[Piglit] [PATCH v2] gl30basic: add some extra suspected extension

[Tapani Pälli]

On 04/27/2017 05:20 PM, Brian Paul wrote:
> On 04/27/2017 05:04 AM, Sandra Koroniewska wrote:
>> This fixes
>> spec/gl30basic test on Windows Intel or Nvidia driver according to 
>> page 693 of openGL 4.4 spec ("In additional to OpenGL extensions, 
>> there are also ARB extensions to the related GLX and WGL APIs.").
>> ---
>>   tests/general/gl30basic.c | 8 ++++++--
>>   1 file changed, 6 insertions(+), 2 deletions(-)
>>
>> diff --git a/tests/general/gl30basic.c b/tests/general/gl30basic.c
>> index 2207787bb..08305a6f8 100644
>> --- a/tests/general/gl30basic.c
>> +++ b/tests/general/gl30basic.c
>> @@ -99,8 +99,12 @@ test_extension_list(void)
>>             ext[0] != 'G' ||
>>             ext[1] != 'L' ||
>>             ext[2] != '_') {
>> -         printf("%s: bad extension string [%d]: %s\n", Prog, k, ext);
>> -         return PIGLIT_FAIL;
>> +          if (ext[0] != 'W' ||
>> +              ext[1] != 'G' ||
>> +              ext[2] != 'L') {
>> +              printf("%s: bad extension string [%d]: %s\n", Prog, k, 
>> ext);
>> +              return PIGLIT_FAIL;
>> +          }
>>         }
>>         if (strchr((char *) ext, ' ')) {
>>            printf("%s: extension string [%d] contains a space: %s\n", 
>> Prog, k, ext);
>>
> 
> Hmm, It looks like NVIDIA's driver does report WGL_EXT_swap_control with 
> glGetStringi(GL_EXTENSIONS, i).  That's really weird.  Is that what 
> Intel's Windows driver does?
> 
> Doing that is actually counter to the extension spec for 
> WGL_EXT_swap_control which says:
> 
> """
> Dependencies on WGL_EXT_extensions_string
> 
>      Because there is no way to extend wgl, these calls are defined in
>      the ICD and can be called by obtaining the address with
>      wglGetProcAddress.  Because this extension is a WGL extension, it
>      is not included in the GL_EXTENSIONS string.  Its existence can be
>      determined with the WGL_EXT_extensions_string extension.
> 
>      For historical reasons, GL_EXT_SWAP_CONTROL is also included in the
>      GL_EXTENSIONS string as return from glGetString.
> """
> 
> I believe the last sentence should read "GL_EXT_swap_control" and not 
> "GL_EXT_SWAP_CONTROL".  The spec for GL_EXT_swap_control make a vague 
> reference to a shortcoming in the WGL extension querying mechanism so 
> I'm not sure what the whole store is there.  NVIDIA's driver does not 
> support GL_EXT_swap_control AFAICT.

I found that WGL_ARB_extensions_string spec has this:

--- 8< ---
Issues

   1. Note that extensions that were previously advertised via
      glGetString (e.g., the swap interval extension) should continue to
      be advertised there so existing applications don't break. They
      should also be advertised via wglGetExtensionsStringARB so new
      applications can make one call to find out which WGL extensions are
      supported.
--- 8< ---

I guess it refers to WGL_EXT_swap_control (?)



> In any case, I think your patch is OK, but let's please put a comment in 
> to the effect of:
> 
> /* Note: normally, glGetStringi(GL_EXTENSIONS, i) should only return
>   * "core" OpenGL extension strings and not window system (WGL/GLX)
>   * extensions.  However, NVIDIA's and Intel's Windows drivers include
>   * WGL_EXT_swap_control in the core extensions list.  This is a special
>   * case for that.
>   */
> 
> 
> -Brian
> 
> _______________________________________________
> Piglit mailing list
> Piglit at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/piglit