Agent was stopped for token budget. Landed here: probe flag through
LiveSessionConnectOptions + the StartCore short-circuit, the mode field
with JsonRequired-to-semantic-validation move, host exit-code mapping,
and 34 passing tests including 3 new probe tests (agent last reported
green before the stop). NOT DONE: the idle-policy unit tests (next
step), full-suite verification, and the WSL run.
Build/test state UNVERIFIED at this commit. Next session: finish idle
policy tests, run Runtime+Headless Release suites Windows and WSL, then
dispatch the Opus dual-lens review.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The second live gate run exposed a real environmental hazard: this shared
ACE dev instance has a THIRD player character online (+Je, guid
0x50000001), and after @teleallto it ended up nearer to the Leader bot than
the actual Recruit bot (+Horan, 0x5000000B). RuntimeFriendlyTargetQuery.
FindClosestOtherPlayer — "nearest ANY other player" — picked +Je, and the
fellowship recruit sent to it obviously never completed (confirmed live:
WaitRecruited/WaitForRecruit both timed out, both bots quarantined and
gracefully logged out cleanly).
RuntimeFriendlyTargetQuery.FindPlayerByName resolves the nearest player
whose streamed name matches exactly, with 3 new conformance tests
(preferring the named player over a closer stranger, returning null when
absent, and case-sensitivity/hidden/no-draw/self rejection).
FellowshipAllegianceGateCoordinator (AcDream.Headless.Policies) is a small
same-process, no-locking (single update thread) carrier for the Recruit
bot's own discovered character name — set by its own HeadlessSessionHost
the instant CharacterList selection resolves it, which IS D8's "discover it
live" mechanism, not a hard-coded value. Constructed once per
HeadlessProcessHost and threaded through HeadlessBotPolicyFactory.Create
into the Leader policy, which now name-matches instead of taking whichever
player entity happens to be closest.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
TWO work products share this commit (a staged-index collision between the
coordinator's docs commit and the OP7 fixer's staged files — content
verified complete and coherent; only this message was wrong before the
amend):
1. OP7 review fixes (all nine findings from
docs/research/2026-08-11-op7-review.md):
- M1: HeadlessSessionDescriptor is a record; WithAccount uses 'with' non-destructive record copy,
so a future property cannot be silently dropped; direct-CLI
regression test proves CharacterOptions survives --user/--password.
- M2 root fix: LiveSessionEventRouter skips BOTH Replace and the
options notification on a trailer-truncated PlayerDescription — a
truncated re-seed can no longer install zeroed words under an armed
latch for OP7's automation to flush into 0x01A1.
- SF1: schema keys validate as ordinal strings against the allowed
names (numeric / comma-combined aliases rejected). SF2: both-true
fellowship exclusion rejected at load, naming both keys. SF3: the
onLoginCompleteSent observer moved after transit.EndTeleport().
SF4: production-hook coverage for all three LoginComplete sites.
SF5: test-script OP7 wire expectation corrected (batched ids ride
only the 0x01A1).
2. docs/research/2026-08-11-op3-rereview.md — OP3 re-review verdict
REOPEN (narrow): M1 byte-decode independently re-verified (6a 07 at
all six sites); residuals R1 (gate script promises a timestamp prefix
acdream doesn't render), R2 (null-controller player-mode still
refuses), R3 (dormancy pin lacks stimulus) — coordinator fixes follow.
Full Release suite at this tree: 12,956 passed / 4 skipped / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Runtime's contract is ONE update thread per session for its whole
lifetime — RuntimePhysicsState.EnsureCollisionMutationThread enforces it
for collision generations (bind-first-mutator, refuse migration), and
the entity directory, physics publication, and placement channel all
document the same assumption without enforcing it. The graphical host
satisfies the contract with its game-loop thread. The headless host
violated it structurally: HeadlessProcessScheduler.RunAsync drove ticks
through await Task.Delay(...).ConfigureAwait(false), and a console app
has no SynchronizationContext, so each resumption could land on a
different ThreadPool worker. Any collision generation spanning two waits
then tripped the guard — reproduced 3/3 against live ACE at
[wake] begin gen=1 (see docs/ISSUES.md #368).
