Γ£ö Campus contains twenty distinct original characters and nine instrumented stations (17.6518ms)
Γ£ö Campus moves only during running frames and pause cancels scheduled motion (4.5269ms)
Γ£ö Reduced motion preserves fixed positions while real stage and commit counts update (2.6414ms)
Γ£ö Restoration and failed persistence update state without fabricating deliveries (5.1558ms)
Γ£ö Actual commit receipts update immediately without accumulating an animation queue (1.2453ms)
Γ£ö Route helper stops at its waypoint and stage aliases cannot invent unknown work (0.3591ms)
Γ£ö Actual stage progress drives original couriers along orthogonal paths with working input dossiers (14.9642ms)
Γ£ö Pause and reduced motion freeze courier positions and attached files while receipts remain current (1.3862ms)
Γ£ö Fast actual calculations coalesce to truthful receipts with no delayed mail work (2.768ms)
Γ£ö Local campus lighting is bounded and cannot change statistical or receipt state (2.0456ms)
Γ£ö No request before explicit Test/Enable; temporary key never enters state, context or request body (3.3141ms)
Γ£ö One bounded request generates exactly five exchanges then chooses another current nearby partner (4.9529ms)
Γ£ö Schema rejects missing turns, wrong speakers, numeric inventions and secret echoes (0.5525ms)
Γ£ö All twenty fictional personalities are distinct; proximity and recency use actual positions without permanent bans (0.2683ms)
Γ£ö Rate budget includes tests and prevents concurrent requests and automatic bursts (0.3268ms)
Γ£ö CORS-hidden rate headers use bounded jitter cooldown and require manual resume (1.6122ms)
Γ£ö Clear aborts active requests and late responses cannot restore key, chatter or memory (0.4282ms)
Γ£ö Timeout and provider errors are sanitized and do not echo credentials or enable fallback (1.1178ms)
Γ£ö Prefixed transport exceptions cannot echo the temporary key into public state (0.4824ms)
Γ£ö Five complete conversations remain session-only and Clear removes all history (3.219ms)
Γ£ö start is idempotent; pause/resume serializes batches and restores paused (26.5105ms)
Γ£ö pause during append waits for one durable batch then stops (1.0778ms)
Γ£ö storage failure retains completed batch; retry does not recompute or announce failure as a commit (2.1846ms)
Γ£ö new run waits for old in-flight commit and uses distinct ID with batch zero (2.0185ms)
Γ£ö export schema validates, valid import resumes and import never overwrites (2.5915ms)
Γ£ö local checkpoint cannot skip hashed batches or reuse an earlier run ID (2.1931ms)
Γ£ö deprecated block/work/byte cap options never stop normal capture, restoration or forks (5.2969ms)
Γ£ö local verification purpose and safe checkpoint indices replace fixed history limits (1.4117ms)
Γ£ö stage observers follow real computation, hashing and durable transaction order (1.6793ms)
Γ£ö failed persistence emits error without complete; retry never repeats sampling or hashing (1.4913ms)
Γ£ö throwing or rejecting stage observers cannot change calculations or durable commits (1.6175ms)
Γ£ö verified fork preserves parent bytes and continues hashed seed/index, ignoring unstarted config (2.9039ms)
Γ£ö terminal fork export/import and repeated forks restore verified child checkpoints (4.1316ms)
Γ£ö empty or corrupt parent cannot produce a fork or write a child checkpoint (0.7775ms)
Γ£ö failed fork persistence retains one sealed fork; retry commits once and stays paused (2.0599ms)
Γ£ö fork waits for in-flight batch and capture continues across deprecated observation caps (4.9627ms)
Γ£ö immediate pause cancels queued start and invalid parameters do not stop an existing run (0.4776ms)
Γ£ö partial cancellation bypasses a waiting pause without recording or advancing the family (1.2298ms)
Γ£ö cancellation during hashing or persistence preserves the completed family and waits for commit (0.8805ms)
Γ£ö real yielded Gaussian progress permits cancellation and deterministic retry (39.3802ms)
Γ£ö restore exposes loading and genuine validator progress without changing saved records (1.0906ms)
Γ£ö canceling restore retains unverified recovery bytes and retry validates the full saved chain (1.0595ms)
Γ£ö failed restore leaves cancellation state settled and preserves records for recovery and retry (0.9013ms)
Γ£ö BH and BY match standard step-up calculations, preserving original order (2.8277ms)
Γ£ö step-up includes equality, tied p-values, all-zero, and no-rejection vectors (0.3625ms)
Γ£ö Gaussian tail accuracy includes small extreme tails without cancellation (0.618ms)
Γ£ö configuration accepts numeric form values and rejects invalid bounds (0.7055ms)
Γ£ö seeded random stream and Box-Muller sampling are repeatable and finite (2.0619ms)
Γ£ö batch reproducibility, known truth counts, and seeded index separation (2.003ms)
Γ£ö all-null and no-null truth follow actual Gaussian means, including zero effect (0.6349ms)
Γ£ö independent mode ignores rho; common factor zero reproduces independence (0.4672ms)
Γ£ö real calculation stages report ordered boundaries and computed metrics without changing results (6.0522ms)
Γ£ö observer throws, rejected promises, and event data mutations cannot affect mathematics (0.9405ms)
Γ£ö canonical JSON sorts object keys recursively and rejects ambiguous non-JSON values (0.3681ms)
Γ£ö SHA-256 matches known vectors in portable and Web Crypto paths (5.5124ms)
Γ£ö sealed blocks copy payloads; multi-run ledger recomputes and validates (5.2714ms)
Γ£ö validation detects un-rehashed content mutations and broken links (2.5226ms)
