RE: Problem with the named constraints example on page 178 of the latest draft

Hello Darren,
          I wasn't sure how to handle my vote given this issue. In general I approve this draft provided we fix this issue but it did not seem to be correct to vote approved with this issue outstanding. What do you think?

thanks

[cid:image002.gif@01CB50C9.DC142F30]

Brett Lammers | Senior Technical Leader, Solutions Deployment Group

P: 303.579.6721 www.cadence.com<http://www.cadence.com>

From: darren.galpin@infineon.com [mailto:darren.galpin@infineon.com]
Sent: Friday, September 10, 2010 1:28 AM
To: Brett Lammers
Cc: Yaron Kashai; ieee1647@eda.org
Subject: RE: Problem with the named constraints example on page 178 of the latest draft

PS. Making data_size undefined with no else error() is consistent then with the example in the next section 10.4.1.1.

________________________________
From: Brett Lammers [mailto:brettl@cadence.com]
Sent: Wednesday, September 08, 2010 5:26 PM
To: Galpin Darren (IFGB ATV MCD CV)
Cc: Yaron Kashai
Subject: Problem with the named constraints example on page 178 of the latest draft
Hello Darren,
                I apologize for the last minute issue but I have been crazy busy and have had to put off my review of the draft until today.

Anyway in my review I was spot checking a few examples listed in the Draft in Specman and ran into some issues with the example mentioned (page 178 "Syntax example (named constraint):")

One issue was just a simple typo (unit=>uint) which I submitted mantis issue 3199
 The other issue I was not as sure whether it was a Specman issue or a Spec issue so I conferred with Yaron and some others in Cadence and Yaron's response was the following:
------------------------------------------------------ From Yaron -----------------------------------------------

Thanks Brett!

This seems broken to me. The example does not adhere to the LRM spec - meaning the LRM is inconsistent, regardless of Specman. In particular:

> keep data_size is undefined else error(...)

should match the syntax:

keep NAME is BOOL-EXP else error(...)

and "undefined" is not a BOOL-EXP.

I think there is no choice but to "stop the press". Let's start by notifying Darren. Someone on the R&D side should investigate the origin of this example and the intention behind it...

As such I will submit another issue...should I do anything else?

[cid:image002.gif@01CB50C9.DC142F30]

Brett Lammers | Senior Technical Leader, Solutions Deployment Group

P: 303.579.6721 www.cadence.com<http://www.cadence.com>

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