[Mesa-dev] [PATCH 5/8] st/mesa: pass the configuration options to st_init_extensions

[Vadim Girlin]

On Tue, 2012-04-10 at 09:53 +0200, Michel Dänzer wrote:
> On Mon, 2012-04-09 at 19:32 +0400, Vadim Girlin wrote: 
> > Signed-off-by: Vadim Girlin <vadimgirlin at gmail.com>
> > ---
> >  src/mesa/state_tracker/st_context.c    |    9 +++++----
> >  src/mesa/state_tracker/st_context.h    |    3 ++-
> >  src/mesa/state_tracker/st_extensions.c |    2 +-
> >  src/mesa/state_tracker/st_extensions.h |    2 +-
> >  src/mesa/state_tracker/st_manager.c    |    2 +-
> >  5 files changed, 10 insertions(+), 8 deletions(-)
> > 
> > diff --git a/src/mesa/state_tracker/st_context.c b/src/mesa/state_tracker/st_context.c
> > index a3fd4db..1e76501 100644
> > --- a/src/mesa/state_tracker/st_context.c
> > +++ b/src/mesa/state_tracker/st_context.c
> > @@ -111,7 +111,7 @@ st_get_msaa(void)
> >  
> > 
> >  static struct st_context *
> > -st_create_context_priv( struct gl_context *ctx, struct pipe_context *pipe )
> > +st_create_context_priv( struct gl_context *ctx, struct pipe_context *pipe, const driOptionCache * optionCache )
> 
> This adds DRI specific code to the mesa state tracker, which is also
> used for non-DRI targets. I'm afraid there would have to be some kind of
> abstraction for this.

Probably we could pass a single boolean value for now
(force_glsl_extensions_warn), maybe in some struct. Does it make sense?

I guess it's better to move the configuration to mesa completely, as
proposed by Kenneth Graunke, but it will require more time for me to do
it.

Vadim