https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93224
--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> --- (In reply to Martin Liška from comment #2) > So a negative zero is reached ;) No it isn't, the value is non-zero: a.load() - 208.0l: -0.000000000000000000000000000000640949485492072091897029459292263811609480228526081191375851631164551 The test just makes assumptions about the FP math which isn't true for power64 long double.
