Implemented EGL_NV_post_sub_buffer.

Implemented EGL_NV_post_sub_buffer.

The change to the sample is just to demonstrate how I tested it. I won't check in the stuff under samples/.

It's not clear to me what eglPostSubBufferNV should do if EGL_POST_SUB_BUFFER_SUPPORTED_NV is false. At the moment it fails silently as though it was called on the wrong surface type:

 "If <surface> is a back-buffered surface, then the requested portion
    of the color buffer is copied to the native window associated with
    that surface. If <surface> is a single-buffered window, pixmap, or
    pbuffer surface, eglSwapBuffers and eglPostSubBufferNV have no
    effect."
Committed: http://code.google.com/p/angleproject/source/detail?r=902

[apatrick1, 2011-11-07 23:58:50]

Hi Daniel.

Please review when you get a chance.

Thanks,

Al

[dgkoch, 2011-11-16 21:43:53]

A few issues, but largely looks good.

BTW for the samples, instead of making throw away changes, how about just
copying one of the samples and adding it in as an example of how to use this,
and for future testing?

http://codereview.appspot.com/5345050/diff/3001/src/libEGL/Display.cpp
File src/libEGL/Display.cpp (right):

http://codereview.appspot.com/5345050/diff/3001/src/libEGL/Display.cpp#newcod...
src/libEGL/Display.cpp:955: mExtensionString += "EGL_NV_post_sub_buffer";
Please keep the extension string sorted.

http://codereview.appspot.com/5345050/diff/3001/src/libEGL/Surface.cpp
File src/libEGL/Surface.cpp (right):

http://codereview.appspot.com/5345050/diff/3001/src/libEGL/Surface.cpp#newcod...
src/libEGL/Surface.cpp:530: if (result == D3DERR_DEVICELOST || result ==
D3DERR_DEVICEHUNG || result == D3DERR_DEVICEREMOVED)
Please see latest source for how we are handling lost context now.

http://codereview.appspot.com/5345050/diff/3001/src/libEGL/libEGL.cpp
File src/libEGL/libEGL.cpp (right):

http://codereview.appspot.com/5345050/diff/3001/src/libEGL/libEGL.cpp#newcode...
src/libEGL/libEGL.cpp:1136: if (surface == EGL_NO_SURFACE)
Use validateSurface here?

[dgkoch, 2011-11-16 21:44:55]

http://codereview.appspot.com/5345050/diff/3001/samples/gles2_book/MipMap2D/M...
File samples/gles2_book/MipMap2D/MipMap2D.c (right):

http://codereview.appspot.com/5345050/diff/3001/samples/gles2_book/MipMap2D/M...
samples/gles2_book/MipMap2D/MipMap2D.c:314: eglPostSubBufferNV =
(PFNEGLPOSTSUBBUFFERNVPROC) eglGetProcAddress("eglPostSubBufferNV");
btw it's unnecessary to call eglGetProcAddress every draw call.. :-)

[apatrick1, 2011-11-16 22:38:37]

Sure, I find the samples a joy to work with. All done I think.

[dgkoch, 2011-11-18 18:47:17]

LGTM, other than the issue in the samples.

http://codereview.appspot.com/5345050/diff/12001/samples/gles2_book/Common/es...
File samples/gles2_book/Common/esUtil.c (right):

http://codereview.appspot.com/5345050/diff/12001/samples/gles2_book/Common/es...
samples/gles2_book/Common/esUtil.c:85: surface = eglCreateWindowSurface(display,
config, (EGLNativeWindowType)hWnd, surfaceAttribs);
problem here is *all* the samples will now be opting into post_sub_buffer
behaviour, which might not be entirely desirable...

[apatrick1, 2011-11-22 19:42:38]

Done.

[dgkoch, 2011-11-23 03:39:20]

one minor item.  LGTM with that fixed.

http://codereview.appspot.com/5345050/diff/17002/src/libEGL/Surface.cpp
File src/libEGL/Surface.cpp (right):

http://codereview.appspot.com/5345050/diff/17002/src/libEGL/Surface.cpp#newco...
src/libEGL/Surface.cpp:514: x, mHeight - y - height,
Should probably be using transformPixelRect() here to handle the RECT flipping.
I should have caught that earlier :-0

[ramasamykalai, 2011-11-23 08:49:34]

just for testing

http://codereview.appspot.com/5345050/diff/17002/src/libEGL/Surface.cpp
File src/libEGL/Surface.cpp (right):

http://codereview.appspot.com/5345050/diff/17002/src/libEGL/Surface.cpp#newco...
src/libEGL/Surface.cpp:310: if (rect.right >
static_cast<LONG>(presentParameters.BackBufferWidth))
why const ?

[apatrick1, 2011-11-23 17:18:43]

http://codereview.appspot.com/5345050/diff/17002/src/libEGL/Surface.cpp
File src/libEGL/Surface.cpp (right):

http://codereview.appspot.com/5345050/diff/17002/src/libEGL/Surface.cpp#newco...
src/libEGL/Surface.cpp:310: if (rect.right >
static_cast<LONG>(presentParameters.BackBufferWidth))
On 2011/11/23 08:49:35, ramasamykalai wrote:
> why const ?

I don't understand the question.

http://codereview.appspot.com/5345050/diff/17002/src/libEGL/Surface.cpp#newco...
src/libEGL/Surface.cpp:514: x, mHeight - y - height,
On 2011/11/23 03:39:20, dgkoch wrote:
> Should probably be using transformPixelRect() here to handle the RECT
flipping.
> I should have caught that earlier :-0

Fair point.