The raised limits' and the walk's comments are marked as the port's, with the unverified HAC2 and MCC comparison left out; the walk's says why the index is compared whole, and the once-only log of a frame reaching the limit gets a comment of its own.
Uninitialised locals get values (debug vertex, datastore result, index
buffer, map tool direction, AI entry points, escape alignment, grenade aim
from a vehicle, game options descriptions, unreachable switch arms), the
AI's squad and platoon records move whole, a speech cause point is read,
the geosphere's out-of-memory path frees itself, a console cond clause
without a result is refused, and the crash reporter rejects truncated map
files. Gameplay the original game has, kept so it plays the same, is noted
in a line rather than a page.
(cherry picked from commit 75afc4a1e57af2ba697e8be0e33b2afeee7639f8)
The Silent Cartographer with 32x co-op enemies (484 actors) went from 53
to about 97 fps on the Linux PC, with the same game and the same picture:
cluster list removals start where the reference was put, full data arrays
are not walked again, object lighting keeps each cluster's drawn lights
and passes over far ones without their distance, the per-texture profile
timer only runs while something reads it, and the renderer binds streams
once per draw, caches its texture, target and shader lookups, and no
longer makes Mesa's GL thread wait on large uploads or visibility tests.
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Do not publish donor research dumps or private assets.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.
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>
Owner ruling Q6: the retail-contradicted `double t` is rejected; the helper
cleanup may land at zero credit. _sphere_intersects_cluster_portal now uses
one shared FALSE exit and the real point_from_line3d call in place of the
hand-expanded rows, with production's names and types kept. It stays
residual at January's 432 bytes, and no other function changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Lane D refresh packet 7: +80 padded bytes strict exact;
structure_detail_objects.obj is now a whole-object match (15/15 functions,
.text/.rdata/.bss all 100%). Zero stable-verdict regressions.
The production `struct detail_object_globals debug_detail_objects`
aggregate and its `#define detail_object_global_runtime_data` alias were
invented. HCEA (DIA2Dump on the structure_detail_objects compiland) lists
the file statics detail_object_global_runtime_data (struct pointer),
fudge_vector (uchar), fudge_offset and final_offset (float), with
debug_detail_objects as a separate global; January's own assert string
uses the bare name detail_object_global_runtime_data. The source now
declares those statics in January's .bss order (initialised, so VC7 keeps
declaration order) and _structure_detail_objects_initialize stores through
the global directly, as the debug build does.
- symbols.json: four "static": true entries after _debug_detail_objects
(line insertion only).
- parked.json: unpark _structure_detail_objects_initialize.
Gate: ninja PASS, parks 124 active / 0 stale / 0 invalid, admission 0
contradicted / 0 revoked, fake scan 0 leads.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Lane D refresh packet 3: HS, effects and structures, +6,896 padded bytes
strict exact (9 functions), zero stable-verdict regressions.
- hs_compile.c: _hs_add_script exact as a goto-free single-exit function.
The worker's discarded hs_function_get call in hs_compile_postprocess
(fuzzy only; the debug build shows a dead local, not a bare call) is not
landed.
- hs_runtime.c: _hs_global_reconcile_read/_write exact. The debug build
repeats the external/internal index test inside the external arm; it is
written with this file's own if/else index idiom instead of the worker's
invented macro name.
- contrails.c: _contrail_add_points exact (contrail_scale_value helper,
corrected indentation).
- particle_systems.c: _particle_system_new_particle_jet, _render and
_update_particle_default exact. The debug build attests the one-field
variables walk (RTC size 4), the render if-block scope and the two-argument
local_random_range call.
- structures.c: _structure_get_planar_fog exact (real scale_vector3d helper;
flags/screen assigned in both arms per the debug build).
- structure_visibility.c: _portal_hull_from_points exact, single exit, real
plane3d_distance_to_point helper. The unused BSP local stays a local
(debug build) rather than a bare call.
- New COMDATs (_scale_vector3d, _local_random_range,
_plane3d_distance_to_point) are byte-identical to January's single folded
copies (all-inlined admission class).
- parked.json: unpark the four newly exact parked functions.
Gate: ninja PASS, parks 140 active / 0 stale / 0 invalid, admission 0
contradicted / 0 revoked, fake scan 0 leads, castcheck clean.
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.
- surface_index_buffer is declared before the frustum selection (January
loads the index buffer before the pvs test)
- material looked up through nested typed tag-block gets; the buffer walk
skips the two index words and post-increments through the count, matching
January's two separate pointer advances
- block-scoped vertex0/1/2 pointers read the compressed vertex data per
surface, matching January's left-to-right index loads before the call
Wave 21 (3 bundle implementers + 1 reviewer, 14 TUs, 22,880 target bytes)
closed only this function. Stable diff vs wave 20: 1 gained, 0 regressions;
active parks 191 -> 190.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
- input.h: struct key_stroke ascii_code is a char (January stores NONE as
-1; XINPUT_DEBUG_KEYSTROKE.Ascii is CHAR); edit_text.c drops its (byte)
cast on the NONE test and stays 5/5. Measured board-wide over all 24
includers: 0 regressions.
