linux_aura: Compile ash into chrome.

linux_aura: Compile ash into chrome.

Running chrome with the "--open-ash" parameter will create an ash
desktop window. Ash on Linux has some problems, but at least we're now
compiling the code into the binary.

TODO: The applist isn't positioned correctly, and some details of the
system tray aren't correct.

BUG=300084, 303862
R=sky@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=269892

[Elliot Glaysher, 2013-09-27 23:04:48]

Enables ash compilation on windows. cpu@ said that I should try to fix the
current compile issues before even more regressions are introduced.

[sky, 2013-09-30 14:45:53]

https://codereview.chromium.org/25108005/diff/21001/chrome/browser/fullscreen...
File chrome/browser/fullscreen_aura.cc (right):

https://codereview.chromium.org/25108005/diff/21001/chrome/browser/fullscreen...
chrome/browser/fullscreen_aura.cc:9: #if defined(OS_LINUX) &&
!defined(OS_CHROMEOS)
Since you're in effect conditionally compiling this file, can it be done at the
gyp level?

https://codereview.chromium.org/25108005/diff/21001/chrome/browser/ui/ash/sol...
File chrome/browser/ui/ash/solid_color_user_wallpaper_delegate.h (right):

https://codereview.chromium.org/25108005/diff/21001/chrome/browser/ui/ash/sol...
chrome/browser/ui/ash/solid_color_user_wallpaper_delegate.h:8: namespace ash {
I thought based on recent discussions no namespaces in c/b/ ?

https://codereview.chromium.org/25108005/diff/21001/chrome/browser/ui/aura/ch...
File chrome/browser/ui/aura/chrome_browser_main_extra_parts_aura.cc (right):

https://codereview.chromium.org/25108005/diff/21001/chrome/browser/ui/aura/ch...
chrome/browser/ui/aura/chrome_browser_main_extra_parts_aura.cc:61:
active_desktop_monitor_.reset(new ActiveDesktopMonitor(GetInitialDesktop()));
How come you're moving initialization of this from ToolkitInitialized to
PostMainMessageLoopStart?

https://codereview.chromium.org/25108005/diff/21001/chrome/chrome_browser_ui....
File chrome/chrome_browser_ui.gypi (right):

https://codereview.chromium.org/25108005/diff/21001/chrome/chrome_browser_ui....
chrome/chrome_browser_ui.gypi:281:
'browser/ui/ash/solid_color_user_wallpaper_delegate.cc',
nit: sort

[Elliot Glaysher, 2013-09-30 20:53:40]

https://codereview.chromium.org/25108005/diff/21001/chrome/browser/ui/ash/sol...
File chrome/browser/ui/ash/solid_color_user_wallpaper_delegate.h (right):

https://codereview.chromium.org/25108005/diff/21001/chrome/browser/ui/ash/sol...
chrome/browser/ui/ash/solid_color_user_wallpaper_delegate.h:8: namespace ash {
On 2013/09/30 14:45:53, sky wrote:
> I thought based on recent discussions no namespaces in c/b/ ?

This is forward declaring an ash type; not putting browser implementation
details inside a namespace.

https://codereview.chromium.org/25108005/diff/21001/chrome/browser/ui/aura/ch...
File chrome/browser/ui/aura/chrome_browser_main_extra_parts_aura.cc (right):

https://codereview.chromium.org/25108005/diff/21001/chrome/browser/ui/aura/ch...
chrome/browser/ui/aura/chrome_browser_main_extra_parts_aura.cc:61:
active_desktop_monitor_.reset(new ActiveDesktopMonitor(GetInitialDesktop()));
On 2013/09/30 14:45:53, sky wrote:
> How come you're moving initialization of this from ToolkitInitialized to
> PostMainMessageLoopStart?

