https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88074
Jakub Jelinek <jakub at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |burnus at gcc dot gnu.org, | |jakub at gcc dot gnu.org --- Comment #26 from Jakub Jelinek <jakub at gcc dot gnu.org> --- Seems to me that the norm2_3.f90 testcase expects infinite precision, I'd say it is broken. Because norm2 ([real(qp) :: 1, 2, huge(3.0_qp)]) needs to be larger than huge(3.0_qp) and as huge(3.0_qp) should be the largest representable finite number in that kind, that norm2 should be really infinity and infinity minus huge(3.0_qp) is still infinity. So, either we acknowledge Fortran wants infinite precision and just set emin/emax temporarily around mpfr/mpc etc. calls in the middle-end temporarily, or we remove the questionable testcase parts.