RE: $finish and final_step

Martin,

of course you're right.
But I think it's an implementation detail: as previously stated, the
simulator has all latitude to re-execute a step in a different mode to
let the final_step events fire, if needs be.

Now, we just need to know what "the powers that be" will decide about
$finish, final_step, $fatal & friends.

Cheers,
	Xav

On Fri, 2008-03-28 at 09:49 -0700, Martin O'Leary wrote:
> Xavier,
> sorry if I am not clear - my point is that you cannot execute the
> @final_step in isolation - you have to execute the entire analog block
> that it is part of.
> This would not be the case if the final_step was its own block.
> Thanks,
> --Martin 
> 
> -----Original Message-----
> From: Xavier Bestel [mailto:Xavier_Bestel@mentor.com] 
> Sent: Friday, March 28, 2008 1:41 AM
> To: Martin O'Leary
> Cc: David Miller; Verilog-AMS LRM Committee
> Subject: RE: $finish and final_step
> 
> On Thu, 2008-03-27 at 17:17 -0700, Martin O'Leary wrote:
> > oleary>final_step is not a block - it is a statement allowed in the
> > analog block (which is part of the problem I think)
> 
> Well, final_step is an event, which you associate with a block:
> 
> @(final_step) begin
>    if (crossings < 2)
>       $strobe("Could not measure period.");
>    else
>       $strobe("period = %g, crossings = %d",
>          latest-previous, crossings);
> end
> 
> [...]
> > oleary>this looks pretty complex and maybe confusing to explain. Also 
> > oleary>it
> > doesn't seem to cover the case of a $finish in a digital block - it 
> > might be better to have a final block like system verilog does with 
> > clear restrictions on what analog statements are allowed in it (e.g. 
> > not contribution statements)
> 
> There should be no difference between a $finish on the digital side and
> on the analogic side - IMHO.
> 
> > But maybe this is becoming too implementation specific.
> 
> Fully agreed Dave. The important question is: do we execute final_step
> blocks or not ?
> 
> 	Xav
> 


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