0001251: 9.7.6 does not explicitly say what happens if "wait" condition is 'x' or 'z'

EDA.org Mantis

Viewing Issue Simple Details [ Jump to Notes ] [ Wiki ]

[ View Advanced ] [ Issue History ] [ Print ]

Category

Severity

Date Submitted

Last Update

0001251

[SystemVerilog P1800] SV-BC

minor

2005-12-18 23:54

2010-02-08 21:46

Reporter

shalom

View Status

public

Assigned To

Priority

normal

Resolution

open

Status

new

Product Version

Summary

0001251: 9.7.6 does not explicitly say what happens if "wait" condition is 'x' or 'z'

Description

From: Steven Sharp <sharp@cadence.com>
To: etf-bugs@boyd.com, Shalom.Bresticker@freescale.com
Cc:
Subject: Re: errata/650: 9.7.6 does not explicitly say what happens if "wait" condition is x or z
Date: Tue, 15 Feb 2005 13:51:01 -0500 (EST)

9.6 does not specify what happens if a "while" condition is x or z either.
It does specify it for a "for" condition.

9.4 does define what it means for an expression to evaluate as "true" or
"false". It does so in the context of "if" statements, but there is
nothing to restrict the definition to that context. So that definition
would apply to these other situations that refer to an expression being
"true" or "false" also. It may not be the best organized, but I think
it is specified.

Steven Sharp
sharp@cadence.com

From: Shalom.Bresticker@freescale.com
To: Steven Sharp <sharp@cadence.com>
Cc: etf-bugs@boyd.com
Subject: Re: errata/650: 9.7.6 does not explicitly say what happens if "wait"
condition is x or z
Date: Tue, 15 Feb 2005 21:06:18 +0200 (IST)

I agree that this was the intent. However, the terms true and false are almost
certainly used in other contexts where x and z are not included. So it would
be more precise to say that x is considered false in the context of a
condition, not an arbitrary expression.

And even there, there are exceptions. Remember issue 237?

Further, if someone looks up what is written in 9.6 or 9.7.6, there is no
reason he should think that 9.4 is relevant.

I see no reason not to specify it more clearly.

Shalom

On Tue, 15 Feb 2005, Steven Sharp wrote:

> 9.6 does not specify what happens if a "while" condition is x or z either.
> It does specify it for a "for" condition.
>
> 9.4 does define what it means for an expression to evaluate as "true" or
> "false". It does so in the context of "if" statements, but there is
> nothing to restrict the definition to that context. So that definition
> would apply to these other situations that refer to an expression being
> "true" or "false" also. It may not be the best organized, but I think
> it is specified.

From: Steven Sharp <sharp@cadence.com>
To: etf-bugs@boyd.com, Shalom.Bresticker@freescale.com
Cc:
Subject: Re: errata/650: 9.7.6 does not explicitly say what happens if "wait" condition is x or z
Date: Tue, 15 Feb 2005 15:54:30 -0500 (EST)

> I agree that this was the intent. However, the terms true and false are almost
> certainly used in other contexts where x and z are not included. So it would
> be more precise to say that x is considered false in the context of a
> condition, not an arbitrary expression.
>
> And even there, there are exceptions. Remember issue 237?

For that issue, the text in 15.6 clearly states how x is treated, and
does not use the terms true and false. The correctness of the text is
questionable, but it does not appear to conflict with the definitions
of true and false.

The conditional operator also describes how x is treated. However, it
uses the terms true and false, and defines them differently from the
definition for "if" statements. This could create confusion.

> Further, if someone looks up what is written in 9.6 or 9.7.6, there is no
> reason he should think that 9.4 is relevant.

I agree that the organization is poor. It may work for someone reading
the text from front to back, but not for someone looking something up.
It could be specified better by repeating the definition every place it
is used, or by moving the definition into a separate section and stating
that it apples to all conditions in procedural statements.

Steven Sharp
sharp@cadence.com

From: Shalom.Bresticker@freescale.com
To: Steven Sharp <sharp@cadence.com>
Cc: etf-bugs@boyd.com
Subject: Re: errata/650: 9.7.6 does not explicitly say what happens if "wait"
condition is x or z
Date: Wed, 16 Feb 2005 04:56:53 +0200 (IST)

Yes to all you wrote.

By chance, I am reviewing Clause 14 (Specify blocks) due to some specify
syntax which appears to be accepted by most of the simulators, but is
not allowed by the BNF. (Expect an errata on it in the near future.)

14.2.4.1 says, "If the conditional expression evaluates to x or z, it shall
be treated as true." Again, in context that is probably the correct behavior,
but it is different than elsewhere.

Shalom

Additional Information

This was imported from BTF650:
http://www.boyd.com/1364_btf/report/full_pr/650.html [^]

Issue History
Date Modified	Username	Field	Change
2005-12-18 23:54	doug_warmke	New Issue
2005-12-18 23:54	doug_warmke	Type	=> Errata
2005-12-18 23:55	doug_warmke	Additional Information Updated
2005-12-18 23:56	doug_warmke	Summary	9.7.6 does not explicitly say what happens if "wait" condition is => 9.7.6 does not explicitly say what happens if "wait" condition is 'x' or 'z'
2006-02-18 13:17	Brad Pierce	Reporter	doug_warmke => shalom
2010-02-08 21:46	shalom	Type	Errata => Clarification
2010-02-08 21:46	shalom	Severity	feature => minor
2010-02-08 21:46	shalom	Category	V-1364 => SV-BC
2010-02-09 00:16	shalom	Issue Monitored: shalom

Relationships