Waveform parity: off-by-one change counts often mean the counting rule differs

When comparing per-signal change counts between two waveform tools (a converted file vs the source reader's own stream), check the counting rule before blaming the converter. One signal differed by exactly one (N vs N-1). Dumping the source reader's raw value stream showed exactly one place where the same value appeared on two consecutive lines. The 'strict' count kept it, a 'normalised' count that also drops repeats removed it, and the converted file matched the strict count exactly. So report every count with its rule named: strict (collapse same-time writes only) vs normalised (also drop consecutive equal values). For any mismatch, first dump the raw stream and look for consecutive repeats. Also check name totals for tool export omissions: one VCD exporter silently skipped a string-typed variable, which looked like a missing signal.