[Bug libfortran/65234] New: Output descriptor (*(1E15.7)) not accepted

[Bug libfortran/65234] New: Output descriptor (*(1E15.7)) not accepted

[vehre at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65234

            Bug ID: 65234
           Summary: Output descriptor (*(1E15.7)) not accepted
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libfortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: vehre at gcc dot gnu.org

An output descriptor of the kind

'(*(1E15.7))' 

is not accepted by the gfortran runtime library, but the error message:

Fortran runtime error: '*' requires at least one associated data descriptor
(*(1E15.7))         
         ^
is emitted. Now, for a repeat specifier of 1 this makes not much sense, besides
that F2003 and F2008 seem to define this kind of output specifier to be valid.
See F2008, 10.3.1,:

R1003 format-items is format-item [ [ , ] format-item ] ...
R1004 format-item is [ r ] data-edit-desc
                  or control-edit-desc
                  or char-string-edit-desc
                  or [ r ] ( format-items )
R1005 unlimited-format-item is * ( format-items )
R1006 r is int-literal-constant

What for example if one wants to emit/read an even number of floats like with:
'(*(2E15.7))'? 

During research Dominique and Tobias found that '(*(2(E15.7)))' is accepted.

The attached example shows that '(*(2(E15.7)))' is accepted while '(*(2E15.7))'
not.

[Bug libfortran/65234] Output descriptor (*(1E15.7)) not accepted

[vehre at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65234

--- Comment #1 from vehre at gcc dot gnu.org ---
Created attachment 34887
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=34887&action=edit
Testcase showing one ok, one fail

[Bug libfortran/65234] Output descriptor (*(1E15.7)) not accepted

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65234

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-02-27
     Ever confirmed|0                           |1

--- Comment #2 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Confirmed.

> What for example if one wants to emit/read an even number of floats like
> with: '(*(2E15.7))'? 

I think this expectation is wrong, i.e., writing an array with an odd number of
elements will print the whole army as demonstrated by the following test

real :: a(11) = [(i,i=1,11)]
print '(*(2(E15.3)))', a
end

which gives at run time

      0.100E+01      0.200E+01      0.300E+01      0.400E+01      0.500E+01    
 0.600E+01      0.700E+01      0.800E+01      0.900E+01      0.100E+02     
0.110E+02

[Bug libfortran/65234] Output descriptor (*(1E15.7)) not accepted

[jvdelisle at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65234

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |jvdelisle at gcc dot gnu.org

--- Comment #3 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> ---
I will have a look at this. I ran into it myself the other day and was
wondering about it. Just have not had time to check the latest standards.

[Bug libfortran/65234] Output descriptor (*(1E15.7)) not accepted

[jvdelisle at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65234

--- Comment #4 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> ---
Created attachment 34906
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=34906&action=edit
Patch tested on x86-64

This patch passes regression testing and NIST testing.

Fairly simple.

[Bug libfortran/65234] Output descriptor (*(1E15.7)) not accepted

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65234

--- Comment #5 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> This patch passes regression testing and NIST testing.
>
> Fairly simple.

Works as advertised! Thanks.

[Bug libfortran/65234] Output descriptor (*(1E15.7)) not accepted

[jvdelisle at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65234

--- Comment #6 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> ---
Author: jvdelisle
Date: Tue Apr 21 18:23:20 2015
New Revision: 222274

URL: https://gcc.gnu.org/viewcvs?rev=222274&root=gcc&view=rev
Log:
2015-04-21 Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    PR libgfortran/65234
    * io/format.c (parse_format_list): Set the seen_dd flag in all
    cases where a data descriptor has been seen.

Modified:
    trunk/libgfortran/ChangeLog
    trunk/libgfortran/io/format.c

[Bug libfortran/65234] Output descriptor (*(1E15.7)) not accepted

[jvdelisle at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65234

--- Comment #7 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> ---
Author: jvdelisle
Date: Tue Apr 21 18:28:39 2015
New Revision: 222276

URL: https://gcc.gnu.org/viewcvs?rev=222276&root=gcc&view=rev
Log:
2015-04-21 Jerry DeLisle  <jvdelisle@gcc.gnu.org>

    PR libgfortran/65234
    * gfortran.dg/fmt_unlimited.f90: New test.

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

[Bug libfortran/65234] Output descriptor (*(1E15.7)) not accepted

[jvdelisle at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65234

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

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

--- Comment #8 from Jerry DeLisle <jvdelisle at gcc dot gnu.org> ---
Fixed on trunk, Closing