Re: [vhdl-200x-ft] FP types and packages

Peter Ashenden wrote:

> function "+" ( L, R : FP;
> round_style : round_type := default_round_style;
> denormalize : BOOLEAN := default_denormalize;
> check_error : BOOLEAN := default_check_error;
> guard_bits : NATURAL := default_guard_bits )
> return FP;
> ...

The basic problem with this approach is that an operator may
have only 1 or two inputs (IEEE-1076-2002 section 2.3.1).

How about this as a compromise here:

Leave the original "fphld_base_pkg" there, with the "floating_point"
type in it. Also leave the packages with the generics.

Then, Jim can use:
use ieee.fphdl_base_pkg.all;
...
signal xxx, yyy, zzz : floating_point (8 downto -23);
...
zzz <= fp_addition( xxx, yyy); -- letting the rest of the parameters default)

and Peter can use:
use ieee.fphdl32_pkg.all;
...
signal xxx, yyy, zzz : fp;
...
zzz <= xxx + yyy;

Will this satisfy both communities?

-- 
David W. Bishop dbishop@vhdl.org       All standard disclaimers apply.