Scale client area, hit test mask and bounds by device_scale_factor.

Scale client area, hit test mask and bounds by device_scale_factor.

Fixes the second TODO of CL https://codereview.chromium.org/2341483004/.

TODO:
1. Chrome browser doesn't show up with window bounds scaling.
2. Also scale window bounds in aura::mus::WindowTreeClient once window
bounds work correctly.

BUG=646926
TEST=manual (with --force-device-scale-factor=2)

Committed: https://crrev.com/075c4290cdfc0cfca43c2114bce282ea9389e6fc
Cr-Commit-Position: refs/heads/master@{#433009}

[riajiang, 2016-10-25 23:20:34]

Description was changed from

==========
Scale client area and hit test mask by device_scale_factor.

Fixes the second TODO of CL https://codereview.chromium.org/2341483004/.

BUG=646926
TEST=manual
==========

to

==========
Scale client area and hit test mask by device_scale_factor.

Fixes the second TODO of CL https://codereview.chromium.org/2341483004/.

BUG=646926
TEST=manual (with --force-device-scale-factor=2)
==========

[riajiang, 2016-10-25 23:27:45]

riajiang@chromium.org changed reviewers:
+ sadrul@chromium.org, sky@chromium.org

[riajiang, 2016-10-25 23:27:45]

Since scaling window bounds in WTC alone didn't fix the window bounds bug, I was
thinking to upload these changes about client area and hit test mask in WTC
first and continue with window bounds related changes in separate CLs. Could you
take a look? Thanks!

[riajiang, 2016-10-25 23:33:51]

Since scaling window bounds in WTC alone didn't fix the window bounds bug, I was
thinking to upload these changes about client area and hit test mask in WTC
first and continue with window bounds related changes in separate CLs. Could you
take a look? Thanks!

[riajiang, 2016-11-08 21:25:26]

Description was changed from

==========
Scale client area and hit test mask by device_scale_factor.

Fixes the second TODO of CL https://codereview.chromium.org/2341483004/.

BUG=646926
TEST=manual (with --force-device-scale-factor=2)
==========

to

==========
Scale client area, hit test mask and bounds by device_scale_factor.

Fixes the second TODO of CL https://codereview.chromium.org/2341483004/.

TODO:
1. Chrome browser doesn't show up with window bounds scaling.
2. Also scale window bounds in aura::mus::WindowTreeClient once window
bounds work correctly.

BUG=646926
TEST=manual (with --force-device-scale-factor=2)
==========

[riajiang, 2016-11-08 21:40:19]

Patchset #3 (id:40001) has been deleted

[riajiang, 2016-11-08 21:42:14]

Updated the patch and CL description. PTAL. Thanks!

[sky, 2016-11-09 00:21:14]

https://codereview.chromium.org/2447303002/diff/60001/services/ui/public/cpp/...
File services/ui/public/cpp/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/60001/services/ui/public/cpp/...
services/ui/public/cpp/window_tree_client.cc:77:
gfx::ConvertRectToDIP(device_scale_factor, window_data->bounds));
How come you adjust bounds in the mus side, but not the aura-mus side?

https://codereview.chromium.org/2447303002/diff/60001/ui/aura/mus/window_tree...
File ui/aura/mus/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/60001/ui/aura/mus/window_tree...
ui/aura/mus/window_tree_client.cc:128: auto iter =
std::find_if(displays.begin(), displays.end(),
When this function is used you always have a window. Can you use
display::Screen::GetScreen()->GetDisplayNearestWindow(window()).device_scale_factor()
?

https://codereview.chromium.org/2447303002/diff/60001/ui/aura/mus/window_tree...
ui/aura/mus/window_tree_client.cc:225:
DeviceScaleFactorForDisplay(GetWindowTreeHostMus(window)->display_id());
GetWindowTreeHostMus might return null.

