[Piglit] [PATCH 1/2] integer-errors: check error detection for integer-related images
Brian Paul
brianp at vmware.com
Fri Jun 8 10:23:17 PDT 2012
On 06/08/2012 11:03 AM, Chad Versace wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 06/07/2012 03:40 PM, Brian Paul wrote:
>> On 06/07/2012 02:32 PM, Chad Versace wrote:
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>> On 06/06/2012 12:40 PM, Brian Paul wrote:
>>>> This is basically code that was in the old integer-texture test.
>>>> This test just validates GL error detection for glReadPixels, glDrawPixels,
>>>> glTexImage, etc. with regard to signed/unsigned integer formats.
>>>> ---
>>>> tests/all.tests | 1 +
>>>> tests/spec/gl-3.0/api/CMakeLists.gl.txt | 1 +
>>>> tests/spec/gl-3.0/api/integer-errors.c | 182 +++++++++++++++++++++++++++++++
>>>> 3 files changed, 184 insertions(+), 0 deletions(-)
>>>> create mode 100644 tests/spec/gl-3.0/api/integer-errors.c
>>>
>>> [snip]
>>>
>>>> + /* Check glTexSubImage for invalid format/type combination */
>>>> + {
>>>> + /* make valid texture image here */
>>>> + glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA32UI_EXT, 8, 8, 0,
>>>> + GL_RGBA_INTEGER, GL_UNSIGNED_INT, NULL);
>>>> +
>>>> + glTexSubImage2D(GL_TEXTURE_2D, 0,
>>>> + 0, 0, 4, 4,
>>>> + GL_RGBA_INTEGER, GL_FLOAT, NULL);
>>>> + if (!piglit_check_gl_error(GL_INVALID_ENUM))
>>>> + return false;
>>>> + }
>>>
>>> For the glTexSubImage case, it looks like the test incorrectly passes if the first
>>> call to glTexSubImage fails. Placing `if(!piglit_check_gl_error(GL_NO_ERROR)) return false`
>>> after the call should fix it.
>>
>> I'm not sure I follow. Are you concerned that the glTexImage2D() call should not generate an error?
>>
>> There's only one call to glTexSubImage2D() in the whole test.
>
> I misspoke. I was concerned that if glTexImage2D produces GL_INVALID_ENUM and glTexSubImage2D
> produces no error, then the test would still pass.
OK. I can fix that. Thanks.
-Brian
More information about the Piglit
mailing list