Thanks so much for your detailed answer, Shonky!

The fact that the behavior is "undefined", but unlikely to be a problem, is excellent news for me.

Double checking the schematic... By Jove, you're right, he lists the 3.3v part number for that linear regulator, but the schematic says its output is 2.5v.

I think the mistake is the part number on the schematic. I'm pretty sure he's using a 2.5v part in that location, and he's doing it on purpose. Not entirely certain why, but my assumption has been that it's to use less battery. I think he just forgot to update the part number when he made that change. I'll notify him about it now.

The reason it still says "3v3" on the pin of the schematic on the "P1" connector is because that's the breakout pin for "VDD_IO" on the BlueGiga chip, which nominally accepts 3.3v. But from what Mark and I could tell, he's deliberately only putting 2.5v there with his choice of linear regulator chip.

I just retested the three boards I have, which have all spent quite a lot of time with 3.3v applied to either side of that linear regulator, and on all three boards, the regulator is still working fine when I switch the board back to its default mode. I think if it was going to short out, all three of them would have done so by now. So no worries!

Thanks so much!