ARGH...

[Steve Kargl]

On Sun, Dec 30, 2018 at 04:53:22PM -0800, Steve Kargl wrote:
> All,
> 
> It is impossible to debug f951 if any of the IEEE modules are
> present in the Fortran code eaten by f951.  I have no idea how
> to fix the situation.  This is true for 6, 7, 8, and trunk!
> 
> This effectively means that ieee_selected_real_kind cannot be
> fixed (at least by me or anyone who wishes to use gdb).  It is
> also not possible to have the interface of ieee_selected_real_kind
> in the ieee_arithmetic module.  It must be generated on the
> fly.
> 
> Well, I'm done ranting now.
> 

Well, I sorta have a work around.

% gdb /path/f951
(gdb) b main
(gdb) b simplify_ieee_selected_real_kind
(gdb) run a.f90
... breaks at main ...
(gdb) call gfc_add_intrinsic_modules_path("/path/to/finclude")
(gdb) c
... breaks at main ...
(gdb) c
... breaks at simplify_ieee_selected_real_kind

Only problem is that if one wants to re-run 'run a.f90', one needs
to reload the intrinsic modules path.

-- 
Steve