re: the last base-solver blocks -- two findings, and a correction to how I read the index

CORRECTION FIRST: a row in name-block-bases.txt is a (function, REGISTER) pair,
not a function.  Every one-line label I carried for these blocks quoted one of
two rows, and the other row of the same function is an unrelated schema:
sub_822E3EC8 r11 = 29 shader constants + techniques, r10 = 15 material map slots;
sub_822AFA50 r11 = XDK shader-compiler tokens, r10 = 13 menu camera tags;
sub_823AE908 r11 = the 43-name S16 boss collision table, r31 = 33 shader tokens
(the dense-block false positive).  Quote the register.

The engine's material vocabulary (into xbg7-mesh.md, which owns the
_col/_spc/_gls/_lum suffixes): nine base map slots -- ColorMap, LuminosityMap,
SpecularityMap, GlossinessMap, ReflectionMap, TransparencyMap, NormalMap,
FresnelMap, EdgeTransparencyMap -- and seven *MapArray twins; 9 - 7 = 2, and the
two without a twin are FresnelMap and EdgeTransparencyMap.  Four of the nine are
exactly the four documented suffixes; the other five have no suffix in the
shipped meshes, so the engine supports more channels than the assets use.  Plus
17 g_* shader constants, three cube samplers, and DepthOnly / MotionVector /
Transparent techniques -- the ...Last matrices with the motion-vector technique
say the renderer keeps the previous frame's per-object transform.

The S16 boss's collision parts (into collisionset.md): sub_823AE908 r11 is an
interleaved (mesh, logical-part) table -- rou_e901_body_02_c ->
Boss16CollisionBody02, five wing pairs, both tails, barrier, rou_e910_core_cmesh
-> Boss16CollisionCore, and mob_n900{,_b01,_b02,_gear}_cmesh -> Room00/01/02 +
Gear00/01/02.  23 logical parts + 20 meshes = 43, differing by exactly three:
Body01 opens with no mesh and the one gear mesh carries three gear names.  Cross-
links the S16 Guardian/Core object found during the PlayerParams work.

Not settled: whether those meshes are CollisionSet entries, why Body01 has no
mesh, and no shader slot was traced to a shipped shader.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PMRJjbxLqZtsb5Vb7KunPE
This commit is contained in:
Claude (auto)
2026-08-27 22:51:53 +00:00
parent e7c89fdcd1
commit dd6b13c751
4 changed files with 128 additions and 9 deletions

View File

@@ -157,3 +157,32 @@ exports any named mesh.
record is exactly counts, positions and indices — so whatever distinguishes a
hull hit from a shield hit is not stored here. And the library is not indexed:
lookup is presumably by name from the unit tables, which has not been traced.
## ✅ The Stage 16 boss's collision parts, named in the executable (2026-08-27)
Found by the base-solver ([[player-tuning-tables]]): **`sub_823AE908`, `r11`,
43 names**, and the list is an **interleaved (mesh, logical-part) table**:
```
rou_e901_body_02_c → Boss16CollisionBody02 rou_e901_wing_0N_{L,R}_c → …Wing0N{L,R}
rou_e901_tail_{L,R}_c→ Boss16CollisionTail{L,R} rou_e901_barrier_cmesh → …Barrier
rou_e910_core_cmesh → Boss16CollisionCore mob_n900_cmesh → …Room00
mob_n900_b01_cmesh → …Room01 mob_n900_b02_cmesh → …Room02
mob_n900_gear_cmesh → …Gear00 / …Gear01 / …Gear02
```
**23 logical parts + 20 meshes = 43**, and the two counts differ by exactly
three: `Boss16CollisionBody01` opens the list with no mesh before it, and
`mob_n900_gear_cmesh` is followed by **three** gear names. A second row
(`sub_823C0260`, 15 names) lists the hull parts alone — `Body01…03`, `TailL/R`,
`Wing01L…Wing05R` — with no meshes.
🔑 **This is the same boss the `PlayerParams` work turned up**: the S16
`Guardian` / `Core` object ([[player-tuning-tables]]) names its weapons
`Shell_S16Boss_*`, and here `rou_e910_core_cmesh → Boss16CollisionCore` gives the
`Core` record a collision hull. `rou_e901` is the boss craft; `mob_n900` is the
station/room structure it sits in, whose three `Gear` parts share one mesh.
🟡 Not settled: whether these meshes are entries of the `CollisionSet` library
above (all 18 blobs are byte-identical, so a name-level join was not attempted),
and why `Body01` has no mesh in the list.