[riajiang, 2016-11-09 21:07:08]

https://codereview.chromium.org/2447303002/diff/60001/services/ui/public/cpp/...
File services/ui/public/cpp/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/60001/services/ui/public/cpp/...
services/ui/public/cpp/window_tree_client.cc:77:
gfx::ConvertRectToDIP(device_scale_factor, window_data->bounds));
On 2016/11/09 00:21:13, sky wrote:
> How come you adjust bounds in the mus side, but not the aura-mus side?

The window bounds adjustment in the mus side of WTC didn't solve all bounds
related bugs (TODOs in the CL description) so we were thinking to add bounds
conversion in aura-mus after everything works (TODO in aura::mus::WTC). Added
the adjustment here in the mus side because the conversion does need to happen
in WTC and I'm still working on finding other places where window bounds need to
be scaled.

https://codereview.chromium.org/2447303002/diff/60001/ui/aura/mus/window_tree...
File ui/aura/mus/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/60001/ui/aura/mus/window_tree...
ui/aura/mus/window_tree_client.cc:128: auto iter =
std::find_if(displays.begin(), displays.end(),
On 2016/11/09 00:21:13, sky wrote:
> When this function is used you always have a window. Can you use
>
display::Screen::GetScreen()->GetDisplayNearestWindow(window()).device_scale_factor()
> ?

Changed to use GetDisplayNearestWindow. Kept the TODO since I'm not sure if it
still makes more sense to use GetDisplayWithDisplayId() in the future
(aura::Window doesn't have a display_id)?

https://codereview.chromium.org/2447303002/diff/60001/ui/aura/mus/window_tree...
ui/aura/mus/window_tree_client.cc:225:
DeviceScaleFactorForDisplay(GetWindowTreeHostMus(window)->display_id());
On 2016/11/09 00:21:13, sky wrote:
> GetWindowTreeHostMus might return null.

Not using GetWindowTreeHostMus anymore since changed to use
GetDisplayNearestWindow to get the display.

[sky, 2016-11-09 23:22:57]

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
File services/ui/public/cpp/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
services/ui/public/cpp/window_tree_client.cc:247: if (device_scale_factor ==
1.f) {
I think the code you have would be simpler if you centralize conversion into a
function. For example, make this code be:

gfx::Rect bounds_pixels = ConvertDipsToPixels(window, bounds);
tree_->SetWindowBounds(...);

That way no conditionals. Do this for all the types you're using.

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
services/ui/public/cpp/window_tree_client.cc:936: InFlightBoundsChange
new_change(window, new_bounds);
Isn't new_bounds in pixels? On 245 I believe the new bounds is dips.

https://codereview.chromium.org/2447303002/diff/80001/ui/aura/mus/window_tree...
File ui/aura/mus/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/80001/ui/aura/mus/window_tree...
ui/aura/mus/window_tree_client.cc:980: // TODO(riajiang): Convert from pixel to
DIP.
Reference bugs.

[kylechar, 2016-11-10 14:30:51]

kylechar@chromium.org changed reviewers:
+ kylechar@chromium.org

[kylechar, 2016-11-10 14:30:52]

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
File services/ui/public/cpp/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
services/ui/public/cpp/window_tree_client.cc:42: float
DeviceScaleFactorForDisplay(int64_t display_id) {
Name this something like ScaleFactorForDisplay()? The effective scale factor for
the display isn't the same as the device scale factor (because of ui scale
factor) for internal displays.

[riajiang, 2016-11-10 19:15:15]

Patchset #5 (id:100001) has been deleted

[riajiang, 2016-11-10 19:16:08]

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
File services/ui/public/cpp/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
services/ui/public/cpp/window_tree_client.cc:42: float
DeviceScaleFactorForDisplay(int64_t display_id) {
On 2016/11/10 14:30:52, kylechar wrote:
> Name this something like ScaleFactorForDisplay()? The effective scale factor
for
> the display isn't the same as the device scale factor (because of ui scale
> factor) for internal displays.

I see, done.

