Makes wm2 reset the bounds immediately after cloning the layer for

Makes wm2 reset the bounds immediately after cloning the layer for
constrained windows. Doing otherwise leads to constrained windows
having an empty bounds and getting into a confused state if someone
asks for them.

BUG=147373
TEST=see bug
R=ben@chromium.org,skuhne@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=158804

[sky, 2012-09-25 23:29:35]

I need to merge this to the branch, so I'm keeping it short. I plan on doing
some changes that mean we don't need aura::client::kConstrainedWindowKey, but
that's bigger and makes me nervous for a merge. And I need a test, but that'll
come after the cleanup. I'll do the cleanup after I merge.

[Mr4D (OOO till 08-26), 2012-09-26 04:25:08]

lgtm

Just thinking: Don't we have a simple way to apply a method to each element of a
sub tree?

http://codereview.chromium.org/10984032/diff/1/ash/wm/workspace/workspace_lay...
File ash/wm/workspace/workspace_layout_manager2.cc (right):

http://codereview.chromium.org/10984032/diff/1/ash/wm/workspace/workspace_lay...
ash/wm/workspace/workspace_layout_manager2.cc:62: BoundsMap::const_iterator i =
bounds_map.find(window);
It is a bit confusing to have to 'i' variables in one procedure which have so
different meaning.

[sky, 2012-09-26 04:51:14]

We don't have a way to apply a method to a subtree, but it seems nice doesn't
it.

http://codereview.chromium.org/10984032/diff/1/ash/wm/workspace/workspace_lay...
File ash/wm/workspace/workspace_layout_manager2.cc (right):

http://codereview.chromium.org/10984032/diff/1/ash/wm/workspace/workspace_lay...
ash/wm/workspace/workspace_layout_manager2.cc:62: BoundsMap::const_iterator i =
bounds_map.find(window);
On 2012/09/26 04:25:08, Mr4D wrote:
> It is a bit confusing to have to 'i' variables in one procedure which have so
> different meaning.

They have totally different scopes. It's no different than having two separate
for loops use 'i'.

[Ben Goodger (Google), 2012-09-26 15:33:50]

lgtm