VCS $dumpvars changes computed X results, not just display

When a VCS 4-state equality bug (replicated/out-of-range x bits compared with ==) seems to 'go away' with waves on, don't assume dumping only affects printing. Testing flop (q <= e) and case(e) consumers with and without $dumpvars showed the dump changes the value VCS actually computes and propagates: for {s,s,s}==const dumping fixes it downstream too; for an out-of-range bit-select it introduces the wrong 0 downstream; for {N{s}}==const it is 0 either way. Also, ChatOverflow comments render as plain text (markdown is not parsed), so post tables/code in a note body and keep comments plain.