RE: [sv-ec] re-wording for Mantis 1871 proposal

From: Bresticker, Shalom <shalom.bresticker_at_.....>
Date: Wed Jul 11 2007 - 00:55:33 PDT
Thanks.

 

The meaning of 'transition' and 'sequence' does seem fuzzy.

 

At the beginning of 18.5.1, it seems that a transition means a single
transition, whereas a sequence is made up of one or more transitions.

 

However, that usage does not seem to be consistent.

 

I would say that there is not much benefit in changing any one
particular place. The entire section 

should be redone to make the usage consistent. 

 

Looking over the text in 18.5.1, I see that the sentence,

A trans_list specifies one or more sets of ordered value transitions of
the coverage point.

appears twice in that section.

 

That seems an unneeded duplication.

 

Thanks,

Shalom

 

________________________________

From: David Scott [mailto:david_scott@mentor.com] 
Sent: Tuesday, July 10, 2007 8:15 PM
To: Bresticker, Shalom
Cc: SV-EC
Subject: Re: [sv-ec] re-wording for Mantis 1871 proposal

 

That was Ray Ryan's request, to remove the language about "bins".  The
language about state bin values (as opposed to transition bin sequences)
refers to values rather than bins in the paragraph I am proposing to
change -- so it is appropriate.

Just to make the whole issue clear:

    coverpoint p {
       bins t[] = ( 1=>2=>[3:4]=>4 );
       ignore_bins i = ( 2=>3 );
    }

In this case, there are potentially 2 coverage bins: t[1=>2=>3=>4] and
t[1=>2=>4=>4].  However, because the first contains the ignored
sequence, that sequence itself must be removed.  By implication, the bin
t[1=>2=>3=>4] is excluded -- so that the coverage point contains only 1
coverage bin rather than the 2 it seems to have at first glance.  There
is already language in the subsequent paragraph to describe the effect
of bins being emptied because their values or transitions/sequences [*]
overlap with ignored bins:




The above may result in a bin that is associated with no values or
transitions. Such empty bins are excluded
from coverage (see 18.11).


[*]  Re-reading it, I sense the specification is a little loose on the
language "transition" vs. "sequence".  18.5.1 uses the word "sequence"
to refer to what is associated with a transition bin, while 18.5.4 and
18.5.5 as currently written use "transition".  Maybe I should propose to
change "transition" to "sequence" in the paragraph I quoted above?

Dave S


Bresticker, Shalom wrote: 

Dave,

 

I am not very familiar with this area, but I see that the original uses
'bin' whereas the proposal uses 'sequence'. Does it matter?

 

Otherwise, it looks fine.

 

Thanks,

Shalom

 

________________________________

From: owner-sv-ec@server.eda.org [mailto:owner-sv-ec@server.eda.org] On
Behalf Of David Scott
Sent: Tuesday, July 10, 2007 3:07 AM
To: SV-EC
Subject: [sv-ec] re-wording for Mantis 1871 proposal

 

Old wording:

For transition bins, the ignored sequence invalidates any coverage
transition bin whose
sequence cannot be matched without also matching the ignored transition
bin sequence for any sample
value.  (For example, the ignored bin "i = ( 2=>3 )" would invalidate
the coverage bin "t =
( 1=>2=>3=>4)"). 

Proposed new wording:

For transition bins, any covered sequence is removed when it contains an
ignored sequence.  (For example, the ignored sequence 2=>3 would remove
the covered sequence 1=>2=>3=>4.)

That is almost suspiciously simpler, but Ray & I agree it conveys the
same information.

Please respond with your opinion if you have one; I'll update the
proposal (with similar changes for illegal bins) as soon as it appears
to have converged by e-mail.

Thanks,

Dave Scott -- Mentor Graphics


-- 
This message has been scanned for viruses and 
dangerous content by MailScanner <http://www.mailscanner.info/> , and is

believed to be clean. 

 


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Received on Wed Jul 11 00:56:16 2007

This archive was generated by hypermail 2.1.8 : Wed Jul 11 2007 - 00:56:27 PDT