Changes

Fix the warning for hyperbolic target orbits

Hyperbolic orbits have an apoapis _radius_ <0; apoapsis _altitude_
between -MainBody.Radius and 0 is still elliptical. It is still wrong to
have the periapsis be greater than the apoapsis, though, but that is
handled by a different check (and shows a different warning message).
 (commit: f9185c9)

Fix clamping of attach altitude

It should be between peR and apR (if apR > 0). Also clamp apR to be <0
or >= peR.
 (commit: f679d4d)

Ignore attach altitude if its checkbox is disabled

Previously, for circular orbits, the attach altitude was still used as
final radius for the terminal constraints, even if the attach altitude
checkbox was _unticked_; it should be using the semi-major axis (or the
periapsis, or the apoapsis; all three are the same) of the orbit in that
case. This causes the final orbit to be incorrect if the target orbit is
(almost) circular.
 (commit: 3da86ba)

Warn when retrograde inclination requires dog-leg

A retrograde near-equatorial orbit requires just as much of a dog-leg as
a prograde near-equatorial orbit. Warn the user in the same way for
those (possibly unreachable) orbits.
 (commit: 53dfc32)

Launch-into-plane target must be in the same SoI

For example; launch into the plane of Mars, when launching from Earth,
takes Mars' inclination and LAN relative to the Sun; and launches to an
orbit with that inclination and LAN relative to Earth. This does not
result in an orbit in the "same plane as" Mars (which is in fact
physically impossible).
 (commit: b52b43b)

Slightly different fix for low eccentricity

Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
 (commit: ea92a83)