Desktop aura: Break aura::Window::SetParent in two.

Desktop aura: Break aura::Window::SetParent in two.

Previously, SetParent will query a global StackingClient object if NULL is
passed to it. Since we want different behavior on the desktop and on ash, we
need to break that. It has been replaced with two methods:

- SetParentTo(), which takes an aura window and does the parenting.
- SetDefaultParentByTargetRoot(), which takes a RootWindow as context and
asks the StackingClient on said RootWindow where it should parent the
window.

The problem is that people have relied for a long time on what amounts to a
global variable. This is the first of several patches that pass around a
context RootWindow. This patch focuses on ash/ unittests mostly. Later patches
will focus on threading context through Widget creation. The last patch in this
series will pull out the global stacking client interface and add DCHECKs so
NULL can't be passed to either of these methods on Window.

BUG=161882
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=170049

[Elliot Glaysher, 2012-11-20 20:59:06]

+ ananta & cpu as CC; this isn't going out for real review yet; still has some
unit test issues.

[Elliot Glaysher, 2012-11-20 23:37:38]

+ ben

Getting some green trybots now.

[Ben Goodger (Google), 2012-11-20 23:46:37]

https://codereview.chromium.org/11421006/diff/3003/ui/aura/window.cc
File ui/aura/window.cc (right):

https://codereview.chromium.org/11421006/diff/3003/ui/aura/window.cc#newcode323
ui/aura/window.cc:323: void Window::SetParentTo(Window* parent) {
Since you're renaming this method, can you just get rid of it and make the
callers use AddChild?

https://codereview.chromium.org/11421006/diff/3003/ui/aura/window.h
File ui/aura/window.h (right):

https://codereview.chromium.org/11421006/diff/3003/ui/aura/window.h#newcode174
ui/aura/window.h:174: // Sets the parent window of the window. Can not be NULL.
Cannot or Must not

https://codereview.chromium.org/11421006/diff/3003/ui/aura/window.h#newcode179
ui/aura/window.h:179: // |context_root| can never be NULL; since desktop aura
and ash can be run at
Don't leak knowledge of downstream components into an upstream comment.

// |target_root| must not be NULL.
void SetDefaultParentByTargetRoot(RootWindow* target_root);

https://codereview.chromium.org/11421006/diff/3003/ui/views/widget/native_wid...
File ui/views/widget/native_widget_aura.cc (right):

https://codereview.chromium.org/11421006/diff/3003/ui/views/widget/native_wid...
ui/views/widget/native_widget_aura.cc:135: parent =
aura::client::GetStackingClient(params.context)->
is there a reason we can't just follow the pattern you use elsewhere of:

if (parent)
  SetParentTo()
else
  SetDefault..()

.. I'm OK with the SetDefault etc. function taking window bounds if that helps.

[Ben Goodger (Google), 2012-11-20 23:47:31]

> .. I'm OK with the SetDefault etc. function taking window bounds if that
helps.

The reason I'm suggesting this btw is that I think it'd be nice if we could hide
StackingClient completely behind Aura... i.e. it's an implementation detail of
Aura and not something that anyone using the Window API should have to worry
about.

[Elliot Glaysher, 2012-11-21 21:20:54]

sky+jam: this is ready for real review right now.

Yesterday, Ben did a preliminary review to check the API changes I made, but
left the whole patch to be reviewed by others.

[sky, 2012-11-21 23:35:58]

https://codereview.chromium.org/11421006/diff/5057/ash/drag_drop/drag_drop_tr...
File ash/drag_drop/drag_drop_tracker.h (right):

https://codereview.chromium.org/11421006/diff/5057/ash/drag_drop/drag_drop_tr...
ash/drag_drop/drag_drop_tracker.h:28: DragDropTracker(aura::RootWindow*
context_root);
explicit

https://codereview.chromium.org/11421006/diff/5057/ash/test/ash_test_base.h
File ash/test/ash_test_base.h (right):

https://codereview.chromium.org/11421006/diff/5057/ash/test/ash_test_base.h#n...
ash/test/ash_test_base.h:77: void AddToRootWindow(aura::Window* window);
Maybe name this SetDefaultParentByPrimaryRootWindow.

