[Bug fortran/19310] unnecessary error for overflowing results

[Bug fortran/19310] unnecessary error for overflowing results

[jvdelisle at gcc dot gnu dot org]

------- Comment #2 from jvdelisle at gcc dot gnu dot org  2006-06-07 03:28 -------
I would like to work on this one.  The range check is only looking for ARITH_OK
when it could also see ARITH_UNDERFLOW or ARITH_OVERFLOW.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] unnecessary error for overflowing results

[kargl at gcc dot gnu dot org]

------- Comment #3 from kargl at gcc dot gnu dot org  2006-06-08 00:05 -------
(In reply to comment #2)
> I would like to work on this one.  The range check is only looking for ARITH_OK
> when it could also see ARITH_UNDERFLOW or ARITH_OVERFLOW.
> 

Jerry,

There is another PR about allowing "real, parameter :: inf = 1. / 0.".
In that PR it was suggested to add a -fallow-inf-nan option.  I think
you should look into this route for this PR.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] unnecessary error for overflowing results

[jvdelisle at gcc dot gnu dot org]

------- Comment #4 from jvdelisle at gcc dot gnu dot org  2006-06-08 01:04 -------
Steve,

I was just about to post an RFC to the list because this PR is requesting a
change of philosophy regarding compile time vs runtime behavior.  Your comment
about -fallow-inf-nan answers my question.  I can see where some users would
want the compile time error, while others may want to use, for example,
exp(1000) to give an infinity.

I think the patch will be straight forward including updating the appropriate
texi.  The -fallow-inf-nan can be applied to range_check in simplify.c and that
will cover most if not all relevant intrinsics.

The default behavior will be to do as is now and give the compile error.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] unnecessary error for overflowing results

[jvdelisle at gcc dot gnu dot org]

-- 

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|2005-12-30 18:59:15         |2006-06-08 01:05:01
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] unnecessary error for overflowing results

[kargl at gcc dot gnu dot org]

------- Comment #5 from kargl at gcc dot gnu dot org  2006-06-08 01:26 -------
(In reply to comment #4)
> 
> I think the patch will be straight forward including updating the appropriate
> texi.  The -fallow-inf-nan can be applied to range_check in simplify.c and that
> will cover most if not all relevant intrinsics.
>

In thinking about it, you might call the option -fno-range-check and
in lang.opt use the RejectNegative key word.  -fallow-inf-nan would
apply to INF and NaN but seems out of place for undeflow.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] unnecessary error for overflowing results

[patchapp at dberlin dot org]

------- Comment #6 from patchapp at dberlin dot org  2006-06-09 05:50 -------
Subject: Bug number PR19310

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/2006-06/msg00435.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] unnecessary error for overflowing results

[jvdelisle at gcc dot gnu dot org]

------- Comment #7 from jvdelisle at gcc dot gnu dot org  2006-06-18 06:36 -------
Subject: Bug 19310

Author: jvdelisle
Date: Sun Jun 18 06:36:45 2006
New Revision: 114752

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=114752
Log:
2006-06-18  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

        PR fortran/19310
        * arith.c (gfc_range_check): Return ARITH_OK if -fno-range-check. Add
        return of ARITH_NAN, ARITH_UNDERFLOW, and ARITH_OVERFLOW.
        (gfc_arith_divide): If -fno-range-check allow mpfr to divide by zero.
        * gfortran.h (gfc_option_t): Add new flag.
        * invoke.texi: Document new flag.
        * lang.opt: Add option -frange-check.
        * options.c (gfc_init_options): Initialize new flag.
        (gfc_handle_options): Set flag if invoked.
        * simplify.c (range_check): Add error messages for
        overflow, underflow, and other errors.
        * trans-const.c (gfc_conv_mpfr_to_tree): Build NaN and Inf from mpfr
        result.

Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/arith.c
    trunk/gcc/fortran/gfortran.h
    trunk/gcc/fortran/invoke.texi
    trunk/gcc/fortran/lang.opt
    trunk/gcc/fortran/options.c
    trunk/gcc/fortran/simplify.c
    trunk/gcc/fortran/trans-const.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] unnecessary error for overflowing results

[jvdelisle at gcc dot gnu dot org]

------- Comment #8 from jvdelisle at gcc dot gnu dot org  2006-06-18 06:42 -------
Subject: Bug 19310

Author: jvdelisle
Date: Sun Jun 18 06:42:26 2006
New Revision: 114753

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=114753
Log:
2006-06-18  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

        PR fortran/19310
        PR fortran/19904
        * gfortran.dg/real_const_3.f90: New test.

Added:
    trunk/gcc/testsuite/gfortran.dg/real_const_3.f90
Modified:
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] [4.1 Only] unnecessary error for overflowing results

[jvdelisle at gcc dot gnu dot org]

------- Comment #9 from jvdelisle at gcc dot gnu dot org  2006-06-18 06:48 -------
Fixed on 4.2


-- 

jvdelisle at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|unnecessary error for       |[4.1 Only] unnecessary error
                   |overflowing results         |for overflowing results


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] [4.1 Only] unnecessary error for overflowing results

[jvdelisle at gcc dot gnu dot org]

------- Comment #10 from jvdelisle at gcc dot gnu dot org  2006-06-28 05:37 -------
Subject: Bug 19310

Author: jvdelisle
Date: Wed Jun 28 05:36:08 2006
New Revision: 115048

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=115048
Log:
2006-06-27  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

        PR fortran/19310
        PR fortran/19904
        * arith.c (gfc_range_check): Return ARITH_OK if -fno-range-check. Add
        return of ARITH_NAN, ARITH_UNDERFLOW, and ARITH_OVERFLOW.
        (gfc_arith_divide): If -fno-range-check allow mpfr to divide by zero.
        * gfortran.h (gfc_option_t): Add new flag.
        * invoke.texi: Document new flag.
        * lang.opt: Add option -frange-check.
        * options.c (gfc_init_options): Initialize new flag.
        (gfc_handle_options): Set flag if invoked.
        * simplify.c (range_check): Add error messages for
        overflow, underflow, and other errors.
        * trans-const.c (gfc_conv_mpfr_to_tree): Build NaN and Inf from mpfr
        result.

Modified:
    branches/gcc-4_1-branch/gcc/fortran/ChangeLog
    branches/gcc-4_1-branch/gcc/fortran/arith.c
    branches/gcc-4_1-branch/gcc/fortran/gfortran.h
    branches/gcc-4_1-branch/gcc/fortran/invoke.texi
    branches/gcc-4_1-branch/gcc/fortran/lang.opt
    branches/gcc-4_1-branch/gcc/fortran/options.c
    branches/gcc-4_1-branch/gcc/fortran/simplify.c
    branches/gcc-4_1-branch/gcc/fortran/trans-const.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] [4.1 Only] unnecessary error for overflowing results

[jvdelisle at gcc dot gnu dot org]

------- Comment #11 from jvdelisle at gcc dot gnu dot org  2006-06-28 05:40 -------
Subject: Bug 19310

Author: jvdelisle
Date: Wed Jun 28 05:39:07 2006
New Revision: 115049

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=115049
Log:
2006-06-27  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

        PR fortran/19310
        PR fortran/19904
        * gfortran.dg/real_const_3.f90: New test.

Added:
    branches/gcc-4_1-branch/gcc/testsuite/gfortran.dg/real_const_3.f90
Modified:
    branches/gcc-4_1-branch/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310

[Bug fortran/19310] [4.1 Only] unnecessary error for overflowing results

[jvdelisle at gcc dot gnu dot org]

------- Comment #12 from jvdelisle at gcc dot gnu dot org  2006-06-28 05:45 -------
Fixed in 4.1 now.


-- 

jvdelisle at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.1.1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19310