https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88248
kargl at gcc dot gnu.org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |janus at gcc dot gnu.org,
| |kargl at gcc dot gnu.org
--- Comment #4 from kargl at gcc dot gnu.org ---
(In reply to G. Steinmetz from comment #0)
> Branching out via END=, ERR= or EOR= specifier in combination with
> CONTINUE is interpreted as a labeled DO loop. Option -std=f2018
> misleadingly flags this with a warning -- (low prio).
>
>
> $ cat z1.f90
> program p
> do
> rewind (1, err=99)
> end do
> 99 continue
> end
>
>
> $ cat z2.f90
> program p
> character(3) :: c = 'abc'
> integer :: k
> do k = 1, 10
> open (k, err=99)
> close (k, err=99)
> backspace (k, err=99)
> endfile (k, err=99)
> rewind (k, err=99)
> flush (k, err=99)
> inquire (k, err=99)
> read (k, '(a)', end=97) c
> read (k, '(a)', eor=98, advance='no') c
> read (k, '(a)', err=99) c
> write (k, '(a)', err=99) c
> wait (k, end=97)
> wait (k, eor=98)
> wait (k, err=99)
> end do
> 97 continue
> 98 continue
> 99 continue
> end
>
>
> $ gfortran-9-20181125 -c z1.f90 -std=f2008
> $
> $ gfortran-9-20181125 -c z1.f90 -std=f2018
> z1.f90:5:2:
>
> 5 | 99 continue
> | 1
> Warning: Fortran 2018 obsolescent feature: Labeled DO statement at (1)
The warning was added by Janus in r260705. I'll need to look
deeper, but I don't think there is an easy way to distinguish
between a labeled statement used as do-loop terminator and
a ordinary labeled statement.
