[PATCH] Rewrite store-motion and invariant motion to use an alias oracle

[Richard Guenther]

On Tue, 25 Mar 2008, Zdenek Dvorak wrote:

> Hi,
> 
> > In its current form it depends on the side-information compiled by
> > the store sinking code, so I think not.
> 
> it does not seem to depend on any such information; perhaps you had some
> other function in mind?

Ah no, but I thought the ttae_cache was populated elsewhere.  I suppose
we can try to consolidate the various (slightly different) implementations
of may alias predicates.

I'll promise to have a look as a followup.

Thanks,
Richard.