[Piglit] [PATCH 7/7] Add occlusion-query tests for various meta operations.

[Eric Anholt]

Carl Worth <cworth at cworth.org> writes:

> Eric Anholt <eric at anholt.net> writes:
>> concurrent
>
> Thanks for the review, Eric. I'll take care of little things like
> "concurrent" etc. throughout, as you mention.
>
>> It would be nice to verify that the correct number of fragments is
>> actually generated, not just != 0.  It's not obvious to me without
>> reading the spec what the right answer would be for glBitmap(), for
>> example.
>
> Yes, that would be a nice improvement to the tests. I don't know the
> actual correct numbers myself, (though I guess I can just sample the
> current implementation and see if the numbers make sense).
>
> It's interesting that you should mention glBitmap. My reading of the
> spec. suggests it should yield a non-zero result, but the current
> implementation on master, (without any of my patches), is yielding a
> value of 0 for this operation. I haven't looked closer to see what's
> going on there, but I'd be glad if anyone has any ideas.

Do not trust anything related to bitmap and OQs in any code path we have
right now.  Looking at the GL 3.0 spec section 3.8:

"Bitmaps are rectangles of zeros and ones specifying a particular
pattern of fragments to be produced. Each of these fragments has the
same associated data. These data are those associated with the current
raster position."

So I think the answer should be OQ incrementing per bit set in your
bitmap.

> Right. The test isn't verifying that. I meant more that I wanted to put
> the results on the framebuffer so that I could manually verify that the
> test was doing what I expected it to. Let me know if you think any of
> this should be removed. (I don't think adding any probing would improve
> the test, but it does seem like making the results visible could help
> anyone trying to decipher the test in the future. Perhaps I should just
> update the comment along those lines.)

Yeah, I guess it's OK as is.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/piglit/attachments/20130111/0de2d9e8/attachment.pgp>