Hi David,
What I mean is that I thought alterBroken just affected the single shot hairpin (wrong). In my score I have about 200 pages where there are hairpins ending on dynamics on the next page, and after the erroneous call to \alterBroken to adjust Y-offset, only some hairpins 80 pages or so away were affected (and no wonder I had forgotten about the alterBroken in the meantime).
It's not a suspect, it's the perpetrator beyond doubt. The issue is resolved for my practical purposes, but I still wonder is this is a bug. Maybe I can cook up an MWE now.
Andrew