This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [tree-ssa] Use of build_binary_op


On Thu, Sep 26, 2002 at 12:08:07PM +0100, Jason Merrill wrote:
> On Thu, 26 Sep 2002 12:43:00 +0200, Pop Sébastian <pop@gauvain.u-strasbg.fr> wrote:
> 
> > On Thu, Sep 26, 2002 at 11:09:28AM +0100, Jason Merrill wrote:
> >> On Thu, 26 Sep 2002 11:57:52 +0200, Pop Sébastian <pop@gauvain.u-strasbg.fr> wrote:
> >> 
> >> > you could just say that GENERIC and GIMPLE trees are of different types.
> >> 
> >> I don't think that's a distinction we want to describe in terms of type.
> 
> > GIMPLE is a subset of GENERIC.  So GIMPLE has all the features GENERIC has,
> > plus some other that are either redefined or added in extension.
> 
> > GIMPLE can be defined as a subtype of GENERIC.
> 
> It could, but I don't think it would be useful.  Transformations on a
> GIMPLE tree can produce something which does not satisfy the GIMPLE
> constraints, and needs to be simplified again.  The inliner, for example.
> 
I have not yet looked at the inliner code, but I have a question:
does the inliner work on the simplified version of GENERIC or just 
on GENERIC trees?

Even if it works on GIMPLE you could generate a GENERIC tree and then
refine to GIMPLE.

> Jason


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]