Invalid client window area when resetting swap chain

Invalidate client window area when resetting swap chain

Resizing a window larger results in the newly exposed region being invalidated but the old region is treated as valid. This can result in the old region no longer updating. This has been observed on Windows 7 64-bit with Aero theme using NVIDIA GeForce GTS 250 and driver version 301.42. Invalidate the entire client window area when resetting the swap chain so that it updates properly.

Original bug report: https://bugreports.qt-project.org/browse/QTBUG-27822

[apatrick1, 2012-11-05 19:59:49]

I'm not convinced that ANGLE should be responsible for invalidating the window.
Can you provide more context about the bug and how this fixes it? Thx.

https://codereview.appspot.com/6812076/diff/5001/src/libEGL/Surface.cpp
File src/libEGL/Surface.cpp (right):

https://codereview.appspot.com/6812076/diff/5001/src/libEGL/Surface.cpp#newco...
src/libEGL/Surface.cpp:313: ASSERT(SUCCEEDED(result));
This assertion is redundant. Did you mean to assert that InvalidateRect
succeeded?

[dgkoch, 2012-11-05 20:27:13]

Please also note that you must sign the CLA and send me the required info as
detailed on http://code.google.com/p/angleproject/wiki/ContributingCode before
we can accept patches.

[Net147, 2012-11-05 22:12:41]

Added more detail to the commit message and fixed the assert. I have also
electronically signed the individual CLA.

[apatrick1, 2012-11-05 22:32:25]

Daniel, thoughts on whether this an appropriate thing for ANGLE to do? It seems
harmless enough. We haven't observed anything like this on Chrome/XP to my
knowledge.

https://codereview.appspot.com/6812076/diff/5005/src/libEGL/Surface.cpp
File src/libEGL/Surface.cpp (right):

https://codereview.appspot.com/6812076/diff/5005/src/libEGL/Surface.cpp#newco...
src/libEGL/Surface.cpp:312: result = InvalidateRect(mWindow, 0, FALSE);
InvalidateRect has return type BOOL. result has type HRESULT and the SUCCEEDED
macro expects HRESULT. Honestly, I'd just ignore the result of InvalidateRect.

[Net147, 2012-11-05 22:45:42]

ASSERT removed.

[apatrick1, 2012-11-05 22:47:55]

Looks okay to me. Over to Daniel.

[dgkoch, 2012-11-06 13:19:35]

On 2012/11/05 22:32:25, apatrick1 wrote:
> Daniel, thoughts on whether this an appropriate thing for ANGLE to do? It
seems
> harmless enough. We haven't observed anything like this on Chrome/XP to my
> knowledge.

Yeah, it seems reasonably harmless enough to include in ANGLE.

Will you need someone to land this for you?

https://codereview.appspot.com/6812076/diff/13001/CONTRIBUTORS
File CONTRIBUTORS (right):

https://codereview.appspot.com/6812076/diff/13001/CONTRIBUTORS#newcode70
CONTRIBUTORS:70: Jonathan Liu
Your name should go in the AUTHORS file since you signed the CLA, not here.

[Net147, 2012-11-06 22:01:22]

On 2012/11/06 13:19:35, dgkoch wrote:
> CONTRIBUTORS:70: Jonathan Liu
> Your name should go in the AUTHORS file since you signed the CLA, not here.

Done.

[Net147, 2012-11-08 10:28:26]

Daniel, will you be able to land the patch?

[dgkoch, 2012-11-08 12:49:54]

On 2012/11/08 10:28:26, Net147 wrote:
> Daniel, will you be able to land the patch?

LGTM.  Yes I'll land it for you.

[Net147, 2012-11-13 09:22:11]

On 2012/11/08 12:49:54, dgkoch wrote:
> On 2012/11/08 10:28:26, Net147 wrote:
> > Daniel, will you be able to land the patch?
> 
> LGTM.  Yes I'll land it for you.

Any update?

[dgkoch, 2012-11-14 14:18:21]

On 2012/11/13 09:22:11, Net147 wrote:
> On 2012/11/08 12:49:54, dgkoch wrote:
> > On 2012/11/08 10:28:26, Net147 wrote:
> > > Daniel, will you be able to land the patch?
> > 
> > LGTM.  Yes I'll land it for you.
> 
> Any update?

Committed at r1392: http://code.google.com/p/angleproject/source/detail?r=1392
Please close review issue.