https://codereview.chromium.org/11421006/diff/5057/ash/wm/workspace/workspace...
File ash/wm/workspace/workspace_layout_manager_unittest.cc (right):

https://codereview.chromium.org/11421006/diff/5057/ash/wm/workspace/workspace...
ash/wm/workspace/workspace_layout_manager_unittest.cc:29: aura::Window* window =
new aura::Window(NULL);
Isn't this CreateTestWindowInShellWithBounds ? In fact makbe you should nuke
CreateTestWindow and replace with CreateTestWindowInShellWithBounds.

https://codereview.chromium.org/11421006/diff/5057/ash/wm/workspace/workspace...
File ash/wm/workspace/workspace_manager_unittest.cc (left):

https://codereview.chromium.org/11421006/diff/5057/ash/wm/workspace/workspace...
ash/wm/workspace/workspace_manager_unittest.cc:1030:
CreateTestWindowWithDelegate(&delegate, 1000, gfx::Rect(10, 11, 250, 251),
Doesn't this still work (well, removing the last arg)?

https://codereview.chromium.org/11421006/diff/5057/content/browser/renderer_h...
File content/browser/renderer_host/render_widget_host_view_aura.cc (right):

https://codereview.chromium.org/11421006/diff/5057/content/browser/renderer_h...
content/browser/renderer_host/render_widget_host_view_aura.cc:327: // TODO(erg):
While I could make sure details of the StackingClient are
Oshima is changing this at https://codereview.chromium.org/11299119/ .

https://codereview.chromium.org/11421006/diff/5057/ui/aura/client/stacking_cl...
File ui/aura/client/stacking_client.h (right):

https://codereview.chromium.org/11421006/diff/5057/ui/aura/client/stacking_cl...
ui/aura/client/stacking_client.h:23: // Called by the Window when its parent is
set to NULL, returns the window
Update this description.

https://codereview.chromium.org/11421006/diff/5057/ui/aura/test/aura_test_base.h
File ui/aura/test/aura_test_base.h (right):

https://codereview.chromium.org/11421006/diff/5057/ui/aura/test/aura_test_bas...
ui/aura/test/aura_test_base.h:30: // Creates a transient base, using the
|root_window()| as context.
base -> window that is transient to |parent| ...

https://codereview.chromium.org/11421006/diff/5057/ui/aura/test/aura_test_bas...
ui/aura/test/aura_test_base.h:33: // Attach this to the current shell's root
window.
This is effectively a cover method for SetDefaultParentByTargetRoot. You should
name it to match that.

https://codereview.chromium.org/11421006/diff/5057/ui/aura/test/test_windows.h
File ui/aura/test/test_windows.h (right):

https://codereview.chromium.org/11421006/diff/5057/ui/aura/test/test_windows....
ui/aura/test/test_windows.h:19: Window* CreateTestWindowWithId(int id, Window*
parent);
Please add a comment that these expect a non-null parent and that if you want a
non-null parent look at similar methods in AshTestBase.

https://codereview.chromium.org/11421006/diff/5057/ui/aura/test/test_windows....
ui/aura/test/test_windows.h:36: Window* CreateTransientChild(int id, Window*
parent, RootWindow* context);
Is this used? I don't see the new implementation which makes me think not.

https://codereview.chromium.org/11421006/diff/5057/ui/aura/window.h
File ui/aura/window.h (right):

https://codereview.chromium.org/11421006/diff/5057/ui/aura/window.h#newcode176
ui/aura/window.h:176: void SetDefaultParentByTargetRoot(RootWindow* target_root,
nit: TargetRoot doesn't mean much, how about SetDefaultParentByRoot or
SetDefaultParentInRoot. Also, clarify the coordinates of |bounds|.

https://codereview.chromium.org/11421006/diff/5057/ui/aura/window_unittest.cc
File ui/aura/window_unittest.cc (right):

https://codereview.chromium.org/11421006/diff/5057/ui/aura/window_unittest.cc...
ui/aura/window_unittest.cc:64: void AddToRootWindow(aura::Window* window) {
Again, I think this should match the method name you're calling.

https://codereview.chromium.org/11421006/diff/5057/ui/views/widget/native_wid...
File ui/views/widget/native_widget_aura.cc (right):

https://codereview.chromium.org/11421006/diff/5057/ui/views/widget/native_wid...
ui/views/widget/native_widget_aura.cc:1037:
native_view->SetDefaultParentByTargetRoot(native_view->GetRootWindow(),
I think this is wrong. In particular this means parenting to NULL leaves
|native_view| in the same root. If the root is destroyed, then |native_view| is
destroyed too, which we don't want.

