9017107960
The P1 "doorway membership lags retail" premise is FALSIFIED. acdream's swept ResolveWithTransition already matches retail's true per-frame curr_cell: the production gate ProductionPath_IndoorCrossings reads 9/9 on the indoor 0170<->0171 crossings with NO code change, once fed an aligned retail golden. Root cause of the false 0/11: CPhysicsObj::SetPositionInternal calls change_cell (acclient_2013_pseudo_c.txt:283456) BEFORE set_frame writes m_position (:283458), so the original golden (find-cell-list-capture.cdb, read at the change_cell BP) paired each frame's NEW cell with the PREVIOUS frame's position — a one-frame skew. Verified 3 ways: the decomp ordering; golden_picked[i] == geom(golden_position[i+1]) for all 22 rows; acdream's static pick == golden_picked[i-1] for all rows. Both retail and acdream pick with center-only point_in_cell on global_sphere[0] (no XY lead; cache_global_sphere @ pc:274196). curr_cell commits via validate_transition (@ pc:272608, curr_cell = check_cell) = the find_cell_list pick, structurally identical to acdream's RunCheckOtherCellsAndAdvance -> FindCellSet -> SetCheckPos. There was nothing to port; a swept advance would make membership LEAD by a frame. - tools/cdb/find-cell-list-capture-aligned.cdb: re-capture reads the committed position from the set_frame that follows change_cell (cell+position same instant). - Fixtures/find-cell-list-threshold.log: replaced with the aligned capture. - ThresholdPortalCrossingReplayTests / FindCellListConformanceTests: rewritten from documents-the-bug to assert retail truth (per-segment / per-indoor-pick equality). - handoff + notes + README + memory: banners correcting the disproven premise. Still open (NOT indoor membership, which is DONE): outdoor->indoor 0031<->0170 entry conformance (needs landcell + building stab in the gate cache); master-plan cleanups (delete CheckBuildingTransit, unify find_env_collisions, demote ResolveCellId) refactor working retail-faithful code -> need explicit user approval. Conformance 60 pass / 1 skip / 0 fail; full Core 1309 pass / 5 fail (pre-existing 2 BSPStepUp + 3 door-collision = P2) / 1 skip. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
61 lines
3.8 KiB
Text
61 lines
3.8 KiB
Text
$$ ============================================================================
|
|
$$ find-cell-list-capture-aligned.cdb — P1 ALIGNED retail golden capture
|
|
$$ ----------------------------------------------------------------------------
|
|
$$ SUPERSEDES find-cell-list-capture.cdb for the P1 gate. The original logged
|
|
$$ this->m_position AT the change_cell breakpoint. But CPhysicsObj::SetPositionInternal
|
|
$$ (acclient_2013_pseudo_c.txt) calls:
|
|
$$ change_cell(this, curr_cell) @ line 283456 <- breakpoint fired here
|
|
$$ set_frame(this, &curr_pos.frame) @ line 283458 <- m_position updated HERE, AFTER
|
|
$$ so the original paired THIS frame's NEW cell with the PREVIOUS frame's position
|
|
$$ — a deterministic one-frame offset (proven: golden_picked[i] == geom(golden_position[i+1])
|
|
$$ for all 22 rows of the original capture; see
|
|
$$ docs/research/2026-06-03-p1-membership-swept-advance-handoff.md follow-up).
|
|
$$
|
|
$$ This script reads the NEW (committed) position from the set_frame call that
|
|
$$ IMMEDIATELY follows change_cell within the same SetPositionInternal, so the cell
|
|
$$ and position are from the SAME instant. The resulting golden lets the P1 gate be a
|
|
$$ true integration test (no position/cell skew).
|
|
$$
|
|
$$ Why the correlation is safe:
|
|
$$ - Physics is single-threaded; SetPositionInternal(obj) runs change_cell(obj) then
|
|
$$ set_frame(obj) back-to-back for the SAME object (no reentrancy).
|
|
$$ - enter_cell (called by change_cell) does NOT call set_frame (verified
|
|
$$ acclient_2013_pseudo_c.txt:278928). So the FIRST set_frame after a change_cell
|
|
$$ is that object's committed frame.
|
|
$$ - $t3 is the pending flag: change_cell arms it, the next set_frame consumes it.
|
|
$$
|
|
$$ Offsets (verified live via discover-types.cdb; reused from the original script):
|
|
$$ change_cell: this=@ecx ; seed = this->cell->id = poi(poi(@ecx+0x90)+0x58)
|
|
$$ new_cell = poi(@esp+4) ; new_cell->id = poi(poi(@esp+4)+0x58)
|
|
$$ set_frame: arg2 (Frame*) = poi(@esp+4) ; Frame.m_fOrigin @ +0x34
|
|
$$ new px/py/pz = poi(poi(@esp+4)+0x34 / +0x38 / +0x3c) (raw IEEE hex)
|
|
$$
|
|
$$ Emits the golden format read by RetailTrace.ParseFindCellList; decode the hex with
|
|
$$ tools/cdb/decode_fcl_capture.py (then filter to the doorway cells 0031/0170/0171):
|
|
$$ [fcl] seed=0x<old> px=0x<hex> py=0x<hex> pz=0x<hex> picked=0x<new>
|
|
$$
|
|
$$ Capture in a spot with NO other moving objects (NPCs crossing cells would also be
|
|
$$ logged); the user paces the PLAYER alone in/out of the doorway. Stray rows are
|
|
$$ filtered post-hoc to seed/picked in {0xA9B40031, 0xA9B40170, 0xA9B40171}.
|
|
$$ ============================================================================
|
|
|
|
.logopen C:\Users\erikn\source\repos\acdream\.claude\worktrees\thirsty-goldberg-51bb9b\find-cell-list-capture-aligned.log
|
|
.sympath C:\Users\erikn\source\repos\acdream\refs
|
|
.symopt+ 0x40
|
|
.reload /f acclient.exe
|
|
|
|
r $t0 = 0
|
|
r $t3 = 0
|
|
r $t4 = 0
|
|
r $t2 = 0
|
|
|
|
$$ BP1: change_cell — capture seed (old cell) + picked (new cell), arm pending.
|
|
$$ Do NOT read position here: m_position is still the PREVIOUS frame's value.
|
|
bp acclient!CPhysicsObj::change_cell "r $t4 = 0; .if (poi(@ecx+0x90) != 0) { r $t4 = poi(poi(@ecx+0x90)+0x58) }; r $t2 = poi(poi(@esp+4)+0x58); r $t3 = 1; gc"
|
|
|
|
$$ BP2: set_frame — the FIRST set_frame after a change_cell carries the NEW
|
|
$$ (committed, aligned) position. arg2 = &curr_pos.frame ; m_fOrigin @ +0x34.
|
|
bp acclient!CPhysicsObj::set_frame ".if (@$t3 == 1) { r $t0 = @$t0 + 1; .printf /D \"[fcl] seed=0x%08x px=0x%08x py=0x%08x pz=0x%08x picked=0x%08x\\n\", @$t4, poi(poi(@esp+4)+0x34), poi(poi(@esp+4)+0x38), poi(poi(@esp+4)+0x3c), @$t2; r $t3 = 0; .if (@$t0 >= 16) { .printf /D \"=== DETACH after %d aligned cell changes ===\\n\", @$t0; qd } }; gc"
|
|
|
|
.printf \"aligned find-cell-list capture armed (change_cell + set_frame). Walk SLOWLY in/out of the doorway now.\\n\"
|
|
g
|