mac: Fix sending of mouse moved events after resizing window's lower left corner

mac: Fix sending of mouse moved events after resizing window's lower left corner

NSTrackingInVisibleRect is apparently just broken (see
http://openradar.appspot.com/radar?id=2773401 ), so keep the tracking area
frame up to date manually.

I tried to write a test, but failed to simulate mouse events that impress
NSTrackingArea, see patch sets on the review.

BUG=176725
TEST=see bug: Open the html file in the first comment, then put the cursor in the inside lower left corner of the window and click-drag-resize the window so that it grows in y direction (x direction doesn't matter). The js-triggered position display should continue updating when moving the cursor to the upper right (further into the window) after the resize operation.
NOTRY=true

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=188430

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=188735

[Nico, 2013-03-12 23:12:54]

asvitkine: I'm having trouble with writing a test. I basically need a test that
simulates mouse events to resize a window. The CL contains what I've tried so
far; it's not working. Do you know if this is possible at all? The internet
suggests that simulating click events might be tricky.

[Alexei Svitkine (slow), 2013-03-13 15:09:10]

Sorry, I don't have any special insight here.

Is your test code actually not succeeding to resize it?

Could you try to simulate the same thing using just -setFrame: (i.e. the
equivalent calls that should get made when the window is resized using the
mouse), or does that actually work correctly? (i.e. does the mouse resize stuff
go through public APIs or use some internal stuff)

[Nico, 2013-03-13 19:59:15]

On 2013/03/13 15:09:10, Alexei Svitkine wrote:
> Sorry, I don't have any special insight here.
> 
> Is your test code actually not succeeding to resize it?
> 
> Could you try to simulate the same thing using just -setFrame: (i.e. the
> equivalent calls that should get made when the window is resized using the
> mouse), or does that actually work correctly? (i.e. does the mouse resize
stuff
> go through public APIs or use some internal stuff)

As it turns out, I can't even simulate mouse move events without resizing the
window. I tried a different way to spin the event loop (see patch set 3), but no
dice. I guess I'll try some more.

[Alexei Svitkine (slow), 2013-03-13 20:14:02]

Try NSApplication's -sendEvent: instead of the NSWindow one?

[Nico, 2013-03-13 20:15:43]

That's what I had originally. (Not that switching to the NSWindow version
helped.)


On Wed, Mar 13, 2013 at 1:14 PM, <asvitkine@chromium.org> wrote:

> Try NSApplication's -sendEvent: instead of the NSWindow one?
>
>
https://codereview.chromium.**org/12330108/<https://codereview.chromium.org/1...
>

[Nico, 2013-03-13 20:38:56]

Making the view first responder and calling setAcceptsMouseMoved: on the
window lets the test pass (
http://developer.apple.com/library/mac/#documentation/cocoa/conceptual/EventO...),
but that kind of defeats the tracking area :-P The test still passes
if
I completely remove the tracking area from BaseView.


On Wed, Mar 13, 2013 at 1:15 PM, Nico Weber <thakis@chromium.org> wrote:

> That's what I had originally. (Not that switching to the NSWindow version
> helped.)
>
>
> On Wed, Mar 13, 2013 at 1:14 PM, <asvitkine@chromium.org> wrote:
>
>> Try NSApplication's -sendEvent: instead of the NSWindow one?
>>
>>
https://codereview.chromium.**org/12330108/<https://codereview.chromium.org/1...
>>
>
>

[Nico, 2013-03-13 21:37:39]

Scrapped the test :-/

Please take a look. (Or 302 to rsesek, if you're not familiar with tracking
areas.)

[Alexei Svitkine (slow), 2013-03-14 00:42:47]

Too bad you couldn't find a way to test this. :\

https://codereview.chromium.org/12330108/diff/12002/ui/base/cocoa/base_view.mm
File ui/base/cocoa/base_view.mm (right):

https://codereview.chromium.org/12330108/diff/12002/ui/base/cocoa/base_view.m...
ui/base/cocoa/base_view.mm:20: trackingArea_ = [[NSTrackingArea alloc]
initWithRect:frame
Could this just call [self refreshTrackingAreaSize], or does [self
removeTrackingArea:nil] cause problems?

https://codereview.chromium.org/12330108/diff/12002/ui/base/cocoa/base_view.m...
ui/base/cocoa/base_view.mm:146: - (void)refreshTrackingAreaSize {
Should this be called refreshTrackingAreaRect, since it's not just the size
you're interested in, right?

https://codereview.chromium.org/12330108/diff/12002/ui/base/cocoa/base_view.m...
ui/base/cocoa/base_view.mm:165: [self refreshTrackingAreaSize];
So this is called because -[NSView setFrame:] doesn't call -setFrameSize: if the
size hasn't changed? Maybe clarify this in the comment ("if the size changed").

