Task 4, re-scoped per the plan's honest-FAIL branch: the c3 class is NOT
retired — every record now states the measured post-ALT truth.
- KNOWN-LIMITATIONS: the cross-metro section is the /cari-shaped record —
WHAT CHANGED (ALT built, D-03 optimality green quoted) / what REMAINS
(the CPU verdict FAIL lines quoted with date; the one-directional-
Lipschitz wall named; the escalation open); the roadmap-pointer ALT row
flips to Built (Phase 8) with the measured verdict — never 'not built'
again
- ROAD-GRAPH: the ladder table gains the (c2) rung (EXECUTED, measured
FAIL, the L=32 probe quoted); the decision paragraph states c2
executed-verdict-fail and names the real remaining options; new 'The
ARG2 binary layout' section (72-byte header, landmarkCount/
landmarkQuantumS, u16 suffix + sentinels, the quantum knob's ≈6,553 s
ceiling, the storage-name-vs-format-truth law); key facts carry the
optimality gate and the ARG2 wording; the fork contract's build step
teaches the landmark stage (no new commands)
- docs/CONFIGURATION.md: verify-only — all path/script claims still true,
no edit needed
- staged as selective hunks; the tree's pre-existing doc-sync edits
(battery paragraph, 'six maintainer scripts' intro) remain uncommitted
greps: 'Next-milestone item, not built' = 0; the measured p95 strings
quoted; (c2)/landmarkQuantumS/ARG2 all present; ARG1 appears only in
layout/router-pending contexts.
Task 3, re-scoped per the plan's honest-FAIL branch: the CPU verdict FAILED
(the ladder ran to its documented stop), so the c3 class is NOT retired —
the record instead carries the executed-and-measured ALT truth.
- budget-gates.mjs: the verdict law documents the Phase-8 outcome (ALT
EXECUTED, optimality green, CPU verdict FAIL 152.788/185.692 ms quoted
with date; L=32 probe measured worse) — the stale 'next-milestone item'
wording is gone; ROUTE_CPU_DRIFT_GUARD_MS re-anchored 200 → 400 (the
post-ALT documented band outgrew the old guard; ≈2x the worst shipped-
config run 185.692 ms); the guard stays a regression trip-wire, the 8 ms
budget stays and its FAIL still prints every run
- budget-audit.test.mjs: CPU3 pins the new law (evidence must name the
EXECUTED-in-Phase-8 verdict and must NEVER say 'next-milestone');
19/19 vectors ok (the count holds — no vector dropped)
- docs/BUDGETS.md: route-cpu row carries the Phase-8 measured numbers +
date; the c3 section is the honest post-ALT record (what changed, the
quoted gate lines, the one-directional-Lipschitz wall, the open
escalation) — staged as selective hunks; the tree's pre-existing
doc-sync edits in other sections remain uncommitted, untouched
grep law: zero 'next-milestone' hits in budget-gates.mjs + BUDGETS.md;
ROUTE_CPU_DRIFT_GUARD_MS still exported and used.
Task 1: the D-03 gate script (measure-optimality-gate.mjs) reads the frozen
pre-ALT record, re-routes the same coordinate-pinned pairs on the current
ARG2 binary, and gates on seconds-equality (±0.1%) + verdict-class
preservation — never on node paths (equal-cost ties are documented behavior).
Rule-1 deviation the gate forced on its first run: routeAstar returned
SUBOPTIMAL real-graph routes (x038: 1234.6s vs the 1214.8s Dijkstra optimum;
x045: 1475.5s vs 1460.6s) — the fixture sweep could not see it. Root cause:
the fromTable bound d̂(ℓ,t) − d̂(ℓ,v) is one-directional Lipschitz (its u-side
term needs the REVERSE edge; a oneway does not provide one — the anchor
reached node 264611 126.8s faster than its own 1.49s upstream neighbour
264610), and the shared Ikeda average p = (h_f − h_b)/2 is consistent only
when every family inside h_f/h_b is two-sided. The toTable bound
d̂(v,ℓ) − d̂(t,ℓ) rides stored traversable edges in BOTH directions; it alone
joins the haversine now. Stop rule, reductions, meet probes, reconstruction:
untouched. landmarkTables/build verified EXACT (0 mismatches vs independent
integer Dijkstras, both rows, all 24 anchors).
- measure-optimality-gate.mjs: exported core + CLI guard (route-corpus
pattern); epsilon mirror + argVersion asserts precede any measurement;
per-pair rows, extract-drift WARNING (never a re-record), honest exit law
- astar.ts: potential() consults landmarkTo only; doc-comments carry the
one-directional-Lipschitz law and the x038/x045 evidence
- astar.test.mjs: A11 re-pinned (landmarkFrom must NOT appear in the
comment-stripped search — re-adding it trips the pin); A12 oneway-ring
sweep added (directed reach asymmetry stays Dijkstra-equal on every pair)
Battery: routing 80/80, bot 126/126, both typechecks green.
