I do think the spigotless barrels are preferable since the spigot does NOT seal the surface, and the clever machining required around the top of the barrel to make the spigot seems suspect to its flatness and RMS and cutter finish.
There is no reason to doubt the spigots were flat and maybe still are. The warping occurs in the cylinder head. I am not sure who introduced sealing by the spigot plus the head gasket, or if it was ever intended. It's almost an impossible feat to make two stepped surfaces exert the same seating pressure.
I believe the original idea was taken from the single racing engines. Those engines could be operated without a head gasket, just careful machining of the head to match the spigot, completed by lapping. This procedure was used on the Matchless G45 and the G80CS (and likely on the 7R/G50) and has the aadvantage of giving a very high area pressure with no risk of blowing a gasket, which could otherwise be a lethal weakness in engines having a high compression ratio.
For a production bike, this procedure is too labor intensive. AMC kept building bikes with spigots, however all roadsters had head gaskets, which indicates the factory eliminated effective sealing by the spigot, but retained the spigot for racing use.
This makes it clear that where a spigot and a copper gasket is present, clamping is NOT to take place at the spigot but rather outside of it, giving unique pressure onto the head gasket. It's all right to fit NM25494, but you have to make sure it doesn't compress.
Piero: Skimming for flatness need to be performed in the ring groove as well as on the outer surface. When fitting the conversion ring(s) with a vertical clearance, the gap will soon be filled with some carbon deposits and there is no need to worry about the clearance.
-Knut