https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122058
--- Comment #36 from Jan Hubicka <hubicka at gcc dot gnu.org> --- One interesting thing is that we redirect to two calls, but both are to the same function (_ZThn8_NK3icu6number4impl15LongNameHandler11getModifierENS1_6SignumENS_14StandardPlural4FormE) which demangles to non-virtual thunk to icu::number::impl::LongNameHandler::getModifier(icu::number::impl::Signum, icu::StandardPlural::Form) const This must be due to inconsistency where we rewrite some of the virtual tables to local alias but not others. By itself this should be harmless - we will just produce more compares. It would be nice to understand why this happens.