- input_xbox (27/27): _input_get_device_states with indexed gamepad and
button loops and a MIN tick clamp (park reopen);
_input_update_keyboard_devices with a short key-code loop and uncast
ascii ternary (park reopen)
- _structure_test_pill2d (624 B, previously unwritten): left/right offset
line tests via point_from_line2d and the nearer collision
- _find_turning_point: normalize2d/point_from_line2d/set_real_vector2d
replace hand-expanded helper math
- _structure_visibility_build_surfaces_traverse_leaf: the static
bounding_rectangles_intersect returns short through one result (park
reopen)
Rejected: _first_person_weapon_update (removed NULL guards, UB hold).
Follow-up: _path_test_pill_endpoint_near_wall_ok_bit is file-local in both
path_smoothing.c and path_structure_bsp.c; its owner is path_structure_bsp.h.
Stable diff: 0 regressions; parks 226 -> 223.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Per-function transplant of historical and worktree donor bodies into the
current tree (whole-blob mining had rejected them for unrelated losses).
Each landing is the minimum hunk, strict exact in its whole-TU gate with
zero inherited losses, and reviewed for house rules and hard holds:
- helper spellings replace hand-expanded math (_device_frontfacing,
_point_in_triangle3d, _actor_look_valid_aim/look_vector)
- _pill_test_vector3d normalizes directly into the output normal
- natural control-flow/declaration forms reopen five parks
(_structure_cluster_mark/unmarked, __rasterizer_widget_draw_sprite2d,
_rasterizer_set_model_skinning, _ai_handle_editing)
- two byte-proven non-UB original bugs kept under the documented
original-bug policy with BUG comments (_ai_handle_editing unscaled
csmemmove counts, _ai_scripting_follow_target_ai repeated name)
Stable diff vs 9062b1e96: +11 functions, 0 regressions; parks 268 -> 263.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Wave n1 of the next-150K lane (fresh structural portfolio at ae10935da).
- structures _structure_render_detail_objects (+1136): reopened with the natural
array initializer, which the w2 verifier asked for when it rejected the
hand-written a[0]=0/memset expansion (R6).
- draw_string _bitmap_draw_character (+944).
- hud_sounds _hud_play_sound (+352): closed by removing the one-use index alias
and reading sound_indices directly (A5), NOT by the volatile/goto body that the
park record shows was deliberately replaced.
- bink_playback _bink_free_texture_cache (+96): the park premise was the repo's
own match_dassert wrapper; the user-supplied Bungie random_math.c spells the
same two sites vassert, and the tree is 874 match_vassert to 14 match_dassert.
- structures _structure_get_planar_fog: zero-credit remeasure. Adding the
discarded fog-plane lookup January provably executes (its 15th
_tag_block_get_element_with_size relocation at +0x16d) closes both structural
tags; 448/448 and 15/15, residual now one register-naming window.
Rejected by the verifiers and not landed: _bink_alloc@4 (exact only with an __asm
block in a pool allocator - house rule 10 allows asm only for evidenced
helper/math routines), the progress_bar _code_000d16d0 rename (standing hold),
and the _portal_hull_from_points remeasure. wind _wind_variance_get is HELD for an
owner ruling (see its ledger).
Full ninja, parks 274 active / 0 stale / 0 invalid, stable diff vs
scratch/opus5-next150-before.json: gained 4 / 2,528 padded, regressions 0; fake
scan 0 leads.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Wave w1 Batch B (adversarially verified): _leaf_map_build_portals_from_leaf is now resolved-address identical to January (480/17, objdiff 100%); the only strict difference is a csplit one-past-the-end relocation alias, so its park is re-measured and reclassified csplit-relocation-alias (validated by tools.parked_functions). Zero strict credit. Full ninja build and stable sweep: 0 regressions.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Wave w2 landing, adversarially verified. park reopens: _render_debug_detail_objects. Full ninja build and per-function stable_verdicts sweep: 0 regressions. 1 park(s) removed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
planes_intersect_rectangle private ABI, single-exit mirror search, established child & LONG_MIN idiom; traverse_leaf park re-measured for the callee ABI (evidence appended). 8 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>
matrix_math.h and real_math_planes.h declare the genuine matrix_math.c / real_math.c signatures for their physics and collisions callers. Declaring either prototype in the broad real_math.h regressed decals::_decals_delete_permanent_from_cluster and draw_primitives::_rasterizer_dynamic_geometry_initialize and drifted four parks at any position (declaration count, bisected), so the precedent of narrow subsystem headers (bitmap_color_conversion.h, rasterizer_xbox_water.h) is followed. structure_fog_plane +0x00/+0x02 are the region index and the signed runtime material type January collision_test_vector reads; scenario.c renames are names only.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Retain 18 strict exact-function gains across Actions, Actor Combat, Decals, Game Engine, Object Lights, and Bipeds after removing donor bodies that regressed exact objects, emitted unowned code, or violated source-credibility rules. Name authenticated private symbols, move declarations to owner headers, record honest residual parks, and preserve units.obj at 189/189.
MrBruh