CommandLine Page Cycler

chrome/browser/page_cycler/page_cycler.h

Index: chrome/browser/page_cycler/page_cycler.h
diff --git a/chrome/browser/page_cycler/page_cycler.h b/chrome/browser/page_cycler/page_cycler.h
new file mode 100644
index 0000000000000000000000000000000000000000..349c0557363b8bc64692ae962610fcaf756464f7
--- /dev/null
+++ b/chrome/browser/page_cycler/page_cycler.h
@@ -0,0 +1,144 @@
+// Copyright (c) 2012 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#ifndef CHROME_BROWSER_PAGE_CYCLER_PAGE_CYCLER_H_
+#define CHROME_BROWSER_PAGE_CYCLER_PAGE_CYCLER_H_
+#pragma once
+
+#include "chrome/browser/ui/browser.h"
+#include "chrome/browser/ui/browser_list.h"
+#include "content/public/browser/navigation_controller.h"
+#include "content/public/browser/notification_observer.h"
+#include "content/public/browser/notification_registrar.h"
+#include "content/public/browser/web_contents_observer.h"
+
+namespace base {
+class TimeTicks;
+}  // namespace base
+
+// Performance test to track the resources used and speed with which chromium
+// fully loads a given set of URLs.  This class is created on the UI thread and
+// does most of its work there. However, some work happens on background threads
+// too; those are named with 'OnBackgroundThread'.
+class PageCycler : public base::RefCountedThreadSafe<PageCycler>,
+                   public BrowserList::Observer,
+                   public content::WebContentsObserver {
+ public:
+  // Called when statistics are not being recorded; typically this means that
+  // the user did not use the record-stats switch in the commandline.
+  PageCycler(Browser* browser, FilePath urls_file, FilePath errors_file);
+
+  // Called when statistics gathering is important, and will be written to file.
+  PageCycler(Browser* browser,
+             FilePath urls_file,
+             FilePath errors_file,

[Aaron Boodman, 2012/03/20 19:28:16]

Consider just making these setters that the caller should specify before calling
Run().

+             FilePath stats_file);
+
+  // Begin running the page cycler; post the ReadURLsFromFile task on the IO

[Aaron Boodman, 2012/03/20 19:28:16]

Avoid duplicating implementation (such as post the blahblah task to blah thread)
details in comments.

Comments should be about what happens, not how.

+  // thread.
+  void Run(unsigned int iterations);
+
+  // content::WebContentsObserver
+  virtual void DidFinishLoad(int64 frame_id,
+                             const GURL& validated_url,
+                             bool is_main_frame) OVERRIDE;
+  virtual void DidFailProvisionalLoad(
+      int64 frame_id,
+      bool is_main_frame,
+      const GURL& validated_url,
+      int error_code,
+      const string16& error_description) OVERRIDE;
+
+  // This method should never be necessary while running PageCycler; this is
+  // for testing purposes only.
+  const std::vector<GURL>* urls() { return &urls_; }

[chebert, 2012/04/09 23:52:01]

I've seen functions that call testing functions with a for_test() at the end. So
like:

urls_for_test()

This pretty much guarantees that no one will use it, especially if they are
scanning through comments quickly.

+
+ private:
+  friend class base::RefCountedThreadSafe<PageCycler>;
+  ~PageCycler();
+
+  // Initialize the PageCyclerNavigationObserver; add a reference.
+  void Init();
+
+  // Read in the urls from |urls_file_| and store them in |urls_|.  Upon finish,
+  // switch thread back to the UI thread and call BeginCycle.
+  void ReadURLsOnBackgroundThread();
+
+  // Perform any initial setup neccessary, and begin visiting the pages.
+  void BeginCycle();
+
+  // If |url_index_| points to a valid position in |urls_|, load the url,
+  // capturing any statistics information. Otherwise, call WriteResults.
+  void LoadNextURL();
+
+  // Complete statistics gathering for the finished visit, and try to load the
+  // next url.
+  void LoadSucceeded();
+
+  // Inidicate that the load failed with an error; try to load the next url.
+  void LoadFailed(const GURL& url, const string16& error_description);
+
+  // Finalize the ouput strings, and post a task on the IO thread for
+  // WriteResultsToFile().
+  void PrepareResults();
+
+  // Write the data stored within output to the file indicated by |stats_file_|,
+  // if |stats_file_| is not empty. Write any errors to |errors_file_|. Switch
+  // thread back to the UI thread, and Finish().
+  void WriteResultsOnBackgroundThread(std::string output);
+
+  // Perform any necessary cleanup, release the reference, and exit |browser_|.
+  void Finish();
+
+  // Called when the Browser to which |browser_| points is closed; exits
+  // PageCycler.
+  void Abort();
+
+  // BrowserList::Observer
+  virtual void OnBrowserAdded(const Browser* browser) OVERRIDE;
+  virtual void OnBrowserRemoved(const Browser* browser) OVERRIDE;
+
+  // The Browser context in which the page cycler is running.
+  Browser* browser_;
+
+  // The path to the file containing the list of urls to visit.
+  FilePath urls_file_;
+
+  // The path to the file to which we write any errors encountered.
+  FilePath errors_file_;
+
+  // The path to the file to which we write the statistics (optional, may be
+  // an empty path).
+  FilePath stats_file_;
+
+  // The list of urls to visit.
+  std::vector<GURL> urls_;
+
+  // The current index into the |urls_| vector.
+  size_t url_index_;
+
+  // The current number of iterations remaining.
+  unsigned int iterations_;
+
+  // The generated string of urls which we have visited; this is built one url
+  // at a time as we iterate through the |urls_| vector. This is primarily
+  // included for interfacing with the previous page_cycler's output style.
+  std::string urls_string_;
+
+  // The generated string of the times taken to visit each url. As with
+  // |urls_string_|, this is built as we visit each url, and is primarily to
+  // produce output similar to the previous page_cycler's.
+  std::string timings_string_;
+
+  // The time at which we begin the process of loading the next url; this is
+  // used to calculate the time taken for each url load.
+  base::TimeTicks initial_time_;
+
+  string16 error_;
+
+
+  DISALLOW_COPY_AND_ASSIGN(PageCycler);
+};
+
+#endif  // CHROME_BROWSER_PAGE_CYCLER_PAGE_CYCLER_H_