https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118874
--- Comment #8 from Iain Sandoe <iains at gcc dot gnu.org> --- the comments in PR117364, lead me to believe that this is a problem down-stream of the FE that happens to be exposed frequently by coroutines (since we need to populate <retval> because of the phasing required of that with the initial suspend. Eric seemed to agree that NVRO could just as easily result in the same circumstance. As of now I am not sure how to proceed... if this is to become a P1 we need to discuss how to meet the constraints.
