Asked for, worn: no seeing into the body through the neck, and the
shoulders and torso following the head in a seat.
In a seat the neck now stays and only the head goes, shrunk into the neck's
base: the neck closes like a dimple instead of opening into the body, and
the seated body is drawn two-sided, its insides dark, so the shoulders and
collar near the eyes show no hole. (On foot the neck still goes, and the
body stays one-sided: there, folded, the collar and shoulders leave thin
sheets over the chest that drawn both ways hid it.)
The seated chest now turns with the head, 0.6 of the head's turn from the
body's way up to 35 degrees, and leans toward where the head has moved
from its place at rest, up to 25 degrees, about the spine's base, before
it goes back out of the near plane; the arms then reach from the moved
shoulders to the controls as before.
Dumps (b30, a Warthog's driver): looking down, a closed dark collar on the
chest plate; the head turned 40 degrees left, the left shoulder solid in
view and the hands still on the wheel. On foot (a30) as before.
At 1.5x the GPU needed 11.5 ms a frame (timestamps, vr.timing_gpu_stamps),
over 90 Hz's 11.1, almost all of it the eyes' pixels. The runtime's
visibility mask (XR_KHR_visibility_mask, 15.6% of each eye's image on the
Frame) is now drawn first in each eye's pass at the near plane, black, so
the scene is not shaded there (vr.visibility_mask, on). Two things it took:
the device binds its target and viewport lazily, so the mask binds the
eye's first (halo_vr_bind_current_target); and its depth goes through the
position, not gl_FragDepth, which turned Adreno's early depth culling off
for the rest of the pass (it saved 0.3 ms that way, 1.5 ms this way).
On the device in a30 at 1.5x: the GPU's frame from 11.5 to 10.0 ms (the
eyes 4.4/5.1 to 3.6/4.1), 85.5 fps with 45 misses in 10 s to 89.4 with 2
to 5. Cutscenes (the 3D screen, not masked) hold 78 to 89.
Also: vr.timing_gpu_passes (each pass waited for; Zink's glFinish costs
too much for it to mean much), glQueryCounterEXT and
glGetQueryObjectui64vEXT for the guest, the mask's share logged.
Halo draws no body of the player's own in first person. With vr.body
"legs" (off by default) the eyes' passes draw the player's unit, on foot
and outside cutscenes, from a copy of its bones' matrices: the head, neck
and upper arms (and all under them) scaled away, the body turned to the
head's yaw (as the first-person shoulders now are) and moved under the
head, 12 cm behind the eyes. What the unit carries (its third-person
weapon) is not drawn with it, nor is its shadow. The body's bone names are
logged once, with which are hidden. The head-yaw direction is now shared
with the shoulders (vr_head_yaw_forward). Built only; not yet run.
In stereo the game's screen flash (a white flash into a scene, a fade to
black) was drawn in the HUD pass, so it covered the HUD's panel: a box in
the middle of the view (reported worn). It is now drawn in each eye's pass,
with the game's own blend over the eye's image. The HUD pass keeps only a
flash that darkens (the darken type, or a lighten toward black), so a fade
to black still takes the HUD with it; one that lightens would light the
panel up as a box again.
While a hand-aimed weapon is zoomed, a pass between the eyes and the HUD
renders the view along the gun into a corner of the back buffer, sized
to the scope's 768x768 image. It goes through a sight mask (round, or
the sniper rifle's wide one) onto a layer held at the gun, at the PC
mod HaloCEVR's per-weapon offsets. The eyes never zoom: the first-person
gun stays in the hand and the HUD layer drops the zoom mask.
vr.scope and vr.scope_size are the settings; vr.diag_zoom_seconds zooms
unattended for dumps (vr-scope.bmp). The docs now cover the VR controls,
gestures, weapon feel and scope, and credit LivingFray's HaloCEVR for
their designs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With the hand aiming, the first-person weapon's camera puts the gun in
the right hand and used to carry both arms along with it. vr.arms "ik"
(the default) gives the arms shoulders where the body is (below and
either side of the eyes, turned with the heading) and solves each arm's
upper arm and forearm, elbows down and out, to reach its hand: the right
the gun's grip as the controller holds it (the wrist and the gun, its
child, stay as they are), the left the left controller, or the gun's
grip for it while the left controller is within 15 cm of it. Out of
reach, the shoulder comes forward rather than the hand letting go.
Fingers and other bones go with the bone above them. "hidden" shows the
gun alone, "animated" the arms as the animation moves them.
Halo's first-person graphs name the hands' bones "wriste"; the gun is the
right one's child.
