RE: [sv-bc] enums in packed arrays

I have the complete opposite opinion. I would rather not see anyone use
an atomic type other than bit, reg, or logic in a packed struct, union
or array. We're trying to get people to think of these predefined types
not in terms individual bits, but as a more abstract value.

Dave


> -----Original Message-----
> From: owner-sv-bc@eda.org [mailto:owner-sv-bc@eda.org] On Behalf Of
Steven
> Sharp
> Sent: Tuesday, April 11, 2006 11:41 AM
> To: sv-bc@eda.org; Brad.Pierce@synopsys.com
> Subject: Re: [sv-bc] enums in packed arrays
> 
> 
> >The change allowed all integral types in packed unions.  Do we then
want
> >to allow all integral types in packed arrays, including bytes?
Sounds
> >good to me, but do we need to exclude 'integer' for backward
> >compatibility with de facto Verilog?
> 
> It sounds reasonable to me also.  As far as I know, the only reason
that
> integer types were not allowed is the one you list: that legacy
designs
> might be relying on tools that ignored the illegal ranges in
> 
> integer [high:low] i;
> 
> I presume that 'time' would also need to be excluded for the same
reason.
> 
> Steven Sharp
> sharp@cadence.com