[Bug fortran/32317] [bounds checking] No warning on bad arguments with explicit interface

["dominiq at lps dot ens.fr" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=32317

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |WAITING

--- Comment #9 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Note that compiling the test

module mod
implicit none

contains

subroutine a(v)
real,dimension(8),intent(in)::v
write(*,*)v
end subroutine a

subroutine b
real,dimension(5)::v
v=0
call a(v(1:4))
end subroutine b

end module mod

gives

pr32317_db_2.f90:14:7:

 call a(v(1:4))
       1

Warning: Actual argument contains too few elements for dummy argument 'v' (4/8)
at (1)

In order to get the same warning with the code in comment 0, the value 8 in

call a(8,v(1:4))

has to be propagated to 'n' in subroutine a(n,v), but the gfortran from-end
does not do constant propagation which is left to the middle-end.

> Can someone with a better understanding of the standard than me comment
> as to whether this is actually legal Fortran when there is an explicit
> interface known?  

I certainly won't claim that I have "a better understanding of the standard",
but I am pretty sure that the for the example the burden is on the user and not
on the compiler: AFAIK the standard does not say that constants should be
propagated.

In addition AFAIU -fcheck=bounds, it is pretty naive, i.e., it checks that v(i)
in subroutine a is not accessed with i<1 or i>n.

Considering that this PR is (IMO) based on wrong expectations and did not have
any activity for more than six years, I'll close it as WONTFIX without (prompt)
feedback.