When I was first learning about these guns, the data I got was to rotate the barrel to the side and observe whether the barrel lug and frame line up evenly. Shim the arbor to make that happen. I don't do that now, except to see how far off the mating is once the arbor is shimmed.
How do you know when the arbor has bottomed out if the barrel isn't cocked to the side?
As an example, I just worked on an Armi San Paolo 1851 with a .035" endshake and a short arbor. The barrel lug and frame were mismatched by about the same amount. Perhaps this gun had been assembled with various parts lying around so nothing fit right. It was pretty gross. But the frame was too long. Removing the excess from the frame brought the endshake within limits at which time the arbor also bottomed.
Was there a different or better way to resolve these issues?
It seems simple now, but it took some thought to understand the force vectors of the barrel/frame/wedge.
I agree - having the arbor bottom into the barrel is the way to go from an engineering standpoint. But I don't see the necessity of having the barrel lug/frame mating much more than a relaxed fit (not stressed or loose). Firing forces want to compress that joint already, so there's no possibility of it opening up. Getting that connection to a point where it's not stressed takes some careful work, so I can see that in a lot of cases it's possibly best just left alone. In my case, I saw no other solution as driving the wedge in cocked the barrel quite a bit, left a big gap at the bottom of the forcing cone and still the endshake wasn't acceptable.