[Bug fortran/50121] New: Labels in a TYPE statement should be put in the derived type's scope

[Bug fortran/50121] New: Labels in a TYPE statement should be put in the derived type's scope

[mikael at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50121

             Bug #: 50121
           Summary: Labels in a TYPE statement should be put in the
                    derived type's scope
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: mikael@gcc.gnu.org


Follow-up to PR 50071, reported by Tobias at:
http://gcc.gnu.org/ml/fortran/2011-08/msg00109.html

gfortran doesn't put labels declared in a derived type in their own scope. 
PR 50071 fixed all cases but the one where the label is in the first line of
the derived type definition:

1 type t
    integer :: i
  end type t

  goto 1
1 print *, 'Hello'
end


This case is more difficult, as at the time we insert the label, we haven't
parsed the TYPE statement yet, thus the derived type scope is not available
yet.

[Bug fortran/50121] Labels in a TYPE statement should be put in the derived type's scope

[mikael at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50121

--- Comment #1 from Mikael Morin <mikael at gcc dot gnu.org> 2011-08-18 23:16:31 UTC ---
The snippet above is rejected with:


pr50121.f90:1.1:

1 type t
 1
pr50121.f90:6.1:

1 print *, 'Hello'
 2
Error: Duplicate statement label 1 at (1) and (2)
pr50121.f90:1.1:

1 type t
 1
pr50121.f90:5.9:

   goto 1
         2
Error: Statement at (1) is not a valid branch target statement for the branch
statement at (2)

[Bug fortran/50121] Labels in a TYPE statement should be put in the derived type's scope

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50121

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011-08-19
                 CC|                            |burnus at gcc dot gnu.org
     Ever Confirmed|0                           |1

--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-08-19 07:33:46 UTC ---
For completeness, the program in comment 0 works with g95, NAG, ifort,
openf95/pathf95.
(Though, in case of g95 seemingly by ignoring such labels in TYPE as adding
another to the same type is not diagnosed. The other compilers do diagnose
this.)

Variant - failing in the same way with gfortran but working with the compiler
mentioned above.

1 integer :: a
1 type t
    integer :: i
  end type t
end