[Mesa-dev] [PATCH] r600g: fix blend setting with multiple render targets all with the same blend

[Jerome Glisse]

On Thu, Mar 31, 2011 at 12:37 PM, Julian Adams <joolsa at gmail.com> wrote:
> inline...
>
> On 31 March 2011 12:07, Henri Verbeet <hverbeet at gmail.com> wrote:
>> On 30 March 2011 23:43, Julian Adams <joolsa at gmail.com> wrote:
>>> ---
>>>  src/gallium/drivers/r600/r600_state.c |   22 +++++++++++++---------
>>>  1 files changed, 13 insertions(+), 9 deletions(-)
>>>
>> Evergreen probably needs the same fix.
>
> Yes, the matching function there looks similar. I can update it, but
> can't test it.
>
>>
>>> +       for (int i = 0, j = 0; i < 8; i++) {
>>> +               /* state->rt entries > 0 only written if independent blending */
>>> +               if (state->independent_blend_enable)
>>> +                       j = i;
>> Not sure about declaring "j" inside the for. I don't think it exactly
>> helps readability, but otoh it doesn't bother me a lot either.
>>
>
> I copied that style from a function up the call stack, which was fixed
> in a similar way in commit: a476ca1fd1b4e76e31c9babfd7fb2a54a09f21d3.
> Mostly I write C++ so probably I'd be inclined to do this:
>
>        for (int i = 0; i < 8; i++) {
>                /* state->rt entries > 0 only written if independent blending */
>                const int j = state->independent_blend_enable ? i : 0;
>
> Let me know if you want the Evergreen fixes, and how "j" is declared and set.


Using for(int loopcountuer;;) construct is a good thing it helps the
compiler to make good decision but gcc is already good on its own to
figure this kind of things by itself.

Cheers,
Jerome