Right now we are thinking that peculiar circumstances led to the extreme effort to resize and annealing the shoulder to solve the problem. The new press and segregating your brass may make it a non problem.
Now we have a theoretical quandary on did the shoulder length change with #213000. I did notice the SAAMI chamber dimension notes the chamber shoulder has a .155 MAX radius. I guess .002 is less than max and would be much easier to measure. I sold my 35/40 reamer to a new KCA member so I can not verify it but I remember a miniscule or nonexistent radius on that corner. May have been a SA practice to make the corner with a minimal radius when new and hone a radius on the corners as it got dull. When .155 R was reached the reamer was discarded.

Just measured the shoulder depth on a loose Krag barrel and got 1.669 calculating the shoulder depth from SAAMI using a .064 rim thickness I got 1.6644 with a plus or minus tolerance of  .015. Kind of odd to have a plus or minus tolerance of .015 from a four place decimal dimension, but 1.669 measured is in tolerance. Way different than the 1.62 on the Frank Mallory sourced drawing. The mystery is getting deeper.
The loose barrel just measured was from #410864 so it was later production, #266389 is similar but it is assembled and the measurement is not exact.

---