[sv-ec] RE: multi-dimensional arrays

I actually think that it is intuitive, logical, and powerful to think of
a multi-dimensional array as having a dual nature. It is both a
collection of elements and an array of arrays. I don't see a
contradiction. How you think of it depends on the context. I agree that
there are some cases which would be ambiguous to the reader unless more
explicitly specified.

Since this issue has come up several times, the LRM must describe this
more clearly.

As an example of a genuine ambiguity in the LRM, see
http://www.eda.org/mantis/bug_view_advanced_page.php?bug_id=0001087.

Shalom 

> -----Original Message-----
> From: owner-sv-ec@eda.org [mailto:owner-sv-ec@eda.org] On
> Behalf Of Brad Pierce
> Sent: Tuesday, January 24, 2006 8:18 PM
> To: sv-ec@eda.org
> Subject: Re: [sv-ec] foreach iterates over unpacked arrays
> 
> Sensible perhaps, but not legal.  Should have been
> 
>     foreach(A[i,j])
> 
> -- Brad
> 
> -----Original Message-----
> From: owner-sv-ec@eda.org [mailto:owner-sv-ec@eda.org] On
> Behalf Of Brad
> Pierce
> Sent: Tuesday, January 24, 2006 10:14 AM
> To: sv-ec@eda.org
> Subject: Re: [sv-ec] foreach iterates over unpacked arrays
> 
> I agree that the text of the LRM is not always as precise as it
> could
> be.  But if A is an SV "multidimensional" array, then
> A[$left(A)] is an
> array.  And, more to the point, if $dimensions(A)==3, then
> foreach(A[i][j]) is both legal and sensible.
> 
> -- Brad
> 
> -----Original Message-----
> From: Bresticker, Shalom [mailto:shalom.bresticker@intel.com]
> Sent: Tuesday, January 24, 2006 9:49 AM
> To: Brad Pierce; sv-ec@eda.org
> Subject: RE: [sv-ec] foreach iterates over unpacked arrays
> 
> So how do you explain statements like,
> 
> "$dimensions shall return the following:
> - The total number of dimensions in the array (packed and
> unpacked,
> static or dynamic)" ?
> 
> Shalom
> 
> > -----Original Message-----
> > From: owner-sv-ec@eda.org [mailto:owner-sv-ec@eda.org] On
> > Behalf Of Brad Pierce
> > Sent: Tuesday, January 24, 2006 7:38 PM
> > To: sv-ec@eda.org
> > Subject: Re: [sv-ec] foreach iterates over unpacked arrays
> >
> > Shalom,
> >
> > I proposed the language I just quoted from P1800.6.9.1.f.
> Yet
> > the fact
> > remains, as noted in http://www.eda.org/sv-bc/hm/3601.html ,
> > that
> > SystemVerilog does not have truly multidimensional arrays.
> >
> > -- Brad
> >
> > -----Original Message-----
> > From: Bresticker, Shalom [mailto:shalom.bresticker@intel.com]
> > Sent: Tuesday, January 24, 2006 8:40 AM
> > To: Brad Pierce; sv-ec@eda.org
> > Subject: RE: [sv-ec] foreach iterates over unpacked arrays
> >
> > Brad,
> >
> > We've been through this before.
> >
> > There are dozens of places in the LRM which refer to arrays
> as
> > multi-dimensional, including a number which contrast them to
> > one-dimensional.
> >
> > Shalom
> >
> >
> > > -----Original Message-----
> > > From: owner-sv-ec@eda.org [mailto:owner-sv-ec@eda.org] On
> > > Behalf Of Brad Pierce
> > > Sent: Tuesday, January 24, 2006 6:10 PM
> > > To: sv-ec@eda.org
> > > Subject: Re: [sv-ec] foreach iterates over unpacked arrays
> > >
> > > According to P1800.10.5.3, "The number of [foreach] loop
> > > variables must
> > > match the number of dimensions of the array variable. Empty
> > > loop
> > > variables can be used to indicate no iteration over that
> > > dimension of
> > > the array, and contiguous empty loop towards the end can be
> > > omitted.".
> > >
> > > I'm not convinced there's a problem with this text.  But if
> > > there is,
> > > it's that it's talking about the dimensions of the array
> > > variable,
> > > instead of the dimensions of the array being iterated over.
> > > Recall that
> > > SystemVerilog does not have true multidimensional arrays,
> it
> > > has arrays
> > > of arrays.  As it says in P1800.6.9.1.f, "The type of the
> > > slowest
> > > varying dimension of a multidimensional array type is
> itself
> > an
> > > array
> > > type."
> > >
> > > -- Brad