Cutscene frames have their own dump trigger (vr.dump_cinema_frame).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Android renderer answered each visibility test (lens-flare
occlusion) by mapping the counter buffer for its one word, and each map
waits for the GPU's queue to drain; on the Steam Frame's Zink it also
flushes and logs a performance warning. In stereo, with the flare tests
of two eyes, that took 45% of the game thread: 13-20 frames a second in
the first level's cryo bay. All counters are now read together into a
copy the first time a result is asked for after a test ends (one wait a
frame), and the same scene runs at the headset's 72.
The headset's HUD pass took the planar fog, whose animation offset only
a structure pass consumes, and the next frame's assertion stopped the
game in fogged clusters; the HUD pass leaves the fog alone.
In VR the head's turns count as looking for the scripts that wait for
the player to look (the first level's look test).
The host's thread sampler takes fractions of a second.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A frame with one player and no cinematic renders the player's window
three times in render_frame's window loop: the left and right eyes, each
from the game's camera turned by its heading and moved by the tracked
eye (vr.world_scale, head reach limited to 0.35 m), with OpenXR's
off-centre field as render_camera_build_frustum's bounds, and each
copied into its swapchain image as it finishes; then the HUD alone on a
transparent ground, shown head-locked (vr.hud_distance, vr.hud_width).
What moves on as it draws does so once a frame: the sky's animation
phase, weather and glow particles run for the left eye only, and the
first-person weapon is posed once, from the head between the eyes. The
fog's screen layers, which assume a symmetric frustum, are left out of
the eyes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Recover transparent geometry submit (+1,285 meaningful code, zero exact losses); reconcile COMMON ownership and Q10 85+EXT1 credit, actions/main data, and separately disclosed denominator corrections. Preserve canonical-only repairs and all held research. C09 is ordinary disclosed cleanup, not a claim of proven scaffolding intent. Clean rebuild, 1,663 tests, parks, audits, provider links and independent EXT1 check pass. Local reconciliation only.
Import only independently reviewed production changes. Preserve canonical physics and virtual-keyboard repairs; keep B3 and all unresolved holds out.
Record exactly four approved canary debits (6987 meaningful bytes), six Halo admissions and two revocations. Keep vendor ioinit/storage work separate. Full build, 8252-row sweep, 1318 tests, park/admission/fake audits and 621-object binding pass within approved debit bounds.
Do not publish donor research dumps or private assets. README and unrelated user research remain untouched. See frontier_publication_reconciliation_20260927.md for complete evidence limits and accounting.
Independently gate storage + K, OQ-4 P-a/P-b1, four static-local packets and comment-only disclosures. Preserve all 7641 exact owner rows, 71 valid parks and 1161 passing tests; code/data credit unchanged. G and game_engine admission remain held because current canonical loses populate_statistic_buffer. Existing collision static-local repair is not duplicated. No donor tree, held candidate, scorer or private asset imported.
The Xbox game state (0x345000 bytes at 0x80061000) ends exactly where the
map-linked tag cache begins and has 948 bytes free, so a 128-player
session cannot even allocate its player datums. The native builds now
place a 16 MB game state (CPU 0xFC0000, GPU 0x40000) at 0x81A00000, above
the tag cache and below everything the Xbox memory window hands out
top-down, and grow the pools that scale with the number of players:
- objects: 2048 -> 8192 headers, 1 MB -> 8 MB pool; the garbage limits
keep January's proportions (5/10/20 % of the pool, 2.5/5/10 % of the
headers, 75/50/30 active garbage objects per 16 players)
- cluster partition reference lists 2048 -> 8192 (collideable and
noncollideable objects, lights); a full list drops objects out of
rendering, collision and sphere queries
- effects 256 -> 2048, effect locations 512 -> 4096
- particles 1024 -> 8192 (render_particles' visible list, filled without
a bound check, follows), particle systems 64 -> 256, their particles
512 -> 4096
- contrails 256 -> 1024, contrail points 1024 -> 8192
(MAXIMUM_CONTRAIL_POINTS stays the pool size for the same reason)
- lights 896 -> 4096, object looping sounds 1024 -> 4096
- cached object render states and rendered objects 256 -> 1024
- script object-list references 128 -> 1024, objects one explosion can
damage 64 -> 256
The values live in port/linux/include/halo_port_capacity.h, included by
halo_port_limits.h. With 128 player datums the CPU part holds 14,256,908
bytes (2,258,164 free). Saved games and cores grow to 16 MB. Every change
is under HALO_LINUX; the MSVC objects are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Approved packets, no owner ruling needed:
- real_math.h plane3d_negate gets its attested body (negate_vector3d plus
the d term; /Od 0x56d630 and the HCEX_Release line table), together with
the render_objects /Od shadow-volume plane pair and the bipeds genuine
helper calls. _biped_find_nearby_support_surface is now exact (+544). A
467-unit header review lost 0 exact rows. The bipeds park is retired, and
render_cameras _render_camera_build_frustum is refreshed to
3408/112/d42e420f (target unchanged).
