Create the BackgroundFetchBatchManager and initiate a download.

content/browser/background_fetch/background_fetch_batch_manager.h

Index: content/browser/background_fetch/background_fetch_batch_manager.h
diff --git a/content/browser/background_fetch/background_fetch_batch_manager.h b/content/browser/background_fetch/background_fetch_batch_manager.h
new file mode 100644
index 0000000000000000000000000000000000000000..9a4c790fec122f941625e53598811f074adb2a54
--- /dev/null
+++ b/content/browser/background_fetch/background_fetch_batch_manager.h
@@ -0,0 +1,56 @@
+// Copyright 2017 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 CONTENT_BROWSER_BACKGROUND_FETCH_BACKGROUND_FETCH_BATCH_MANAGER_H_
+#define CONTENT_BROWSER_BACKGROUND_FETCH_BACKGROUND_FETCH_BATCH_MANAGER_H_
+
+#include <unordered_map>
+#include <vector>

[Peter Beverloo, 2017/02/24 02:05:15]

IWYU: <string>

[harkness, 2017/02/24 11:47:10]

Done.

+
+#include "content/browser/background_fetch/background_fetch_data_manager.h"
+#include "content/browser/background_fetch/fetch_request.h"
+#include "content/common/content_export.h"
+
+namespace content {
+
+class BrowserContext;
+class StoragePartition;
+
+// The BatchManager will be responsible for coordinating communication with the
+// DownloadManager. It will get requests from the BatchData and dispatch them to

[Peter Beverloo, 2017/02/24 02:05:15]

from the -> as a?

[harkness, 2017/02/24 11:47:10]

It queries the BatchData for the next requests, so "from the" seems right to me.

+// the DownloadManager.
+// TODO(harkness): The BatchManager should also observe downloads.
+class CONTENT_EXPORT BackgroundFetchBatchManager {
+ public:
+  BackgroundFetchBatchManager(BrowserContext* browser_context,
+                              StoragePartition* storage_partition);
+  ~BackgroundFetchBatchManager();
+
+  // Start processing on a batch of requests. Some of these may already be in
+  // progress or completed from a previous chromium instance.
+  void ProcessBatch(const std::string& batch_guid,
+                    BackgroundFetchDataManager::BatchData* batch_data);
+
+ private:
+  void ProcessRequest(const std::string& batch_guid,
+                      const FetchRequest& request);
+
+  // Pointer to the browser context. The BackgroundFetchBatchManager is owned by
+  // the BrowserContext via the StoragePartition.
+  BrowserContext* browser_context_;
+
+  // Pointer to the storage partition. This object is owned by the partition
+  // (through a sequence of other classes).
+  StoragePartition* storage_partition_;
+
+  // The fetch_map holds any requests which have been sent to the
+  // DownloadManager indexed by the batch_guid.
+  std::unordered_map<std::string, FetchRequest> fetch_map_;

[Peter Beverloo, 2017/02/24 02:05:15]

In the longer term, are we going to have to hold on to this in memory? We're
going to have to enable resumption/reattaching to the notification anyway in
case of OOM.

[harkness, 2017/02/24 11:47:10]

I think we'll probably want to hold onto it, to have somewhere to hang the
DownloadItem* which we need once I set up pause/resume/cancel. We can make all
of that queryable on demand, but it seems easier to hold it here. I'm open to
changing it though if it doesn't end up being helpful.

+
+  DISALLOW_COPY_AND_ASSIGN(BackgroundFetchBatchManager);
+};
+
+}  // namespace content
+
+#endif  // CONTENT_BROWSER_BACKGROUND_FETCH_BACKGROUND_FETCH_BATCH_MANAGER_H_