[sky, 2012-11-21 23:37:11]

I'm adding Oshima to the cc list. Oshima, you might want to look at the changes
to render_widget_host_view_aura.cc as well as plumbing through bounds to
Window::SetDefaultParentByTargetRoot.

[Elliot Glaysher, 2012-11-22 00:26:55]

https://codereview.chromium.org/11421006/diff/5057/ui/views/widget/native_wid...
File ui/views/widget/native_widget_aura.cc (right):

https://codereview.chromium.org/11421006/diff/5057/ui/views/widget/native_wid...
ui/views/widget/native_widget_aura.cc:1037:
native_view->SetDefaultParentByTargetRoot(native_view->GetRootWindow(),
On 2012/11/21 23:35:59, sky wrote:
> I think this is wrong. In particular this means parenting to NULL leaves
> |native_view| in the same root. If the root is destroyed, then |native_view|
is
> destroyed too, which we don't want.

So currently this works because the StackingClient is a singleton (and will be
for a few more patches). Are you suggesting using |new_parent|'s root window for
context here?

[sky, 2012-11-22 00:59:50]

https://codereview.chromium.org/11421006/diff/5057/ui/views/widget/native_wid...
File ui/views/widget/native_widget_aura.cc (right):

https://codereview.chromium.org/11421006/diff/5057/ui/views/widget/native_wid...
ui/views/widget/native_widget_aura.cc:1037:
native_view->SetDefaultParentByTargetRoot(native_view->GetRootWindow(),
On 2012/11/22 00:26:56, Elliot Glaysher wrote:
> On 2012/11/21 23:35:59, sky wrote:
> > I think this is wrong. In particular this means parenting to NULL leaves
> > |native_view| in the same root. If the root is destroyed, then |native_view|
> is
> > destroyed too, which we don't want.
> 
> So currently this works because the StackingClient is a singleton (and will be
> for a few more patches). Are you suggesting using |new_parent|'s root window
for
> context here?

The problem is new_parent is NULL. I'm assuming SetDefaultParentbyTargetRoot is
going to result in leaving native_view in the same RootView, right? That is what
I'm saying is wrong. Maybe I'm missing something.

[Elliot Glaysher, 2012-11-26 21:04:54]

https://codereview.chromium.org/11421006/diff/5057/ui/views/widget/native_wid...
File ui/views/widget/native_widget_aura.cc (right):

https://codereview.chromium.org/11421006/diff/5057/ui/views/widget/native_wid...
ui/views/widget/native_widget_aura.cc:1037:
native_view->SetDefaultParentByTargetRoot(native_view->GetRootWindow(),
On 2012/11/22 00:59:50, sky wrote:
> On 2012/11/22 00:26:56, Elliot Glaysher wrote:
> > On 2012/11/21 23:35:59, sky wrote:
> > > I think this is wrong. In particular this means parenting to NULL leaves
> > > |native_view| in the same root. If the root is destroyed, then
|native_view|
> > is
> > > destroyed too, which we don't want.
> > 
> > So currently this works because the StackingClient is a singleton (and will
be
> > for a few more patches). Are you suggesting using |new_parent|'s root window
> for
> > context here?
> 
> The problem is new_parent is NULL. I'm assuming SetDefaultParentbyTargetRoot
is
> going to result in leaving native_view in the same RootView, right? That is
what
> I'm saying is wrong. Maybe I'm missing something.

After talking to scottmg and jam, I'm convinced that this is  equivalent to what
we do today, and the best we can do without some more major overhauls.

