[Bug fortran/31201] New: wrong code generated with gfortran

[Bug fortran/31201] New: wrong code generated with gfortran

[jv244 at cam dot ac dot uk]

With a recent gfortran, the following compiles, but generates the wrong
results:

      integer*8  :: k=2_8**36+10
      integer*4  :: j=10
      logical  ex,op
      INQUIRE(unit=k, exist=ex,opened=op)
      IF (ex) THEN
         OPEN(unit=k)
         INQUIRE(unit=j, opened=op)
         IF (op) CALL ABORT()
      ENDIF
      end


-- 
           Summary: wrong code generated with gfortran
           Product: gcc
           Version: 4.3.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: jv244 at cam dot ac dot uk


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[fxcoudert at gcc dot gnu dot org]

------- Comment #1 from fxcoudert at gcc dot gnu dot org  2007-03-16 20:29 -------
Unit numbers are passed and stored internally as GFC_INTEGER_4, but we don't
error out on too large numbers (like the one that don't fit inside a
GFC_INTEGER_4).


-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
           Keywords|                            |wrong-code
      Known to fail|                            |4.1.3 4.2.0 4.3.0
   Last reconfirmed|0000-00-00 00:00:00         |2007-03-16 20:29:34
               date|                            |
            Summary|wrong code generated with   |Too large unit number
                   |gfortran                    |generates wrong code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[jvdelisle at gcc dot gnu dot org]

------- Comment #2 from jvdelisle at gcc dot gnu dot org  2007-03-16 23:01 -------
I will fix this one.


-- 

jvdelisle at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |jvdelisle at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED
   Last reconfirmed|2007-03-16 20:29:34         |2007-03-16 23:01:41
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[jvdelisle at gcc dot gnu dot org]

------- Comment #3 from jvdelisle at gcc dot gnu dot org  2007-03-17 06:28 -------
The front end is converting the integer8 to integer4 and not checking the range
of the number first.  The integer value gets wrapped around and the library
sees unit=10 for this test case.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[fxcoudert at gcc dot gnu dot org]

------- Comment #4 from fxcoudert at gcc dot gnu dot org  2007-03-17 13:01 -------
(In reply to comment #3)
> The front end is converting the integer8 to integer4 and not checking the range
> of the number first.

Every time a unit number is used, there should be a check to be sure it's in
the integer4 range, and otherwise emit an error.

Thanks for taking it, Jerry.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[jvdelisle at gcc dot gnu dot org]

------- Comment #5 from jvdelisle at gcc dot gnu dot org  2007-03-18 02:22 -------
Heres the real problem:

Try compiling this with and without -fno-range-check.

      integer*8  :: k= 68719476746
      integer*8  :: k2 = 2_8**36 + 10
      print *, "k=",k, k2, int(k2, kind=8), int(k2, kind=4)
      end

With checking:

overflow.f:1.34:

      integer*8  :: k= 68719476746                                      
                                 1
Error: Integer too big for its kind at (1)

Without checking:

$ gfc -fno-range-check overflow.f 
$ ./a.out
 k=          68719476746          68719476746          68719476746          10

So our range checking does not catch the case when too big integer is hidden in
some arithmetic and conversion of kinds does not either.

I don't think we want integer kind conversion to catch this, its no different
then a cast.  The programmer needs to be aware of what they are doing.

In tha case of the overly large unit number, We do catch non disguised cases.

Any opinions on whether we need to really 'fix' this?  (I am not taking sides
here, the answer too my question has broader implications, ie int intrinsics)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[jvdelisle at gcc dot gnu dot org]

------- Comment #6 from jvdelisle at gcc dot gnu dot org  2007-04-03 14:50 -------
I have a patch testing


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[patchapp at dberlin dot org]

------- Comment #7 from patchapp at dberlin dot org  2007-04-04 05:55 -------
Subject: Bug number PR31201

A patch for this bug has been added to the patch tracker.
The mailing list url for the patch is
http://gcc.gnu.org/ml/gcc-patches/2007-04/msg00126.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[patchapp at dberlin dot org]

------- Comment #8 from patchapp at dberlin dot org  2007-04-14 19:15 -------
Subject: Bug number PR31201

