Kathy, I think the group had talked about not just leaving out the data type
entirely, but also leaving out the base type for a packed array, as in
var [15:0] vw; // equivalent to "var logic [15:0] vw;"
This would make it completely consistent with net declarations. Is that
consistent with what other people wanted? If so, it would be good to
add an example of this also.
This might create a little complication in the BNF. I don't think Brad's
BNF allows leaving out the data type (partially or entirely), despite the
fact that Brad was the one who brought up this short-hand potential in
the meeting :-). There is a little complication in this, since you can't
allow leaving out both the data type and the "var" keyword. The production
[var] data_type_or_implicit list_of_variable_decl_assignments;
could expand into
[15:0] r;
or just
my_variable;
So there would need to be a semantic restriction against leaving out
both "var" and the data type. Or the grammar would need to exclude it
by having two separate productions. Perhaps
[var] data_type list_of_variable_decl_assignments; |
var [signing] {packed_dimension} list_of_variable_decl_assignments;
or
[var] data_type list_of_variable_decl_assignments; |
var data_type_or_implicit list_of_variable_decl_assignments;
Steven Sharp
sharp@cadence.com
Received on Thu Nov 18 15:34:48 2004
This archive was generated by hypermail 2.1.8 : Thu Nov 18 2004 - 15:34:51 PST