View File

@@ -236,15 +236,22 @@ census that followed is in that document. The same base `0x8208583C` also serves
`sub_82338EE0` (97 names, `Weapon TargetType SpecialWeaponType ReticleType
IsCharging …`) — the **weapon** datasheet loader, likewise not previously named.
**Other unowned blocks the index surfaces** (measured, not investigated):
`sub_822215D0` (205/206, `PGHUD_*` / `PGREMAIN_NUM%d` / `PGTIMER_NUM%d` — HUD
part names; [[mission-script-manifest]] mentions `PGHUD_SUBTARGET` only),
`sub_822814D8` + `sub_8227A3A0` (`STAGE_RESULT`, `stage_num_shoot_down_aircrafts`,
`EX_OVERVIEW` — results/leaderboard fields), `sub_822E3EC8`
(`g_mWorldViewProjection`, `NormalMap`, `GlossinessMap` — engine shader/material
slots), `sub_823C0260` + `sub_823AE908` (`Boss16CollisionBody01…`,
`rou_e901_body_02_c` — the S16 boss collision parts, cross-linking the `Guardian`
object above), and `sub_822AFA50` (`roh_n001_menu1_cam_pos` — menu camera tags).
⚠️ **CORRECTION (2026-08-27) — a row is a (function, REGISTER) pair, not a
function.** My one-line labels for those blocks each quoted **one of two** rows,
and the other row of the same function is a different schema entirely:
| function | one row | the other row |
|---|---|---|
| `sub_822E3EC8` | `r11` 29 — shader constants + techniques | `r10` 15 — the material **map slots** |
| `sub_822AFA50` | `r11` 16 — XDK shader-compiler tokens (`__vspltw`, `__vsel`) | `r10` 13 — the **menu camera tags** |
| `sub_823AE908` | `r11` 43 — the **S16 boss collision** table | `r31` 33 — shader tokens (the dense-block false positive) |
**Read `docs/re/data/name-block-bases.txt` by row, and quote the register.**
The three findings themselves are written up where they belong:
[[xbg7-mesh]] (material slots + shader constants) and [[collisionset]] (the S16
boss's 23 collision parts over 20 meshes). `sub_822AFA50`'s camera-tag row is
13 `roh_n001_menu{1,2}_*_cam_{pos,tag}` names — the menu scene's camera
positions and look-at tags, with `menu2` having sub-shots `2_0``2_4b`.
🟡 None of those five were opened; the index says what each names, not what each
means.

View File

@@ -1727,3 +1727,39 @@ resident but the unit never appeared).
So the anchor line of work is **done for practical purposes**: every real mesh the
engine draws in three captured missions decodes, at the offsets the GPU used, with
index runs matching byte for byte (93/93 and 128/128).
## ✅ The engine's material slots and shader constants (2026-08-27)
This document names the **TX2D channel suffixes**`_col` albedo, `_spc`
specular, `_gls` gloss, `_lum` luminance. What it did not say is what the engine
calls those slots. `sub_822E3EC8` reads two blocks (base-solver,
[[player-tuning-tables]]):
**`r10` — 15 material map slots.** Eight base maps, `LuminosityMap`,
`SpecularityMap`, `GlossinessMap`, `ReflectionMap`, `TransparencyMap`,
`NormalMap`, `FresnelMap`, `EdgeTransparencyMap`, plus **seven `*MapArray`
twins** — `ColorMapArray`, `LuminosityMapArray`, `SpecularityMapArray`,
`GlossinessMapArray`, `ReflectionMapArray`, `TransparencyMapArray`,
`NormalMapArray`. ⚠️ `ColorMap` itself sits in the *other* block, so the base set
is **nine**; **`FresnelMap` and `EdgeTransparencyMap` are the two with no array
twin** (9 base 7 arrays = 2, named in full).
🔑 **Four of the nine slots are exactly the four suffixes above**`ColorMap`
(`_col`), `SpecularityMap` (`_spc`), `GlossinessMap` (`_gls`), `LuminosityMap`
(`_lum`). The other five (`NormalMap`, `ReflectionMap`, `TransparencyMap`,
`FresnelMap`, `EdgeTransparencyMap`) have **no suffix observed in the meshes this
document decodes** — the engine supports more channels than the shipped assets
use.
**`r11` — 29 shader constants and techniques.** `g_mWorldViewProjection`,
`g_mWorldViewProjectionLast`, `g_mObjectToView`, `g_mObjectToViewLast`,
`g_mWorld`, `g_Params`, `g_UVParams`, `g_LayerParams`, `g_MapEnable`,
`g_MapLayerEnable`, `g_vWorldEyePos`, `g_vLightDirection`, `g_LightColor`,
`g_EffectLightPosition`, `g_EffectLightColor`, `g_fogEnable`, `g_fogParams`;
cube samplers `DiffuseCube`, `SpecularCube`, `GlossCube`; and the techniques
`DepthOnlyTechnique`, `MotionVectorTechnique`, `TransparentTechnique` (+ a
"Normal technique found" / "Shadow map technique found" log pair).
🔑 The `…Last` matrices and `MotionVectorTechnique` together say the renderer
keeps the **previous frame's transform per object** — i.e. a motion-vector pass.
🟡 None of this was traced to a shipped shader; it is the slot vocabulary only.