re: the consistency figure is mostly composite bounding boxes -- real disagreement is 1

consensus_check names the minority container instead of just flagging
disagreement: 89 minority decodes across 477 resources with a majority, and 88
are scene composites. A composite's descriptor carries a 24-vertex bounding box
(e_rou_e106 -> 22x22x22, e_rou_f106 -> 745x718x718); those boxes look
interchangeable so the assignment shuffles per container. 1141 of 6209 decoded
resources have scene nodes. Counting only real geometry the disc has ONE
disagreement: e101_wep_01_l in Stage_S25.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

crates/sylpheed-formats/examples/composite_geometry.rs

@@ -0,0 +1,34 @@
+//! How much of the "decoded geometry" belongs to SCENE COMPOSITES rather than to
+//! real meshes? A composite (`rou_*`, `e_rou_*`) carries node transforms; the
+//! anchor scan nevertheless finds a block for it, and that pseudo-geometry lands
+//! in coverage and consistency counts.
+//! Usage: composite_geometry <resource3d_dir>
+use sylpheed_formats::mesh::{scene_world_nodes, Xbg7Model};
+fn main() {
+    let dir = std::env::args().nth(1).expect("resource3d dir");
+    let mut files: Vec<_> = std::fs::read_dir(&dir)
+        .unwrap()
+        .flatten()
+        .map(|e| e.path())
+        .filter(|p| p.extension().and_then(|s| s.to_str()) == Some("xpr"))
+        .collect();
+    files.sort();
+    let (mut total, mut composite, mut composite_named) = (0usize, 0usize, 0usize);
+    for f in &files {
+        let Ok(bytes) = std::fs::read(f) else { continue };
+        for m in Xbg7Model::anchor_models_cancellable(&bytes, 0.0, &|| false) {
+            total += 1;
+            let has_nodes = !scene_world_nodes(&bytes, &m.name).is_empty();
+            let named = m.name.starts_with("rou_") || m.name.starts_with("e_rou_");
+            if has_nodes {
+                composite += 1;
+            }
+            if named {
+                composite_named += 1;
+            }
+        }
+    }
+    println!(
+        "{total} decoded; {composite} have scene nodes (a composite), {composite_named} are named rou_/e_rou_"
+    );
+}

