https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101118
--- Comment #11 from Iain Sandoe <iains at gcc dot gnu.org> --- (In reply to Iain Sandoe from comment #10) > Hmm... maybe I am being too hasty here. > > If the coroutine has a definition in a header, then the coroutine frame type > _should_ be the same for each instance of it. So maybe this is actually > reporting a genuine ODR violation? ah (I still misread) ... so the reported error is that the frame type *is* the same (which is correct) but the field names differ (which is probably a consequence of using static counters to produce them). So we need a stable way to generate the field names.