A patch for this bug has been added to the patch tracker.
The mailing list url for the patch is
http://gcc.gnu.org/ml/gcc-patches/2007-04/msg00863.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[jvdelisle at gcc dot gnu dot org]

------- Comment #9 from jvdelisle at gcc dot gnu dot org  2007-05-06 23:28 -------
Subject: Bug 31201

Author: jvdelisle
Date: Sun May  6 22:28:31 2007
New Revision: 124479

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=124479
Log:
2007-05-06  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

        PR libfortran/31201
        * runtime/error.c (runtime_error_at): New function.
        (generate_error): Export this function.
        * gfortran.map: Add _gfortran_generate_error and
        _gfortran_runtime_error_at.
        * libgfortran.h: Add comment to reference error codes in front end.
        (library_start): Locate prototype with library_end macro and add
        a new comment.  Add prototype for runtime_error_at. Export prototype
for
        generate_error.
        * io/lock.c (library_start): Fix check for error condition.
        * io/transfer.c (data_transfer_init): Add library check.

Modified:
    trunk/libgfortran/ChangeLog
    trunk/libgfortran/gfortran.map
    trunk/libgfortran/io/lock.c
    trunk/libgfortran/io/transfer.c
    trunk/libgfortran/libgfortran.h
    trunk/libgfortran/runtime/error.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[jvdelisle at gcc dot gnu dot org]

------- Comment #10 from jvdelisle at gcc dot gnu dot org  2007-05-06 23:32 -------
Subject: Bug 31201

Author: jvdelisle
Date: Sun May  6 22:32:33 2007
New Revision: 124480

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=124480
Log:
2007-05-06  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
            Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>

        PR fortran/31201
        * gfortran.h: Add runtime error codes from libgfortran.h. Define
        MAX_UNIT_NUMBER.
        * trans.c (gfc_trans_runtime_check): Update the format of runtime error
        messages to match library runtime errors.  Use call to new library
        function runtime_error_at().
        * trans.h: Add prototype for new function gfc_trans_io_runtime_check.
        Add declaration for library functions runtime_error_at and
        generate_error.
        * trans_io.c (gfc_trans_io_runtime_check): New function.
        (set_parameter_value): Add error checking for UNIT numbers.
        (set_parameter_ref): Initialize the users variable to zero. 
        (gfc_trans_open): Move setting of unit number to after setting of
common
        flags so that runtime error trapping can be detected.
        (gfc_trans_close): Likewise. (build_filepos): Likewise.
        (gfc_trans_inquire): Likewise. (build_dt): Likewise.
        * trans-decl.c: Add declarations for runtime_error_at and
        generate_error. (gfc_build_builtin_function_decls): Build function
        declarations for runtime_error_at and generate_error.

Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/gfortran.h
    trunk/gcc/fortran/trans-decl.c
    trunk/gcc/fortran/trans-io.c
    trunk/gcc/fortran/trans.c
    trunk/gcc/fortran/trans.h


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[jvdelisle at gcc dot gnu dot org]

------- Comment #11 from jvdelisle at gcc dot gnu dot org  2007-05-06 23:37 -------
Subject: Bug 31201

Author: jvdelisle
Date: Sun May  6 22:37:27 2007
New Revision: 124481

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=124481
Log:
2007-05-06  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

        PR fortran/31201
        * gfortran.dg/large_unit_1.f90: New.
        * gfortran.dg/large_unit_2.f90: New.
        * gfortran.dg/iostat_4.f90: New
        * gfortran.dg/bounds_check_fail_1.f90: Updated.

Added:
    trunk/gcc/testsuite/gfortran.dg/iostat_4.f90
    trunk/gcc/testsuite/gfortran.dg/large_unit_1.f90
    trunk/gcc/testsuite/gfortran.dg/large_unit_2.f90
Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gfortran.dg/bounds_check_fail_1.f90


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201

[Bug fortran/31201] Too large unit number generates wrong code

[jvdelisle at gcc dot gnu dot org]

------- Comment #12 from jvdelisle at gcc dot gnu dot org  2007-05-07 00:04 -------
Fixed on mainline.


-- 

jvdelisle at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.3.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31201