[Bug fortran/46010] New: reading of structure

[Bug fortran/46010] New: reading of structure

[marco at hulten dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

           Summary: reading of structure
           Product: gcc
           Version: 4.5.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: marco@hulten.org


Created attachment 22032
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22032
program with the struct and input namelist file

I compiled the NEMO ocean model (revision 1162) with gfortran.  If I now run
the program, I get nonsense (random memory) data written in the 2nd, 4th, 6th
etc. ocean tracer entry in my output.  The relevant code and namelist file
needed for compilation and the relevant output are attached to this e-mail.

This problem arises with gfortran 4.1.2 (RHEL), 4.4.5 (Debian) and 4.5.1
(Arch), but it doesn't with the Intel Fortran Compiler (ifort).  Please let me
know if you need other information.

[Bug fortran/46010] I/O: Namelist-reading bug

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

Tobias Burnus <burnus at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code
                 CC|                            |burnus at gcc dot gnu.org,
                   |                            |jvdelisle at gcc dot
                   |                            |gnu.org
            Summary|reading of structure        |I/O: Namelist-reading bug
      Known to fail|                            |4.2.4, 4.6.0

--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> 2010-10-13 21:42:19 UTC ---
Jerry, can you have a look?

The first 12 items look OK (cf. "fort.11") if one compiles with NAG or ifort.
With gfortran, the items 1 looks OK, two has garbage, three looks OK - but has
the short/long name belonging to item 2 etc.

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

Jerry DeLisle <jvdelisle at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2010.10.14 03:20:03
         AssignedTo|unassigned at gcc dot       |jvdelisle at gcc dot
                   |gnu.org                     |gnu.org
     Ever Confirmed|0                           |1

--- Comment #2 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-10-14 03:20:03 UTC ---
I am on it. Thanks for test case

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #3 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-10-14 03:33:07 UTC ---
Try compiling your program with -std=f95.

What we have here is what I believe is a non-standard extension byting us in
the you know where.  Regardless, I will investigate further. Stay tuned.

[Bug fortran/46010] I/O: Namelist-reading bug

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #4 from Tobias Burnus <burnus at gcc dot gnu.org> 2010-10-14 08:09:04 UTC ---
Reduced test case:

----------nml.dat-----------------------
&namtoptrc
   tracer(1)   = 'Aa', .true.
   tracer(2)   = 'Bb', .true.
   tracer(3)   = 'Cc', .true.
/
----------------------------------------

----------test.f90----------------------
      TYPE PTRACER
         CHARACTER(len = 2)  :: sname
         LOGICAL              :: lini
      END TYPE PTRACER
      TYPE(PTRACER) , DIMENSION(3) :: tracer
      NAMELIST/namtoptrc/  tracer

open (99, file='nml.dat')
read (99, nml=namtoptrc)
write (*, nml=namtoptrc)
end
----------------------------------------

Works with -std=f90/f2003/f2008 but fails with without -std=f* (i.e. with
-std=gnu) as follows:

At line 10 of file test.f90 (unit = 99, file = 'nml.dat')
Fortran runtime error: Cannot match namelist object name 'bb'

The issue is really the ".true." ("T","true", ".false.", "F", "false"). If the
last item (in the namelist and in the derived type) is an integer or a string
it works.

[Bug fortran/46010] I/O: Namelist-reading bug

[marco at hulten dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #5 from Marco van Hulten <marco at hulten dot org> 2010-10-14 18:15:59 UTC ---
(In reply to comment #4)
> [...]
> The issue is really the ".true." ("T","true", ".false.", "F", "false"). If the
> last item (in the namelist and in the derived type) is an integer or a string
> it works.

No, that doesn't seem to be the issue.  I tried both your minimal test case as
well as my test case, with an added character as well as an added number, but
all give the following result:

marco@graviton testje $ gfortran -std=f2003 test.f90
marco@graviton testje $ ./a.out 
&NAMTOPTRC
 TRACER(1)%SNAME="Aa",
 TRACER(1)%LINI=T,
 TRACER(1)%GETAL=          1,
 TRACER(2)%SNAME="Bb",
 TRACER(2)%LINI=T,
 TRACER(2)%GETAL=          1,
 TRACER(3)%SNAME="Cc",
 TRACER(3)%LINI=T,
 TRACER(3)%GETAL=          1,
 /
marco@graviton testje $ gfortran test.f90 
marco@graviton testje $ ./a.out 
At line 13 of file test.f90 (unit = 99, file = 'nml.dat')
Fortran runtime error: Bad data for namelist object tracer%getal

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #6 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-10-14 19:13:30 UTC ---
This one is on the top of my list.  I usually take a few days to sort it out. 
I already have a general idea what the problem is.  I will use your added test
cases when I nail it.  Ther is a possibility we have more then one issue going
on. The namelist reading has been very fertile ground for weeds to grow in. :)

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #7 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-10-25 05:47:58 UTC ---
Here is a patch.

Please try this if you can.

Index: list_read.c
===================================================================
--- list_read.c    (revision 165900)
+++ list_read.c    (working copy)
@@ -2105,8 +2105,9 @@ nml_parse_qualifier (st_parameter_dt *dtp, descrip

           /*  If -std=f95/2003 or an array section is specified,
               do not allow excess data to be processed.  */
-                  if (is_array_section == 1
-              || !(compile_options.allow_std & GFC_STD_GNU))
+          if (is_array_section == 1
+              || !(compile_options.allow_std & GFC_STD_GNU)
+              || !dtp->u.p.ionml->touched)
             ls[dim].end = ls[dim].start;
           else
             dtp->u.p.expanded_read = 1;

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #8 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-10-25 13:12:33 UTC ---
The patch in comment #7 works for the original test case but fails for the one
in comment #4.  I have a revised patch that also works for comment #4 and
passes regression testing.  I will submit it later tonight.

