Re: [Piglit] [PATCH] shader_runner: fix uniform array name lookups

2015-03-04 Thread Arthur Huillet
On 2015-03-04 15:58, Ilia Mirkin wrote: On Wed, Mar 4, 2015 at 4:53 AM, Arthur Huillet wrote: From: Arthur Huillet Don't look up uniform names for non-zero array elements, as this is illegal per GL4.5. From the discussion of GetProgramResourceIndex in the GL4.5 spec: If

[Piglit] [PATCH] DSA: fix error value for *TextureSubImage* when target doesn't match

2015-04-22 Thread Arthur Huillet
From: Arthur Huillet Section 8.6 of the OpenGL 4.5 compatibility specification states: "An INVALID_OPERATION error is generated by *TextureSubImage* if the effective target of texture does not match the command, as shown in ta- ble 8.23." Piglit was expecting INVALID_ENUM. Sig

[Piglit] "Back up your tests with spec" policy proposal

2015-04-22 Thread Arthur Huillet
From: Arthur Huillet Hello, there are a few things being tested by the direct_state_access* series that don't appear correct to me. I suggest that future tests in Piglit should have a comment that quotes the part of the specification they are testing - this will make it easie

Re: [Piglit] [PATCH] DSA: fix error value for *TextureSubImage* when target doesn't match

2015-04-22 Thread Arthur Huillet
Hi, On 2015-04-22 18:37, Ilia Mirkin wrote: From the ARB_dsa spec: An INVALID_ENUM error is generated by *TexSubImage* if does not match the command, as shown in table 8.subtarg. An INVALID_OPERATION error is generated by *TextureSubImage* if the effective target of does no

Re: [Piglit] [PATCH] DSA: fix error value for *TextureSubImage* when target doesn't match

2015-04-23 Thread Arthur Huillet
Hi, On Thu, 23 Apr 2015 14:26:32 +0300 Martin Peres wrote: > >>> Does the patch look OK to you? > >> Yeah, patch looks perfectly fine. But I'd like to hear from Laura (or > >> Martin perhaps, who reviewed a lot of these) about whether this was > >> done on purpose or not, pending a spec fix. >

[Piglit] [PATCH 2/2] DSA: CreateTextures() with a NULL argument is allowed to crash

2015-04-24 Thread Arthur Huillet
From: Arthur Huillet The pointer passed to CreateTextures() isn't required by the specification to be checked for validity, and the implementation is allowed to crash (as happens with NVIDIA's). Remove the bogus tests. Signed-off-by: Arthur Huillet --- tests/spec/arb_direct_st

[Piglit] [PATCH 1/2] DSA: fix incorrect errors expected for BindTextureUnit()

2015-04-24 Thread Arthur Huillet
From: Arthur Huillet Updated BindTextureUnit() to match the OpenGL specification: GL_INVALID_OPERATION is expected when the texture object doesn't exist, not GL_INVALID_ENUM. When the texture unit doesn't exist, the specification doesn't strictly speaking define the erro

Re: [Piglit] [PATCH 1/2 v2] DSA: fix incorrect errors expected for BindTextureUnit()

2015-04-24 Thread Arthur Huillet
Fix typo and indentation, specify GL spec version the quotes are from. Thanks! ___ Piglit mailing list Piglit@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/piglit

[Piglit] [PATCH] DSA: fix incorrect errors expected for BindTextureUnit()

2015-04-24 Thread Arthur Huillet
From: Arthur Huillet Updated BindTextureUnit() to match the OpenGL specification: GL_INVALID_OPERATION is expected when the texture object doesn't exist, not GL_INVALID_ENUM. When the texture unit doesn't exist, the specification doesn't strictly speaking define the erro

Re: [Piglit] [PATCH 1/2 v2] DSA: fix incorrect errors expected for BindTextureUnit()

2015-04-24 Thread Arthur Huillet
On 2015-04-24 17:58, Arthur Huillet wrote: Fix typo and indentation, specify GL spec version the quotes are from. ... and I apologize for publicly shooting myself in the foot with git-send-email. :) -- Greetings, A.H. ___ Piglit mailing list

Re: [Piglit] [PATCH v2] arb_framebuffer_no_attachments: add params set&get test

2015-04-26 Thread Arthur Huillet
On Wed, 22 Apr 2015 08:06:51 +0300 Tapani Pälli wrote: > All other tests except invalid_enum_check pass on Nvidia > binary driver (version 346.35), What is failing on the NVIDIA binary driver, and can you prove that what Piglit is testing is correct? I believe quoting the spec in invalid_enum_

Re: [Piglit] [PATCH] arb_framebuffer_no_attachments: check GL_EXT_direct_state_access

2015-04-29 Thread Arthur Huillet
I'm not entirely following the first part, so get approval from somebody else, but for the second part: Reviewed-by: Arthur Huillet (And actually even Suggested-by :p) Thanks! On 2015-04-29 13:46, Tapani Pälli wrote: Check for GL_EXT_direct_state_access which was the original purpo

Re: [Piglit] [PATCH 0/5] Randomized UBO tests of doom

2015-11-10 Thread Arthur Huillet
On 2015-11-10 2:08, Ian Romanick wrote: On 11/09/2015 05:19 AM, Ilia Mirkin wrote: On Thu, Sep 25, 2014 at 4:39 PM, Ian Romanick wrote: On 09/24/2014 09:47 AM, Ian Romanick wrote: So, here it is. Finally. The first two patches provide the infrastructure for generating randomized UBO tests.

