[patch,ping??] Fix PR17917 - module equivalences

[THOMAS Paul Richard 169137]

This patch was submitted in
http://gcc.gnu.org/ml/fortran/2005-06/msg00307.html

It works fine but has the unfortunate characteristic of producing a
redundant union for each and every module equivalence.  This has me
researching how module symbols are declared and equivalences are built.  I
expect that this will, in the words of Hitchhiker's Guide to the Galaxy,
"take a little while".  Should I commit the patch, as it is, on the grounds
that it is better than nothing and will, of itself, need little or no
eventual modfication?

Paul T