Re: [PATCH] fortran/50514 -- Fix static chekcing of ISHFT[C] arguments.

[Steve Kargl <sgk at troutmask dot apl dot washington dot edu>]

On Mon, Oct 17, 2011 at 12:22:03PM +0200, Tobias Burnus wrote:
> On 10/15/2011 11:21 PM, Steve Kargl wrote:
> >OK for trunk?
> >
> >2011-10-15  Steven G. Kargl<kargl@gcc.gnu.org>
> >
> >	* gfortran.dg/ishft_3.f90:  Update test.
> 
> I am not so happy with complete test replacements. How about adding it 
> as new test case?
> 

Well, the old testcase is 

%cat ishft_3.f90
! { dg-do compile }
program ishft_3
  integer i, j
  write(*,*) ishftc( 3, 2, 3 )
  write(*,*) ishftc( 3, 2, i )
  write(*,*) ishftc( 3, i, j )
  write(*,*) ishftc( 3, 128 )     ! { dg-error "exceeds BIT_SIZE of first" }
  write(*,*) ishftc( 3, 0, 128 )  ! { dg-error "exceeds BIT_SIZE of first" }
  write(*,*) ishftc( 3, 0, 0 )    ! { dg-error "Invalid third argument" }
  write(*,*) ishftc( 3, 3, 2 )    ! { dg-error "exceeds third argument" }
end program

1) It's misnamed.  There is no testing of ishft.
2) i and j are undefined, so lines 4 and 5 are invalid.
   Even if i and j were defined, there is nothing special
   about those lines in that gfortran generates a function
   call to a runtime routine.  Note, this a "dg-do compile"
   test.
3) The four dg-error strings would need to be updated to
   the new error messages.

The only line that would survive is the first line, which
is covered in the updated testcase.

> >2011-10-15  Steven G. Kargl<kargl@gcc.gnu.org>
> >
> >	* check.c (less_than_bitsize1): Check|shift|  <= bit_size(i).
> >	(gfc_check_ishftc):  Check|shift|  <= bit_size(i) and check
> >	that size is positive.
> 
> I somehow find less_than_bitsize1's
> 
> +      if (strncmp (arg2, "ISHFT", 5) == 0)
> 
> not that elegant and would prefer another argument, which tells the 
> function that it should take the absolute value of the argument; 
> however, given that ISHFT seems to be the only function which allows 
> negative arguments, one could also leave it.
> 

In looking at the other uses of less_than_bitsize1() I
could pass arg2=NULL for ISHFT[C], and then the code
would become

     if (arg2 == NULL) { /* Special case for ISHFT[C].  */

Would that be better?

-- 
Steve