Re: [Piglit] [PATCH] winsys-framework Default to showing window

2015-11-17 Thread Arthur Huillet
Hi Alex, You're right on the theory, but your patch seems to have an issue (read on). On 2015-08-13 19:49, agoins wrote: winsys-framework incorrectly assumed that it will always have ownership of the pixels in its own buffers. If using the default framebuffer (i.e., when Piglit is not runnin

Re: [Piglit] [PATCH v2] winsys-framework Default to showing window

2015-11-17 Thread Arthur Huillet
Thanks :) Reviewed by: Arthur Huillet On Tue, 17 Nov 2015 11:33:31 -0800 Alex Goins wrote: > From: agoins > > winsys-framework incorrectly assumed that it will always have ownership of the > pixels in its own buffers. If using the default framebuffer (i.e., when Piglit > is

[Piglit] [PATCH] GL_ARB_draw_instanced/dlist: Fix crash

2016-12-06 Thread arthur . huillet
From: Arthur Huillet The test was passing a NULL indices pointer without an element buffer bound. Nothing in the GL spec requires a NULL check, so drivers were free to crash before checking the error condition that the test attempts to check. Fix this by passing an index buffer. Also swap

[Piglit] [PATCH] GL_ARB_draw_instanced/dlist: Fix crash

2016-12-08 Thread arthur . huillet
From: Arthur Huillet The test was passing a NULL indices pointer without an element buffer bound. Nothing in the GL spec requires a NULL check, so drivers were free to crash before checking the error condition that the test attempts to check. Fix this by passing an index buffer. Also swap

Re: [Piglit] [PATCH] arb_shader_storage_buffer_object: don't do out-of-bounds writes

2017-08-25 Thread Arthur Huillet
Reviewed-by: Arthur Huillet The test now passes with our change of behavior to out-of-bounds writes (redirected to 0 instead of ignored). That change isn't present in any currently available NVIDIA driver release. Thanks! On 25.08.2017 08:13, Samuel Iglesias Gonsálvez wrote: From GLG

Re: [Piglit] [PATCH] arb_texture_multisample: stress test of very large textures (v3)

2017-12-26 Thread Arthur Huillet
On Sun, 24 Dec 2017 15:41:38 -0700 Brian Paul wrote: > + * Some drivers/GPUs may fail this test. NVIDIA, for example, appears to > + * only store the MSAA coverage info, not the sample colors, for samples > + * beyond the 8th sample. We may tune the way this test operates over time > + * to be

Re: [Piglit] [PATCH 1/3] arb_get_texture_sub_image: fix expected error when querying a level which hasn't been explicitly defined

2018-03-28 Thread Arthur Huillet
I'm confused why the test is passing a bad enum when its purpose seems to be to test a non-existent level. In any case, your change looks correct to me. Reviewed-by: Arthur Huillet On 28.03.2018 17:15, Anthony Pesch wrote: From: Anthony Pesch Change expected error from INVALID_OPER

Re: [Piglit] [PATCH 2/3] arb_get_texture_sub_image: fix depth parameter when performing zero-sized texture queries

2018-03-28 Thread Arthur Huillet
Reviewed-by: Arthur Huillet Thanks -- Arthur On 28.03.2018 17:15, Anthony Pesch wrote: From: Anthony Pesch Change zero-sized texture queries to pass a valid depth parameter, otherwise they should error as per the OpenGL 4.6 Core spec: "An INVALID_VALUE error is generated if the effe

Re: [Piglit] [PATCH 3/3] arb_get_texture_sub_image: update cube map tests to complete the cube map

2018-03-28 Thread Arthur Huillet
I'm less confident than for the others but Reviewed-by: Arthur Huillet On 28.03.2018 17:15, Anthony Pesch wrote: From: Anthony Pesch Update cube map tests to complete the cube map before performing the final query. This final query is expected to succeed, however, querying a cub

[Piglit] [PATCH] all.py: add run_concurrent=False where needed

2018-11-14 Thread arthur . huillet
From: Arthur Huillet Commit 57537d45b75218438716506594e16b91dade968f removed run_concurrent=False from a bunch of tests, saying there was no reason for them to be marked as such. This is wrong for at least two tests, which require a displayed window (cannot render to an FBO). As such, when run

Re: [Piglit] [PATCH] all.py: add run_concurrent=False where needed

2019-01-22 Thread Arthur Huillet
on NVIDIA driver is that we have a unified backbuffer under some circumstances. Thanks -- Arthur On 14.11.2018 13:41, arthur.huil...@free.fr wrote: From: Arthur Huillet Commit 57537d45b75218438716506594e16b91dade968f removed run_concurrent=False from a bunch of tests, saying there was no

Re: [Piglit] [PATCH] all.py: add run_concurrent=False where needed

2019-01-22 Thread Arthur Huillet
are a LOT more tests that will have this problem. -ilia On Tue, Jan 22, 2019 at 5:40 AM Arthur Huillet wrote: Ping. Pixel ownership test makes it impossible for Piglit to blindly assume that it can read back all the pixels of its backbuffer, so tests that don't use an FBO can't