Rich, Dave wrote:
> This should be related to http://www.eda.org/svdb/bug_view_page.php?bug_id=0000026
>
> I think the answer should be yes. See the thread starting at http://www.eda.org/sv-bc/hm/1783.html
>
> Dave
First I think we should agree to extend the BNF so that enum literals can
be accessed via their type name. Where we now allow "class_scope", we can
read the enum's type identifier ("t" in Paul's example) as a class name
and thus permit t::blue to be an unambiguous reference to the enum literal.
Of course we cannot confuse enums and classes, so a separate production rule
for enum_scope is probably necessary, one which includes finding enum types
wherever types are sold.
With this syntax in hand, we are then free to say NO to letting an enum
type parameter accidentally pollute the module namespace. I think Paul's
choice of "blue" was apt: identifiers whose meaning emerges "out of the blue"
are a menace to readability. This choice also saves the committee from
writing about what happens when TWO enum types become type parameters of
the same module.
Greg Jaxon
>
>
> -----Original Message-----
> From: owner-sv-bc@eda.org [mailto:owner-sv-bc@eda.org] On Behalf Of Paul Graham
> Sent: Wednesday, October 06, 2004 12:24 PM
> To: sv-bc@eda.org
> Subject: [sv-bc] visibility with parameterized types
>
> Does using a parameterized type have an effect on visibility? For instance:
>
> module sub #(parameter type t = int)
> (input t x, ...);
> ...
> endmodule
>
> module top();
> typedef struct { bit f1, f2; } mystruct;
> typedef enum {red, blue, green} color;
>
> sub #(mystruct) u1((...);
> sub #(color) u2(...);
> endmodule
>
> If sub is instantiated with type mystruct, then within sub are the fields f1 and f2 visible?
>
> assign y = x.f1; // ??
>
> If sub is instantiated with type color, then within sub are the colors red, blue, and green visible?
>
> assign y = (x == blue); // ??
>
> These components of the types are not directly visible within sub, but I wonder if they are somehow imported through the type parameter.
>
> Paul
>
>
Received on Wed Oct 6 14:26:16 2004
This archive was generated by hypermail 2.1.8 : Wed Oct 06 2004 - 14:26:20 PDT