>>2862859surprisingly, almost exactly as planned (in the second try, first try the automatic weight renormalization fucked it up, second try I set weight=uv/4 so the sum would never exceed 0.5)
the rest I can adjust in UV Editor, so I'm happy with it