RE: [sv-bc] const constants

From: Bresticker, Shalom <shalom.bresticker_at_.....>
Date: Wed Feb 15 2006 - 03:45:07 PST
Brad,

Clause 7 is Classes. The subject of 7.18 is constants in classes.
Information about const's which is not specific to classes should not
appear only in 7.18. When I study SV design constructs, I don't read
about classes.

The same problem occurs with respect to statement labels, where the
information that they create scopes only appears in the clause on
Assertions.

7.18 says that const variables "cannot be assigned a value anywhere
other than in the declaration", which is a stronger restriction than you
stated. That information should be stated in 6.3.5.

Thanks,
Shalom
 

> -----Original Message-----
> From: owner-sv-bc@eda.org [mailto:owner-sv-bc@eda.org] On
> Behalf Of Brad Pierce
> Sent: Wednesday, February 15, 2006 1:21 AM
> To: sv-bc@eda.org
> Subject: Re: [sv-bc] const constants
> 
> Shalom,
> 
> The LRM would be clearer if it referred to 'const'-qualified
> variables
> as "write-once variables", instead of as "constants" or "read-
> only
> variables" (7.18).
> 
> Each instance of an SV write-once variable must be written
> exactly once,
> after the design has been elaborated and before the variable
> instance is
> read.  (By saying 'instance' here, I am trying to take into
> account that
> a new instance of an automatic variable is created each time
> its scope
> is entered at run-time.)  It would be a help to state this rule
> more
> explicitly.  In particular, 7.18 should clarify that is a run-
> time error
> to read an "instance constant" before it is written.
> 
> -- Brad
Received on Wed Feb 15 03:45:15 2006

This archive was generated by hypermail 2.1.8 : Wed Feb 15 2006 - 03:48:41 PST