Consider row readout spoilers

[taldcroft]

Potential C&R spoilers are planets, stars, and FID lights (hot pixels?), which are brighter than the candidate by a TBS magnitude.

From email earlier:

The case of a downstream column spoiler is the most common and reasonably easy to model (pixels go under the spoiler star and integrate for the row readout time). I don't actually understand the physical mechanism for Row spoilers and upstream column spoilers, but presumably relates to non-linear effects in the readout electronics. Maybe? Probably just a simple rule like nothing more than 5 mags brighter might do the trick. The cases of upstream spoiling that I can recall were Jupiter which is both extremely bright and extended (lots of electrons get injected, as opposed to a star that just saturates a few pixels).

[jeanconn]

For guide star selection, I've got the check for column spoilers in https://github.com/sot/proseco/pull/23/commits/6748e1c3f3039b87f726a42bd839e13f95a2984c (but that is only checking for stars from the AGASC that are spoilers).

I note that, try as I might, I could not construct a case that would hit the readout register spoiler check in SAUSAGE. There, the checks are implemented as

rhs_register = ( (max(ypixlim) - Register.Width - register_pad) < ypos ) & ~off_chip;
lhs_register = ( (min(ypixlim) + Register.Width + register_pad) > ypos ) & ~off_chip;
in_register  = find(rhs_register | lhs_register);

but as far as I could tell, the padding on the readout register for the check and the padding when defining "off_chip" meant that this check never actually caught any stars (any star actually in the readout register was marked off_chip and not caught by the check). So if we want to check for readout register spoilers for guide and acq we'll need to look at this from first principles again.