test(measure): the spine is the backbone in 7 of 8 strates -- and not in

VerticalShafts

Diagnosis only. No generation or placement changed.

The hypothesis was that the (0,0) spine (OriginSpineRadius 14) is the
universal connector, so "can this mouth reach the spine" would be the
real primordial-law test. Measured at margin 0, step 2, radius 192:

  strate           spine share  largest?  arr->spine   dep->spine
  TunnelNetwork      0.979        YES     CONNECTED    not connected
  FlatPlain          1.000        YES     CONNECTED    CONNECTED
  CrystalChamber     1.000        YES     CONNECTED    CONNECTED
  Maze               0.997        YES     coarse-lied  CONNECTED
  SurfaceWorld       1.000        YES     CONNECTED    CONNECTED
  VerticalShafts     0.096        NO      not conn.    not connected
  FloatingIslands    1.000        YES     CONNECTED    CONNECTED
  Underwater         0.999        YES     CONNECTED    n/a

The spine forms one uninterrupted open interior run in 8/8 strates, and
holds essentially ALL the air in 7 of them. VerticalShafts is the lone
anomaly: the spine punches through as an isolated column holding 9.6% of
the air and is NOT the largest component.

So the archetype has three mutually unreachable things -- the spine, the
arrival network, the departure network. It has no lateral connectivity at
all. Over 16 seeds: spine open 16/16, spine largest only 8/16,
arrival->spine 1 connected, departure->spine 0 connected.

Neither extreme was right. The spine is not a universal connector, but it
is far from irrelevant: it is the de-facto backbone everywhere except
here, which reframes the bug as an ANOMALY to bring in line rather than a
missing feature to invent.

Ordering hazard confirmed on the record: making passage i depend on
passage i-1 would break LayoutOrderIndependence. Both candidate fixes
avoid it, because the spine is a fixed deterministic anchor that two
mouths can target independently.

Build clean, 19 succeeded / 0 warnings / 0 failed.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Bp7FxoQcSoiuhxjuLr6aBM
This commit is contained in:
Fr0zkaandClaude Opus 5 committed 2026-08-30 15:10:00 +02:00
1 parent 0bbfdeb7a7
commit 5519b323c8
4 files changed
+903 -146

No files matched your search

+2 -2
View File
@@ -320,7 +320,7 @@ or PIE is required.
|--------|------|
| `FVoxelStrateMeasureSettings` / `FVoxelStrateMetrics` | Plain settings/result structs for bounded strate sampling and derived measurements; callers may override the interior margin (including zero to sample the seal), and results identify the resolved margin/Z window and deterministic component representatives. |
| `VF_MeasureStrate` | One-grid/one-flood-fill strate metrics; refuses invalid bounds or a grid over `MaxCells`. |
| `VF_AreConnected` | Coarse 6-connected BFS plus one full-resolution recheck of its recovered route, with explicit endpoint-solid, out-of-window, `CoarseLied`, `NotConnectedAtThisResolution`, and connected outcomes. |
| `VF_AreConnected` | Coarse 6-connected BFS plus one full-resolution recheck of its recovered route, with explicit endpoint-solid, out-of-window, `CoarseLied`, `NotConnectedAtThisResolution`, and connected outcomes; the diagnostics overload returns endpoint component facts from that same grid. |
| `VF_DiagnoseConnectivity` | The same query plus each mouth's air-component size/share and exact geometric distance to the nearest cell of the other component; proximity is measured in coarse-cell coordinates and does not claim a route through solid. |
### 3.7 Cave morphology (SDF rooms/tunnels) — `Public/VoxelCaveMorphology.h` + `.cpp`
@@ -486,7 +486,7 @@ The plugin's first tests (`OPSTACK-PLAN.md` Phase 0.5). Run them from the editor
| `VoxelForgeStrateParamCoverageTest.cpp` | `VoxelForge.Determinism.StrateParamBlendCoverage` | **The X-macro guard** (added 2026-08-17). `FStrateGenerationParams::Lerp` blends the hand-written `VF_STRATE_PARAM_FIELDS` list, **not** the struct — so a field added to one and not the other compiles, tests green, and silently takes its **default** inside every Gradient/Interleaved transition band. This expands the X-macro a **third** way (after LERP and SNAP), into a name list, and diffs it against the struct's UObject reflection. Pure shape test: no fixture, no world, instant. `GExemptFieldNames` is **empty** — every reflected field is covered today, and any exemption must be written down as a decision. Stakes rise with the world composer, which intends to invent parameter sets through this same `Lerp` (`COMPOSER-NOTES.md`). |
| `VoxelForgeLayoutOrderIndependenceTest.cpp` | `VoxelForge.Determinism.LayoutOrderIndependence` | Builds a known transient soft-pointer pool, then rebuilds it in original, reversed, and swapped orders. Requires a non-empty layout and passage set, and compares every slot's definition/Z/height plus passage endpoints, radius, type, control geometry, and bounds bit-for-bit. |
| `VoxelForgePassageOpenSpaceTest.cpp` | `VoxelForge.Determinism.PassageLandsInOpenSpace` | Uses the real fixture density path to check every generated inter-strate passage whose destination query answers: a 16-point ring outside the mouth's carve/blend band has at least half its samples in destination air, and the endpoint matches the pure open-point result within the mouth's float envelope. This is a connectivity proxy, not a flood-fill proof. Reports checked passages and false/unanswerable archetypes; fails if it inspects zero passages. |
| `VoxelForgeStrateConnectivityTest.cpp` | `VoxelForge.Generation.StrateConnectivity` / `VoxelForge.Generation.StrateConnectivityRefinement` | Bounded strate metrics with density-polarity and solid-gap controls, deterministic route rechecks, and a VerticalShafts refinement sweep. The refinement test explicitly checks margins 0/2/4/8 at step 2/radius 192, then conditionally reports mouth components/proximity and a 16-seed arrival→departure result. |
| `VoxelForgeStrateConnectivityTest.cpp` | `VoxelForge.Generation.StrateConnectivity` / `VoxelForge.Generation.StrateConnectivityRefinement` | Bounded strate metrics with density-polarity and solid-gap controls, deterministic route rechecks, and refinement sweeps. The refinement test explicitly checks margins 0/2/4/8 at step 2/radius 192, reports the `(0,0)` spine's full-window/interior vertical run and component for all eight fixture strates plus arrival/departure→spine and arrival→departure enums, then repeats those spine/mouth measurements across 16 VerticalShafts seeds. |
## 4. The density pipeline (most-edited hot path)