https://gcc.gnu.org/bugzilla/show_bug.cgi?id=37643
Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED CC| |fxcoudert at gcc dot gnu.org Resolution|--- |WONTFIX --- Comment #1 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> --- I don't think anyone on the gfortran team has access to a VAX machine to build, and so much has changed in the compiler and library since 2008 that I am going to close this. If someone can reproduce on modern versions, reopen and I will have a look. The middle-end seems to be guilty for emitting QNaN in the first place.