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
services/ui/public/cpp/window_tree_client.cc:247: if (device_scale_factor ==
1.f) {
On 2016/11/09 23:22:57, sky wrote:
> I think the code you have would be simpler if you centralize conversion into a
> function. For example, make this code be:
> 
> gfx::Rect bounds_pixels = ConvertDipsToPixels(window, bounds);
> tree_->SetWindowBounds(...);
> 
> That way no conditionals. Do this for all the types you're using.

I did conditionals to avoid extra work if DSF is 1. Now I added early returns
for conversion functions in dip_util.cc and just call them from tree_->Set...
(kept the one for client area since there's a vector) What do you think of this
instead?

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
services/ui/public/cpp/window_tree_client.cc:936: InFlightBoundsChange
new_change(window, new_bounds);
On 2016/11/09 23:22:57, sky wrote:
> Isn't new_bounds in pixels? On 245 I believe the new bounds is dips.

new_bounds is in pixels, but since DSF is 1 here, new_bounds has the same value
in pixels and in dips? Updated this function to be clearer.

https://codereview.chromium.org/2447303002/diff/80001/ui/aura/mus/window_tree...
File ui/aura/mus/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/80001/ui/aura/mus/window_tree...
ui/aura/mus/window_tree_client.cc:980: // TODO(riajiang): Convert from pixel to
DIP.
On 2016/11/09 23:22:57, sky wrote:
> Reference bugs.

Done.

[sky, 2016-11-10 23:55:15]

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
File services/ui/public/cpp/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/80001/services/ui/public/cpp/...
services/ui/public/cpp/window_tree_client.cc:247: if (device_scale_factor ==
1.f) {
On 2016/11/10 19:16:08, riajiang wrote:
> On 2016/11/09 23:22:57, sky wrote:
> > I think the code you have would be simpler if you centralize conversion into
a
> > function. For example, make this code be:
> > 
> > gfx::Rect bounds_pixels = ConvertDipsToPixels(window, bounds);
> > tree_->SetWindowBounds(...);
> > 
> > That way no conditionals. Do this for all the types you're using.
> 
> I did conditionals to avoid extra work if DSF is 1. Now I added early returns
> for conversion functions in dip_util.cc and just call them from tree_->Set...
> (kept the one for client area since there's a vector) What do you think of
this
> instead? 

This isn't a performance sensitive area, so going with more maintainable code is
better.

https://codereview.chromium.org/2447303002/diff/120001/services/ui/public/cpp...
File services/ui/public/cpp/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/120001/services/ui/public/cpp...
services/ui/public/cpp/window_tree_client.cc:1485: server_id(window),
offset.x(), offset.y(),
Similar comment about converting offset.x/y.

https://codereview.chromium.org/2447303002/diff/120001/ui/aura/mus/window_tre...
File ui/aura/mus/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/120001/ui/aura/mus/window_tre...
ui/aura/mus/window_tree_client.cc:219: if (device_scale_factor == 1.f) {
Again, this isn't performance sensitive code, so I would rather you go on the
side of maintainable code. By that I mean don't have separate branches in any of
this code. Always do the conversion, even if it ends up a no-op.

https://codereview.chromium.org/2447303002/diff/120001/ui/aura/mus/window_tre...
ui/aura/mus/window_tree_client.cc:1508: WindowMus::Get(window)->server_id(),
offset.x(), offset.y(),
Don't offset.x/y need to be convered?

[riajiang, 2016-11-11 19:32:35]

Patchset #6 (id:140001) has been deleted

[riajiang, 2016-11-11 19:41:21]

https://codereview.chromium.org/2447303002/diff/120001/ui/aura/mus/window_tre...
File ui/aura/mus/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/120001/ui/aura/mus/window_tre...
ui/aura/mus/window_tree_client.cc:219: if (device_scale_factor == 1.f) {
On 2016/11/10 23:55:15, sky wrote:
> Again, this isn't performance sensitive code, so I would rather you go on the
> side of maintainable code. By that I mean don't have separate branches in any
of
> this code. Always do the conversion, even if it ends up a no-op.