Because using ActiveDesktopMonitor required aura::Env which requires a
MessageLoop, at least on linux. Truncated stack of when the initialization is up
there:

[14600:14600:0930/133646:FATAL:message_loop.h(584)] Check failed: loop. 
 [0x7f1bd21f1dbe] base::debug::StackTrace::StackTrace()
 [0x7f1bd224e8ff] logging::LogMessage::~LogMessage()
 [0x7f1bd21c3602] base::MessageLoopForUI::current()
 [0x7f1bd21c260d] base::MessagePumpX11::Current()
 [0x7f1bcadd89e9] aura::Env::Init()
 [0x7f1bcadd89b5] aura::Env::GetInstance()
 [0x000001e6a999] ActiveDesktopMonitor::ActiveDesktopMonitor()
 [0x000001b5c6a1] ChromeBrowserMainExtraPartsAura::ToolkitInitialized()
 [0x000000e8fb6d] ChromeBrowserMainParts::ToolkitInitialized()
 [0x7f1bc8627cd6] content::BrowserMainLoop::InitializeToolkit()
 [0x7f1bc8630446] content::BrowserMainRunnerImpl::Initialize()

https://codereview.chromium.org/25108005/diff/21001/chrome/chrome_browser_ui....
File chrome/chrome_browser_ui.gypi (right):

https://codereview.chromium.org/25108005/diff/21001/chrome/chrome_browser_ui....
chrome/chrome_browser_ui.gypi:281:
'browser/ui/ash/solid_color_user_wallpaper_delegate.cc',
On 2013/09/30 14:45:53, sky wrote:
> nit: sort

I really  wish tools/git/move_source_file.py was smart enough to do this on its
own...

[sky, 2013-09-30 21:05:33]

LGTM

[commit-bot: I haz the power, 2013-10-01 01:05:42]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/erg@chromium.org/25108005/84001

[commit-bot: I haz the power, 2013-10-01 03:15:19]

Retried try job too often on linux_chromeos for step(s) browser_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_chro...

[Elliot Glaysher, 2014-04-24 22:52:40]

6 month revival. PTAL.

I've fixed all the desktop linux failing tests, and made
theme_image_mapper_aura_win.cc work across desktop platforms.

The linux_chromium_chromeos_rel bot seems to be crazy. I've reproduced several
of the browser_test failures locally without my patch.

[sky, 2014-04-24 23:14:32]

LGTM

https://codereview.chromium.org/25108005/diff/359001/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/25108005/diff/359001/build/common.gypi#newcode91
build/common.gypi:91: 'use_ash%': 1,
Can we get rid of use_ash now?

https://codereview.chromium.org/25108005/diff/359001/chrome/browser/fullscree...
File chrome/browser/fullscreen_aurax11.cc (right):

https://codereview.chromium.org/25108005/diff/359001/chrome/browser/fullscree...
chrome/browser/fullscreen_aurax11.cc:36: 
nit: remove empty line.

https://codereview.chromium.org/25108005/diff/359001/chrome/browser/ui/aura/c...
File chrome/browser/ui/aura/chrome_browser_main_extra_parts_aura.cc (right):

https://codereview.chromium.org/25108005/diff/359001/chrome/browser/ui/aura/c...
chrome/browser/ui/aura/chrome_browser_main_extra_parts_aura.cc:52: if
(command_line->HasSwitch(switches::kOpenAsh)) {
nit: no {}

[Elliot Glaysher, 2014-05-12 21:35:41]

https://codereview.chromium.org/25108005/diff/359001/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/25108005/diff/359001/build/common.gypi#newcode91
build/common.gypi:91: 'use_ash%': 1,
On 2014/04/24 23:14:32, sky wrote:
> Can we get rid of use_ash now?

Theoretically yes, but I'd prefer to not do it in this already giant patch.

[Elliot Glaysher, 2014-05-12 22:18:56]

Committed patchset #35 manually as r269892 (presubmit successful).

[Nico, 2014-05-13 08:02:44]

As said on the bug, not lgtm. Please reply to the concerns on the bug before
relanding.