1 detector on detector seems to spuriously set B1 or B2 or at least
2 panic DR when we honour it
4 point setting doesn't seem to work
5 residual stuff from CDU is present and changes pt0 when you
7 * say `+' to master to turn on CDU
8 * switch to programming mode
9 pt0 on pic2 is the one nearest the switch box, moving to hacker's right
10 (we think - I haven't checked this yet)
12 crash readout: master i2c should go into known state
14 layout polarity diagram colourful segment encoding:
15 make moveable feature selection in subseg2display bitwise rather than
18 safety system no crashes at crossovers please