[Mesa-dev] [PATCH v2 01/12] mesa: glGet: fix indentation of _mesa_init_get_hash

[Imre Deak]

On Tue, 2012-09-11 at 17:10 -0600, Brian Paul wrote:
> On 09/11/2012 03:55 PM, Imre Deak wrote:
> > On Tue, 2012-09-11 at 09:30 -0600, Brian Paul wrote:
> >> On 09/10/2012 12:41 AM, Imre Deak wrote:
> >>> No functional change.
> >>>
> >>> Signed-off-by: Imre Deak<imre.deak at intel.com>
> >>> ---
> >>>    src/mesa/main/get.c |   18 +++++++++---------
> >>>    1 file changed, 9 insertions(+), 9 deletions(-)
> >>>
> >>> diff --git a/src/mesa/main/get.c b/src/mesa/main/get.c
> >>> index 24d2a18..57457ef 100644
> >>> --- a/src/mesa/main/get.c
> >>> +++ b/src/mesa/main/get.c
> >>> @@ -1442,20 +1442,20 @@ void _mesa_init_get_hash(struct gl_context *ctx)
> >>>
> >>>       for (i = 0; i<   Elements(values); i++) {
> >>>          if (values[i].type == TYPE_API_MASK) {
> >>> -	 api_mask = values[i].offset;
> >>> -	 continue;
> >>> +         api_mask = values[i].offset;
> >>> +         continue;
> >>>          }
> >>>          if (!(api_mask&   api_bit))
> >>> -	 continue;
> >>> +         continue;
> >>>
> >>>          hash = (values[i].pname * prime_factor)&   mask;
> >>>          while (1) {
> >>> -	 index = hash&   mask;
> >>> -	 if (!table[index]) {
> >>> -	    table[index] = i;
> >>> -	    break;
> >>> -	 }
> >>> -	 hash += prime_step;
> >>> +         index = hash&   mask;
> >>> +         if (!table[index]) {
> >>> +            table[index] = i;
> >>> +            break;
> >>> +         }
> >>> +         hash += prime_step;
> >>>          }
> >>>       }
> >>>
> >>
> >> Reviewed-by: Brian Paul<brianp at vmware.com>
> >>
> >> Do you need someone to commit these patches for you?
> >
> > Yes, please.
> >
> > I don't know who have commit right or what's the workflow in general, so
> > I just posted these in the hope that someone will commit them
> > eventually.
> 
> OK, I'll push 1-4 tonight or tomorrow AM.

Thanks.

> 
> 
> > Patches 01-04 are trivial and stand on their own, so they could be
> > committed right away.
> >
> > 05-06 containing the actual fix is also independent of the subsequent
> > ones, but I will post a new version with the comments addressed.
> >
> > 07-12 is moving the currently run-time hash generation to build time and
> > hasn't received any reviewed-by lines, though Ian looked through at
> > least some of them already.
> 
> I'd like to have a second set of eyes look that over too.  

Ok, I've now made the changes you asked and talked to Ian about the
rest. He's ok with merging it to master and 9.0. Do you want me to post
a new version of it to the list? Otherwise I'd just add your Reviewed-by
lines and send a pull request to Ian.

> It's funny that we're moving to a python-generated get.c.  That's the way I had 
> it a few years ago.

I would have probably do it in C, but as I understood cross-compilation
is not fully supported. At least in the sense that you can't build stuff
for the build environment. The CC_FOR_BUILD autoconf thing would solve
this, but the attempt to add support for it to Mesa wasn't successful
([1]), so I thought I'll try another path.

One advantage though of using a script is that we can catch cases where
we add an enum to the C headers (gl.h, glext.h) but not to the XML spec.

--Imre

[1]
http://www.mail-archive.com/mesa-dev@lists.freedesktop.org/msg16878.html