Okay makes sense, done.

https://codereview.chromium.org/2447303002/diff/120001/ui/aura/mus/window_tre...
ui/aura/mus/window_tree_client.cc:1508: WindowMus::Get(window)->server_id(),
offset.x(), offset.y(),
On 2016/11/10 23:55:15, sky wrote:
> Don't offset.x/y need to be convered?

When I was converting offset, browser tabs were not at the right positions tho.
I'll look into it more in the window positioning bug (http://crbugs.com/646932).
Added a TODO.

[sky, 2016-11-11 21:16:53]

LGTM

https://codereview.chromium.org/2447303002/diff/160001/services/ui/public/cpp...
File services/ui/public/cpp/window_tree_client.cc (right):

https://codereview.chromium.org/2447303002/diff/160001/services/ui/public/cpp...
services/ui/public/cpp/window_tree_client.cc:921:
WindowPrivate(window).LocalSetBounds(old_bounds_in_dip, new_bounds_in_dip);
Good thing we're getting rid of this. I worry rounding errors could easily
trigger DCHECKs in LocalSetBounds.

https://codereview.chromium.org/2447303002/diff/160001/services/ui/public/cpp...
services/ui/public/cpp/window_tree_client.cc:1474: // TODO(riajiang): Figure out
if |offset| needs to be converted.
I will be quite surprised if it doesn't need to be.

[riajiang, 2016-11-11 21:48:21]

The CQ bit was checked by riajiang@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-11 21:48:52]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-11 22:47:28]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-11-11 22:47:29]

Dry run: Try jobs failed on following builders:
  android_n5x_swarming_rel on master.tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_n5x_...)

[sadrul, 2016-11-12 02:20:04]

Change looks good. But the try failures look relevant. Perhaps the Screen
instance hasn't been yet been set when the set-up happens?

[riajiang, 2016-11-15 00:22:02]

Patchset #7 (id:180001) has been deleted

[riajiang, 2016-11-16 22:47:39]

Patchset #10 (id:260001) has been deleted

[riajiang, 2016-11-17 01:02:01]

On 2016/11/12 02:20:04, sadrul wrote:
> Change looks good. But the try failures look relevant. Perhaps the Screen
> instance hasn't been yet been set when the set-up happens?

Passed all tests. Please take another look. Thanks!

[kylechar, 2016-11-17 14:16:38]

https://codereview.chromium.org/2447303002/diff/280001/services/ui/demo/mus_d...
File services/ui/demo/mus_demo.h (right):

https://codereview.chromium.org/2447303002/diff/280001/services/ui/demo/mus_d...
services/ui/demo/mus_demo.h:83: std::unique_ptr<display::test::TestScreen>
test_screen_;
Was there a reason not to use ScreenBase?

[riajiang, 2016-11-17 15:54:21]

The CQ bit was checked by riajiang@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-17 15:54:43]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-17 16:06:34]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-11-17 16:06:35]

Dry run: Try jobs failed on following builders:
  chromeos_x86-generic_chromium_compile_only_ng on
master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromeos_x86-ge...)
  linux_chromium_chromeos_ozone_rel_ng on master.tryserver.chromium.linux
(JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[riajiang, 2016-11-17 16:36:06]

The CQ bit was checked by riajiang@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-17 16:36:53]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-17 18:27:16]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-11-17 18:27:17]

Dry run: This issue passed the CQ dry run.

[riajiang, 2016-11-17 18:30:25]

https://codereview.chromium.org/2447303002/diff/280001/services/ui/demo/mus_d...
File services/ui/demo/mus_demo.h (right):

https://codereview.chromium.org/2447303002/diff/280001/services/ui/demo/mus_d...
services/ui/demo/mus_demo.h:83: std::unique_ptr<display::test::TestScreen>
test_screen_;
On 2016/11/17 14:16:37, kylechar wrote:
> Was there a reason not to use ScreenBase?

