[Bug fortran/51652] New: [F03] ICE with allocatable scalars

[Bug fortran/51652] New: [F03] ICE with allocatable scalars

[davidgkinniburgh at yahoo dot co.uk]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51652

             Bug #: 51652
           Summary: [F03] ICE with allocatable scalars
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: davidgkinniburgh@yahoo.co.uk


module settings

type keyword
!  character(60), allocatable :: c(:)   ! works but should it?
!  character(80), allocatable :: c(:)   ! works
   character(:), allocatable :: c(:)
end type keyword

type(keyword) :: kw(10)

contains

subroutine save_kw

!allocate(kw(1)%c(10))
allocate(character(80) :: kw(1)%c(10))

kw(1)%c(1) = 'abcd'

if (kw(1)%c(1).eq.'abcd') then  ! problem here
   print *, 'yes'
else
   print *, 'no'
endif

end subroutine save_kw

end module settings

!*************************************************************************

program ice
use settings

call save_kw

end program ice


gcc version 4.7.0 20111212 (experimental) [trunk revision 182257] (GCC)

=========================================================================
f951.exe: internal compiler error: in gfc_get_derived_type, at
fortran/trans-types.c:2401

I'm also interested in testing substrings like

if (kw(1)%c(1)(1:1).eq.'a') then ...

The above program works ok with non-deferred scalar lengths. However, it also
works when the component is assigned different lengths in the subroutine and
main body (e.g. 60 and 80) whereas Intel Fortran reports this as an error: "In
an ALLOCATE statement the type parameter values in type specification must be
the same as a corresponding nondeferred values of the objects being
allocated.".

[Bug fortran/51652] [F03] ICE with allocatable scalarstype parameter

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51652

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |accepts-invalid,
                   |                            |diagnostic, wrong-code
                 CC|                            |burnus at gcc dot gnu.org
            Summary|[F03] ICE with allocatable  |[F03] ICE with allocatable
                   |scalars                     |scalarstype parameter

--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-12-22 08:40:10 UTC ---
(In reply to comment #0)
> type keyword
> !  character(60), allocatable :: c(:)   ! works but should it?
> !  character(80), allocatable :: c(:)   ! works
>    character(:), allocatable :: c(:)
> end type keyword

Unfortunately, deferred-length type parameters (i.e. allocatable string
lengths) are not yet supported for components of derived types. The current
version does not crash (ICE, internal compiler error) but prints an error:
    character(:), allocatable :: c(:)
                                     1
Error: Deferred-length character component 'c' at (1) is not yet supported

As this is already tracked in PR 51550, PR 47545, PR 49050, PR 51075 - and PR
45170, I ignore the issue in this PR and concentrate on your second issue.


 * * *

Regarding the allocation:
  character(60), allocatable :: c(:)
  allocate(character(80) :: c(10))
That's invalid; currently, the value in the allocate statement is ignored by
gfortran. 

The standard requires that this is diagnosed:

"If type-spec appears and the value of a type parameter it specifies differs
from the value of the corresponding nondeferred type parameter specified in the
declaration of any allocate-object, an error condition occurs. If the value of
a nondeferred length type parameter of an allocate-object differs from the
value of the corresponding type parameter of source-expr, an error condition
occurs."

Note that this cannot always diagnosed at compile time, given that also the
following is valid:

subroutine sub(n)
  integer :: n
  character(len=n), allocatable :: str(:)
  allocate(character(n) :: str(1))
  print *, len(str), size(str)
end
call sub(4)
call sub(5)
end

[Bug fortran/51652] Allocate with type-spec and source-expr: check whether length type-parameter is the same is lacking

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51652

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[F03] ICE with allocatable  |Allocate with type-spec and
                   |scalarstype parameter       |source-expr: check whether
                   |                            |length type-parameter is
                   |                            |the same is lacking

--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-12-22 08:44:08 UTC ---
(In reply to comment #1)
> If the value of a nondeferred length type parameter of an allocate-object
> differs from the value of the corresponding type parameter of source-expr,
> an error condition occurs."

The postscripts: Remember that source-expr covers both SOURCE= and MOLD=. And I
forgot to mention that the source of the quote: Fortran 2008, "6.7.1 ALLOCATE
statement", paragraph 6.

[Bug fortran/51652] Allocate with type-spec and source-expr: check whether length type-parameter is the same is lacking

[davidgkinniburgh at yahoo dot co.uk]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51652

--- Comment #3 from David Kinniburgh <davidgkinniburgh at yahoo dot co.uk> 2011-12-22 16:21:09 UTC ---
Sorry, I wasn't up-to-date on the long history of allocatable scalars and dt
components, and hadn't checked the documentation in detail recently. Anyway
thanks for all your work on this - I greatly appreciate your effort.

[Bug fortran/51652] Allocate with type-spec and source-expr: check whether length type-parameter is the same is lacking

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51652

--- Comment #4 from Tobias Burnus <burnus at gcc dot gnu.org> 2012-01-10 11:22:24 UTC ---
Author: burnus
Date: Tue Jan 10 11:22:16 2012
New Revision: 183061

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=183061
Log:
2012-01-10  Tobias Burnus  <burnus@net-b.de>

        PR fortran/51652
        * resolve.c (resolve_allocate_expr): For non-deferred char
        lengths, check whether type-spec matches declaration.

2012-01-10  Tobias Burnus  <burnus@net-b.de>

        PR fortran/51652
        * gfortran.dg/allocate_with_typespec_5.f90: New.


Added:
    trunk/gcc/testsuite/gfortran.dg/allocate_with_typespec_5.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/resolve.c
    trunk/gcc/testsuite/ChangeLog

[Bug fortran/51652] Allocate with type-spec and source-expr: check whether length type-parameter is the same is lacking

[mikael at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51652

Mikael Morin <mikael at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mikael at gcc dot gnu.org

--- Comment #5 from Mikael Morin <mikael at gcc dot gnu.org> 2012-06-29 18:19:56 UTC ---
Fixed?

[Bug fortran/51652] Allocate with type-spec and source-expr: check whether length type-parameter is the same is lacking

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=51652

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-07-20
     Ever confirmed|0                           |1

--- Comment #6 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> Fixed?

Mostly for 4.8.3, 4.9.0, and trunk (4.10.0). The test in comment 0 does not
compile with 4.8 (Deferred-length character component 'c' at (1) is not yet
supported) and, when compiled with 4.9 or 4.10, executing the code gives 'no':
'kw(1)%c(1)' is empty.

[Bug fortran/51652] Allocate with type-spec and source-expr: check whether length type-parameter is the same is lacking

[mikael at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=51652

--- Comment #7 from Mikael Morin <mikael at gcc dot gnu.org> ---
(In reply to Dominique d'Humieres from comment #6)
> > Fixed?
> 
> Mostly for 4.8.3, 4.9.0, and trunk (4.10.0). The test in comment 0 does not
> compile with 4.8 (Deferred-length character component 'c' at (1) is not yet
> supported) and, when compiled with 4.9 or 4.10, executing the code gives
> 'no': 'kw(1)%c(1)' is empty.

Current trunk(6) and 5 branch now print 'yes'.
4.9.x prints 'no'.