Update screen rects when showing an infobar.

Update screen rects when showing an infobar.


BUG=161748
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=170725

[keishi, 2012-11-30 01:42:38]

On Mac and Linux the screen rects weren't updated when the info bar was shown.
And it was causing the autofill popup and page popups to show up in the wrong
position. It seems impossible to completely keep the screen rects in sync on
these two platforms so this patch adds enough calls to UpdateScreenRect so that
we can fix this bug. I also added checks to see if the rects changed so we don't
send unnecessary IPCs.

[jam, 2012-11-30 02:12:06]

are you sure all these calls that you added are necesary?

https://codereview.chromium.org/11421166/diff/1/content/browser/renderer_host...
File content/browser/renderer_host/render_widget_host_view_mac.mm (right):

https://codereview.chromium.org/11421166/diff/1/content/browser/renderer_host...
content/browser/renderer_host/render_widget_host_view_mac.mm:333: NSWindow*
popupWindow = [[NSWindow alloc] initWithContentRect:NSMakeRect(10,100,400,300)
testing code?

[jam, 2012-11-30 22:44:20]

btw since you're in a different timezone, and this regression is my fault, I
patched this in to test on mac and linux.

on mac, in render_widget_host_view_mac.mm if you add the 
renderWidgetHostView_->render_widget_host_->SendScreenRects();

call to the existing setFrameSize method then it works

on linux, only the change on line 712 is needed

I don't know about Android or the guest. I don't think for Android this matters
since this data was never sent anyways afaik.

am i missing any other scenarios?

[keishi, 2012-12-03 05:19:10]

On 2012/11/30 22:44:20, John Abd-El-Malek wrote:
> btw since you're in a different timezone, and this regression is my fault, I
> patched this in to test on mac and linux.
> 
> on mac, in render_widget_host_view_mac.mm if you add the 
> renderWidgetHostView_->render_widget_host_->SendScreenRects();
> 
> call to the existing setFrameSize method then it works
> 
> on linux, only the change on line 712 is needed
> 
> I don't know about Android or the guest. I don't think for Android this
matters
> since this data was never sent anyways afaik.
> 
> am i missing any other scenarios?

The changes to android and guest were speculative. (I couldn't even figure out
what guest is used for.)

In this new patch I made the change minimal, like you suggested. My testing also
agrees that this should be enough to fix the bug.

[jam, 2012-12-03 08:13:08]

lgtm

[commit-bot: I haz the power, 2012-12-03 09:08:02]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/keishi@chromium.org/11421166/7001

[commit-bot: I haz the power, 2012-12-03 10:47:31]

Change committed as 170725