[Bug fortran/26201] New: [4.1/4/2 regression, fortran] __convert_i4_i8 written to a module.

[Bug fortran/26201] New: [4.1/4/2 regression, fortran] __convert_i4_i8 written to a module.

[kargl at gcc dot gnu dot org]

I have been contacted about compiling MUMPS.
http://graal.ens-lyon.fr/MUMPS/MUMPS_4.6.tar.gz

For some reason, gfortran 4.1.0 writes out __convert_i4_i8 
as a function within a module.  The question is why?

gfc -O2 -Wfatal-errors -I../libseq -Dpord -I. -I../include -c dmumps_ooc.F
gfc -O2 -Wfatal-errors -I../libseq -Dpord -I. -I../include -c dmumps_part1.F
gfc -O2 -Wfatal-errors -I../libseq -Dpord -I. -I../include -c dmumps_part2.F
 In file dmumps_part2.F:962

      END SUBROUTINE DMUMPS_80                                          
                                                                       1
Error: Name '__convert_i4_i8' at (1) is an ambiguous reference to
'__convert_i4_i8' from module 'dmumps_ooc'

troutmask:sgk[209] grep -2 convert src/dmumps_ooc.mod
()

(2 '__convert_i4_i8' 'dmumps_ooc' 1 ((PROCEDURE UNKNOWN-INTENT
UNKNOWN-PROC UNKNOWN INTRINSIC FUNCTION ELEMENTAL PURE) (INTEGER 8 ()) 0
0 () () 0 () ())
--

('max_ooc_file_size' 0 329 'dmumps_609' 0 282 'dmumps_579' 0 216
'db_do_io_and_change_hbuf' 0 55 'buf_io' 0 17 '__convert_i4_i8' 0 2
'current_pos' 0 16 'cur_pos_sequence' 0 4 'cur_hbuf' 0 3
'db_copy_factor_to_buffer' 0 6 'current_solve_read_zone' 0 5


-- 
           Summary: [4.1/4/2 regression, fortran] __convert_i4_i8 written to
                    a module.
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: kargl at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4/2 regression, fortran] __convert_i4_i8 written to a module.

[eedelman at gcc dot gnu dot org]

------- Comment #1 from eedelman at gcc dot gnu dot org  2006-02-09 19:12 -------
Possibly related to the patch for PR 25716 (committed to mainline 2005-01-25).


-- 

eedelman at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4/2 regression, fortran] __convert_i4_i8 written to a module.

[jb at gcc dot gnu dot org]

------- Comment #2 from jb at gcc dot gnu dot org  2006-02-12 21:25 -------
FWIW, this same bug (or a very similar one) blocks compiling VASP too:

 In file cl_shift.f90:1377

END MODULE
         1
Error: Name '__convert_r4_r8' at (1) is an ambiguous reference to
'__convert_r4_r8' from module 'constant'

well, it's r4_r8 and not i4_i8 but otherwise it seems to be the same..


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4.2 regression, fortran] __convert_i4_i8 written to a module.

[pinskia at gcc dot gnu dot org]

-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu dot
                   |                            |org
           Priority|P3                          |P5
            Summary|[4.1/4/2 regression,        |[4.1/4.2 regression,
                   |fortran] __convert_i4_i8    |fortran] __convert_i4_i8
                   |written to a module.        |written to a module.
   Target Milestone|---                         |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4.2 regression] __convert_i4_i8 written to a module.

[eedelman at gcc dot gnu dot org]

------- Comment #3 from eedelman at gcc dot gnu dot org  2006-02-15 21:20 -------
I can't reproduce the problem with MUMPS on Linux/x86 (mainline revision
111111, 4.1 revision 110989).  Assuming that Steve used a 64 bit processor, it
seems that for this code the bug is visible only on 64 bit systems. (The fact
that problem occurs for the function __convert_i4_i8 also indicates that, since
conversions from 4 byte to 8 byte integers happens more often on 64 bit systems
than on 32 bit ones.)

The VASP code I do not have available, so I can't test that one. (Since it is
about reals rather than integers, it's not unlikely that that would be
reproducable on 32 bit systems.)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4.2 regression] __convert_i4_i8 written to a module.

[sgk at troutmask dot apl dot washington dot edu]

------- Comment #4 from sgk at troutmask dot apl dot washington dot edu  2006-02-15 21:30 -------
Subject: Re:  [4.1/4.2 regression] __convert_i4_i8 written to a module.

