[sv-ec] RE: [sv-champions] 5-day email vote 1723

This discussion is from the champion's rejection of mantis 1723

I do believe the text in 7.9.4 should be modified. Index expressions
should simply be cast to the declared index type of the array. We don't
need separate local definitions of casting to int.

I still think the example is correct. If you sign extend an unsigned
literal, it gets zero filled. 7.9.4 does not consider what it means to
sign extend an unsigned type.

Since this mantis in now re-opened, I will try to address this without
opening up a new mantis item


Dave


P.S. The reason I changed the index type from [*] to [int] was because I
was hoping to discourage the use of wildcard index types as I think they
are no longer useful with the text that was added by mantis 1457.



> -----Original Message-----
> From: owner-sv-champions@server.eda.org [mailto:owner-sv-
> champions@server.eda.org] On Behalf Of Brad Pierce
> Sent: Tuesday, August 28, 2007 8:49 AM
> To: sv-champions@server.eda.org
> Subject: RE: [sv-champions] 5-day email vote 1723
> 
> Dave,
> 
> Why do we need to open/resolve another Mantis item, just to add a
single
> line to the example?  That's a lot of overhead for a simple editorial
> change.  Why not just upload an updated version of the proposal with
> that editorial change?
> 
> Thanks for pointing me to the 7.9.4 text about indices.  According to
> that text, all of these unsigned integral values are sign extended to
32
> bits, becoming negative ints.  Was that the intent of your change?
> 
> That 7.9.4 text also does not say how an indexing expression such
(1'b1
> + 1'b1) should be handled.
> 
> To me it looks like the text in 7.9.4 is wrong, and the indexing
> expression should be implicitly static cast to integer or int, that
is,
> the indexing expression should be evaluated in the context of an
> assignment to integer or int.
> 
> -- Brad
> 
> -----Original Message-----
> From: Bresticker, Shalom [mailto:shalom.bresticker@intel.com]
> Sent: Tuesday, August 28, 2007 4:01 AM
> To: Rich, Dave; Brad Pierce; sv-champions@eda.org
> Subject: RE: [sv-champions] 5-day email vote 1723
> 
> I agree with Dave on this one.
> 
> Shalom
> 
> > -----Original Message-----
> > From: owner-sv-champions@server.eda.org
> > [mailto:owner-sv-champions@server.eda.org] On Behalf Of Rich, Dave
> > Sent: Tuesday, August 28, 2007 9:40 AM
> > To: Brad Pierce; sv-champions@server.eda.org
> > Subject: RE: [sv-champions] 5-day email vote 1723
> >
> > >
> > >     http://www.eda-stds.org/svdb/view.php?id=1723
> > >
> > >         6) Needs an example of the new method.
> > [DR] This would be another mantis item.
> > >
> > >         7) Not clear that that the indexing expressions are still
> > legal
> > > after the change from [*] to [int].
> > >
> > [DR]From the LRM 7.9.4 Integer (or int) index
> > - Indices can be any integral expression.
> > - Indices are signed.
> > - A 4-state index containing X or Z is invalid.
> > - Indices smaller than integer are sign extended to 32 bits.
> > - Indices larger than integer are truncated to 32 bits.
> > - The ordering is signed numerical.
> >
> >
> > What's not legal?
> >
> > Dave
> >
> >
> > --
> > This message has been scanned for viruses and dangerous content by
> > MailScanner, and is believed to be clean.
> >
> 
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
> 


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