Make GTK and Aura parts orthogonal to OS parts

content/public/browser/browser_main_parts.h

Index: content/public/browser/browser_main_parts.h
diff --git a/content/public/browser/browser_main_parts.h b/content/public/browser/browser_main_parts.h
new file mode 100644
index 0000000000000000000000000000000000000000..3df496decf0e0d821164569caa696af3a8b3e4f6
--- /dev/null
+++ b/content/public/browser/browser_main_parts.h
@@ -0,0 +1,56 @@
+// Copyright (c) 2011 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_PUBLIC_BROWSER_CONTENT_BROWSER_BROWSER_MAIN_PARTS_H_

[jam, 2011/10/24 19:36:00]

nit: CONTENT_PUBLIC_BROWSER_BOWSER_MAIN_PARTS_H_

[stevenjb, 2011/10/25 02:51:04]

Oops. Done.

+#define CONTENT_PUBLIC_BROWSER_CONTENT_BROWSER_BROWSER_MAIN_PARTS_H_
+#pragma once
+
+#include "content/common/content_export.h"
+#include "content/common/result_codes.h"
+
+namespace content {
+
+// Classes deriving fromc BrowserMainParts should ontain singleton member
+// variables specific to the browser process, and virtual methods representing
+// different "stages" to be executed from |BrowserMain()|, mostly to do
+// initialization and shutdown.

[jam, 2011/10/24 19:36:00]

we don't really need to tell embedders where to put their global variables etc.
it's also obvious that they'll have to implement the virtual functions that they
care about, so i think this comment (line 14-17) isn't necessary

[stevenjb, 2011/10/25 02:51:04]

Replaced with comment block from browser_main.cc

+//
+// There can be any number of "Parts". These should be constructed in
+// ContentBrowserClient::CreateBrowserMainParts. Each stage will be called
+// for each part in the order it was added. Destruction is in the inverse order.
+

[jam, 2011/10/24 19:36:00]

nit: extra line

[stevenjb, 2011/10/25 02:51:04]

Done.

+class CONTENT_EXPORT BrowserMainParts {
+ public:
+  BrowserMainParts() {};
+  virtual ~BrowserMainParts() {};
+
+  virtual void PreEarlyInitialization() = 0;
+
+  virtual void PostEarlyInitialization() = 0;
+
+  virtual void PreMainMessageLoopStart() = 0;
+
+  // Allows an embedder to do any extra toolkit initialization.
+  virtual void ToolkitInitialized() = 0;
+
+  virtual void PostMainMessageLoopStart() = 0;
+
+  // Returns a result code (see /common/result_codes.h).
+  // Default return value should be content::RESULT_CODE_NORMAL_EXIT.
+  virtual int PreMainMessageLoopRun() = 0;

[jam, 2011/10/24 19:36:00]

why did you change this function to return a value? it seems confusing that a
function which runs before the message loop returns an exit code.

i believe you're doing this because the chrome impl sometimes doesn't want to
run the main message loop, and stashes the result code in its
PreMainMessageLoopRun function. so why not just have it return true there? it
can also have a chance to return the error code, i.e.

bool MainMessageLoopRun(int* exit_code)

[stevenjb, 2011/10/25 02:51:04]

That's much better. Done.

+
+  // Returns true if the message loop was run, false otherwise.
+  // If no BrowserMainParts implementations return true, the default
+  // implementation will be run.
+  virtual bool MainMessageLoopRun() = 0;
+
+  virtual void PostMainMessageLoopRun() = 0;
+
+private:

[jam, 2011/10/24 19:36:00]

nit: space before private

[stevenjb, 2011/10/25 02:51:04]

Done.

+  DISALLOW_COPY_AND_ASSIGN(BrowserMainParts);
+};
+
+}  // namespace content
+
+#endif  // CONTENT_PUBLIC_BROWSER_CONTENT_BROWSER_BROWSER_MAIN_PARTS_H_