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #9 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-10-26 01:37:04 UTC ---
Created attachment 22156
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22156
Patch to fix this.

I think my email service is out. I attach the patch here for review.  It may
show up on the list.  I tried sending it three times.

[Bug fortran/46010] I/O: Namelist-reading bug

[marco at hulten dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #10 from Marco van Hulten <marco at hulten dot org> 2010-10-26 09:13:28 UTC ---
On Tue, 26 Oct 2010 01:37:08 +0000 jvdelisle wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010
> 
> --- Comment #9 from Jerry DeLisle <jvdelisle at gcc dot gnu.org>
> 2010-10-26 01:37:04 UTC --- Created attachment 22156
>   --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22156
> Patch to fix this.
> 
> I think my email service is out. I attach the patch here for review.
> It may show up on the list.  I tried sending it three times.

Thank you for the patch.

I have no time right now to apply and check the patch, but maybe in the
weekend.

Cheers,

Marco

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #11 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-10-26 19:05:14 UTC ---
Author: jvdelisle
Date: Tue Oct 26 19:05:08 2010
New Revision: 165979

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=165979
Log:
2010-10-26  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    PR libgfortran/46010
    * io/list_read.c (nml_parse_qualifier): Add additional conditions for
    setting the end index for loop specification. Fix some whitespace.
    * io/write.c (write_default_char4): Const-ify the source argument.

Modified:
    trunk/libgfortran/ChangeLog
    trunk/libgfortran/io/list_read.c
    trunk/libgfortran/io/write.c

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #12 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-10-26 19:34:29 UTC ---
Author: jvdelisle
Date: Tue Oct 26 19:34:21 2010
New Revision: 165981

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=165981
Log:
2010-10-26  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    PR libgfortran/46010
    * gfortran.dg/namelist_66.f90: New test.
    * gfortran.dg/fmt_t_6.f: Fix whitespace.

Added:
    trunk/gcc/testsuite/gfortran.dg/namelist_66.f90
Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gfortran.dg/fmt_t_6.f

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #13 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-10-26 19:37:39 UTC ---
Fixed on trunk.  I will backport to 4.4 and 4.5 in the next few weeks since
this is a wrong code problem.

[Bug fortran/46010] I/O: Namelist-reading bug

[marco at hulten dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #14 from Marco van Hulten <marco at hulten dot org> 2010-11-02 20:29:17 UTC ---
I succesfully compiled the trunk version and hereafter compiled and executed my
test case without problems (giving the righ results).  I am looking forward to
the backport into gcc 4.4.

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #15 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-04 00:16:55 UTC ---
Author: jvdelisle
Date: Thu Nov  4 00:16:47 2010
New Revision: 166288

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166288
Log:
2010-11-03  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    PR libgfortran/46010
    Backport from mainline:
    * io/list_read.c (nml_parse_qualifier): Add additional conditions for
    setting the end index for loop specification. Fix some whitespace.

Modified:
    branches/gcc-4_5-branch/libgfortran/ChangeLog
    branches/gcc-4_5-branch/libgfortran/io/list_read.c

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #16 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-04 00:25:15 UTC ---
Author: jvdelisle
Date: Thu Nov  4 00:25:10 2010
New Revision: 166293

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166293
Log:
2010-11-03  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    Backport from mainline:
    PR libgfortran/46010
    * gfortran.dg/namelist_66.f90: New test.
    * gfortran.dg/fmt_t_6.f: Whitespace.

Added:
    branches/gcc-4_5-branch/gcc/testsuite/gfortran.dg/namelist_66.f90
Modified:
    branches/gcc-4_5-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_5-branch/gcc/testsuite/gfortran.dg/fmt_t_6.f

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #17 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-04 00:34:20 UTC ---
Author: jvdelisle
Date: Thu Nov  4 00:34:16 2010
New Revision: 166294

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166294
Log:
2010-11-03  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    PR libgfortran/46010
    Backport from mainline:
    * io/list_read.c (nml_parse_qualifier): Add additional conditions for
    setting the end index for loop specification. Fix some whitespace.

Modified:
    branches/gcc-4_4-branch/libgfortran/ChangeLog
    branches/gcc-4_4-branch/libgfortran/io/list_read.c

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

--- Comment #18 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-04 00:38:21 UTC ---
Author: jvdelisle
Date: Thu Nov  4 00:38:17 2010
New Revision: 166295

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166295
Log:
2010-11-03  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    Backport from mainline:
    PR libgfortran/46010
    * gfortran.dg/namelist_66.f90: New test.

Added:
    branches/gcc-4_4-branch/gcc/testsuite/gfortran.dg/namelist_66.f90
Modified:
    branches/gcc-4_4-branch/gcc/testsuite/ChangeLog

[Bug fortran/46010] I/O: Namelist-reading bug

[jvdelisle at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46010

Jerry DeLisle <jvdelisle at gcc dot gnu.org> changed:

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

--- Comment #19 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> 2010-11-04 00:42:40 UTC ---
Now back ported to 4.5 and 4.4