Fix shape (headless-only; zero shared Runtime changes, so the graphical
host is untouched by construction):
- HeadlessProcessScheduler.Run(CancellationToken) replaces RunAsync: the
same deadline math, counters, and NormalizeTimerDelay clamp, but fully
synchronous on the calling thread. Waits go through one rearmed
TimeProvider timer signalling an event (WaitHandle.WaitAny with the
cancellation handle), so the loop never leaves its thread and returns
normally on cancellation.
- HeadlessProcessHost.RunAsync now spawns one named dedicated thread
("acdream-headless-update") that owns Start (the live connect
transaction), every scheduler turn, and the post-loop resource
captures, bridged to the same Task<HeadlessExitCode> via a
TaskCompletionSource. Start had to move too: the first
collision-mutating call can happen during connect, and binding the
guard on the caller's thread would trip the very first dedicated tick.
Disposal stays on the lifecycle thread, which the Runtime teardown
path explicitly supports (ResetSessionPhysics's doc comment) and every
prior graceful-teardown run exercised.
New test ProcessHostRunsStartAndEveryTickOnOneDedicatedUpdateThread
pins the contract: Start and every tick share one thread that is not
the RunAsync caller's, across real timer waits (RED pre-fix — Start ran
on the caller's thread). SystemTimerCadenceDoesNotBusyLoopBetweenTurns
moved to the synchronous seam and still bounds WaitCount.
Verification: Headless suite 97/97; full Release suite 12,554 passed /
4 skipped / 0 failed; three live jump-probe runs against local ACE
(ACDREAM_PROBE_PARK=1) each crossed the collision generation cleanly
(205 entities hydrated, zero faults, policy completion, ACE-confirmed
graceful logout, converged disposed sample, exit 0) — pre-fix the same
recipe quarantined 3/3. The jump-airborne timeout persists 3/3 on the
fixed tree, refuting the #365 diagnosis's "threading artifact"
hypothesis for it — filed separately as #370.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
SHOULD-FIX 1: RetailClientCommandCatalog's ~45 catalog leaf verbs were
showing acdream-authored Summary text for /help <verb> instead of
retail's own Detail_HelpType(2) text. Byte-swept every Help* handler
against the PDB-paired acclient.exe (verified MATCH), confirmed each
Detail/Summary branch by reading the actual decompiled if/else shape
(address order and string length both proved unreliable alone), and
fixed a sweep_weenie_strings.py 800-char truncation bug that silently
dropped several longer Detail branches. Resolved every ambiguous
CmdHashData-registered verb (hor/hr/hom/hoa/alh/ah/friends_add/
friends_remove/squelch/unsquelch) by reading for Binary Ninja's
nullptr-4th-arg decompiler artifact instead of trusting it. Coverage:
42 of 47 distinct catalog Definitions verbatim-extracted, 4
confirmed-null (index/clist/on/off register with a genuinely null help
pointer — DoHelp falls to UnknownCommand for these, now reproduced),
1 honest UNVERIFIED (messagetypes builds its text from a runtime enum
table, not a static string). ChatCommandRouter now prefers retail
Detail text over the catalog summary; RetailCommandHelpTable's class
doc no longer overclaims its own scope.
SHOULD-FIX 2: extracted the a5a7eb4f-class OnInterfaceText wiring into
a testable CreateChatViewModel method and added
ComposedChatViewModelWiresOnInterfaceTextToSpewBox, which the prior
FakeFactory-based test suite could never exercise.
