feat(astro): specific orbital angular momentum from elements

[nochallenge]

What

Adds orbit::specific_angular_momentum_from_elements(semi_major_axis, eccentricity) — the conserved specific orbital angular momentum:

// h = sqrt(mu * a * (1 - e^2)) = sqrt(mu * p)   (m^2/s)
pub fn specific_angular_momentum_from_elements(semi_major_axis: f64, eccentricity: f64) -> Result<f64, AstroError>

Why

The orbit.rs conic-elements cluster has eccentricity_from_apsides (#372), semi_major_axis_from_apsides (#378), and semi_latus_rectum_from_elements (#384), but not the angular momentum. h = sqrt(mu*a*(1-e^2)) is the conserved angular momentum per unit mass (twice the areal velocity, Kepler's 2nd law). Equivalently sqrt(mu*p) with p the semi-latus rectum; at the apsides h = r*v (r and v are perpendicular there); and for a circle (e=0) it reduces to circular_angular_momentum sqrt(mu*a) (#348). Guard mirrors semi_latus_rectum_from_elements (a > 0, 0 <= e < 1).

Test

specific_angular_momentum_from_elements_is_sqrt_mu_p:

• (a) worked — a=7e6, e=0.1 -> h = sqrt(mu*6.93e6);
• (b) cross-check threading semi_latus_rectum_from_elements (chore(deps)(deps): bump sha2 from 0.10.9 to 0.11.0 #384) — h^2 = mu*p across three (a,e);
• (c) circular cross-check threading circular_angular_momentum (feat(astro): add orbit::circular_angular_momentum (h = sqrt(mu*r)) #348) — h(a,0) = sqrt(mu*a);
• (d) periapsis cross-check threading orbital_speed — h = r_p*v_p;
• (e) error guards — a<=0, e>=1, non-finite.

Free fn in orbit.rs (reached via valenx_astro::orbit::), no lib.rs change. valenx-astro 221 lib tests (was 220); cargo clippy -p valenx-astro --all-targets -- -D warnings clean. Research-grade two-body primitive.