Occasionally a clock smith will repair a broken spring using rivets
(scab on). I suspect the stored E is mainly shear (bending) instead
of tension. Enormous, comparatively, force is necessary to stretch
w/ such a large cross section of spring steel. Therefore, not much
stress on the broken part . The rivets will be under mostly
tension? Am I correct? The reason I ask is a poster on a clocks
group thinks "this is a recipe for disaster". The energy stored is
rather great, and when one goes, if not ina barrel, ruins the clock.
Occasionally barrels are broken.
p.s. We'll possibly know as a Polish list member has done the repair
using screws in a (rare?) multiple face clock (master w/ slave faces)
being unable to find a replacement.