Subject: RE: [sv-bc] Erratum and PROPOSAL (BNF) -- timeunits_declaration
From: Brad Pierce (Brad.Pierce@synopsys.com)
Date: Wed Nov 19 2003 - 09:13:40 PST
Peter,
Maybe I have misunderstood Section 18.7 of the package proposal
which defines time scopes and states that, if specified, the
timeunit_declaration must precede any other items in the current
time scope. It also states that the timeunit_declaration can
then be repeated, but must match the previous declaration.
I don't see how a timeunit_declaration could legally occur half way
through a module without having been preceded by a matching
timeunit_declaration in the same time scope.
-- Brad
-----Original Message-----
From: Peter Flake [mailto:flake@synopsys.COM]
Sent: Wednesday, November 19, 2003 8:37 AM
To: Brad.Pierce@synopsys.COM
Cc: sv-bc@eda.org
Subject: Re: [sv-bc] Erratum and PROPOSAL (BNF) -- timeunits_declaration
Brad,
This requires some wording to say what happens if a timeunits_declaration
occurs half way through a module but not at the start.
Are the delays before the timeunits_declaration scaled according to the
previously valid declaration or to the forward reference? I would think
the latter.
Peter.
At 19:22 13/11/2003 -0800, Brad Pierce wrote:
>The current BNF does not allow timeunits_declarations
>in modules, interfaces, etc., except as the first item.
>But this clearly conflicts with 18.6 of the LRM.
>(Repeated timeunits_declarations are very useful
>as constancy checks, especially when the time scope
>is split among multiple files.)
>
>The attached proposal fixes that.
>
>-- Brad
>
This archive was generated by hypermail 2b28 : Wed Nov 19 2003 - 09:14:33 PST