re(flight): rows pinned by world-Y, and nothing yaws -- the yaw gap is answered

The probe now pins which non-forward row is up and which is right, instead of
taking the D3D convention on faith:

    row world-Y means: [0.469, 0.883, -0.000]   forward = row 2
      -> up = row 1, right = row 0   CONFIDENT

That is the OPPOSITE of the assumption the previous run used, so that run's yaw
and pitch columns were swapped -- under the correct labels its ly+ reading of
154.1 deg/wall-s is PITCH, which is what a left-stick Y should do.

Measured with the unknown inputs FIRST, each passing a liveness check:

    rx  0.0  0.0  0.0      ry  0.0  0.0  0.0
    LB  0.0  0.0  0.0      RB  0.0  0.0  0.0
    lx  roll 209.8         ly  pitch 154.1

These zeros are trustworthy where the previous run's were not: the craft was
verified alive between inputs, and the probe aborted the moment it stopped moving
rather than reporting the clean zeros a destroyed craft produces (it did abort,
after lx+, which is why lx/ly are carried from the earlier run rather than
re-measured).

So NO PAD INPUT YAWS THE CRAFT. AV_Yaw_* (45/25) exists in the definitions but
nothing on the right stick or the shoulders drives it, which upgrades the old
"yaw: no input found" from a failure to find one into a measurement that the
remaining candidates do nothing. The natural reading is that yaw is a consequence
of banking rather than a commanded axis.

Not covered, and not claimed: the d-pad (tactical map) and the face buttons
(fire/weapon select). Neither is a plausible flight axis; neither was measured.

docs/re/captures/axis-probe-rows-pinned.csv

@@ -0,0 +1,5 @@
+input,roll_deg_s,yaw_deg_s,pitch_deg_s
+rx+,0.0,0.0,0.0
+ry+,0.0,0.0,0.0
+LB,0.0,0.0,0.0
+RB,0.0,0.0,0.0