I don't think anyone disagrees with the need to standardize the conventions for generating names of variables declared in unnamed blocks. That should be an issue for the sv-cc. There is no way to use those names from the source text. But I disagree with the notion that the occurrence of declaring a variable in an unnamed block creates a scope. Consider a 'for' loop with a local declaration. I don't think users expect that to hide the nested named blocks in the 'for' loop. It's too late to debate whether it was a good idea to allow variables to be declared in unnamed blocks. It will be in the IEEE-1800 standard. You can choose not to implement it, and another PAR can remove it later. Dave > -----Original Message----- > From: Krishna Garlapati [mailto:krishna@synplicity.com] > Sent: Thursday, May 19, 2005 1:31 PM > To: Steven Sharp > Cc: Rich, Dave; mac@verisity.com; stuart@sutherland-hdl.com; > Mark.Hartoog@synopsys.com; Shalom.Bresticker@freescale.com; sv-bc@eda.org > Subject: Re: [sv-bc] Naming of unnamed sequential blocks > > > I agree with Steven Sharp. My personal opinion is that declarations > should not be allowed in unnamed blocks. And, if declarations are > allowed, the LRM should specify a naming convention that is consistent > across the whole language. > > Inconsistent naming has (and still is) been big problem here when a > user runs equivalence check on the synthesis output and it fails due > to a name mismatch. > > - Krishna. > > > Steven Sharp wrote: > > Mark Hartoog wrote: > > > >>The hierarchical name 't.blk.y' is perfectly valid. Now if I modify this > >>design to System Verilog and add a variable in the enclosing unnamed > block: > >> > >>module test; > >>task t; > >>input x; > >>begin > >> reg z; > >> if (x) begin : blk > >> reg y; > >> end > >>end > >>endtask > >>initial $display(t.blk.y); > >>endmodule > >> > >>Is the hierarchical name 't.blk.y' now illegal? > > > > > > In my opinion, yes. > > > > Declaring reg z is short-hand for naming the block and then > > declaring reg z inside it. You save the effort of naming > > the block yourself, but give up the ability to refer to the > > name. > > > > Steven Sharp > > sharp@cadence.com > > > > > > > > > -- > Krishna > 408-215-6152Received on Thu May 19 14:01:22 2005
This archive was generated by hypermail 2.1.8 : Thu May 19 2005 - 14:01:27 PDT