RE: [sv-bc] 19.12.5: array of instances connection to packed array port

I agree. 

By the way, I had started to file a Mantis on this section, but stopped
when I found that it looked even more confusing than I first thought.
The section is titled "Unpacked array ports and arrays of instances",
but there is discussion of packed arrays also and of array port
connections as well as array ports. It is not always clear just what is
being discussed and what is being connected to what. There are 3 arrays
involved: arrays of instances, arrays ports, and array port connections,
and all can be involved simultaneously, and the ports and port
connections can be packed or unpacked.

Shalom


> -----Original Message-----
> From: owner-sv-bc@server.eda.org [mailto:owner-sv-bc@server.eda.org]
On
> Behalf Of Jonathan Bromley
> Sent: Wednesday, September 20, 2006 1:05 PM
> To: sv-bc@server.eda-stds.org
> Subject: RE: [sv-bc] 19.12.5: array of instances connection to packed
> array port
> 
> > I think the phrase could be made more successful if it were phrased
in
> > terms of bit significance, starting from LSB to MSB (indirectly
> > referring to the vector equivalent Steven mentioned)
> 
> Would it not be better still if we had an explicit notion of the
> "equivalent vector" of any packed object?  I suspect that would be
> useful in other situations too.  By defining the equivalent vector
> to have a normalized subscript range [N-1:0] you could make it
> rather easy to rewrite 19.12.5.
> 
> I'm not suggesting that "equivalent vector" be defined in 19.12.5,
> but rather that it be a definition quite early in the LRM that
> can be used in discussion of *any* packed object.
> 
> Nor am I suggesting that the equivalent vector be required to
> exist as a real physical object.  It would be only a convenient
> (and well-defined) fiction in the LRM.