RE: [sv-bc] query regarding array assignment pattern

Both assignments are legal. You can think of integer as built-in
shortcut for

 

typedef logic signed [31:0] integer;

 

Type matching means you can treat an integer as if were a packed array
even though you can't declare it as one.

 

 

The restriction on packed arrays of the predefined types is entirely
lexical for two reasons

1.	De-facto implementation of Verilog ignore the range on a
declaration like
	integer [31:0] A;
	This was a Verilog-XL bug turned into a feature that many people
used to denote the size of the integer, but is ignored by most
implementations. Changing it to mean a packed dimension would not be
backwards compatible with the de-facto standard
2.	Declaring a packed dimension on any fixed sized data type would
be visually confusing.
	int [15:0] A;
	looks too much like declaring a 16-bit integer

 

Dave

 

 

 

 

 

________________________________

From: owner-sv-bc@server.eda.org [mailto:owner-sv-bc@server.eda.org] On
Behalf Of Subhamoy Pal
Sent: Thursday, June 01, 2006 9:56 PM
To: sv-bc@server.eda.org
Subject: Re: [sv-bc] query regarding array assignment pattern

 

Sharmistha Rakshit wrote:

Hi, 
I have query regarding the following testcase:: 
module top; 
   integer a1 ,a2 ; 
   initial 
   begin 
       a1= '{0:1'b0,3:1'b1,default:1'b0}; 

I have a more basic query regarding this LRM testcase. The right hand
side expression is an array assignment pattern. For array assignment
pattern LRM 8.13.1 says 
"Each expression item shall be evaluated in the context of an assignment
to the type of the corresponding element in the array". 

From this statement it can be inferred that the left hand side for this
perticular case is an array. But LRM 5.2 also says:
"Integer types with predefined widths cannot have packed array
dimensions declared. These types are byte, shortint, int, longint, and
integer. Although an integer type with a predefined width n is not a
packed array, it matches (see 6.9.2), and can be selected from as if it
were, a packed array type with a single [n-1:0] dimension."

My question is that whether this LRM example (of assigning an array
literal to a integer type) is really correct one or not? 


       a2 ='{32{1'b1}}; 
   end 
endmodule 

Now by LRM , the  first assignment is valid. 
and LRM also states that 
"A syntax resembling multiple concatenations can be used in array
assignment patterns as well" 
So , shouldn't the second assignment be also valid ? 

Thanks, 
Sharmistha 


-- Subhamoy