RE: [sv-bc] $bits question

IMHO, it is way too late to change this syntax. It's been there since
3.0.

I am in favor of restricting the syntax to disallow anonymous types,
i.e. allow only simple_types.

I would also support an enhancement at a later date to allow 'parameter
typedef'

Dave


> -----Original Message-----
> From: owner-sv-bc@eda.org [mailto:owner-sv-bc@eda.org] On Behalf Of
> Michael (Mac) McNamara
> Sent: Monday, January 23, 2006 11:21 AM
> To: Greg Jaxon; Steven Sharp
> Cc: sv-bc@eda.org
> Subject: RE: [sv-bc] $bits question
> 
> I like where your thoughts are going here, Greg.
> 
> Having the "=" in what really is a parameterized typedef makes it look
> too much like the familiar parameter assignment; and with all the new
> data keywords, and the introduction of typedef, as Steve points out,
it
> could _really_ look like a parameter assignment!!
> 
> Removing the "=" and using "typedef" instead of "type" removes that
> un-useful familiarity; and in truth shifts the familiarity to remind
one
> of the typedef, which is quite apt.
> 
> Perhaps, noting your last comment, it is too late to make such a
change?
> 
> 
> Michael McNamara
> mcnamara@cadence.com
> 408-914-6808 work
> 408-348-7025 cell
> 
> 
> -----Original Message-----
> From: owner-sv-bc@eda.org [mailto:owner-sv-bc@eda.org] On Behalf Of
Greg
> Jaxon
> Sent: Monday, January 23, 2006 10:52 AM
> To: Steven Sharp
> Cc: sv-bc@eda.org
> Subject: Re: [sv-bc] $bits question
> 
> Steven Sharp wrote:
> > ... I don't think that the data type declaration syntax
> > belongs anywhere but in a declaration (typedef or object
declaration).
> In
> > other contexts, the syntax is ambiguous.  For example:
> >
> > typedef logic [N-1:0] T1;
> > typedef logic T2 [N-1:0];
> >
> > parameter type p = logic [N-1:0];
> >
> > It is clear what types T1 and T2 represent, because the position of
> the
> > name in the declaration syntax disambiguates them.  But it is not
> clear
> > what type p is supposed to represent.  Is it similar to T1 or T2?
And
> > putting type() around it doesn't help.
> >
> > Steven Sharp
> > sharp@cadence.com
> >
> 
> 
> Steven,
> 
>     So would you favor eliminating the assignment "=" from the syntax
> of parameter type altogether?
> 
> parameter type my_type P;   /* was: parameter type P = my_type */
> 
> parameter type logic [N-1:0] P1;  /* matches T1 */
> 
> parameter type logic P2 [N-1:0];  /* matches T2 */
> 
> Notice the similarity, then, between a "parameter type" and a
"typedef".
> To me this argues for replacing the keyword "type" here and using
> "typedef".
> The keyword "parameter" becomes just a modifier indicating that this
> typedef can be overridden at elaborate-time.
> 
> Greg Jaxon
> 
> Disclaimer:  The above daydream has nothing to do with anyone's
> product plans or LRM proposals - it's simply design speculation.
>