On Wed, Feb 15, 2006 at 09:20:15PM -0000, eedelman at gcc dot gnu dot org
wrote:
> 
> I can't reproduce the problem with MUMPS on Linux/x86 (mainline revision
> 111111, 4.1 revision 110989).  Assuming that Steve used a 64 bit processor, it
> seems that for this code the bug is visible only on 64 bit systems. (The fact
> that problem occurs for the function __convert_i4_i8 also indicates that, since
> conversions from 4 byte to 8 byte integers happens more often on 64 bit systems
> than on 32 bit ones.)
> 
> The VASP code I do not have available, so I can't test that one. (Since it is
> about reals rather than integers, it's not unlikely that that would be
> reproducable on 32 bit systems.)
> 

I see the problem on amd64-*-freebsd.  For the record,

gfc4x -O2 -Wfatal-errors -v -I../libseq -Dpord -I. -I../include -c
dmumps_part2.F
Using built-in specs.
Target: amd64-unknown-freebsd7.0
Configured with: ../gcc4x/configure --prefix=/mnt1/sgk/work/4x
--enable-languages=c,fortran
Thread model: posix
gcc version 4.2.0 20060215 (experimental)
 /mnt1/sgk/work/4x/libexec/gcc/amd64-unknown-freebsd7.0/4.2.0/cc1 -E
-lang-fortran -traditional-cpp -D_LANGUAGE_FORTRAN -quiet -v -I../libseq -I.
-I../include -Dpord dmumps_part2.F -mtune=generic -Wfatal-errors -O2 -o
/mnt1/sgk/tmp/cc6Lln73.f
ignoring nonexistent directory
"/mnt1/sgk/work/4x/lib/gcc/amd64-unknown-freebsd7.0/4.2.0/../../../../amd64-unknown-freebsd7.0/include"
#include "..." search starts here:
#include <...> search starts here:
 ../libseq
 .
 ../include
 /usr/local/include

 /mnt1/sgk/work/4x/include
 /mnt1/sgk/work/4x/lib/gcc/amd64-unknown-freebsd7.0/4.2.0/include
 /usr/include
End of search list.
 /mnt1/sgk/work/4x/libexec/gcc/amd64-unknown-freebsd7.0/4.2.0/f951
/mnt1/sgk/tmp/cc6Lln73.f -ffixed-form -quiet -dumpbase dmumps_part2.F
-mtune=generic -auxbase dmumps_part2 -O2 -Wfatal-errors -version -I../libseq
-I. -I../include -fpreprocessed -I
/mnt1/sgk/work/4x/lib/gcc/amd64-unknown-freebsd7.0/4.2.0/finclude -o
/mnt1/sgk/tmp/ccSOkFPo.s
GNU F95 version 4.2.0 20060215 (experimental) (amd64-unknown-freebsd7.0)
        compiled by GNU C version 4.2.0 20060215 (experimental).
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
 In file dmumps_part2.F:962

      END SUBROUTINE DMUMPS_80                                          
                                                                       1
Error: Name '__convert_i4_i8' at (1) is an ambiguous reference to
'__convert_i4_i8' from module 'dmumps_ooc'
 In file dmumps_part2.F:962


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4.2 regression] __convert_i4_i8 written to a module.

[fxcoudert at gcc dot gnu dot org]

-- 

fxcoudert at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2006-02-16 12:25:32
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4.2 regression] __convert_i4_i8 written to a module.

[P dot Schaffnit at access dot rwth-aachen dot de]

------- Comment #5 from P dot Schaffnit at access dot rwth-aachen dot de  2006-02-16 14:08 -------
Created an attachment (id=10862)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=10862&action=view)
Sources which exhibit this behaviour (under 32-bit linux)

Hi!

Here's something that exhibits this behaviour... If I may say so, I find it
*really* weird: removing things which do not seem to be related to that seem to
make it disappear though...

I hope this helps, good luck!

Philippe

PS: do not look for any meaning in these sources: I tried to remove as much as
I could from the original sources (...)

PPS: the way I do it:

.../Gcc/bin/gfortran -c Sources.f90
 In file Sources.f90:84

      END SUBROUTINE toolModifyAnzKorn
                                     1
Error: Name '__convert_i4_r4' at (1) is an ambiguous reference to
'__convert_i4_r4' from module 'module_randbed'


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4.2 regression] __convert_i4_i8 written to a module.

[eedelman at gcc dot gnu dot org]

------- Comment #6 from eedelman at gcc dot gnu dot org  2006-02-16 14:12 -------
With Philippe's testcase, I can reproduce the problem on my system now.


-- 

eedelman at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4.2 regression] __convert_i4_i8 written to a module.

[eedelman at gcc dot gnu dot org]

------- Comment #7 from eedelman at gcc dot gnu dot org  2006-02-18 19:52 -------
Here's a small one-liner patch that I think fixes the bug. I'll post it to the
mailing list later tonight if/when it passes regression testing.


Index: gcc/fortran/intrinsic.c
===================================================================
--- gcc/fortran/intrinsic.c     (revision 111236)
+++ gcc/fortran/intrinsic.c     (working copy)
@@ -3428,6 +3428,7 @@ gfc_convert_type_warn (gfc_expr * expr, 
   new->symtree->n.sym->attr.elemental = 1;
   new->symtree->n.sym->attr.pure = 1;
   new->symtree->n.sym->attr.referenced = 1;
+  gfc_intrinsic_symbol(new->symtree->n.sym);
   gfc_commit_symbol (new->symtree->n.sym);

   *expr = *new;


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4.2 regression] __convert_i4_i8 written to a module.