D-03 gate: GATE PASS — 53/53 routed pairs equal within ±0.1% (max |Δ| 0.005%),
verdict classes preserved (6 unreachable + 41 unsnappable, 0 re-snapped).
- The stage between edge compaction and encode feeds the graph's edges to
landmark.ts (farthestPointAnchors + landmarkTables + landmarkCoverage) —
the script quantizes and implements nothing itself (single-authority law)
- D-02 scaling rule L = clamp(round(N/13000), 16, 32) (~24 at KL's 311k
nodes) with a --landmarks <n> measured knob (positive int, hard cap 128,
parsed before --check)
- Honest gates before any write: quantum-ceiling throw -> fail with the
raise-quantum hint (Pitfall 5); per-anchor coverage >= 0.9 (the SAME
fraction for any L — the override cannot lower it, T-08-11); anchor
shortfall fails loud instead of encoding a weaker ALT
- encodeGraph receives landmarks: { quantumS, nodes, to, from } — the one
ARG2 authority; no local fallback encode
- Stats gate prints L/quantum/coverage min-mean/table bytes/sentinel totals
BEFORE the size gates; --check mirrors the facts from the decoded binary
(landmarkCount/quantum/anchors/coverage/sentinels, no network)
- A1 ladder history gains rung (c2): EXECUTED in this preprocessing stage
- RO1 upload-format-agnostic: upload-road-graph.mjs keeps KEY
'kl-roadgraph.v1.bin' + BUCKET 'araharah-graphs'; comment-stripped code
contains ZERO format tokens (landmark/encodeGraph/decodeGraph/ARG1/ARG2
magics/GRAPH_VERSION) — the migration is byte-shipping of the same key,
forever (A8 grep-shape)
- RO2 one decode path: worker-router.ts keeps GRAPH_KEY unchanged, exactly
ONE decodeGraph call in the loadGraph flow, no version-conditional branch
(GRAPH_VERSION / magics / '=== 2' absent) — an undecodable binary flows
to the existing null -> router-pending path (D-01 at the consumer)
- full battery at this commit — the FIRST where routeAstar consults real
landmark tables on the bot suite's ARG2 corridor graphs: routing 79/79
(A1 seal + A9/A10/A11 + RO1/RO2), bot 126/126 (RT trunks + cron),
routing+bot typechecks clean; worker-router.ts and upload-road-graph.mjs
untouched (git diff --stat empty)
- routeAstar's ONLY change: the h_f/h_b bodies become max(ALT triangle
bounds, haversine) — per-search Float64Array(L) constant tables in quanta
(0xFFFF written as NaN = search-scoped landmark drop), lazy per-node
max over the four triangle families with RAW u16 sentinel compares
BEFORE any float conversion, haversine max-fallback on sentinel-flooded
nodes (Pattern 3 containment)
- the (h_f - h_b)/2 combination, reductions, balanced alternation, meet
probes, BH stop rule, and real-seconds reconstruction are byte-identical
(research Pattern 1 — the survey's (pf-pr)/2 is what p_f already is)
- no Map/Object allocations (A8 green), no new imports, one search path
(L >= 1 is the decodeGraph guarantee, D-01)
- A1's 625-pair Dijkstra-equivalence sweep re-pinned under ALT: the
standing optimality proof; A9 island + A10 skip laws sealed; A11
structural pin green
- module header + routeAstar doc gain the ALT paragraph (c-metric
exactness, sentinel laws, Ikeda history intact)
- A9: two-component synthetic (3x3 grid + 4-node path), anchors HAND-PINNED
inside component A only via landmarkTables; full 169-pair sweep asserts
seconds-only Dijkstra equivalence (A-internal with ALT live, B-internal
on the haversine fallback, 72 cross pairs -> null)
- A10: origin AND goal inside the anchorless component — all families drop
at SEARCH scope, the 16 B-internal pairs still route Dijkstra-equal
- A11: comment-stripped astar.ts must reference landmarkTo/landmarkFrom/
landmarkQuantumS in CODE; one sentinel spelling (0xffff, never 65535);
A8's no-Map law unchanged — FAILS now (RED: the potential is still
pure haversine)
- case-index doc-comment extended A1-A8 -> A1-A11; buildIslandFixture
helper routes fixture tables through the ONE landmark module
webhook corridor/sparse builders + guidance-cron unnamed way compute
tiny tables through the ONE preprocessing core (@araharah/routing
farthestPointAnchors + landmarkTables) and pass the landmarks block
into the same single encodeGraph call — call-site updates, not a
second encode function. Graphs with no degree>=2 candidate (sparse,
2-node) pin their anchor explicitly; tables still come from the
integer-Dijkstra core.
