> > >and thus we will hapilly merge types with different TYPE_ALIGN. > > >Should func_checker::types_compatible be extended to compare these? > > >Clearly TYPE_ALIGN matters for vectorizer and other plaes... > > > > But it matters on MEM_REFs only, and you can't use canonical types for that. > > And we don't. Tree merging correctly takes TYPE_ALING into account. > > Yep, I wonder about ICF that will match the different MEM_REFs based on > useless > type conversion assumptions. In fact here is a wrong code testcase:
Now tracked as PR65270 Honza