According to those two, it's an invariant that we have to be attached to a
RootWindow, which is what the previous SetParent(NULL) call does: it asks a
global where to place the window in some RootWindow.

This will usually be the same RootWindow as |native_view|.

[oshima, 2012-11-26 21:24:45]

https://codereview.chromium.org/11421006/diff/8005/content/browser/renderer_h...
File content/browser/renderer_host/render_widget_host_view_aura.cc (right):

https://codereview.chromium.org/11421006/diff/8005/content/browser/renderer_h...
content/browser/renderer_host/render_widget_host_view_aura.cc:330: gfx::Point
origin_in_screen(bounds_in_display.origin());
bounds_in_display is in screen coordinates, so you shouldn't have to convert it
to screen. 

SetBounds, on the other hand, takes the bounds in the parent's coordinates,
so you need to convert it to the parent's coordinate.

Please see https://chromiumcodereview.appspot.com/11299119.

https://codereview.chromium.org/11421006/diff/8005/content/browser/renderer_h...
content/browser/renderer_host/render_widget_host_view_aura.cc:336:
window_->SetDefaultParentByRoot(
Does this take screen coordinate? If so, SetDefaultParentByScreen or
SetDefaultParentInScreen would be better. "Root" can mean aura's root, ash's
root or platform's root which are all different unfortunately.

[sky, 2012-11-26 22:29:08]

https://codereview.chromium.org/11421006/diff/5057/ui/views/widget/native_wid...
File ui/views/widget/native_widget_aura.cc (right):

https://codereview.chromium.org/11421006/diff/5057/ui/views/widget/native_wid...
ui/views/widget/native_widget_aura.cc:1037:
native_view->SetDefaultParentByTargetRoot(native_view->GetRootWindow(),
On 2012/11/26 21:04:54, Elliot Glaysher wrote:
> On 2012/11/22 00:59:50, sky wrote:
> > On 2012/11/22 00:26:56, Elliot Glaysher wrote:
> > > On 2012/11/21 23:35:59, sky wrote:
> > > > I think this is wrong. In particular this means parenting to NULL leaves
> > > > |native_view| in the same root. If the root is destroyed, then
> |native_view|
> > > is
> > > > destroyed too, which we don't want.
> > > 
> > > So currently this works because the StackingClient is a singleton (and
will
> be
> > > for a few more patches). Are you suggesting using |new_parent|'s root
window
> > for
> > > context here?
> > 
> > The problem is new_parent is NULL. I'm assuming SetDefaultParentbyTargetRoot
> is
> > going to result in leaving native_view in the same RootView, right? That is
> what
> > I'm saying is wrong. Maybe I'm missing something.
> 
> After talking to scottmg and jam, I'm convinced that this is  equivalent to
what
> we do today, and the best we can do without some more major overhauls.
> 
> According to those two, it's an invariant that we have to be attached to a
> RootWindow, which is what the previous SetParent(NULL) call does: it asks a
> global where to place the window in some RootWindow.
> 
> This will usually be the same RootWindow as |native_view|.

I agree. I think it's just your comment that needs updating here.

[Elliot Glaysher, 2012-11-26 23:02:30]

https://codereview.chromium.org/11421006/diff/8005/content/browser/renderer_h...
File content/browser/renderer_host/render_widget_host_view_aura.cc (right):

https://codereview.chromium.org/11421006/diff/8005/content/browser/renderer_h...
content/browser/renderer_host/render_widget_host_view_aura.cc:336:
window_->SetDefaultParentByRoot(
On 2012/11/26 21:24:46, oshima wrote:
> Does this take screen coordinate? If so, SetDefaultParentByScreen or
> SetDefaultParentInScreen would be better. "Root" can mean aura's root, ash's
> root or platform's root which are all different unfortunately.

Root in this case is referring to RootWindow, which is the important parameter.
The bounds are optional, even in ash most places. This is one of the few places
where we previously passed in some set of bounds.

A question: Once I plumb aura::RootWindow context through the rest of ash and
views, will it be safe to remove this piece of location context? Are there cases
where a window is spawned from one RootWindow onto another by coordinates?

