- Original Message -
> >
> > That is, lp_bld_arit.c functions such as lp_build_add() are be
> > confident that passed the values and types are consistent, and all
> > assertions to test that should be preserved.
> >
> > I hope this makes sense.
>
> Thanks Jose, this seems sensible and I'
>
> That is, lp_bld_arit.c functions such as lp_build_add() are be confident that
> passed the values and types are consistent, and all assertions to test that
> should be preserved.
>
> I hope this makes sense.
Thanks Jose, this seems sensible and I'll try and go in that
direction, some issues
Hi Dave,
Thanks for looking at this.
The gist is there, but I'd prefer that the ambiguity of "ints encoded as
floats" stayed contained inside lp_bld_tgsi_* and never bleeds into
lp_bld_arith or other generic LLVM emiting helper functions. That is, when
lp_bld_arith and other functions are cal
Hi,
So I've been playing a bit more with integers in gallivm, but I'm not
100% sure how the original design though about incorporating them.
The lp_build_contexts seems to take a base type of a float and have
some sort of int types associated it with it, should I add more to
this struct for uints
