[Bug fortran/40019] New: LEADZ and TRAILZ give wrong results

[Bug fortran/40019] New: LEADZ and TRAILZ give wrong results

[fxcoudert at gcc dot gnu dot org]

LEADZ and TRAILZ can give wrong results for kinds other than 4 and 8. For
example, the following code:

  integer(kind=1) :: i1
  integer(kind=2) :: i2
  integer(kind=4) :: i4
  integer(kind=8) :: i8
  integer(kind=16) :: i16

  i1 = -1
  i2 = -1
  i4 = -1
  i8 = -1
  i16 = -1
  print *, leadz(i1), leadz(i2), leadz(i4), leadz(i8), leadz(i16)

gives "        -24         -16           0           0          64" as results,
while it should only yield zeros!

There are a few reasons:

  - for kinds 1 and 2, we should not cast directly to (unsigned int), but do a
double cast: (unsigned int)(unsigned char) for kind=1, for example; otherwise,
negative values are screwed.

  - trans-intrinsic.c uses a correspondance, kinds == 1, 2 and 4 use
BUILT_IN_CLZ, kind 8 uses BUILT_IN_CLZL and kind 16 uses BUILT_IN_CLZLL; in
reality, there is no such correspondance, as kind==16 is larger than long long!


This, of course, makes it harder, because there is no function ready for use in
kind==16.


-- 
           Summary: LEADZ and TRAILZ give wrong results
           Product: gcc
           Version: 4.5.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: fxcoudert at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40019

[Bug fortran/40019] LEADZ and TRAILZ give wrong results

[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
                URL|                            |http://gcc.gnu.org/ml/gcc-
                   |                            |patches/2009-
                   |                            |05/msg01267.html
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
           Keywords|                            |patch
   Last reconfirmed|0000-00-00 00:00:00         |2009-05-19 22:24:12
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40019

[Bug fortran/40019] LEADZ and TRAILZ give wrong results

[burnus at gcc dot gnu dot org]

------- Comment #1 from burnus at gcc dot gnu dot org  2009-05-27 07:44 -------
I don't think your patch fixes the following,

  print *, leadz(-1_1), leadz(-1_2), leadz(-1_4), leadz(-1_8), leadz(-1_16)

which yields

           7          15          31          63         127


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40019

[Bug fortran/40019] LEADZ and TRAILZ give wrong results

[fxcoudert at gcc dot gnu dot org]

------- Comment #2 from fxcoudert at gcc dot gnu dot org  2009-05-27 10:27 -------
(In reply to comment #1)
> I don't think your patch fixes the following,
> 
>   print *, leadz(-1_1), leadz(-1_2), leadz(-1_4), leadz(-1_8), leadz(-1_16)
> 
> which yields
> 
>            7          15          31          63         127

I think it does:

$ cat h.f90 
  print *, leadz(-1_1), leadz(-1_2), leadz(-1_4), leadz(-1_8), leadz(-1_16)
  end
$ gfortran h.f90 && ./a.out 
           0           0           0           0           0


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40019

[Bug fortran/40019] LEADZ and TRAILZ give wrong results

[burnus at gcc dot gnu dot org]

------- Comment #3 from burnus at gcc dot gnu dot org  2009-05-29 21:28 -------
Subject: Bug 40019

Author: burnus
Date: Fri May 29 21:27:54 2009
New Revision: 147987

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=147987
Log:
2009-05-29  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>

        PR fortran/40019
        * trans-types.c (gfc_build_uint_type): Make nonstatic.
        * trans.h (gfor_fndecl_clz128, gfor_fndecl_ctz128): New
        * prototypes.
        * trans-types.h (gfc_build_uint_type): Add prototype.
        * trans-decl.c (gfc_build_intrinsic_function_decls): Build
        gfor_fndecl_clz128 and gfor_fndecl_ctz128.
        * trans-intrinsic.c (gfc_conv_intrinsic_leadz,
        gfc_conv_intrinsic_trailz): Call the right builtins or library
        functions, and cast arguments to unsigned types first.
        * simplify.c (gfc_simplify_leadz): Deal with negative arguments.

2009-05-29  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>

        PR fortran/40019
        * intrinsics/bit_intrinsics.c: New file.
        * gfortran.map (GFORTRAN_1.2): New list.
        * Makefile.am: Add intrinsics/bit_intrinsics.c.
        * Makefile.in: Regenerate.

2009-05-29  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>

        PR fortran/40019
        * gfortran.dg/leadz_trailz_1.f90: New test.
        * gfortran.dg/leadz_trailz_2.f90: New test.


