry.in_toolbar = in_toolbar;
+    entry.is_folder = is_folder;
+    entry.url = url;
+    entry.path = path;
+    entry.title = title;
+    entry.creation_time = date_updated;
+    return entry;
+  }
+};
+
+struct EdgeFavoriteEntryComparator {
+  bool operator()(const EdgeFavoriteEntry* lhs,
+                  const EdgeFavoriteEntry* rhs) const {
+    return std::tie(lhs->order_number, lhs->title) <
+           std::tie(rhs->order_number, rhs->title);
+  }
+};
+
+// The name of the database file is spartan.edb, however it isn't clear how
+// the intermediate path between the DataStore and the database is generated.
+// Therefore we just do a simple recursive search until we find a matching name.
+base::FilePath FindSpartanDatabase(const base::FilePath& profile_path) {
+  base::FilePath data_path =
+      profile_path.empty() ? importer::GetEdgeDataFilePath() : profile_path;
+  if (data_path.empty())
+    return base::FilePath();
+
+  base::FileEnumerator enumerator(data_path.Append(L"DataStore\\Data"), true,
+                                  base::FileEnumerator::FILES);
+  base::FilePath path = enumerator.Next();
+  while (!path.empty()) {
+    if (base::EqualsCaseInsensitiveASCII(path.BaseName().value(),
+                                         kSpartanDatabaseFile))
+      return path;
+    path = enumerator.Next();
+  }
+  return base::FilePath();
+}
+
+struct GuidComparator {
+  bool operator()(const GUID& a, const GUID& b) const {
+    return memcmp(&a, &b, sizeof(a)) < 0;
+  }
+};
+
+bool ReadFaviconData(const base::FilePath& file,
+                     std::vector<unsigned char>* data) {
+  std::string image_data;
+  if (!base::ReadFileToString(file, &image_data))
+    return false;
+
+  const unsigned char* ptr =
+      reinterpret_cast<const unsigned char*>(image_data.c_str());
+  return importer::ReencodeFavicon(ptr, image_data.size(), data);
+}
+
+void BuildBookmarkEntries(const EdgeFavoriteEntry* current_entry,

[Ilya Sherman, 2015/12/01 07:31:18]

nit: Pass by const-ref?

[forshaw, 2015/12/01 11:12:12]

Acknowledged.

+                          bool is_toolbar,
+                          std::vector<ImportedBookmarkEntry>* bookmarks,
+                          favicon_base::FaviconUsageDataList* favicons,
+                          std::vector<base::string16>* path) {
+  for (const EdgeFavoriteEntry* entry : current_entry->children) {
+    if (entry->is_folder) {
+      // If the favorites bar then load all children as toolbar items.
+      if (base::EqualsCaseInsensitiveASCII(entry->title, kFavoritesBarTitle)) {
+        // Replace name with Links similar to IE.
+        path->push_back(L"Links");
+        BuildBookmarkEntries(entry, true, bookmarks, favicons, path);
+        path->pop_back();
+      } else {
+        path->push_back(entry->title);
+        BuildBookmarkEntries(entry, is_toolbar, bookmarks, favicons, path);
+        path->pop_back();
+      }
+    } else {
+      bookmarks->push_back(entry->ToBookmarkEntry(is_toolbar, *path));
+      favicon_base::FaviconUsageData favicon;
+      if (entry->url.is_valid() && !entry->favicon_file.empty() &&
+          ReadFaviconData(entry->favicon_file, &favicon.png_data)) {
+        // As the database doesn't provide us a favicon URL we'll fake one.
+        GURL::Replacements path_replace;
+        path_replace.SetPathStr("/favicon.ico");
+        favicon.favicon_url =
+            entry->url.GetWithEmptyPath().ReplaceComponents(path_replace);
+        favicon.urls.insert(entry->url);
+        favicons->push_back(favicon);
+      }
+    }
+  }
+}
+
+}  // namespace
+
+EdgeImporter::EdgeImporter() {}
+
+void EdgeImporter::StartImport(const importer::SourceProfile& source_profile,
+                               uint16 items,
+                               ImporterBridge* bridge) {
+  bridge_ = bridge;
+  bridge_->NotifyStarted();
+  source_path_ = source_profile.source_path;
+
+  if ((items & importer::FAVORITES) && !cancelled()) {
+    bridge_->NotifyItemStarted(importer::FAVORITES);
+    ImportFavorites();
+    bridge_->NotifyItemEnded(importer::FAVORITES);
+  }
+  bridge_->NotifyEnded();
+}
+
+EdgeImporter::~EdgeImporter() {}
+
+void EdgeImporter::ImportFavorites() {
+  std::vector<ImportedBookmarkEntry> bookmarks;
+  favicon_base::FaviconUsageDataList favicons;
+  ParseFavoritesDatabase(&bookmarks, &favicons);
+
+  if (!bookmarks.empty() && !cancelled()) {
+    const base::string16& first_folder_name =
+        l10n_util::GetStringUTF16(IDS_BOOKMARK_GROUP_FROM_EDGE);
+    bridge_->AddBookmarks(bookmarks, first_folder_name);
+  }
+  if (!favicons.empty() && !cancelled())
+    bridge_->SetFavicons(favicons);
+}
+
+// From Edge 13 (released with Windows 10 TH2), Favorites are stored in a JET
+// database within the Edge local storage. The import uses the ESE library to
+// open and read the data file. The data is stored in a Favorites table with
+// the following schema.
+// Column Name          Column Type
+// ------------------------------------------
+// DateUpdated          LongLong - FILETIME
+// FaviconFile          LongText - Relative path
+// HashedUrl            ULong
+// IsDeleted            Bit
+// IsFolder             Bit
+// ItemId               Guid
+// OrderNumber          LongLong
+// ParentId             Guid
+// RoamDisabled         Bit
+// RowId                Long
+// Title                LongText
+// URL                  LongText
+void EdgeImporter::ParseFavoritesDatabase(
+    std::vector<ImportedBookmarkEntry>* bookmarks,
+    favicon_base::FaviconUsageDataList* favicons) {
+  base::FilePath database_path = FindSpartanDatabase(source_path_);
+  if (database_path.empty())
+    return;
+
+  EdgeDatabaseSession database;
+  if (!database.OpenDatabase(database_path.value())) {
+    DVLOG(ERROR) << "Error opening database " << database.GetErrorMessage();

