Re: [PATCH, gfortran testsuite]: Do not load denormals in gfortran.fortran_torture/intrinsic_nearest.f90

[IainS <developer at sandoe-acoustics dot co dot uk>]

Thanks Steve,

for an extremely helpful explanation.

On 28 Aug 2009, at 20:23, Steve Kargl wrote:
<snip>
> symbol.c:#include "iso-c-binding.def"
> symbol.c:#include "iso-c-binding.def"
> trans-types.c:#include "iso-c-binding.def"
>
> > Is this the model that should be used for "ieee_arithmetic"? (i.e.
> > declare the c functions in a __ieee_arithmetic_ namespace).
> > Is this the manner in which a USE statement decides how to form a
> > function name precursor ?
>
> The prefix should be _gfortran_.  I haven't given it much thought, but
> I suspect the IEEE modules can be built on the fly as well.

I had found the .def files but hadn't got any further...
I'll look into replication of this approach for  IEEE_*

--

The current issue for me is that, if I read the spec correctly,  
adoption of IEEE_* requires that (at minimum) overflow and div-0 IEEE  
exceptions are enabled.
Enabling them in the chip isn't likely to be hard .. but whether/how  
they percolate through to libgfortran's runtime error handler is not  
clear (yet).

thanks again,
Iain