[Bug fortran/11220] New: g77: hiccup on variable declaration interchange

[anlauf@hep.tu-darmstadt.de]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11220

           Summary: g77: hiccup on variable declaration interchange
           Product: gcc
           Version: 3.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: anlauf@hep.tu-darmstadt.de
                CC: gcc-bugs@gcc.gnu.org
  GCC host triplet: i586-suse-linux

The following code makes g77 puke:

% cat g77-bug.f
      subroutine good (a, n)
      implicit none
      integer n, a(n)
      a(1)=n
      end
      subroutine bug (a, n)
      implicit none
      integer a(n), n
      a(1)=n
      end

% g77 -c g77-bug.f 
g77-bug.f: In subroutine `bug':
g77-bug.f:6: 
         subroutine bug (a, n)
                            1
g77-bug.f:8: (continued):
         integer a(n), n
                   2
Invalid declaration of or reference to symbol `n' at (2) [initially seen at (1)]
g77-bug.f:6: 
         subroutine bug (a, n)
                            1
g77-bug.f:8: (continued):
         integer a(n), n
                       2
Invalid declaration of or reference to symbol `n' at (2) [initially seen at (1)]
g77-bug.f:6: 
         subroutine bug (a, n)
                         1
g77-bug.f:9: (continued):
         a(1)=n
         2
Invalid declaration of or reference to symbol `a' at (2) [initially seen at (1)]

% g77 --version |head -1
GNU Fortran (GCC) 3.3 20030226 (prerelease) (SuSE Linux)

Replacing
      integer a(n), n
by
      integer n, a(n)
or removing
      implicit none
fixes the symptom.

I reported this for g77 from 0.5.19.1 through 0.5.24-19981002
already on 16 Feb 2000, but nobody appeared to track this report.
I consider this an annoying inconvenience for legacy code.
Could somebody please look into this?