[Bug fortran/31820] New: Warning if case label value exceeds maximum value for type

[Bug fortran/31820] New: Warning if case label value exceeds maximum value for type

[burnus at gcc dot gnu dot org]

integer(kind=1) i          ! kind = 1, -128 <= i < 127
    select case (i)
    case (200)         ! kind = 4, unreachable because of range of i

Taken from gfortran.dg/select_5.f90.

This is valid, but one should print a warning. One may also print an error
since having such a case label does not make sense.

NAG f95:  warning: case label value exceeds maximum value for type

g95: Error: Integer overflow in CASE statement at (1)

ifort:  Error: select_5.f90, line 11: The case-value is out-of-range.


-- 
           Summary: Warning if case label value exceeds maximum value for
                    type
           Product: gcc
           Version: 4.3.0
            Status: UNCONFIRMED
          Keywords: diagnostic
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: burnus at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820

[Bug fortran/31820] Warning if case label value exceeds maximum value for type

[dfranke at gcc dot gnu dot org]

------- Comment #1 from dfranke at gcc dot gnu dot org  2007-06-21 17:49 -------
http://gcc.gnu.org/onlinedocs/gfortran/Error-and-Warning-Options.html

-Wsurprising
[...] This currently produces a warning under the following circumstances: 
  * An INTEGER SELECT construct has a CASE that can never be matched as its
    lower value is greater than its upper value. 
  * A LOGICAL SELECT construct has three CASE statements.

The latter is implemented and gives:
    Warning: Logical SELECT CASE block at (1) has more that two cases

but the former?


-- 

dfranke at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dfranke at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820

[Bug fortran/31820] Warning if case label value exceeds maximum value for type

[dfranke at gcc dot gnu dot org]

------- Comment #2 from dfranke at gcc dot gnu dot org  2007-06-21 19:05 -------
>  * An INTEGER SELECT construct has a CASE that can never be matched as its
>    lower value is greater than its upper value. 

In these cases, no error is shown (integer(kind=1) :: i):
select case (i)
  case (300)
  case (301:399)
  case (400)
end select

But:
  select case (i)
    case (400:300)
  end select

results in:

  case (400:300)
       1
Error: Arithmetic overflow converting INTEGER(4) to INTEGER(1) at (1)

(and an "Internal Error" as a follow-up).


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820

[Bug fortran/31820] Warning if case label value exceeds maximum value for type

[fxcoudert at gcc dot gnu dot org]

-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |fxcoudert at gcc dot gnu dot
                   |dot org                     |org
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2007-08-08 17:10:37
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820

[Bug fortran/31820] Warning if case label value exceeds maximum value for type

[fxcoudert at gcc dot gnu dot org]

------- Comment #3 from fxcoudert at gcc dot gnu dot org  2007-08-15 14:38 -------
(In reply to comment #1)
>   * An INTEGER SELECT construct has a CASE that can never be matched as its
>     lower value is greater than its upper value. 

It is also implemented:

$ cat u1.f90
  integer :: i, j
  select case (i)
    case (251:250)
  end select
end
$ gfortran u1.f90 -Wsurprising
u1.f90:3.10:

    case (251:250)
         1
Warning: Range specification at (1) can never be matched


When an error occurs, though, we will not display the warning, which is fine
IMHO. We still need to add a check for the initial testcase.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820

[Bug fortran/31820] Warning if case label value exceeds maximum value for type

[tobi at gcc dot gnu dot org]

------- Comment #4 from tobi at gcc dot gnu dot org  2007-10-04 21:58 -------
It should be sufficient to convert all case-selectors to the same kind as the
case-expression.  We should then be giving exactly the same error as g95,
unless I'm missing something.


-- 

tobi at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tobi at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820

[Bug fortran/31820] Warning if case label value exceeds maximum value for type

[fxcoudert at gcc dot gnu dot org]

-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|fxcoudert at gcc dot gnu dot|unassigned at gcc dot gnu
                   |org                         |dot org
             Status|ASSIGNED                    |NEW


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820

[Bug fortran/31820] Warning if case label value exceeds maximum value for type

[dfranke at gcc dot gnu dot org]

------- Comment #5 from dfranke at gcc dot gnu dot org  2010-05-09 09:11 -------
(In reply to comment #4)
> It should be sufficient to convert all case-selectors to the same kind as the
> case-expression.  We should then be giving exactly the same error as g95,
> unless I'm missing something.

This is directly contradicting the conclusion reached in PR19168 where it was
decided that the case-expr should be promoted to the largest kind of the
case-values (short summary: case values shall be of the same type, but only
character expressions are required to be of same kind, hence apply the same
conversion rules as for direct comparison; see F2003, C805 (R808); F2008, C832
(R838)).

In the example below, the comparison done for the IF statement is promoted to
default integer kind (see dump), no warning is generated for the unreachable
branch. Same holds for the equivalent SELECT. 

  INTEGER(KIND=1) i
  i = 100

  IF (i == 200) THEN
      print *, "unreachable!"
  ELSE
      print *, "default!"
  END IF

  SELECT CASE (i)
    CASE (200)
      print *, "unreachable!"

    CASE DEFAULT
      print *, "default!"
  END SELECT
END

Nonetheless, I also think that it is counter-intuitive. The least that could be
done: add a warning (-Wconversion? -Wsurprising?) before converting the
case-expr, mentioning that at least one case-value has a kind larger than the
case-expressions kind, which thus can not be reached.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820

[Bug fortran/31820] Warning if case label value exceeds maximum value for type

[dfranke at gcc dot gnu dot org]

------- Comment #6 from dfranke at gcc dot gnu dot org  2010-05-09 12:59 -------
(In reply to comment #5)
> Nonetheless, I also think that it is counter-intuitive. The least that could be
> done: add a warning (-Wconversion? -Wsurprising?) before converting the
> case-expr, mentioning that at least one case-value has a kind larger than the
> case-expressions kind, which thus can not be reached.

Suggested patch:
    http://gcc.gnu.org/ml/fortran/2010-05/msg00084.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820

[Bug fortran/31820] Warning if case label value exceeds maximum value for type

[dfranke at gcc dot gnu dot org]

------- Comment #7 from dfranke at gcc dot gnu dot org  2010-05-11 15:43 -------
Subject: Bug 31820

Author: dfranke
Date: Tue May 11 15:43:16 2010
New Revision: 159278

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=159278
Log:
gcc/fortran/:
2010-05-11  Daniel Franke  <franke.daniel@gmail.com>

        PR fortran/31820
        * resolve.c (validate_case_label_expr): Removed FIXME.
        (resolve_select): Raise default warning on case labels out of range
        of the case expression.

gcc/testsuite/:
2010-05-11  Daniel Franke  <franke.daniel@gmail.com>

        PR fortran/31820
        * gfortran.dg/select_5.f90: Updated.


Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/resolve.c
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gfortran.dg/select_5.f90


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820

[Bug fortran/31820] Warning if case label value exceeds maximum value for type

[dfranke at gcc dot gnu dot org]

------- Comment #8 from dfranke at gcc dot gnu dot org  2010-05-11 15:46 -------
Fixed in trunk. Closing.


-- 

dfranke at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.6.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31820