RE: [sv-bc] E-mail Ballot: Respond by 8AM PST, Mon, Dec 10, 2007

From: Steven Sharp <sharp_at_.....>
Date: Tue Dec 04 2007 - 15:18:57 PST
>>SVDB 1339 _X_Yes   ___No
>>http://www.eda.org/svdb/view.php?id=1339
>>
>>SVDB 1345 _X_Yes   ___No
>>http://www.eda.org/svdb/view.php?id=1345

I do think that 12.4.2 should say "after finding a true condition" instead
of "after finding a matching condition".  Can we make this a friendly
amendment?

Case-statements look for matches, while if-statements look for true (though
that could be regarded as "matching" non-zero).  The rest of this section
refers to true conditions.  I assume that this was just a copying error from
the case section.

I am also not sure that 12.4.2 should be saying "evaluate and compare"
instead of just "evaluate".  This may have come in from the case section
also.  Yes, there is an implicit compare to zero, but that is just the
definition of whether the condition is true or not, so I don't think it
needs specific mention.  It seems more likely to confuse than help.
However, I acknowledge that it is not wrong.

It is still possible that this text could be misinterpreted as meaning
that it executes the first true/matching statement *of the subset that it
evaluated before finding a violation*.  However, it is hard to make it
completely clear while leaving the flexibility Gord requested.  Overall,
I think that the proposed text is quite good, and a definite improvement.


>>SVDB 1583 _X_Yes   ___No  
>>http://www.eda.org/svdb/view.php?id=1583
>>
>>SVDB 1809 ___Yes   _X_No
>>http://www.eda.org/svdb/view.php?id=1809

Should wait until Dec 10 for Francoise's revised version.


>>SVDB 2222 _X_Yes   ___No  
>>http://www.eda.org/svdb/view.php?id=2222
>>
>>SVDB 2225 ___Yes   _X_No  
>>http://www.eda.org/svdb/view.php?id=2225

There are a number of other problems in this section that should be
fixed, but I will confine my comments to the text being changed.

In the first text change, the name of a parent module declaration is
allowed instead of its instance name.  The same would presumably apply
to a program or interface name.  However, this was not mentioned
in this part of the text before, so maybe it is a separate issue.

In the second text change, the revised text would read

"If not found and the current scope is not the module scope, look for
the name in the enclosing scope, repeating as necessary until the name
is found, or the compilation unit scope is reached."

Notice the "and the current scope is not the module scope".  That would
imply that we don't look in the compilation unit scope if the current
scope is the module scope.  If it is assumed to apply to each repetition,
it would prevent us from ever looking in the compilation unit scope.

This was basically a combination of a bottom-testing loop with an
if-statement, to make it act like a while loop.  That has a problem
if you change the condition on one without the other.

The simplest fix is to strike the "and the current scope is not the module 
scope".  We don't need to prevent a look upward starting from the module
scope any more, since we will always be going up at least once before we
reach the compilation unit scope.

In the final text added to the end of item b), it should not say
"design unit scope is reached".  It should say "module scope is reached".
I think we have agreed that upward hierarchical searches do not look in
the compilation unit scope of parent instances, only in the compilation
unit scope of the reference.

If this was just an error, and it should say "module scope is reached",
then that should fix it.  If you really think that it should be looking
in the compilation unit scope of parent instances, then we have a bigger
issue.


>SVDB 1571 ___Yes   ___No  _X_Abstain
>http://www.eda.org/svdb/view.php?id=1571

The new proposal addresses the specific concerns that I had with the
text.

I still have more general concerns about adding this functionality at
this stage of the process, when there are still open issues with the
existing macro functionality.  However, since nobody else appears to
share that concern, I will not vote against the proposal.  There is
no point in wasting meeting time on a live vote if nobody else votes
against this in the email vote.

Steven Sharp
sharp@cadence.com


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Received on Tue Dec 4 15:19:26 2007

This archive was generated by hypermail 2.1.8 : Tue Dec 04 2007 - 15:19:51 PST