Re: [vhdl-200x-ft] Bit rules

Jim Lewis wrote:
> David,
>
>> You can't use bit string literals in the fixed_pkg.
>> The reason for this is because the "ufixed" and "sfixed"
>> type are not bounded on either end.
>
>
> I still have an action item to discuss "101010.100" as
> a potential fixed point format with John Ries (owner of
> FT09). I also have an action item for floating point
> formats.
>
> If you accept proposal 4, is there any reason I should
> pursue the above? Under proposal 4, this format is
> not usable in expressions and would only be used in
> assignments and initializations. Hence the "." would
> only be for user readability and not for determining
> bounds.

Correct. The ability to add a bit literal was lost when
we added the ability to do:
ufixed := ufixed + ufixed;
and
ufixed := ufixed + unsigned;

So adding any more won't hurt anything.

Also correct. The "." is just there for readability.
The same thing goes for the floating point packages, as they
are overloaded for:
fp32 := fp32 + fp32
and
fp32 := fp32 + signed
Thus you have the same problem with a bit literal, it must be
type casted.

-- 
NAME:     David W. Bishop           INTERNET: dbishop@vhdl.org