https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86281
--- Comment #7 from Steve Kargl <sgk at troutmask dot apl.washington.edu> --- On Fri, Jun 22, 2018 at 10:36:04PM +0000, pault at gcc dot gnu.org wrote: > > I don't know how this got past my regtesting... > > The regression on assumed_charlen_function_7.f90 was entirely my > fault. I found the failure, corrected on my tree after saving > the diff and then cleaned the tree for another commit. > Result - the legacy option was lost. Thanks for the quick fix. I know the joy of juggling patches for several different bugs leading to this type of problem.