From: Steven Sharp (sharp@cadence.com)
Date: Tue Mar 01 2005 - 19:18:25 PST
>I agree with your point, at least V2K LRM doesn't support the hier-ref in
>constant expression,
>I am not sure of SV.
I don't believe that SV has changed this. The rule is not an arbitrary
one, but exists for good reasons, which apply just as well to SV.
>But the following example is passed by MTI simulator.
So MTI is not compliant with the standard. That doesn't change the
standard.
Out of curiosity, what does MTI do with your example? What happens
if you remove the instance outside the generate? Since you put it in,
I am guessing that it must resolve the I1.p into that instance before
expanding the generate-if.
I note that MTI is not erroring out and claiming that I1 is declared
twice in the same scope. Either it is treating the generate-if as
creating a nested scope (as per the LRM modifications for erratum 113),
or the implementors were trying to match the original LRM text and failed.
Steven Sharp
sharp@cadence.com
This archive was generated by hypermail 2.1.4
: Tue Mar 01 2005 - 19:00:09 PST
and
sponsored by Boyd Technology, Inc.