Re: 0003177: Real numbers with scale factors in digital delays

Hi, Ken -
I had a quick look through the 1364 Verilog LRMs, and I found
that 1364-1995 (on which V-AMS was originally built) and -2001,

delay_value ::=
     unsigned_number
     | parameter_identifier
     | specparam_identifier
     | mintypmax_expression

whereas in 1364-2005,

delay_value ::=
     unsigned_number
     | real_number
     | identifier

I assume the semantics in 2.3.1 just weren't updated to reflect
the addition of real_number in the BNF when the new BNF was
brought over from 1364. I support your proposal.

-Geoffrey

Ken Kundert wrote:
> All,
> I have found an inconsistency in the 2.3.1 version of the LRM. In
> particular on page 15 is says ...
>
> Scale factors are not allowed to be used in defining digital delays
> (e.g., #5u).
>
> However, the BNF uses delay_value to represent a digital delay, and on
> page 331 it says that delay_value may be a real number, and on page 351
> a real number is allowed to have a scale factor.
>
> This conflict within the LRM should be eliminated. I see no reason for
> this restriction, and in fact one existing commercial implementation
> allows this, seemingly without difficulty. I believe having arbitrary
> restrictions like this in the language reduces its usability and in this
> case it reduces it readability. I propose that we delete the restriction
> on page 15.
>
> I would like to get this issue put on the agenda for an upcoming
> meeting. It's ID in Mantis is 0003177.
>
> -Ken
>

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.