[Mesa-dev] Mesa (master): mesa: always generate error in glColorTableParameter[fi]v()

[Brian Paul]

On 03/01/2011 05:50 PM, Ian Romanick wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 03/01/2011 01:30 PM, Brian Paul wrote:
>> On 03/01/2011 12:12 PM, Ian Romanick wrote:
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>> On 02/28/2011 05:34 PM, Brian Paul wrote:
>>>> Module: Mesa
>>>> Branch: master
>>>> Commit: 7d8db55148b0861e35ec6bb6323db6dad4c8f17f
>>>> URL:
>>>> http://cgit.freedesktop.org/mesa/mesa/commit/?id=7d8db55148b0861e35ec6bb6323db6dad4c8f17f
>>>>
>>>>
>>>> Author: Brian Paul<brianp at vmware.com>
>>>> Date:   Mon Feb 28 18:24:30 2011 -0700
>>>>
>>>> mesa: always generate error in glColorTableParameter[fi]v()
>>>
>>> Do we even need to keep these functions?
>>
>> I elected to stub them out for now as Eric did with the glHistogram*()
>> and glConvolution*() functions.  All these stubs could probably be
>> removed but we should test to be sure nothing goes wrong if someone does
>> call them.  The API dispatch no-op code _should_ handle it.
>
> That's what I was thinking.
>
>> If we can get rid of paletted texture support (which I think is only
>> used in swrast and a couple of old DRI drivers) we could remove all the
>> other glColorTable functions too.
>
> I thought I gutted all of the paletted texture support last
> September-ish.  It would be good to see as much of that unused code as
> possible go away.

Looks like tdfx and via/unichrome use it.  It could be disabled in 
both drivers, if we don't outright kill tdfx (and via/unichrome??).

For swrast, the bits for paletted texture are hidden in 
main/texfetch_tmp.h

I have no objection to removing this feature (and those drivers).

-Brian