RE: [sv-ac] #1381

Hi Doron, Bassam,

Currently it is also an implication rather than a followed_by. I guess
it does not matter if the reporting and execution of action blocks would
ignore vacuous successes. So, do we need to change to followed_by?

ed

> -----Original Message-----
> From: Doron Bustan [mailto:dbustan@freescale.com] 
> Sent: Wednesday, May 10, 2006 4:26 PM
> To: Eduard Cerny
> Cc: Bassam Tabbara; sv-ac@eda.org
> Subject: Re: [sv-ac] #1381
> 
> Ed, Bassam,
> 
> I don't think it is so important, (but it should be defined).
> So if no-one else will comment on that in the next few days,
> I will change the proposal, such that there is an attempt for every
> evaluation of the condition in the procedural code.
> 
> do you agree that for coverage, the semantic should be
> 
> cover property (@(clk) not (a |-> not p); //(a followed_by p) ?
> 
> 
> thanks
> 
> Doron
> 
> Eduard Cerny wrote:
> 
> >I am not sure that not countying "attempts" for embedded 
> assert property
> >when the condition is false is a good idea. This is just a syntactic
> >structure and the effective one is the extracted one into a 
> concurrent
> >assertion. If one extracts it manually or let the compiler do it, the
> >behavior should be the same. I.e., have the same evaluation attempts,
> >even if vacuous. 
> >
> >ed 
> >  
> >
> 
> 
> Bassam Tabbara wrote:
> 
> Doron,
> 
> What Ed says below is exactly where 1381 and 805 do not mesh. Putting
> aside whether it's a disabled (vacuous) success/fail/nothing, that's a
> don't care, 805 does count these as disabled *attempts*. 1381 
> in current
> form does not even consider these attempts.
> 
> From your email we have more agreement than difference so again this
> binning *outside* of an attempt seems wrong, any kind of 
> tracking has to
> have an *attempt*. Sorry for repeating myself in many 
> different ways --
> thinking aloud, and can't converge on an understanding of 
> where you are
> at :). 
> 
> Thx.
> -Bassam.
> 
> 
> 
>