SHOULD-FIX 3: retires register row AP-113. DoLifestone/DoMarketplace
print their own 0x1A refusal text (byte-recovered, UTF-16LE) instead
of falling through to the generic 0x26 fallback; ChatCommandRouter's
comment corrected to state the fallback's real scope.
SHOULD-FIX 4: corrected the divergence register's stale AP section
header sentence about AP-190's opacity default (refuted by cc582899).
NITs: (a) HeadlessStaticStateAudit routes through the injected
HeadlessDiagnosticWriter instead of Console.WriteLine; (b) a bounded
300-pump liveness diagnostic on the IsQuiescent conductor gate (no
retry, no behavior change); (c) fixed the #365 hydration test's doc
comment contradiction against diagnosis §8; (d) the 0x26 fallback
dispatches on WeenieErrorMessages' own Type instead of hardcoding
ClientLocal.
Full Release suite: 12,553 passed / 4 skipped / 0 failed (baseline
03404b71: 12,542/4/0; net +11 tests, zero regressions).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Root cause (measured live via ACDREAM_PROBE_PARK=1): HeadlessSessionWorldProjection
drove the first-entry conductor unconditionally, including while
HeadlessCollisionNeighborhood's own 3x3 publication plan held a genuinely open
RuntimeCollisionAdmission for the local player's landblock. Every
TrySealCollisionEvaluationAuthority attempt during that window failed
(IsCollisionEvaluationPrefixAdmissible false) and retried forever without
recovering — measured verdict: "seal-refused" repeating with no preceding
[rearm] verdict= line (the operation never even reached the AwaitingCell park).
This is the diagnosis doc's "structural half" mechanism; no evidence of the
"circular HasOldPrefixPlacementDebt" hypothesis was observed, so that shape
was not needed.
Step 1 (enabler): HeadlessStaticStateAudit.ValidateProcessIsolation now takes
sessionCount and only refuses process-global physics probes for
sessionCount > 1 — its own multi-root-attribution rationale never applied to
a single session, and it was blocking the exact probe built to diagnose this
class of stall.
Step 3a (root cause): new IHeadlessCollisionNeighborhood.IsQuiescent gates
ProjectSpawn/ProjectPosition/PumpFirstEntry's conductor-drive calls — the
conductor is never driven while the neighborhood's own publication owns
collision authority for that tick.
Step 4 (defense-in-depth): HeadlessLocalPlayerFrameHost.CanAdvancePlayer now
requires Controller.CanExecuteLiveMovement instead of just a non-null
controller — the headless-only gap that turned the (now-fixed) hydration
stall into a hard crash reaching SuspendObjectUpdate on a dormant controller.
RuntimeLocalPlayerFrameController's three shared entry points gained the same
guard, contract-preserving for the graphical host.
Verified end-to-end against live ACE (jump-probe policy, three runs):
hydration succeeds cleanly (136 entities load vs. 0 before), no seal-refused
spam, no crash from the original bug, graceful logout every time. Full
airborne-transition confirmation is blocked by a separate, newly-discovered,
pre-existing defect filed as #368 (the headless scheduler's
Task.Delay(...).ConfigureAwait(false) tick loop can resume on a different
ThreadPool thread mid collision-generation, tripping
EnsureCollisionMutationThread) — explicitly out of scope here, not mentioned
anywhere in the #365 diagnosis, and unsafe to fix without graphical-host
verification this session was constrained not to perform.
New tests: the real-admission hydration test (fails on the pre-Step-3a tree,
verified by temporarily reverting the three gates and confirming failure,
then restoring), the PumpFirstEntry quiescence-gate test, the
CanAdvancePlayer publication-lifecycle test, the dormant-controller
sabotage tests for RuntimeLocalPlayerFrameController, and the audit
single/multi-session tests. RuntimeLocalPlayerPhysicsPublicationStateTests
is untouched.
Full Release suite: 12,343 passed / 4 skipped / 0 failed (baseline ~12,330/4
plus 11 new tests).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
