The broken version was 1.05, but I "slipstreamed" the fix into the same release without updating the version stamp.
I really should've bumped the version number when I posted the fix, but it'll be pretty obvious if your copy has the bug. It will appear to work, but it will always start the PN sweep at 0 Hz rather than the minimum offset you specify. That makes the sweep take much longer than necessary, and the trace information at offsets below 10x the analyzer's minimum RBW is meaningless.
Basically, if you downloaded it around the time I posted all those auctions, you probably got the bad version.
Heh... I've wondered where Microsoft came up with that; it might have come from the Unix world. There's a lot of weirdness in the calendar business, and it's usually less arbitrary, but more obscure, than it seems at first.
-- john, KE5FX