- biped_falling_danger, update_airborne, snap_facing, update_moving and
fudge_vectors become file-static, as the Sept-2001 map STATIC list and the
retail private ABI show. Their prototypes leave bipeds.h and symbols.json
marks them static. This is zero credit; bipeds pdb_storage goes to 0. The
parked ai _ai_test_ballistic_line_of_fire is rebaselined for the resulting
declaration-count drift; no exact row is lost.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Zero credit; every packet was independently adversarially reviewed (22/22 approved).
- critic_3: 72 renames of placeholder/invented static function names across 20
objects to first-party names. Each rename has Sept-2001 linker-map alignment
(cachebeta xbe / retail cache exe, several also Aug-2001), relocation-masked
body identity and matching callers; cachebeta publics show every row was a
January static. hs's radiosity_hack_{start,save,find_point}_evaluate pairing
is proven by the Sept .rdata hs_function_definition pointers; players'
create_weapon also matches HCEX.pdb. Objects: ai_debug (reviewer-amended: a
dead stub removed, its comment re-attached), transport_endpoint_winsock,
transport_endpoint_set_winsock (+ parked.json key renames), profile,
progress_bar, network_connection, saved_game_files, rasterizer_xbox_profile,
xbox_sound_cache, game_engine_king, players, hs, hs_runtime, main, decals,
rasterizer_lights, rasterizer_xbox_dynavobgeom,
rasterizer_xbox_environment_fog, rasterizer_xbox_hardware_bitmaps,
render_cameras.
- hs_compile: whitespace_characters, eol_characters and hs_type_primitive_parsers
are external (cachebeta publics, HCEX DataIsGlobal); hs_compile_finish is
static with a TU-local prototype (no public; HCEX static function).
- bink_playback: bink_playback_idle -> bink_playback_update__internal
(first-party). The bink_decompress_audio_frame rename is NOT landed: it needs
an added call whose position is unattested, the same class canonical held for
hardware_geometry in 4d1ebf17.
Gate (full ninja, objdiff 3.3.1): objects 383, code 1,577,237, data 2,586,915
(all unchanged); stable diff 0 regressions; parks 83/0/0; admission 11/0/2/0;
fake scan 26; pytest 1161 passed; diff --check clean.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
All packets passed a precedent audit and an independent adversarial review.
No object changes status.
- Storage per cachebeta publics (source static + symbols.json rows):
- bink_playback: four globals;
- weapons: 11 rows + weapon_magazine_finish_reload;
- object_lights: 4 helpers; should_render_lights is the /Od-attested call and
its prototype leaves the public header;
- vehicles: _update_turret_physics;
- weather_particle_systems: one_over_char_max is external (cachebeta public,
HCEX Global); its globals are static;
- rasterizer_xbox_debug: 8 D3DINLINE wrapper rows (already Matching: audit
FAIL(8) -> PASS, pdb 8 -> 0).
- draw_string: the invented external draw_string_globals aggregate (pads, blob,
#define aliases) becomes four HCEX-attested file statics in /Od declaration
order with = { 0 } (a595bbc2 precedent); typed multitexture params.
- render_debug: three rule-6 hand expansions become real point_from_line3d
calls (ruling #1/#2); the opaque vehicle view gives way to struct
vehicle_datum + vehicle_datum_try_and_get; the six debug_* toggles become
tentative definitions (COMMON evidence).
- vehicles: the TU-wide REAL_MATH_EXTERNAL_POINT_FROM_LINE3D define is gone and
the pathfinding hand expansion becomes the genuine call (ruling #1).
- rasterizer_xbox_profile V1R (reviewer-amended): byte-inert /Od parameter
order, the BUG marker comment, detail_objects via the rasterizer_xbox.h API,
and a rejection text that keeps the SDK-table policy question open.
- semantic_data_matches.json: single-section '$'-defect entries with
allow_incomplete_unit for projectiles _projectile_update_section (1,548 B) and
object_lights _lights_section (1,541 B) (leaf_map / editor_flying_camera /
race precedents).
