ModuleROEjectionSeat logspam/framerate drop on EVA failure.

[Dunbaratu]

To recreate:

1. Put something on the cockpit that will cause the pilot to be unable to EVA, like obstructing the hatch with a thermometer.
2. Fly a plane with that cockpit.
3. Trigger abort action group in flight (which causes ModuleROEjectionSeat to try to EVA the pilot).

Result:

This message gets log spammed:

"ROEEjectionSeat] It appears that nobody was EVA'd"

While this message gets screen-spammed:

"Hatch is Obstructed, Cannot Exit"

and the frame rate drops. A LOT. The more seconds you let it run the slower it gets. Something about re-attempting the EVA again and again very Update is tanking the frame rate. (Since it gets worse the longer it goes, I wonder if it's the accumulated screen messages being created faster than they decay.)

Should the system be changed to only fire off when there's an edge-detect in the boolean state of Abort? (only when it first becomes true, or first become false, so that one keypress = one attempt and you press the key again to try again rather than it retrying again on its own in the background?)

(I Noticed this in RP-1. I am assuming the issue belongs here, because RealismOverhaul has ModuleROEjectionSeat, but it's possible it belongs in the RP-0 github repo. I have a hard time telling when something is caused by RealismOverhaul versus caused by a thing RP-1 did on top of RealismOverhaul.)