> OK, so you think the test above should check > && (!TYPE_NONALIASED_COMPONENT (t1) || TYPE_STRUCTURAL_EQUALITY_P (t1)) > && (!TYPE_NONALIASED_COMPONENT (t2) || TYPE_STRUCTURAL_EQUALITY_P (t2))
I'm not sure about TYPE_STRUCTURAL_EQUALITY_P (does it matter in LTO mode?) but, yes, TYPE_NONALIASED_COMPONENT should be checked I think. > Do we have wrong code issue with lto20.adb? No, it's only about the warning. As a matter of fact, I'm ready to clear TYPE_NONALIASED_COMPONENT in the Ada FE for this particular case (array of pointers) so that gnat.dg/lto20.adb still passes after the above change. -- Eric Botcazou