[oshima, 2012-11-27 00:14:29]

https://codereview.chromium.org/11421006/diff/8005/content/browser/renderer_h...
File content/browser/renderer_host/render_widget_host_view_aura.cc (right):

https://codereview.chromium.org/11421006/diff/8005/content/browser/renderer_h...
content/browser/renderer_host/render_widget_host_view_aura.cc:336:
window_->SetDefaultParentByRoot(
On 2012/11/26 23:02:30, Elliot Glaysher wrote:
> On 2012/11/26 21:24:46, oshima wrote:
> > Does this take screen coordinate? If so, SetDefaultParentByScreen or
> > SetDefaultParentInScreen would be better. "Root" can mean aura's root, ash's
> > root or platform's root which are all different unfortunately.
> 
> Root in this case is referring to RootWindow, which is the important
parameter.
> The bounds are optional, even in ash most places. This is one of the few
places
> where we previously passed in some set of bounds.

I see, my mistake.

> 
> A question: Once I plumb aura::RootWindow context through the rest of ash and
> views, will it be safe to remove this piece of location context? Are there
cases
> where a window is spawned from one RootWindow onto another by coordinates?

We still need it because there is no context when you restore window with the
stored location, and you can open a window at given location using
window.open().
(This looks broken now though)

https://codereview.chromium.org/11421006/diff/5115/content/browser/renderer_h...
File content/browser/renderer_host/render_widget_host_view_aura.cc (right):

https://codereview.chromium.org/11421006/diff/5115/content/browser/renderer_h...
content/browser/renderer_host/render_widget_host_view_aura.cc:336:
SetBounds(gfx::Rect(origin_in_screen, bounds_in_screen.size()));
This will break chromeos/win_ash. The bounds_in_screen should be in
screen coordinates, and SetBounds should take parent's coordinates
(as it sets by window_->SetBounds()). If this works on linux_aura,
wrong coordinates may be used somewhere else.

https://codereview.chromium.org/11421006/diff/5115/ui/aura/window.cc
File ui/aura/window.cc (right):

https://codereview.chromium.org/11421006/diff/5115/ui/aura/window.cc#newcode324
ui/aura/window.cc:324: const gfx::Rect& bounds) {
nit: indent & bounds -> bounds_in_screen

https://codereview.chromium.org/11421006/diff/5115/ui/views/widget/native_wid...
File ui/views/widget/native_widget_aura.cc (right):

https://codereview.chromium.org/11421006/diff/5115/ui/views/widget/native_wid...
ui/views/widget/native_widget_aura.cc:122: if (window_bounds == gfx::Rect()) {
parent && window_bounds == gfx::Rect()

[Elliot Glaysher, 2012-11-27 00:33:50]

https://codereview.chromium.org/11421006/diff/5115/content/browser/renderer_h...
File content/browser/renderer_host/render_widget_host_view_aura.cc (right):

https://codereview.chromium.org/11421006/diff/5115/content/browser/renderer_h...
content/browser/renderer_host/render_widget_host_view_aura.cc:336:
SetBounds(gfx::Rect(origin_in_screen, bounds_in_screen.size()));
On 2012/11/27 00:14:30, oshima wrote:
> This will break chromeos/win_ash. The bounds_in_screen should be in
> screen coordinates, and SetBounds should take parent's coordinates
> (as it sets by window_->SetBounds()). If this works on linux_aura,
> wrong coordinates may be used somewhere else.

I changed this to be what it used to be, and grabbing the parent from |window_|
here since that variable was removed.

Also,I've been doing my testing with a chromeos=1 build.

[jam, 2012-11-27 00:37:11]

do you need me to review this? sky is an owner in chrome and content/browser

[oshima, 2012-11-27 00:45:55]

RWHVA lgtm.

https://codereview.chromium.org/11421006/diff/2019/ui/aura/window.cc
File ui/aura/window.cc (right):

https://codereview.chromium.org/11421006/diff/2019/ui/aura/window.cc#newcode324
ui/aura/window.cc:324: const gfx::Rect& bounds) {
would you mind renaming |bounds| to |bounds_in_screen|?

