[Piglit] [RFC 1/3] hack: enable shader runner on 4.2.0 NVIDIA 304.88

[Pohjolainen, Topi]

On Wed, Oct 16, 2013 at 10:25:00AM -0500, Ken Phillis Jr wrote:
>    On Wed, Oct 16, 2013 at 12:43 AM, Pohjolainen, Topi
>    <topi.pohjolainen at intel.com> wrote:
> 
>      I don't know why it fixes it, I just tried it without the visual
>      settings and
>      found out that I could run the shader_runner by relaxing the settings.
>      I'm
>      running 7.1 debian and I thought that the gl-driver and X-server I get
>      via the
>      distribution should be just fine - shader_runner is quite central test
>      in
>      piglit after all. That's why I wrote in the cover letter that:
> 
>           First patch I included just to ask if there something amiss in
>           my system. I have
> 
>           X.Org X Server 1.12.4
>           Release Date: 2012-08-27
> 
>           running on NVIDIA (304.88 on GTX 660). Without the patch I just
>           wasn't able to run the shader_runner at all:
> 
>           piglit: info: Failed to create GL 3.2 core context
>           piglit: info: Falling back to GL 3.2 compatibility context
>           piglit: info: Failed to create GL 3.2 compatibility context
>           piglit: info: Failed to create any GL context
>           PIGLIT: {'result': 'skip' }
>      -Topi
> 
>    I Think this is a bug to a specific nvidia driver release. So is it
>    possible to retry this test against the Nvidia 325.08 driver and Nvidia
>    331.13 drivers?

I'm looking into this but I really wanted to avoid removing the distribution
packages and installing the driver directly (even debian unstable provides only
304.108 which isn't any different w.r.t. the problem in hand).

> 
>    Also, what is the exact test that gives this result?

It is the shader_runner binary that won't get passed the configuration phase.
The exact *.shader_test file being used doesn't really make any difference as
it's contents are considered only after the configuration succeeds.