https://gcc.gnu.org/bugzilla/show_bug.cgi?id=54753
--- Comment #5 from sandra at gcc dot gnu.org --- Patch to add the diagnostic posted here: https://gcc.gnu.org/pipermail/fortran/2021-October/056656.html There's still a problem with the bogus diagnostic arising from deallocation/initialization when you have just a subroutine definition instead of an interface. That has nothing to do with this diagnostic or other caller/callee argument mismatch so I opened a new issue (PR102641) for that.
