Flight Computer erratic when switching control to non-root parts of docked vessels.

[ProtonM]

trafficstars

In a docked vessel, when switching to a core that has an orientation other than the vessel's root the Flight Computer's commands cause the vessel to spin uncontrollably. Does not occur on non-docked vessels (radial or reversed secondary cores work properly). Issue occurs in .23.5 and .24 in clean savegames in clean installs of KSP with RT build 32.

[Starstrider42]

Possibly related to #25. If you could post a savegame file that would make investigating it a lot easier.

[ProtonM]

Here's the link: https://www.dropbox.com/s/torxso0n8lluyhf/RT%20Issue%20134.sfs 0.24.2, clean game, develop branch of RT. And a couple more things. The flight computer will hold orientation correctly in some cases namely:

• If you set the attitude to GRD+ or NML+, then disable the FC, then switch to the opposite pod, then enable FC with the opposite setting (GRD- or NML-, respectively) the vessel will start rolling but will keep the nose pointed along the correct vector.
• If you set the attitude to RAD+, then disable the FC, then switch to the opposite pod, then enable FC with the RAD- setting the vessel will correctly keep pointing the nose at RAD- without any abnormal attitude changes. If you disable FC and change your attitude by a small amount the vessel will still return to pointing at RAD-. However if you point your nose more than exactly 90 degrees away from RAD- the vessel will start spinning along all three axes and never reach the desired attitude.

[JDPKSP]

This might be due to the flight computer using the wrong reference transform. It shouldn't use the vessel transform, but the vessel.referencetransform transform.

[Starstrider42]

Possibly related to KSP-KOS/KOS#272, since we share attitude control code with them.