crates/sylpheed-formats/examples/consensus_check.rs

@@ -0,0 +1,87 @@
+//! Minority report: which container decodes a shared resource differently from
+//! all the others?
+//!
+//! Cross-container consistency has been measured as "do the spans agree", which
+//! only says a resource is inconsistent — not which copy is wrong. With three or
+//! more copies the majority is the reference, and the minority names the
+//! container AND the resource to look at. That is what caught the `n054`/`n056`
+//! shift chain after the exact-coverage fix.
+//!
+//! Usage: consensus_check <resource3d_dir> [--list]
+use sylpheed_formats::mesh::Xbg7Model;
+use std::collections::{BTreeMap, HashMap};
+
+fn main() {
+    let dir = std::env::args().nth(1).expect("resource3d dir");
+    let list = std::env::args().any(|a| a == "--list");
+    let mut files: Vec<_> = std::fs::read_dir(&dir)
+        .unwrap()
+        .flatten()
+        .map(|e| e.path())
+        .filter(|p| p.extension().and_then(|s| s.to_str()) == Some("xpr"))
+        .collect();
+    files.sort();
+
+    // name -> [(container, verts, tris, span)]
+    let mut seen: BTreeMap<String, Vec<(String, usize, usize, [i64; 3])>> = BTreeMap::new();
+    for f in &files {
+        let Ok(bytes) = std::fs::read(f) else { continue };
+        let where_ = f.file_name().unwrap().to_string_lossy().to_string();
+        for m in Xbg7Model::anchor_models_cancellable(&bytes, 0.0, &|| false) {
+            let (mut lo, mut hi) = ([f32::MAX; 3], [f32::MIN; 3]);
+            for s in &m.meshes {
+                for q in &s.positions {
+                    for k in 0..3 {
+                        lo[k] = lo[k].min(q[k]);
+                        hi[k] = hi[k].max(q[k]);
+                    }
+                }
+            }
+            if lo[0] == f32::MAX {
+                continue;
+            }
+            let v: usize = m.meshes.iter().map(|s| s.positions.len()).sum();
+            let t: usize = m.meshes.iter().map(|s| s.indices.len() / 3).sum();
+            let span = [
+                (hi[0] - lo[0]).round() as i64,
+                (hi[1] - lo[1]).round() as i64,
+                (hi[2] - lo[2]).round() as i64,
+            ];
+            seen.entry(m.name.clone()).or_default().push((where_.clone(), v, t, span));
+        }
+    }
+
+    let (mut resources, mut minority) = (0usize, 0usize);
+    let mut rows: Vec<String> = Vec::new();
+    for (name, list_) in &seen {
+        // Only compare decodes that agree on how much geometry they found, and
+        // only where a majority can exist.
+        if list_.len() < 3 || !list_.iter().all(|e| e.1 == list_[0].1 && e.2 == list_[0].2) {
+            continue;
+        }
+        let mut votes: HashMap<[i64; 3], usize> = HashMap::new();
+        for e in list_ {
+            *votes.entry(e.3).or_default() += 1;
+        }
+        let (best, n) = votes.iter().max_by_key(|(_, n)| **n).unwrap();
+        if *n * 2 <= list_.len() {
+            continue; // no majority — cannot call anyone the odd one out
+        }
+        resources += 1;
+        for e in list_.iter().filter(|e| e.3 != *best) {
+            minority += 1;
+            rows.push(format!(
+                "{:<22} {name:<26} {:?} vs the other {n} containers' {:?}",
+                e.0, e.3, best
+            ));
+        }
+    }
+    if list {
+        for r in &rows {
+            println!("{r}");
+        }
+    }
+    println!(
+        "{minority} minority decodes across {resources} resources that have a majority"
+    );
+}

docs/re/structures/xbg7-mesh.md

@@ -1097,6 +1097,36 @@ that carry it) and `t901_e01_D` is 58×59×3013 (a mast). And `f101_bdy_01` flag
 at 6× while being **capture-verified exact** in the truth table — a useful
 reminder that a bulky hull is not a bug.
 
+### The consistency figure is mostly bounding boxes — real disagreement is ONE resource
+
+`examples/consensus_check.rs` sharpens the cross-container test: with three or
+more copies of a resource, the majority span is the reference and the **minority
+names the container that is wrong**, not just "these disagree".
+
+Disc-wide: **89 minority decodes across 477 resources that have a majority** —
+and **88 of the 89 are scene composites** (`rou_*` / `e_rou_*`), not drawable
+geometry. Inspecting one shows why: `e_rou_e106` decodes a **24-vertex,
+12-triangle box** (span 22×22×22), `e_rou_f106` another (745×718×718). A
+composite's descriptor carries a **bounding box**, the anchor scan finds it, and
+because those boxes are interchangeable-looking the assignment shuffles between
+containers. `mesh::scene_world_nodes` identifies them structurally: **1 141 of
+the 6 209 decoded resources have scene nodes**, i.e. are composites (478 are
+named `rou_`/`e_rou_`).
+
+So the headline number this file has been quoting — cross-container inconsistency
+— is **dominated by composite bounding boxes**. Counting only real geometry, the
+whole disc now has **one** disagreement:
+
+```
+Stage_S25.xpr   e101_wep_01_l   [779, 5769, 5769]  vs the other 3 containers' [206, 545, 545]
+```
+
+That is the sharper metric to work against, and the concrete next target. (The
+composite boxes are harmless — nothing draws them — but they should be excluded
+from any consistency figure quoted in future, and the ignored
+`shared_resources_decode_identically_in_every_container` test measures the mixed
+population, so its number is not comparable to this one.)
+
 ### What the exact-coverage fix actually reached
 
 The fix was justified on one resource (`e106_bdy_03`) and one render. Comparing