[Elliot Glaysher, 2012-11-27 00:46:00]

I didn't know sky was an owner in content/

-jam

[sky, 2012-11-27 01:45:51]

LGTM

https://codereview.chromium.org/11421006/diff/2021/ash/test/ash_test_base.h
File ash/test/ash_test_base.h (right):

https://codereview.chromium.org/11421006/diff/2021/ash/test/ash_test_base.h#n...
ash/test/ash_test_base.h:76: // Attach this to the current shell's root window.
nit: this -> '|window|'

https://codereview.chromium.org/11421006/diff/2021/ui/aura/test/aura_test_base.h
File ui/aura/test/aura_test_base.h (right):

https://codereview.chromium.org/11421006/diff/2021/ui/aura/test/aura_test_bas...
ui/aura/test/aura_test_base.h:33: // Attach this to the current shell's root
window.
nit: 'this' -> '|window|'

https://codereview.chromium.org/11421006/diff/2021/ui/aura/window_unittest.cc
File ui/aura/window_unittest.cc (right):

https://codereview.chromium.org/11421006/diff/2021/ui/aura/window_unittest.cc...
ui/aura/window_unittest.cc:68: 
nit: remove one newline.

[commit-bot: I haz the power, 2012-11-27 17:51:28]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/11421006/5172

[commit-bot: I haz the power, 2012-11-27 17:52:00]

Presubmit check for 11421006-5172 failed and returned exit status 1.


Running presubmit commit checks ...

** Presubmit Warnings **
License must match:
.*? Copyright (\(c\) )?(2012|2011|2010|2009|2008|2007|2006|2006-2008) The
Chromium Authors\. All rights reserved\.\n.*? Use of this source code is
governed by a BSD-style license that can be\n.*? found in the LICENSE file\.(?:
\*/)?\n
Found a bad license header in these files:
  ui/aura/test/test_windows.h

Presubmit checks took 4.2s to calculate.

[commit-bot: I haz the power, 2012-11-27 17:53:47]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/11421006/5173

[commit-bot: I haz the power, 2012-11-28 01:25:30]

Retried try job too often for step(s) base_unittests, compile, crypto_unittests,
googleurl_unittests, sql_unittests

[commit-bot: I haz the power, 2012-11-28 17:26:47]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/11421006/5173

[commit-bot: I haz the power, 2012-11-28 17:27:10]

Failed to apply patch for ash/drag_drop/drag_drop_controller.cc:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  patching file ash/drag_drop/drag_drop_controller.cc
  Hunk #1 FAILED at 65.
  1 out of 1 hunk FAILED -- saving rejects to file
ash/drag_drop/drag_drop_controller.cc.rej

Patch:       ash/drag_drop/drag_drop_controller.cc
Index: ash/drag_drop/drag_drop_controller.cc
diff --git a/ash/drag_drop/drag_drop_controller.cc
b/ash/drag_drop/drag_drop_controller.cc
index
da309195917babcbf29575f3792000c17d90681a..eb8c2228dc79d733fb99a6e4bf2b9cfc70b0ae79
100644
--- a/ash/drag_drop/drag_drop_controller.cc
+++ b/ash/drag_drop/drag_drop_controller.cc
@@ -65,7 +65,7 @@ int DragDropController::StartDragAndDrop(
     ui::DragDropTypes::DragEventSource source) {
   DCHECK(!IsDragDropInProgress());
 
-  drag_drop_tracker_.reset(new DragDropTracker);
+  drag_drop_tracker_.reset(new DragDropTracker(root_window));
 
   drag_data_ = &data;
   drag_operation_ = operation;

[commit-bot: I haz the power, 2012-11-28 17:55:27]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/11421006/4072

[commit-bot: I haz the power, 2012-11-28 18:40:27]

Sorry for I got bad news for ya.
Compile failed with a clobber build on linux_chromeos.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_chro...
Your code is likely broken or HEAD is junk. Please ensure your
code is not broken then alert the build sheriffs.
Look at the try server FAQ for more details.

[commit-bot: I haz the power, 2012-11-28 18:42:41]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/11421006/18022

[commit-bot: I haz the power, 2012-11-28 21:00:09]

Change committed as 170049