I approve the patch. If no driver expects the callback, it shouldn't be called. The long term goal should be to get rid of all of these driver state callbacks (I've already removed a few dozens of them), but some classic drivers still rely on them and that's why mesa/main has to call them.
Marek On Fri, Jun 2, 2017 at 2:52 PM, Samuel Pitoiset <[email protected]> wrote: > > > On 06/02/2017 02:44 PM, Ilia Mirkin wrote: >> >> On Fri, Jun 2, 2017 at 8:40 AM, Samuel Pitoiset >> <[email protected]> wrote: >>> >>> >>> >>> On 06/02/2017 02:34 PM, Ilia Mirkin wrote: >>>> >>>> >>>> As I mentioned to you on IRC... my understanding of the Driver.Scissor >>>> callback is that it's a guarantee by the mesa core that whenever >>>> Scissor.* changes, Driver.Scissor() gets called if it's there. >>>> >>>> Just because nothing cares on the other side of the abstraction >>>> barrier doesn't implicitly mean it should be removed. >>> >>> >>> >>> Well, the window rectangle fields could be moved outside of the scissor >>> struct state actually. >> >> >> They're part of the "scissor" attrib group (see the spec), which was >> why I originally put it there. There doesn't *have* to be a 1:1 >> mapping, but life is easier when it is. (That's the main reason for >> the existence of those structs in the first place.) > > > Right. Anyway, this patch can be discarded because it's not the most > important one. > > >> >> -ilia >> > _______________________________________________ > mesa-dev mailing list > [email protected] > https://lists.freedesktop.org/mailman/listinfo/mesa-dev _______________________________________________ mesa-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/mesa-dev
