On Sat, Apr 26, 2014 at 2:53 AM, Emil Velikov <emil.l.veli...@gmail.com> wrote:
> On 25/04/14 11:41, Chia-I Wu wrote:
>> Hi,
>>
>> This series cleans up GL_KHR_debug support, fixes message control for some
>> corner cases, and makes one enhancement: cheap
>> glPushDebugGroup/glPopDebugGroup for command stream annotation.
>>
>> Patch 1-12 refactor the code. There should be no functional difference.
>> These patches are also part of the series I sent the other day for threaded
>> glCompileShader.
>>
>> Patch 13 makes an interface change to hide struct gl_debug_state from the
>> rest
>> of the driver. I need this for threaded glCompileShader support. But I
>> think
>> this is a good idea anyhow because, unlike other context states,
>> gl_debug_state is created on demand. This prevents others from having to do
>> an if-check.
>>
>> Patch 14-16 refactor more of the code. No functional difference is expected
>> either.
>>
>> Patch 17 makes debug group pushing copy-on-write. This makes the operation
>> very cheap in a common scenario.
>>
>> Patch 18 simplifies the debug namespace code a lot and fixes a few corner
>> cases.
>>
> Hi Chia-I,
>
> While you're around you may want to fold gl_debug_state::Defaults inside
> gl_debug_state::Namespaces as done in the attached patch.
>
> I must admit I've never looked at the spec to confirm if it would make sense
> to do so, although it saves us ~13K per gl_debug_state struct on x86-64
> builds.
I made a similar change as part of patch 15
Turn the 4-dimensional array, Defaults, in struct gl_debug_state to a
1-dimensional array in struct gl_debug_namespace.
>
> -Emil
>
>> Please review.
>> _______________________________________________
>> mesa-dev mailing list
>> mesa-dev@lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>>
>
--
o...@lunarg.com
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
