https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125391
--- Comment #10 from Mikael Morin <mikael at gcc dot gnu.org> --- (In reply to Mikael Morin from comment #8) > > The third one is that we evaluate the lhs a second time for finalization, > which can be problematic in case of side-effects. It should be possible to reuse what is available in gfc_trans_assignment_1 to get the lhs without reevaluating. In array context, the lhs descriptor is in lss->info->data.array.descriptor. In scalar context, the lhs is lse->expr.
