[Patch, gfortran] Tidy up repeated symbols in .mod files (cf. patch for PR 16861)

[Paul Thomas]

:ADDPATCH fortran:

Steve,

Thanks for clearing http://gcc.gnu.org/ml/fortran/2005-09/msg00316.html 
so quickly.  I'll commit it this afternoon.

As mentioned in the above, such indices for contained assumed shape 
arrays cause multiple entries of the same symbol
in .mod files, when the reference to the index is resolved with a use 
associated symbol.  The patch below cures this
"problem"(ie. nothing is broken) by eliminating extraneous symbol 
entries.  The comment speaks for itself.

read_module was already capable of dealing with the resulting 
streamlined .mod files, as a bit of judicious editing
of the file foobar.mod from nested_modules_2.f90 will demonstrate.

Regtested on FC3/Athlon 1700.

OK for mainline and 4.03, when open?

Paul T

===================================================================

2005-09-15  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/16861
    * module.c (mio_symtree_ref): Resolve "unique" symbol reference
    as soon as a corresponding symbol is available in the main
    namespace.


*** module.c    2005-09-17 10:54:31.743944128 +0200
--- gcc/gcc/fortran/module.c    2005-09-17 10:37:25.677929872 +0200
*************** mio_symtree_ref (gfc_symtree ** stp)
*** 2091,2097 ****
 
    if (iomode == IO_OUTPUT)
      {
!       mio_symbol_ref (&(*stp)->n.sym);
      }
    else
      {
--- 2091,2105 ----
 
    if (iomode == IO_OUTPUT)
      {
!       /* If this is a symtree for a symbol that came from a contained
!      module namespace, it has a "unique" name and we should look in
!      the current namespace to see if the required, non-contained
!      symbol is available yet.  */
!       if ((*stp)->n.sym && check_unique_name((*stp)->name))
!     ns_st = gfc_find_symtree (gfc_current_ns->sym_root,
!                   (*stp)->n.sym->name);
!
!       mio_symbol_ref (ns_st ? &ns_st->n.sym : &(*stp)->n.sym);
      }
    else
      {