Forgot about ScreenBase... changed to use ScreenBase and changed back MusDemo to
be not test-only. Thanks!

[riajiang, 2016-11-17 18:30:31]

The CQ bit was checked by riajiang@chromium.org

[riajiang, 2016-11-17 18:30:31]

The patchset sent to the CQ was uploaded after l-g-t-m from sky@chromium.org
Link to the patchset: https://codereview.chromium.org/2447303002/#ps300001
(title: "ScreenBase")

[commit-bot: I haz the power, 2016-11-17 18:31:04]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[kylechar, 2016-11-17 18:31:32]

lgtm

[sadrul, 2016-11-17 18:32:28]

lgtm

[commit-bot: I haz the power, 2016-11-17 18:40:10]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-11-17 18:40:13]

Failed to apply patch for services/ui/public/cpp/window_tree_client.cc:
While running git apply --index -p1;
  error: patch failed: services/ui/public/cpp/window_tree_client.cc:25
  error: services/ui/public/cpp/window_tree_client.cc: patch does not apply

Patch:       services/ui/public/cpp/window_tree_client.cc
Index: services/ui/public/cpp/window_tree_client.cc
diff --git a/services/ui/public/cpp/window_tree_client.cc
b/services/ui/public/cpp/window_tree_client.cc
index
3005f4433bbea3b2289ece42abd4ba9beafd9ca5..b0d2402ab4537140fc85d1bd3288f244f398fbcb
100644
--- a/services/ui/public/cpp/window_tree_client.cc
+++ b/services/ui/public/cpp/window_tree_client.cc
@@ -25,7 +25,9 @@
 #include "services/ui/public/cpp/window_tree_client_delegate.h"
 #include "services/ui/public/cpp/window_tree_client_observer.h"
 #include
"services/ui/public/interfaces/window_manager_window_tree_factory.mojom.h"
+#include "ui/display/screen.h"
 #include "ui/events/event.h"
+#include "ui/gfx/geometry/dip_util.h"
 #include "ui/gfx/geometry/insets.h"
 #include "ui/gfx/geometry/size.h"
 
@@ -35,6 +37,22 @@ Id MakeTransportId(ClientSpecificId client_id,
ClientSpecificId local_id) {
   return (client_id << 16) | local_id;
 }
 
+// Helper function to get the device_scale_factor() of the display::Display
+// with |display_id|.
+float ScaleFactorForDisplay(int64_t display_id) {
+  // TODO(riajiang): Change to use display::GetDisplayWithDisplayId() after
+  // https://codereview.chromium.org/2361283002/ is landed.
+  std::vector<display::Display> displays =
+      display::Screen::GetScreen()->GetAllDisplays();
+  auto iter = std::find_if(displays.begin(), displays.end(),
+                           [display_id](const display::Display& display) {
+                             return display.id() == display_id;
+                           });
+  if (iter != displays.end())
+    return iter->device_scale_factor();
+  return 1.f;
+}
+
 // Helper called to construct a local window object from transport data.
 Window* AddWindowToClient(WindowTreeClient* client,
                           Window* parent,
@@ -50,7 +68,10 @@ Window* AddWindowToClient(WindowTreeClient* client,
       window_data->properties
           .To<std::map<std::string, std::vector<uint8_t>>>());
   client->AddWindow(window);
-  private_window.LocalSetBounds(gfx::Rect(), window_data->bounds);
+  private_window.LocalSetBounds(
+      gfx::Rect(),
+      gfx::ConvertRectToDIP(ScaleFactorForDisplay(window->display_id()),
+                            window_data->bounds));
   if (parent)
     WindowPrivate(parent).LocalAddChild(window);
   return window;
@@ -218,7 +239,10 @@ void WindowTreeClient::SetBounds(Window* window,
   DCHECK(tree_);
   const uint32_t change_id = ScheduleInFlightChange(
       base::MakeUnique<InFlightBoundsChange>(window, old_bounds));
-  tree_->SetWindowBounds(change_id, server_id(window), bounds);
+  tree_->SetWindowBounds(
+      change_id, server_id(window),
+      gfx::ConvertRectToPixel(ScaleFactorForDisplay(window->display_id()),
+                              bounds));
 }
 