Γ£ö re-hashed fabricated results still fail seeded calculation verification (2.1647ms)
Γ£ö strict ledger schema rejects extras, malformed timestamps, duplicate indices, and changed run config (2.5931ms)
Γ£ö makeBlock rejects malformed payloads, sequence, previous hash, and timestamp (0.4273ms)
Γ£ö import preflight rejects excessive work before expensive ledger recomputation (50.004ms)
Γ£ö v2 fork preserves parent bytes and records hash-covered provenance for seeded continuation (2.9181ms)
Γ£ö v2 nested forks inherit the immediate seeded checkpoint without inventing batches (1.0663ms)
Γ£ö fork validation rejects rehashed false provenance, checkpoint, parameters, and reused IDs (4.0778ms)
Γ£ö chunked simulation preserves exact v1 values for both dependence modes and larger configurations (81.8554ms)
Γ£ö chunk cancellation returns no partial batch and a fresh retry reproduces the full seeded result (21.9611ms)
Γ£ö completed calculations return their result when cancellation arrives at the completion event (4.9041ms)
Γ£ö local verification supports uncapped history while imports retain defensive limits and cancellation is read-only (41.0238ms)
Γ£ö large indices are safe integers and impossible array allocations fail clearly before generating data (0.2931ms)
Γ£ö p-value preview uses actual stable ranks, bounds display points, and preserves input (18.3367ms)
Γ£ö preview yields for cancellation and Gaussian preparation announces start before allocating (1.9914ms)
Γ£ö Default sync processes at most 500 raw blocks and retains all accumulated families (8.0955ms)
Γ£ö Same-array append reads the added tail without rescanning completed block payloads or test arrays (6.0506ms)
Γ£ö Displayed points are bounded while their endpoints and running mean indices retain full history (6.0022ms)
Γ£ö Constructor limits bound each sync and the displayed points independently (0.1928ms)
Γ£ö Separate runs accumulate independently and fork blocks count only toward the raw cursor (0.8964ms)
Γ£ö Power uses rounded synthetic truth and treats zero effect and all-null families as undefined (0.189ms)
Γ£ö Welford metrics match the established family-level reference, including no-variation intervals (0.7624ms)
Γ£ö Reading and aggregating a deeply frozen ledger never changes the stored evidence (0.5923ms)
Γ£ö Replacing the source array resets cached history and replays within the per-sync bound (0.3829ms)
Γ£ö Truncating the same source array resets counts instead of retaining removed families (0.3219ms)
Γ£ö Replacing the processed tail resets metrics and replays the corrected source (0.2312ms)
Γ£ö Asynchronous rebuild yields between chunks and finishes the entire ledger without a display cap (3.7421ms)
Γ£ö A newer source takes ownership while an older asynchronous rebuild is yielding (0.1777ms)
Γ£ö Returned chart point objects are detached from the cached display state (0.1701ms)
Γ£ö Fork provenance is available per branch before any batches without entering its metrics or latest batch (0.2622ms)
Γ£ö Replacing, resetting, or truncating the source removes stale fork provenance (0.1804ms)
Γ£ö Both former parameter ceilings are removed while unsafe arithmetic is rejected (0.9567ms)
Γ£ö One bounded 513 by 201 batch yields, records every test and preserves its seeded result (119.8278ms)
Γ£ö Partial computation can be canceled at its first tiny chunk and resumes deterministically (483.9754ms)
Γ£ö Deprecated block, replay and byte options cannot stop capture or trim saved history (2.3288ms)
Γ£ö Canceling cannot discard a completed batch awaiting a storage retry (1.0102ms)
Γ£ö Actual progress updates the campus without pretending a stage or block completed (3.3411ms)
Γ£ö All local hours use dawn 5-7, day 7-18, dusk 18-20, and night 20-5 (18.8863ms)
Γ£ö Lighting changes at exact local-hour boundaries and handles midnight (0.715ms)
Γ£ö Overlay is restrained and lamp intensity increases into night (1.5384ms)
Γ£ö Browser UMD export uses local getHours and a default-local Intl formatter (1.0544ms)
Γ£ö Default date uses the current local clock without changing simulation state (1.2698ms)
Γ£ö The same instant follows actual subprocess local timezones (167.9042ms)
Γ£ö Lighting is deterministic, does not mutate input, and rejects invalid dates (0.6173ms)
Γ£ö Monte Carlo SE uses family-level unbiased variance divided by replication count (0.8254ms)
Γ£ö No-alternative power and one-batch uncertainty are undefined (0.1171ms)
Γ£ö Approximate interval requires 30 replicated families and is bounded (0.1999ms)
Γ£ö Nonzero family variation yields the usual Monte Carlo interval after 30 families (0.929ms)
Γ£ö append atomically saves canonical block and next checkpoint; reload restores both (11.4332ms)
Γ£ö quota abort rolls back block and checkpoint without hidden partial writes (2.222ms)
Γ£ö two connections cannot overwrite a newer run or append from a stale head (1.2898ms)
Γ£ö wrong checkpoint and broken head are rejected before writes (1.2488ms)
Γ£ö import is atomic and requires both ledger and checkpoint empty (1.2483ms)
Γ£ö missing, blocked and closed IndexedDB report explicit failures (0.6757ms)
Γ£ö fork append checks actual hashed parent and atomically installs child checkpoint (2.3787ms)
Γ£ö invalid or failed fork append preserves every parent record and checkpoint (2.738ms)
Γ£ö paged reads yield between chunks and load preserves its captured immutable prefix (4.1439ms)
Γä╣ tests 113
Γä╣ suites 0
Γä╣ pass 113
Γä╣ fail 0
Γä╣ cancelled 0
Γä╣ skipped 0
Γä╣ todo 0
Γä╣ duration_ms 774.9249