It would also be nice to at-least have a test for what you say in the comment
(that setFrame: does call setFrameSize:), to make sure we don't break if the
implementation of -setFrame: changes.

[Nico, 2013-03-14 22:09:39]

I found NSView's -updateTrackingAreas and I'm now overriding that.

[Alexei Svitkine (slow), 2013-03-14 22:38:25]

LGTM


On Thu, Mar 14, 2013 at 6:09 PM, <thakis@chromium.org> wrote:

> I found NSView's -updateTrackingAreas and I'm now overriding that.
>
>
https://codereview.chromium.**org/12330108/<https://codereview.chromium.org/1...
>

[commit-bot: I haz the power, 2013-03-14 22:51:09]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/thakis@chromium.org/12330108/18001

[commit-bot: I haz the power, 2013-03-14 22:51:13]

Presubmit check for 12330108-18001 failed and returned exit status 1.

INFO:root:Found 3 file(s).

Running presubmit commit checks ...
Running /b/commit-queue/workdir/chromium/PRESUBMIT.py

** Presubmit Warnings **
Banned functions were used.
    ui/base/cocoa/base_view.mm:144:
      The use of NSTrackingAreas is prohibited. Please use CrTrackingArea
      instead.
      http://dev.chromium.org/developers/coding-style/cocoa-dos-and-donts

Presubmit checks took 2.8s to calculate.

[commit-bot: I haz the power, 2013-03-18 14:47:21]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/thakis@chromium.org/12330108/18001

[commit-bot: I haz the power, 2013-03-18 14:47:23]

Failed to apply patch for
content/browser/renderer_host/render_widget_host_view_mac.mm:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  patching file content/browser/renderer_host/render_widget_host_view_mac.mm
  Hunk #2 FAILED at 23.
  1 out of 3 hunks FAILED -- saving rejects to file
content/browser/renderer_host/render_widget_host_view_mac.mm.rej

Patch:       content/browser/renderer_host/render_widget_host_view_mac.mm
Index: content/browser/renderer_host/render_widget_host_view_mac.mm
diff --git a/content/browser/renderer_host/render_widget_host_view_mac.mm
b/content/browser/renderer_host/render_widget_host_view_mac.mm
index
f4131bfd3753b5215531f4cb5b5d014d5e653a27..8a4ed0b326b75ce3b96b745624e30e17f48b5aa2
100644
--- a/content/browser/renderer_host/render_widget_host_view_mac.mm
+++ b/content/browser/renderer_host/render_widget_host_view_mac.mm
@@ -13,7 +13,6 @@
 #include "base/logging.h"
 #include "base/mac/mac_util.h"
 #include "base/mac/scoped_cftyperef.h"
-#import "base/mac/scoped_nsautorelease_pool.h"
 #import "base/memory/scoped_nsobject.h"
 #include "base/message_loop.h"
 #include "base/metrics/histogram.h"
@@ -24,12 +23,10 @@
 #include "base/utf_string_conversions.h"
 #import "content/browser/accessibility/browser_accessibility_cocoa.h"
 #include "content/browser/accessibility/browser_accessibility_manager.h"
-#include "content/browser/plugin_process_host.h"
 #import "content/browser/renderer_host/accelerated_plugin_view_mac.h"
 #include "content/browser/renderer_host/backing_store_mac.h"
 #include "content/browser/renderer_host/backing_store_manager.h"
 #include "content/browser/renderer_host/compositing_iosurface_mac.h"
-#include "content/browser/renderer_host/render_process_host_impl.h"
 #include "content/browser/renderer_host/render_view_host_impl.h"
 #import
"content/browser/renderer_host/render_widget_host_view_mac_editcommand_helper.h"
 #import "content/browser/renderer_host/text_input_client_mac.h"
@@ -48,7 +45,6 @@
 #include
"third_party/WebKit/Source/WebKit/chromium/public/mac/WebInputEventFactory.h"
 #include
"third_party/WebKit/Source/WebKit/chromium/public/mac/WebScreenInfoFactory.h"
 #import "third_party/mozilla/ComplexTextInputPanel.h"
-#include "third_party/skia/include/core/SkColor.h"
 #import "ui/base/cocoa/fullscreen_window_manager.h"
 #import "ui/base/cocoa/underlay_opengl_hosting_window.h"
 #include "ui/base/keycodes/keyboard_codes.h"

[Nico, 2013-03-18 14:51:41]

Committed patchset #6 manually as r188735 (presubmit successful).