 void WindowTreeClient::SetCapture(Window* window) {
@@ -250,12 +274,25 @@ void WindowTreeClient::SetClientArea(
     const gfx::Insets& client_area,
     const std::vector<gfx::Rect>& additional_client_areas) {
   DCHECK(tree_);
-  tree_->SetClientArea(window_id, client_area, additional_client_areas);
+  float device_scale_factor =
+      ScaleFactorForDisplay(GetWindowByServerId(window_id)->display_id());
+  std::vector<gfx::Rect> additional_client_areas_in_pixel;
+  for (const gfx::Rect& area : additional_client_areas) {
+    additional_client_areas_in_pixel.push_back(
+        gfx::ConvertRectToPixel(device_scale_factor, area));
+  }
+  tree_->SetClientArea(
+      window_id, gfx::ConvertInsetsToPixel(device_scale_factor, client_area),
+      additional_client_areas_in_pixel);
 }
 
 void WindowTreeClient::SetHitTestMask(Id window_id, const gfx::Rect& mask) {
   DCHECK(tree_);
-  tree_->SetHitTestMask(window_id, mask);
+  tree_->SetHitTestMask(
+      window_id,
+      gfx::ConvertRectToPixel(
+          ScaleFactorForDisplay(GetWindowByServerId(window_id)->display_id()),
+          mask));
 }
 
 void WindowTreeClient::ClearHitTestMask(Id window_id) {
@@ -872,11 +909,16 @@ void WindowTreeClient::OnWindowBoundsChanged(Id window_id,
   if (!window)
     return;
 
-  InFlightBoundsChange new_change(window, new_bounds);
+  float device_scale_factor = ScaleFactorForDisplay(window->display_id());
+  gfx::Rect old_bounds_in_dip =
+      gfx::ConvertRectToDIP(device_scale_factor, old_bounds);
+  gfx::Rect new_bounds_in_dip =
+      gfx::ConvertRectToDIP(device_scale_factor, new_bounds);
+
+  InFlightBoundsChange new_change(window, new_bounds_in_dip);
   if (ApplyServerChangeToExistingInFlightChange(new_change))
     return;
-
-  WindowPrivate(window).LocalSetBounds(old_bounds, new_bounds);
+  WindowPrivate(window).LocalSetBounds(old_bounds_in_dip, new_bounds_in_dip);
 }
 
 void WindowTreeClient::OnClientAreaChanged(
@@ -885,9 +927,15 @@ void WindowTreeClient::OnClientAreaChanged(
     mojo::Array<gfx::Rect> new_additional_client_areas) {
   Window* window = GetWindowByServerId(window_id);
   if (window) {
+    float device_scale_factor = ScaleFactorForDisplay(window->display_id());
+    std::vector<gfx::Rect> new_additional_client_areas_in_dip;
+    for (const gfx::Rect& area : new_additional_client_areas) {
+      new_additional_client_areas_in_dip.push_back(
+          gfx::ConvertRectToDIP(device_scale_factor, area));
+    }
     WindowPrivate(window).LocalSetClientArea(
-        new_client_area,
-        new_additional_client_areas.To<std::vector<gfx::Rect>>());
+        gfx::ConvertInsetsToDIP(device_scale_factor, new_client_area),
+        new_additional_client_areas_in_dip);
   }
 }
 
