Posibly the hard way, but is ther any mileage in measuring the shaft diameter, and finding a set of steel backed shell bearings from, say a car or motorcycle engine?
Having found your bearings, you then machine up a bush to fit the parent bore in the Headstock (Shim for a perfectly round bore when measuring ) You then need to make a bush, to tolerance of "Headstock Bore +0.001 / 0.000" which can be split and clamped in a collet bore which is machined to the OD of the shell bearings + 0.000 / -0.001"round or -0.001" ) .
The shell bearings are then placed in the split bush (which will need slots filing for the retaining tangs of the shells ), and the half bush / shell assemblies placed in the headstock, oiled, and the caps tightened down.
This will not be suitable if the journals are badly scored or have worn to a taper. Correcting that by turning down undersize puts you back in Catch 22 again, plus the need to find shell bearings of a suitable size.
Which brings us back to using the lathe "as is" to machine bushes, from phosphor bronze perhaps, to fit the Headstock parent bores, before being split, clamped in a collet and fine bored to fit the shaft with minimum clearance.
Helical oil grooves could be carefully "screwcut" before being removed for fitting, before finish boring, possibly.
The grooves do not need to be particularly deep, 0.020" should be quite adequate, to spread the oil.
If this works you are in business again. Non standard, but as already said, How frequently do you expect to have to replace the bearings? If it is a ML7 it could have taken over 70 years to wear out the originals
Howard