[gomp merge] Fortran FE changes from gomp branch not dependent on OpenMP

[Paul Brook]

On Wednesday 18 January 2006 19:53, Jakub Jelinek wrote:
> Hi!
>
> On Tue, Jan 17, 2006 at 05:04:09PM +0100, Paul Thomas wrote:
> > What is stunning is how different the code looks with
> > -fdump-tree-original. See original.old and original.new.  First there
> > are no changes in scope or binding level whatsoever and second, there is
> > a proliferation of temporaries, many of which do nothing and, as far as
> > I can tell, get gimplified away.  Do these features matter?  Are there
> > any implications for compilation times or for compiled code efficiency?
>
> Without revamping fortran/trans-*.c, so that it builds GENERIC trees
> that are later on gimplified without already splitting the expressions
> multiple statements and adding temporaries I believe there is no way
> to avoid that.  

I'll take a look at the patch.
Could you explain in general terms what you're trying to achieve?
The fortran frontend DOES create GENERIC trees. We only gimplify before 
passing to the backend because middle-end bits broke when we didn't.

Paul