Added:
    trunk/gcc/testsuite/gfortran.dg/leadz_trailz_1.f90
    trunk/gcc/testsuite/gfortran.dg/leadz_trailz_2.f90
    trunk/libgfortran/intrinsics/bit_intrinsics.c
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/simplify.c
    trunk/gcc/fortran/trans-decl.c
    trunk/gcc/fortran/trans-intrinsic.c
    trunk/gcc/fortran/trans-types.c
    trunk/gcc/fortran/trans-types.h
    trunk/gcc/fortran/trans.h
    trunk/gcc/testsuite/ChangeLog
    trunk/libgfortran/ChangeLog
    trunk/libgfortran/Makefile.am
    trunk/libgfortran/Makefile.in
    trunk/libgfortran/gfortran.map


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40019

[Bug fortran/40019] LEADZ and TRAILZ give wrong results

[burnus at gcc dot gnu dot org]

------- Comment #4 from burnus at gcc dot gnu dot org  2009-05-29 21:33 -------
Checked into the trunk (4.5). Keep open for potential back porting to 4.4.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40019

[Bug fortran/40019] [4.4 only] LEADZ and TRAILZ give wrong results

[burnus at gcc dot gnu dot org]

------- Comment #5 from burnus at gcc dot gnu dot org  2009-06-03 19:39 -------
Subject: Bug 40019

Author: burnus
Date: Wed Jun  3 19:39:09 2009
New Revision: 148143

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=148143
Log:
2009-06-03  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>

        PR fortran/40019
        * trans-types.c (gfc_build_uint_type): Make nonstatic.
        * trans.h (gfor_fndecl_clz128, gfor_fndecl_ctz128): New
        * prototypes.
        * trans-types.h (gfc_build_uint_type): Add prototype.
        * trans-decl.c (gfc_build_intrinsic_function_decls): Build
        gfor_fndecl_clz128 and gfor_fndecl_ctz128.
        * trans-intrinsic.c (gfc_conv_intrinsic_leadz,
        gfc_conv_intrinsic_trailz): Call the right builtins or library
        functions, and cast arguments to unsigned types first.
        * simplify.c (gfc_simplify_leadz): Deal with negative arguments.


2009-06-03  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>

        PR fortran/40019
        * intrinsics/bit_intrinsics.c: New file.
        * gfortran.map (GFORTRAN_1.2): New list.
        * Makefile.am: Add intrinsics/bit_intrinsics.c.
        * Makefile.in: Regenerate.


2009-06-03  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>

        PR fortran/40019
        * gfortran.dg/leadz_trailz_1.f90: New test.
        * gfortran.dg/leadz_trailz_2.f90: New test.

-- Diese und die folgenden Zeilen werden ignoriert --

M    libgfortran/Makefile.in
A    libgfortran/intrinsics/bit_intrinsics.c
M    libgfortran/ChangeLog
M    libgfortran/gfortran.map
M    libgfortran/Makefile.am
M    gcc/testsuite/ChangeLog
A    gcc/testsuite/gfortran.dg/leadz_trailz_2.f90
A    gcc/testsuite/gfortran.dg/leadz_trailz_1.f90
M    gcc/fortran/ChangeLog
M    gcc/fortran/trans-types.c
M    gcc/fortran/trans.h
M    gcc/fortran/trans-types.h
M    gcc/fortran/trans-decl.c
M    gcc/fortran/trans-intrinsic.c
M    gcc/fortran/simplify.c

Added:
    branches/gcc-4_4-branch/gcc/testsuite/gfortran.dg/leadz_trailz_1.f90
    branches/gcc-4_4-branch/gcc/testsuite/gfortran.dg/leadz_trailz_2.f90
    branches/gcc-4_4-branch/libgfortran/intrinsics/bit_intrinsics.c
Modified:
    branches/gcc-4_4-branch/gcc/fortran/ChangeLog
    branches/gcc-4_4-branch/gcc/fortran/simplify.c
    branches/gcc-4_4-branch/gcc/fortran/trans-decl.c
    branches/gcc-4_4-branch/gcc/fortran/trans-intrinsic.c
    branches/gcc-4_4-branch/gcc/fortran/trans-types.c
    branches/gcc-4_4-branch/gcc/fortran/trans-types.h
    branches/gcc-4_4-branch/gcc/fortran/trans.h
    branches/gcc-4_4-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_4-branch/libgfortran/ChangeLog
    branches/gcc-4_4-branch/libgfortran/Makefile.am
    branches/gcc-4_4-branch/libgfortran/Makefile.in
    branches/gcc-4_4-branch/libgfortran/gfortran.map


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40019

[Bug fortran/40019] LEADZ and TRAILZ give wrong results

[burnus at gcc dot gnu dot org]

------- Comment #6 from burnus at gcc dot gnu dot org  2009-06-03 19:40 -------
Now also FIXED on the 4.4 branch. -> CLOSE


-- 

burnus at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED
            Summary|[4.4 only] LEADZ and TRAILZ |LEADZ and TRAILZ give wrong
                   |give wrong results          |results


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40019

[Bug fortran/40019] LEADZ and TRAILZ give wrong results

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|                            |4.4.0
   Target Milestone|---                         |4.4.1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40019