Closed GoogleCodeExporter closed 9 years ago
Original comment by g...@chromium.org
on 15 Nov 2012 at 9:55
It depends on exactly what operation is being attempted. If the scissor test
is enabled, this does potentially reduce the effective area of the region being
copied, as the area to be copied is the intersection of the requested rect and
the scissor rect.
The ANGLE_framebuffer_blit extension does not support partial multisample
resolves or partial depth/stencil blits, so
- if the scissor test is enabled and the scissor rect only specifies a portion of the surface, and
- the source is multi-sampled, or
- the mask includes DEPTH_BUFFER_BIT or STENCIL_BUFFER_BIT
this will result in INVALID_OPERATION, by design. Otherwise it might be a bug, in which case a minimal test case would be appreciated.
Original comment by dan...@transgaming.com
on 16 Nov 2012 at 1:35
Is that specified in somewhere that with SCISSOR test it does work with
multi-sample but does otherwise?
Seems like it would be better if just ignored SCISSOR test no?
Original comment by g...@chromium.org
on 16 Nov 2012 at 1:41
I'm having a hard time parsing the first question. Feels like it's missing a
"not" somewhere, but I'm not sure in which clause.
The scissor test is applied in *all* cases, whether multi sampled or not.
However, BlitFramebufferANGLE only supports partial buffer copies for
non-multisampled color-only copies. This is all specified in
http://code.google.com/p/angleproject/source/browse/trunk/extensions/ANGLE_frame
buffer_blit.txt
The SCISSOR test is not ignored in any other GL implementations, so I don't
think it should be in the ANGLE one either.
In all the BlitFramebuffer specifications (EXT_framebuffer_blit,
ARB_framebuffer_objects, GL 3.0-4.3, GLES 3.0) it says something to the effect
of:
"The pixel copy bypasses the fragment pipeline. The only fragment
operations which affect the blit are the pixel ownership test and
the scissor test."
(Recent specs also mention sRGB conversions, but that is not relevant here).
Original comment by dan...@transgaming.com
on 16 Nov 2012 at 1:53
ok, sorry for the mis-understanding
Original comment by g...@chromium.org
on 19 Nov 2012 at 4:54
Original issue reported on code.google.com by
g...@chromium.org
on 15 Nov 2012 at 9:55