https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78746
--- Comment #3 from kargl at gcc dot gnu.org --- The issue with charlen_15.f90 was fixed here: https://gcc.gnu.org/ml/gcc-bugs/2016-12/msg01451.html The issues with both charlen_03.f90 and charlen_10.f90 have nothing at all to do with a charlen. The problem is referencing an uninitialized array bound. For *_03.f90 we have program p type t character(1), allocatable :: x(n) end type end The variable 'n' is used, but not defined. The backtrace is 0x87bac2a crash_signal /mnt/kargl/gcc/gcc/toplev.c:333 0x826ec1b resolve_variable /mnt/kargl/gcc/gcc/fortran/resolve.c:5048 0x826ec1b gfc_resolve_expr(gfc_expr*) /mnt/kargl/gcc/gcc/fortran/resolve.c:6492 0x81d7152 resolve_array_bound /mnt/kargl/gcc/gcc/fortran/array.c:327 0x81d7c6d gfc_resolve_array_spec(gfc_array_spec*, int) /mnt/kargl/gcc/gcc/fortran/array.c:368 0x826c9f0 resolve_component /mnt/kargl/gcc/gcc/fortran/resolve.c:13629 0x826d378 resolve_fl_derived0 /mnt/kargl/gcc/gcc/fortran/resolve.c:13746 0x826d79e resolve_fl_derived0 /mnt/kargl/gcc/gcc/fortran/resolve.c:13831 0x826d79e resolve_fl_derived /mnt/kargl/gcc/gcc/fortran/resolve.c:13823 0x826a226 resolve_symbol /mnt/kargl/gcc/gcc/fortran/resolve.c:14146 0x82898dd do_traverse_symtree /mnt/kargl/gcc/gcc/fortran/symbol.c:3986 0x8264f0d resolve_types /mnt/kargl/gcc/gcc/fortran/resolve.c:15999 0x826969e gfc_resolve(gfc_namespace*) /mnt/kargl/gcc/gcc/fortran/resolve.c:16112 0x825860b resolve_all_program_units /mnt/kargl/gcc/gcc/fortran/parse.c:5947 0x825860b gfc_parse_file() /mnt/kargl/gcc/gcc/fortran/parse.c:6194 0x829bc12 gfc_be_parse_file /mnt/kargl/gcc/gcc/fortran/f95-lang.c:202