[sgk at troutmask dot apl dot washington dot edu]

------- Comment #8 from sgk at troutmask dot apl dot washington dot edu  2006-02-18 23:54 -------
Subject: Re:  [4.1/4.2 regression] __convert_i4_i8 written to a module.

On Sat, Feb 18, 2006 at 07:52:27PM -0000, eedelman at gcc dot gnu dot org
wrote:
> 
> 
> Index: gcc/fortran/intrinsic.c
> ===================================================================
> --- gcc/fortran/intrinsic.c     (revision 111236)
> +++ gcc/fortran/intrinsic.c     (working copy)
> @@ -3428,6 +3428,7 @@ gfc_convert_type_warn (gfc_expr * expr, 
>    new->symtree->n.sym->attr.elemental = 1;
>    new->symtree->n.sym->attr.pure = 1;
>    new->symtree->n.sym->attr.referenced = 1;
> +  gfc_intrinsic_symbol(new->symtree->n.sym);
>    gfc_commit_symbol (new->symtree->n.sym);
> 
>    *expr = *new;
> 

This fixes the problem of compiling MUMPS on amd64-*-freebsd.
OK to commit to trunk.  You'll need approval from markm to
commit to 4.1.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1/4.2 regression] __convert_i4_i8 written to a module.

[eedelman at gcc dot gnu dot org]

------- Comment #9 from eedelman at gcc dot gnu dot org  2006-02-19 17:23 -------
Subject: Bug 26201

Author: eedelman
Date: Sun Feb 19 17:23:07 2006
New Revision: 111270

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=111270
Log:
fortran/
2006-02-19  Erik Edelmann  <eedelman@gcc.gnu.org>

        PR fortran/26201
        * intrinsic.c (gfc_convert_type_warn): Call
        gfc_intrinsic_symbol() on the newly created symbol.

testsuite/
2006-02-19  Erik Edelmann  <eedelman@gcc.gnu.org>

        PR fortran/26201
        * gfortran.dg/convert_1.f90: New.


Added:
    trunk/gcc/testsuite/gfortran.dg/convert_1.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/intrinsic.c
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1 regression] __convert_i4_i8 written to a module.

[jb at gcc dot gnu dot org]

------- Comment #10 from jb at gcc dot gnu dot org  2006-02-20 18:34 -------
With this fix, trunk is now once again able to compile VASP. Too bad it seems
to be a bit late to get into 4.1. Maybe 4.1.1?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1 regression] __convert_i4_i8 written to a module.

[pinskia at gcc dot gnu dot org]

------- Comment #11 from pinskia at gcc dot gnu dot org  2006-02-20 18:36 -------
(In reply to comment #10)
> With this fix, trunk is now once again able to compile VASP. Too bad it seems
> to be a bit late to get into 4.1. Maybe 4.1.1?
Actually it looks like it was approved for 4.1 (after RC1 was out which it is
already):
http://gcc.gnu.org/ml/gcc-patches/2006-02/msg01574.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1 regression] __convert_i4_i8 written to a module.

[eedelman at gcc dot gnu dot org]

------- Comment #12 from eedelman at gcc dot gnu dot org  2006-02-20 18:50 -------
(In reply to comment #11)
> (In reply to comment #10)
> > With this fix, trunk is now once again able to compile VASP. Too bad it seems
> > to be a bit late to get into 4.1. Maybe 4.1.1?
> Actually it looks like it was approved for 4.1 (after RC1 was out which it is
> already):
> http://gcc.gnu.org/ml/gcc-patches/2006-02/msg01574.html

Yep.  The patch was small and simple enough that I dared to ask RM Mark M if it
was OK. And he said yes, so I intend to commit it to 4.1 later tonight.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1 regression] __convert_i4_i8 written to a module.

[eedelman at gcc dot gnu dot org]

------- Comment #13 from eedelman at gcc dot gnu dot org  2006-02-20 20:06 -------
Subject: Bug 26201

Author: eedelman
Date: Mon Feb 20 20:06:49 2006
New Revision: 111311

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=111311
Log:
fortran/
2006-02-20  Erik Edelmann  <eedelman@gcc.gnu.org>

        PR fortran/26201
        * intrinsic.c (gfc_convert_type_warn): Call
        gfc_intrinsic_symbol() on the newly created symbol.

testsuite/
2006-02-20  Erik Edelmann  <eedelman@gcc.gnu.org>

        PR fortran/26201
        * gfortran.dg/convert_1.f90: New.



Added:
    branches/gcc-4_1-branch/gcc/testsuite/gfortran.dg/convert_1.f90
Modified:
    branches/gcc-4_1-branch/gcc/fortran/ChangeLog
    branches/gcc-4_1-branch/gcc/fortran/intrinsic.c
    branches/gcc-4_1-branch/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201

[Bug fortran/26201] [4.1 regression] __convert_i4_i8 written to a module.

[pinskia at gcc dot gnu dot org]

------- Comment #14 from pinskia at gcc dot gnu dot org  2006-02-20 23:37 -------
Fixed.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26201