FULL BATTERY at this commit: routing:test 74/74, bot 126/126, both
typechecks clean; the regenerated fixture is deterministic (byte-
identical re-run); mini-graph-arg1.bin is byte-stable vs its committed
08-01 bytes.
graph.ts versioned IN PLACE (no second codec): GRAPH_MAGIC 0x41524732,
GRAPH_VERSION 2, HEADER_SIZE 72 (72%8==0 keeps f64 views aligned);
header gains landmarkCount u32 @60 (the old pad), landmarkQuantumS f32
@64, reserved u32 @68; the landmark suffix (u16 toTable, u16 fromTable,
u32 landmarkNodes) extends layoutFor's total arithmetic — exact-size
law covers the variable suffix BEFORE any view allocation (T-08-01).
D-01 fail-honest: ARG1 bytes / version!=2 / landmarkCount=0 / hostile
quantum / size mismatch all decode null (committed mini-graph-arg1.bin
is the standing rejection vector). GraphInput.landmarks REQUIRED and
loud-validated (T-08-04). RoadGraph gains the five landmark fields
(Plan 08-03's ALT potentials consume them).
Fixture regenerated as ARG2 through the SAME encode authority: the
builder gains the tiny real landmark stage (L=2, anchors [0,18],
quantum 0.1s, coverage 1.0 — identical bytes every re-run).
Rule-3 call-site updates forced by the required-landmarks law (missed
by the plan's files list): astar.test A3 + snap.test S3 synthetics
carry inert hand landmark blocks. Fixture regen pulled forward from
task 3 so this commit's own battery is green (74 passed, 0 failed).
farthestPointAnchors + landmarkTables + landmarkCoverage over the
ĉ metric: edge weights floor-quantized once (Math.floor(seconds/q)),
exact integer Dijkstras (Int32 labels, binary heap, lazy deletion),
0xFFFF sentinel, 65000-quantum loud overflow. Traversal law mirrors
astar.ts verbatim: bit0 edges never relaxed in EITHER direction.
Test-vector corrections caught at GREEN (expectations were wrong,
implementation was right):
- L4: rows index by anchor POSITION (row-major [l*N+v]), not node id
- L6: the c-metric farthest node from corner 0 is interior node 18
(297q: interior hops cost 2x perimeter), not corner 24 (264q) —
verified against an independent brute-force Dijkstra
A 2-node graph has no degree>=2 candidates, so farthestPointAnchors
returns empty before any Dijkstra runs and never reaches the overflow
throw. Node 1 on a 3-node path is a candidate, so the selection pass
actually exercises the law.
Hand-computed vectors for the pure landmark compute core: deterministic
degree-filtered farthest-point anchors, exact integer quanta
(floor(2.55/0.1)=25, two edges=50), 0xFFFF sentinel + self-cell laws,
oneway asymmetry, the 65000-quantum overflow throw, and the 5x5-grid
corner-to-corner 264-quanta vector with coverage 1.0.
- Part A: the ONE consolidated walk (D-05) — four clients × seven
interactions, ⬜ cells, PASS/FAIL/NOT RUN vocabulary defined in-file
- Part B: the Android battery record (D-04) — 11 required fields, the
no-extrapolation rule in the file's own voice
- Part C: the 12 automated-half commands with ⬜ result cells + the
--live NOT-RUN-is-not-a-pass note
- Part D: the six D-09 checklist items, each naming its evidence
- Executor notes: the /cari defect ROOT-CAUSED+FIXED (script param
name), the c3 posture, the R2 graph state, the KV-list over-ceiling
options (user decision), the kv-expiry credential caveat, never-fake
- FINAL VERDICT: ⬜ NOT YET SIGNED — the human fills the verdicts
- the Bot API parameter is 'scope' (verified against
core.telegram.org/bots/api/methods/setmycommands); the script's
'bot_command_scope' was an unknown field Telegram silently ignores,
so every call hit the default scope and the two clear calls wiped the
default set entirely — the ok:true lines proved nothing about scope
- registration re-run: default (agnostic/en/ms) + all_private_chats
(agnostic/ms) read back help, privacy, cari via getMyCommands with
the correct 'scope' param; group scopes verified cleared
- docs/KNOWN-LIMITATIONS.md: the 'Telegram-side defect' attribution
corrected to the actual script bug; residual client menu-caching
caveat + type-the-command workaround kept
- bot/README.md command row updated to the real scope set
- banner takes one step down the SAME type-token scale in landscape
(headline-md distance / label-lg street) + xs vertical padding so the
measures fit the 72px min-height instead of overflowing it
- guide-list chrome offset corrected to its own margin stack
(sm + 72 + 1 + 24 + sm = 113px, was 104px)
- bottom register: the portrait 64px control-clearance composite already
seats the guidance/confirm card inside the landscape height — the
landscape override is removed (measured card top 159 vs banner bottom