Gate (full ninja, objdiff 3.3.1): objects 384 (unchanged), code unchanged,
data +3,089; stable diff 0 regressions; parks 83/0/0; admission 10/0/2/0; fake
scan 26; pytest 1159 passed; diff --check clean.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- symbols.json: IDirect3DDevice8_IsBusy/KickPushBuffer and IDirect3DBaseTexture8_IsBusy
are static (D3DINLINE is static __forceinline; none is a cachebeta public).
- Three invented static names replaced by first-party names from the 2001-09-25
cachebeta-xbe linker map (atlas 7eacac85, re-anchored on 20 named symbols):
texture_cache_name_block_proc, compare (also January atlas exact tier) and
texture_cache_initialize_hardware_format.
- bitmap_group.h (shared, 45 includers): struct bitmap_data fields per HCEX
(mipmap_pad, long pixels_offset/pixels_size/tag_index/cache_block_index);
bink_playback.c and bitmaps.c adopt the names. The TU-local bitmap_data view
and raw tag_get('bitm') prefix view in xbox_texture_cache.c are gone.
- texture_cache.h declares texture_cache_debug_render (render_debug.c had no
prototype in scope). The Xbox-only public format mappers get the object's own
header cache/xbox_texture_cache.h. Adding them to texture_cache.h regresses
declaration-count-sensitive consumers (decals); this is disclosed in the header.
- Reviewer amendment: the invented pad member reserved006[2] in the TU-private
texture struct is removed (natural alignment; HCEX texture_datum has no pad).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
objects.h: both bounding-sphere inline assertions report OBJECTS_H_FILE, which
defaults to January's relative "..\objects\objects.h". object_lights defines the
absolute spelling first, as January's string census shows. action_vehicle loses
its NODUP hand copy of object_get_bounding_sphere, so the header inline is the only
definition. render_objects' lighting_storage becomes the static local of
object_get_cached_render_lighting (symbols.json).
path.h: the opaque __unknown140A0[0x5C] span is HCEX's `struct path_result
result`, and path.c stores through the field instead of a punning cast.
Consumer sweep (full ninja, objdiff 3.3.1): stable diff shows 0 regressions and no
moved rows beyond the batch-1/Layer-2 gains. Parks 93/0/0, admission 15/0/7/0,
pytest 1157 passed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Structural lane batch 3: +4,128 padded / +4,101 meaningful bytes strict
exact (3 functions), zero stable-verdict regressions.
- rasterizer_xbox_active_camouflage.c: _rasterizer_active_camouflage_draw.
The debug build (0x7cd9c0) writes each normal/hyper-stealth blend as one
(1-p)*base + p*hyper expression. The texture-animation vertex constants
are a block-scoped aggregate initializer, so both constant arrays share
a slot as in January.
- render_particles.c: _render_particles. The draw pass indexes
group_particle_counts[group_index] with a short loop, and the inner loop
is while (particle_count-- > 0), advancing the particle at the body end
and before the delete-path continue (/Od 0x84f800).
- rasterizer_xbox_shadows.c: __rasterizer_environment_shadow_begin. The
debug build's boolean success = TRUE is returned once, and the
translation terms use dot_product3d (/Od 0x80d870; HCEX array name
vsh_constants__screenproj).
Rule-24 disclosure (owner per-site admission, 2026-09-21): two new
(real_vector3d const *)&shadow_matrix->position casts, one per
dot_product3d call. The /Od build calls dot_product3d at 0x80db50 and
0x80dbc2 with the matrix position. Deleting both casts leaves the
function strict exact, and point/vector share the three-real layout.
parked.json: unpark _render_particles and __rasterizer_environment_shadow_begin.
The owner packet gains the aggregate 3D->2D view-copy question.
Gate: ninja PASS, parks 112/0/0, admission 0 contradicted/0 revoked,
provider link PASS, fake scan unchanged, pytest 1152 passed / 5 skipped
/ 26 subtests.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Integrate verified source and data progress; exclude unproven COMMON ownership, held UB/asm forms, and address-named data credit. Stable sweep: 23 exact functions gained, zero lost.
Integrate six accepted exact functions, two whole-object admissions, evidence-backed zero-credit source corrections, owner-header cleanup, and curated campaign records. Preserve zero regressions and exclude generated scratch archives from canonical history.
