Re: [Fortran] Cleanup patch.

[Steve Kargl <sgk at troutmask dot apl dot washington dot edu>]

On Thu, May 07, 2009 at 06:09:44PM +0200, Paul Richard Thomas wrote:
> Steve,
> 
> This is long overdue.
> 
> > OK for trunk?
> 
> OK
> 
> Thanks for the patch, which I have to say was one of the most
> monotonous I have ever seen :-)
> 

A large amount of the patch was mechanical.  I hope to 
have a follow-on patch to consistently use code->expr#
throughout the source code.  Currently, we have c->expr,
code->expr, cblock->expr, and at least one occurrence
of p->expr.

> Paul
> 
> PS expr3????

allocate(real(kind=4) :: x(3), status=stat, errmsg=msg, source=sqrt(42))

code->expr1 = status
code->expr2 = errmsg
code->expr3 = source

The one-to-one correspondence above simplifies the resolution and
translation stages.

I have the parsing of the above working with only 2 regressions:
auto_char_dummy_array_1.f90 and auto_char_pointer_array_result_1.f90.
The problem comes from parsing the optional typespec.

-- 
Steve