[Ilya Sherman, 2015/12/01 07:31:18]

nit: Does this compile?  I think DVLOG expects a number rather than "ERROR" etc.
 That is, I'd expect this to be DVLOG(1).

[forshaw, 2015/12/01 11:12:13]

Acknowledged.

+    return;
+  }
+
+  scoped_ptr<EdgeDatabaseTableEnumerator> enumerator =
+      database.OpenTableEnumerator(L"Favorites");
+  if (!enumerator) {
+    DVLOG(ERROR) << "Error opening database table "
+                 << database.GetErrorMessage();
+    return;
+  }
+
+  std::map<GUID, EdgeFavoriteEntry, GuidComparator> database_entries;
+  base::FilePath favicon_base =
+      source_path_.empty() ? importer::GetEdgeDataFilePath() : source_path_;
+  favicon_base = favicon_base.Append(L"DataStore");
+
+  if (!enumerator->Reset())
+    return;

[Ilya Sherman, 2015/12/01 07:31:18]

nit: I'd move this up, above the favicon_base computation.

[forshaw, 2015/12/01 11:12:13]

Acknowledged.

+  do {
+    EdgeFavoriteEntry entry;
+    bool is_deleted = false;
+    if (!enumerator->RetrieveColumn(L"IsDeleted", &is_deleted))
+      continue;
+    if (is_deleted)
+      continue;
+    if (!enumerator->RetrieveColumn(L"IsFolder", &entry.is_folder))
+      continue;
+    base::string16 url;
+    if (!enumerator->RetrieveColumn(L"URL", &url))
+      continue;
+    entry.url = GURL(url);
+    if (!entry.is_folder && !entry.url.is_valid())
+      continue;
+    if (!enumerator->RetrieveColumn(L"Title", &entry.title))
+      continue;
+    base::string16 favicon_file;
+    if (!enumerator->RetrieveColumn(L"FaviconFile", &favicon_file))
+      continue;
+    if (!favicon_file.empty())
+      entry.favicon_file = favicon_base.Append(favicon_file);
+    if (!enumerator->RetrieveColumn(L"ParentId", &entry.parent_id))
+      continue;
+    if (!enumerator->RetrieveColumn(L"ItemId", &entry.item_id))
+      continue;
+    if (!enumerator->RetrieveColumn(L"OrderNumber", &entry.order_number))
+      continue;
+    FILETIME data_updated;
+    if (!enumerator->RetrieveColumn(L"DateUpdated", &data_updated))
+      continue;
+    entry.date_updated = base::Time::FromFileTime(data_updated);
+    database_entries[entry.item_id] = entry;
+  } while (enumerator->Next() && !cancelled());
+
+  // Build simple tree.
+  EdgeFavoriteEntry root_entry;
+  for (auto& entry : database_entries) {
+    auto found_parent = database_entries.find(entry.second.parent_id);
+    if (found_parent == database_entries.end() ||
+        !found_parent->second.is_folder) {
+      root_entry.children.push_back(&entry.second);
+    } else {
+      found_parent->second.children.push_back(&entry.second);
+    }
+  }
+  // With tree built sort the children of each node including the root.
+  std::sort(root_entry.children.begin(), root_entry.children.end(),
+            EdgeFavoriteEntryComparator());
+  for (auto& entry : database_entries) {
+    std::sort(entry.second.children.begin(), entry.second.children.end(),
+              EdgeFavoriteEntryComparator());
+  }
+  std::vector<base::string16> path;
+  BuildBookmarkEntries(&root_entry, false, bookmarks, favicons, &path);
+}