103 at 844x390); attribution chip + recenter control clear
- device-walk feedback on 07-02's committed geometry; tokens binding
- captures re-run: CAPTURE OK — 22/22 non-blank
- /start sendPhoto now points at MASTER_FILE-arah2_logo.png (deployed
from web/public/ — the untracked asset lands here so the reference
resolves on deploy)
- README + AGENTS.md + the 4 mockup screens swapped the Stitch
googleusercontent logo placeholder for the local master logo
- pre-existing on the tree through 07-03..07-05 execution; verified by
this session's baselines (bot 126/126, web 165/165, captures 22/22)
- source-side runtime-hosts gate gains DOC_POINTER_STATIC_TEXT_HOSTS (the
MapLibre-attribution precedent): the published repo origin may appear
ONLY as a bare URL literal in bot copy (D-08/D-09), never fetched
- class is NARROW and pinned by SA6: a fetch( target naming the same
host is still a violation
- Rule 1 (found by SA6's RED): the fetch( handler's declaration
heuristic misclassified any quoted absolute URL as a TS declaration —
quoted strings are targets; they now reach the host check
- docs/PRIVACY-AUDIT.md: adjudication documented + dated re-run addendum
- deferred-items.md #1 closed: npm run privacy-audit exits 0 again
(PRIVACY AUDIT CLEAN — 5 gate(s) passed, 0 violations; suite 13/13)
- asset-header probe reads HTTP status BEFORE interpreting headers — a
non-2xx is deployment drift (the notRun exit-2 class), never a cache
violation; the violations.length count bug fixed
- XHTML namespace URI joins the SVG spec-identifier adjudication
- probeUnavailable carries the reason only (the CLI prints NOT RUN)
- CP2/AU1 suite vectors updated to the corrected contracts
Residue note: these edits were on the tree during 07-03/07-04 execution
but escaped those plans' commits (07-04-SUMMARY residue caution).
- 3 tasks (TDD RED→GREEN), 7 commits, 19/19 budget-audit suite vectors
- BUDGET AUDIT CLEAN (exit 0); --live exits 2 on deployment drift (honest class)
- 2 deviations documented: the c3-aware route-cpu verdict (plan contradiction),
the golden-corpus environment-drift class
- open items for 07-06 flagged: the KV-list cadence finding + the privacy-audit
RED (07-05's gitmirah strings) — both in deferred-items.md
- a missing package.json in the audit root is T-07-22's environment-drift class,
not a corpus failure — the corpora cannot even be spawned there
- AU1's flag-skip assertion matches the printed 'NOT RUN — (pass --live)' shape
- the full ceilings table (Cloudflare Workers limits / KV limits / pricing, sourced)
- current measurements: A* p95 83.989 ms vs 8 ms ladder budget (the accepted c3 class,
negative headroom printed), BA9 79,815/204,800 gz, app bundle 479,178/500,000 gz,
KV write classes (cron = 0, source-proven), golden corpora exit 0
- the per-request-class subrequest inventory (source-derived, not a live Workers Logs metric)
- the DISCOVERED finding: KV list rate 1,440/day (60 s cron x 1 list/tick) OVER the
1,000/day ceiling — documented with options, decision deferred to the 07-06 launch gate
- the c3 cross-metro escalation restated; honest Statement; unsigned signature block
- npm run budget-audit: 5 required gates + the --live cache-header probe; one evidence line per gate
- the FREE-TIER COMPARISON table: every documented Cloudflare free-plan ceiling beside the
current value where measured, `not measured` where not — the console form of docs/BUDGETS.md
- exit 0 CLEAN / 1 FAILED / 2 NOT RUN — a skip is never a pass (T-07-22)
- no secret, no coordinate in the output (D-14's category-only law, applied to budgets)
- route-cpu: RE-RUNS routing/scripts/measure-route-cpu.mjs as a child process; p95 printed
against the 8 ms product budget + 10 ms free ceiling; the child's GATE FAIL line carried
verbatim; the accepted c3 band has a 200 ms drift guard — uncharted regression fails
- bundle-budget: BA9 route-payload gz sum + whole-app bundle ceiling (500000 gz bytes);
Worker bundle honest NOT MEASURED when bot/dist is absent (never zero)
- kv-writes: 4 write sites inventoried by family, every write TTL-named (PG1); the cron
zero-write law is source-proven — runGuidanceCronTick's body must contain no .put(/.delete(
- subrequests: per-request-class external-subrequest inventory derived from the route table;
uninventoried classes FAIL; source-derived honesty sentence verbatim; never a fabricated zero
- cache-headers: reuses cloudflare-probes' runAssetHeaderProbe (imported, never forked)
- golden-corpus: RE-RUNS routing:corpus + search:corpus as child processes (SC4.a/SC4.b)
araharah-executor