> I didn't yet relax some of its restrictions (we can safely > skip real/imag-part and array[-range?]-refs - we don't have > to stop at them. There is also the question whether > aliasing_component_refs_p still does sth useful after this > (if those restrictions are removed). > > Bootstrap / regtest pending on x86_64-unknown-linux-gnu. > > Eric, does this look reasonable?
Yes, although I wonder whether this machinery is still that useful these days, given that we already have aliasing_component_refs_p (for the indirect/direct case) and nonoverlapping_component_refs_of_decl_p (for the direct/direct case) but, on the other hand, your testcases don't look too far-fetched. We'll have 2 nonoverlapping_component_refs[_of_decls]_p predicates with a very similar implementation, I guess unifying them would be the next logical step. -- Eric Botcazou