@@ -1268,12 +1316,14 @@ void WindowTreeClient::WmSetBounds(uint32_t change_id,
   bool result = false;
   if (window) {
     DCHECK(window_manager_delegate_);
-    gfx::Rect bounds = transit_bounds;
+    gfx::Rect transit_bounds_in_dip = gfx::ConvertRectToDIP(
+        ScaleFactorForDisplay(window->display_id()), transit_bounds);
+    gfx::Rect bounds = transit_bounds_in_dip;
     result = window_manager_delegate_->OnWmSetBounds(window, &bounds);
     if (result) {
       // If the resulting bounds differ return false. Returning false ensures
       // the client applies the bounds we set below.
-      result = bounds == transit_bounds;
+      result = bounds == transit_bounds_in_dip;
       window->SetBounds(bounds);
     }
   }
@@ -1421,8 +1471,12 @@ void
WindowTreeClient::SetUnderlaySurfaceOffsetAndExtendedHitArea(
     const gfx::Vector2d& offset,
     const gfx::Insets& hit_area) {
   if (window_manager_internal_client_) {
+    // TODO(riajiang): Figure out if |offset| needs to be converted.
+    // (http://crbugs.com/646932)
    
window_manager_internal_client_->SetUnderlaySurfaceOffsetAndExtendedHitArea(
-        server_id(window), offset.x(), offset.y(), hit_area);
+        server_id(window), offset.x(), offset.y(),
+        gfx::ConvertInsetsToDIP(ScaleFactorForDisplay(window->display_id()),
+                                hit_area));
   }
 }

[riajiang, 2016-11-17 19:02:31]

The CQ bit was checked by riajiang@chromium.org

[riajiang, 2016-11-17 19:02:31]

The patchset sent to the CQ was uploaded after l-g-t-m from
kylechar@chromium.org, sadrul@chromium.org, sky@chromium.org
Link to the patchset: https://codereview.chromium.org/2447303002/#ps320001
(title: "rebase")

[commit-bot: I haz the power, 2016-11-17 19:03:46]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-17 23:14:29]

Description was changed from

==========
Scale client area, hit test mask and bounds by device_scale_factor.

Fixes the second TODO of CL https://codereview.chromium.org/2341483004/.

TODO:
1. Chrome browser doesn't show up with window bounds scaling.
2. Also scale window bounds in aura::mus::WindowTreeClient once window
bounds work correctly.

BUG=646926
TEST=manual (with --force-device-scale-factor=2)
==========

to

==========
Scale client area, hit test mask and bounds by device_scale_factor.

Fixes the second TODO of CL https://codereview.chromium.org/2341483004/.

TODO:
1. Chrome browser doesn't show up with window bounds scaling.
2. Also scale window bounds in aura::mus::WindowTreeClient once window
bounds work correctly.

BUG=646926
TEST=manual (with --force-device-scale-factor=2)
==========

[commit-bot: I haz the power, 2016-11-17 23:14:30]

Committed patchset #12 (id:320001)

[commit-bot: I haz the power, 2016-11-17 23:19:05]

Description was changed from

==========
Scale client area, hit test mask and bounds by device_scale_factor.

Fixes the second TODO of CL https://codereview.chromium.org/2341483004/.

TODO:
1. Chrome browser doesn't show up with window bounds scaling.
2. Also scale window bounds in aura::mus::WindowTreeClient once window
bounds work correctly.

BUG=646926
TEST=manual (with --force-device-scale-factor=2)
==========

to

==========
Scale client area, hit test mask and bounds by device_scale_factor.

Fixes the second TODO of CL https://codereview.chromium.org/2341483004/.

TODO:
1. Chrome browser doesn't show up with window bounds scaling.
2. Also scale window bounds in aura::mus::WindowTreeClient once window
bounds work correctly.

BUG=646926
TEST=manual (with --force-device-scale-factor=2)

Committed: https://crrev.com/075c4290cdfc0cfca43c2114bce282ea9389e6fc
Cr-Commit-Position: refs/heads/master@{#433009}
==========

[commit-bot: I haz the power, 2016-11-17 23:19:07]

Patchset 12 (id:??) landed as
https://crrev.com/075c4290cdfc0cfca43c2114bce282ea9389e6fc
Cr-Commit-Position: refs/heads/master@{#433009}