RE: [sv-ac] Final proposal for 196

Adam,

Thanks for your input. I added disallowing the anonymous declaration as you state.
The example shows how to work around the fact that 'pass by value' does not exist,
why do you recommend dropping it.

Thanks
Hillel

-----Original Message-----
From: Adam Krolnik [mailto:krolnik@lsil.com]
Sent: Monday, November 22, 2004 7:17 PM
To: Miller Hillel-R53776
Cc: sv-ac@eda.org
Subject: Re: [sv-ac] Final proposal for 196

Hi Hillel;

Thanks for continuing on with this proposal - it still is a good thing to have.

You wrote:

data_type ::=
   integer_vector_type [ signing ] { packed_dimension }
   | integer_atom_type [ signing ]
   | struct_union [ packed [ signing ] ] { struct_union_member { struct_union_member } }
     { packed_dimension }
   | enum [ enum_base_type ] { enum_name_declaration { , enum_name_declaration } }
   | string
   | type_identifier { packed_dimension }

I do not believe it will be useful to allow defining anonymous struct, union or
enumerations. If I am correct, the type matching rules do not allow independent
anonymous declarations to match.

For example, I believe that this should result in a warning or an error (which I don't
recall.)

struct packed { logic f1 } var_a;

property simple_prop(
   struct packed { logic f1} param_1;
   )
  ...
endproperty

assert property (simple_prop(var_a)); // This is a type casting/matching error.

These anonymous structures/enumerations would only be useful within the sequence or
property as a declaration for a local variable.

Also, I recommend dropping the example about pass by value as it is not relevant to
the proposal as it stands now.

    Thanks.

    Adam Krolnik
    Verification Mgr.
    LSI Logic Corp.
    Plano TX. 75074
    Co-author "Assertion-Based Design"
Received on Mon Nov 22 09:39:07 2004