test(physics): #116 shape-2 — un-skip D4 airborne wall hard-stop pin

Per docs/research/2026-07-30-ts4-116-oracle-plan.md §3, §3.3 step 1.
Test-only change with zero production code in this commit: TS-4's
retirement (5e2be19b, the previous commit) is what actually unblocks
this test's routing.

The plan's confirming instrumentation (probes added in 5e2be19b: which
BSPQuery.cs path fires, whether Path 4's FindWalkableInternal finds a
walkable candidate) traced D4's tall-vertical-wall scenario: Path 6 fires
(SetCollide, no reposition, Adjusted) -> the retry routes to Path 4
(find_walkable), which finds NO candidate for this sheer wall
(changed=false, confirmed) -> Path 4 returns OK -> TransitionalInsert's
Phase 3 sp.Collide block runs (reachable now that Phase 1/2 both return
OK): ContactPlaneValid is false (first airborne contact) so the reset
branch fires, LastKnownContactPlaneValid is false too (first frame), so
SetCollisionNormal(sp.StepUpNormal) runs and the function returns
Collided -- a hard stop, in place, with the wall's real normal. This
exactly reproduces the test's original (pre-Skip) expectation, confirming
the oracle plan's §3.1/§3.2 structural finding without a live cdb trace.

Full AcDream.Core.Tests suite: 4061 passed / 1 skipped (the remaining
skip is the unrelated Pvs_CottageInterior_MatchesRetailCellDrawList),
no regressions.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

tests/AcDream.Core.Tests/Physics/BSPStepUpTests.cs

@@ -549,18 +549,33 @@ public class BSPStepUpTests
     /// does this with transient_state bit 2 + InitSlidingNormal. Without that,
     /// every frame replays the same hard stop and the character hangs in falling
     /// animation until another correction breaks the loop.
+    ///
+    /// <para>
+    /// #116 shape-2 RESOLVED (Campaign P final physics slice, 2026-07-30;
+    /// docs/research/2026-07-30-ts4-116-oracle-plan.md §3, §3.3 step 1). This
+    /// pin was `Skip`-tagged because the engine slid IN-FRAME to Z=1.92 on
+    /// the first airborne wall frame instead of hard-stopping at Z=2.0. The
+    /// confirming instrumentation (a probe on which BSPQuery.cs path fires
+    /// and whether Path 4's FindWalkableInternal finds a candidate) showed:
+    /// Path 6 fires (`SetCollide`, no reposition, `Adjusted`) → the retry
+    /// routes to Path 4 (`find_walkable`), which finds NO candidate for this
+    /// sheer vertical wall (`changed=false`, confirmed) → Path 4 returns `OK`
+    /// → TransitionalInsert's Phase 3 `sp.Collide` block runs (finally
+    /// reachable since Phase 1/2 both returned OK): `ContactPlaneValid` is
+    /// false (first airborne contact), so the reset branch fires,
+    /// `LastKnownContactPlaneValid` is false too (first frame), so
+    /// `SetCollisionNormal(sp.StepUpNormal)` runs and the function returns
+    /// `Collided` — a hard stop, in place, with the wall's real normal. This
+    /// exactly matches this test's ORIGINAL (pre-Skip) expectation. Un-skipped
+    /// as a test-only change (landed one commit after TS-4's own retirement,
+    /// 5e2be19b, which is what actually unblocks Path 6's faithful
+    /// `SetCollide` for a single-sphere mover on a steep/vertical wall):
+    /// the routing this test now pins was already correct in
+    /// `BSPQuery.cs`/`TransitionTypes.cs`; only the now-deleted shortcut was
+    /// masking it.
+    /// </para>
     /// </summary>
-    [Fact(Skip = "Issue #116 shape-2 — the engine slides IN-FRAME to Z=1.92 " +
+    [Fact]
-        "on the first airborne wall frame; this pin expects an L.2c hard stop " +
-        "at Z=2.0. Ghidra (2026-06-12) confirms retail CSphere::slide_sphere " +
-        "(0x00537440) applies the slide IN-FRAME (add_offset_to_check_pos → " +
-        "SLID_TS), so our 1.92 is faithful TO slide_sphere and the Z=2.0 " +
-        "expectation is the SUSPECT half — but whether retail's first " +
-        "airborne frame REACHES slide_sphere (→1.92) or hard-stops upstream " +
-        "(collide_with_environment dispatch / no last-known plane) needs a " +
-        "cdb trace of an airborne wall hit before flipping the assertion. The " +
-        "#116 threshold fix (EpsilonSq→F_EPSILON) did NOT change this — the D4 " +
-        "offset is a real slide, not degenerate. See docs/ISSUES.md #116.")]
     public void D4_AirborneMover_TallWall_PersistsSlidingNormalAcrossFrames()
     {
         var (root, resolved) = BSPStepUpFixtures.TallWall();