- _render_debug_add_cache_entry: success-first single exit, string case
if/else, one case body per debug type as January emits them
- _ctf_engine_weapon_update: long return-team locals from the owner team
field (January's 32-bit movsx locals)
- _action_uncover_perform: timeslice-block locals, one evaluation-mode
store (removes HEAD's duplicated arm stores), field-order target stores
- _row_dither: indexed channel loops with PIN clamps replace pointer
cursors and a hand-written ternary clamp
- _lruv_debug_to_file: removes a goto web; MIN(9999, age) and the
NULL-name fallback in the fprintf arguments (integer compare order
disclosed; constant-first MIN is attested in Bungie source)
- _action_guard_perform: short firing-position index matching the
actor_select_firing_position return type
render_debug, game_engine_ctf, action_uncover, bitmaps_quantitize,
lruv_cache and action_guard are now all-function exact.
Stable diff: 0 regressions; active parks 219 -> 213.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Reviewed reconstructions, each strict exact in its whole-TU gate with
zero inherited losses, guard passing and no new owners:
- rasterizer_xbox_screen_effect (13/13): complete per-arm shader setups,
load-before-call locals, January assert texts, PIN clamp; screen_flash
projection spelling and the defined inverse-pixel default arm
- hs_compile: _hs_parse_begin (per-branch result type; January's extra
sprintf name argument kept with a BUG comment), _hs_parse_cond_recursive
(single exit; January's always-true negation test kept with a BUG
comment, byte-proven and recorded in HCEA)
- geometry: definitions in January object order (assert lines ascend),
duplicate prototype removed, _get_edge_vertex single exit
- render_cameras: view_to_screen and build_frustum_bounds simplifications
- sound_manager: _sound_dispose_from_old_map loop-exit flag
Rejected in review: _encounter_test_rule (redundant default stores),
_render_camera_mirror (point_from_line3d COMDAT), _sound_set_definition_end
(prototype contradicting its definition).
Stable diff vs 9062b1e96: 0 regressions; parks 255 -> 250.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Both admitted by explicit owner rulings on 2026-09-16; each lands as the minimum
hunk after a whole-TU gate, and neither involves undefined behaviour or steering.
render_debug _render_debug_player (+266, park reopened and removed). LAW D, from
the research campaign's frame study, proves January's `sub esp,0x404` cannot come
from a dead local or any other declaration - a dead whole object costs 0 frame -
and only a declared text extent of 1025-1028 yields it. Written in the house
idiom as a named capacity plus its terminator, added to the file's existing
constants enum rather than a new block: `MAXIMUM_RENDER_DEBUG_PLAYER_TEXT_LENGTH =
1024` and `char text[MAXIMUM_RENDER_DEBUG_PLAYER_TEXT_LENGTH + 1]`. The tree
already uses `[X_LENGTH+1]` string buffers 70+ times. This ruling supersedes the
pre-LAW-D park rejection of enlarging the buffer. Whole-TU gate 34 -> 35 with only
this row changing; the sibling park _render_debug_add_cache_entry is unchanged; the
added enumerator shifted no declaration-count-keyed row; fake scan 0 leads.
hud_draw _real_rgb_color_to_pixel32 (+181, previously unwritten). A genuine
bitmaps_inlines.h color-conversion helper, admitted under house rule 10. The owner
ruling rests on affirmative hand-written-assembly evidence in January's bytes -
dword fistp and fmul on non-top x87 stack slots (fmul st(3),st(0)) - not merely on
asm being able to force a match. The permission covers this one body only; it does
not authorize _real_argb_color_to_pixel32, _fast_ftol_C or assembly elsewhere.
Whole-TU gate 20 -> 21 exact, including _hud_draw_meter, which calls it three
times. fake_match_scan: 3 leads, 2 of them pre-existing at HEAD (the naked
get_return_eip); the only new lead is this admitted block.
Full ninja, parks 271 active / 0 stale / 0 invalid, stable diff vs
scratch/opus5-next150-before.json: gained 12 / 8,992 padded, regressions 0.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Wave w1 landing, adversarially verified. exact: _render_object_shadow_end, _object_render_state_refresh. Full build of the combined batch: 54 strict gains / 29,440 padded, 0 regressions, parks 355/0/0.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The edge factors are spelled (-inverse_z) inside the MIN/MAX arguments (macro argument duplication) instead of hoisted positive locals. 1 park(s) removed. Adversarially verified (wave D park re-investigation); full build of the combined wave D set: 0 regressions, parks 402/0/0.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
+1 strict exact function, 384 padded bytes (previously residual). Adversarial
review accepted; no top-level declaration was added and the other three
render_sprite residuals keep their tags. A proposed spelling for
build_sprite_transform_origin_and_direction that would write a vector over its
own parameter home was not landed (owner/UB ruling required).
render_sprite: 7 exact / 3 residual.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Madison