Explains and fixes the Phase C partial failure WITHOUT changing a single wire field. The operator saw a difference between the one-item probe (Time Remaining "-") and the generalized rows (Time Remaining "Expired"). Cause: route coverage, not encoding. /tradePile advertised the unlisted tradeIds while ISVIEWTRADE (GET .../trade/status) resolved ids from the market store only -- and an unlisted pile member has no listing row, so the poll returned an empty auctionInfo. Observed live as `route=market-status requested=1 returned=0` repeating for the row the operator had selected, while that same id was present in /tradePile. The client polls status for the row it displays and degrades it when the answer is empty, which is also why no actions were offered. The probe showed "-" only because the client had not yet polled that id (logs of the time show only tradeIds=1000000097). So expires, tradeState, itemData.itemState and pile were all innocent. Nothing was guessed and no field changed: both routes now share one pile enumeration (Server::resolve_trade_pile), so an id advertised by /tradePile always resolves on /trade/status. The corpus predicted exactly this -- tradeId must resolve across /transfermarket, /tradePile, /watchList AND /trade/status; we had stability but not coverage. Same defect class as the original empty-trade/status bug. Status still answers only the ids actually asked about, and a real auction always wins over an inactive row for the same tradeId. Regression test covers all four cases. Records the downgraded conclusion: "inactive" is a CONFIRMED section/lifecycle discriminator; whether the full actionable contract is now complete is the operator's next test. itemState/pile recovery was queued on the assumption the encoding was incomplete -- neither was touched, and both remain the next candidates if actions are still absent. 342 tests pass, 0 failed, clippy clean. Verified live: the six inactive ids went from returned=0 to returned=6.
openfut-utas-host
The first live FIFA 17 UTAS migration host. It fronts the client-visible UTAS port and migrates one route at a time to OpenFUT Core, proxying everything else to the Python UTAS oracle so the rest of FUT keeps working unchanged.
FIFA 17 ──HTTP──▶ openfut-utas-host
├── GET …/club ──▶ FIFA17 adapter ──▶ OpenFUT Core (/collection)
└── everything else ──▶ Python UTAS oracle (verbatim reverse proxy)
What it owns / does not own
Owns: socket + HTTP/1.1 keep-alive transport, route classification, the Core
access client, the Python passthrough, and diagnostics. It owns no game
domain state — filtering/pagination is Core's; wire parsing/shaping is the
adapter's. The adapter never learns how Core is reached (the architecture rule):
the host holds the [CoreAccess] boundary (GET {core_url}/collection?… today).
Safety model
- Classification happens once, before execution. Exact
GET /ut/game/<title>/club→ Rust; everything else → Python. No shared path, no "try Rust then Python". - A Core failure on
/clubdegrades to a valid empty{"itemData":[]}and logs an error — it never falls back to Python (which could double-apply a mutation on other routes)./clubis read-only, but the rule is absolute. - Mutating routes (PUT/POST,
/squad,/purchased, quick-sell, market, auth, SBC,/club/stats/*,/clubUser) all classify to passthrough and are untouched.
Configuration (env)
| Var | Required | Default | Meaning |
|---|---|---|---|
OPENFUT_UTAS_HOST_ADDR |
yes | — | where this host listens (client-visible UTAS addr) |
OPENFUT_UTAS_PYTHON_URL |
yes | — | Python UTAS oracle base URL for fallback (must differ from this host) |
OPENFUT_FIFA17_CATALOG |
yes | — | FIFA 17 card-definition identity catalog (Fifa17CardCatalog JSON: card id → asset id) |
OPENFUT_IDENTITY_STORE |
yes | — | persistent external-identity store file (owned instance → stable wire id) |
OPENFUT_PERSONA_ID |
yes | — | FIFA persona id stamped on GET /squad/active (must match the persona LSX/Blaze/POW/UTAS agree on) |
OPENFUT_CORE_URL |
no | http://127.0.0.1:8080 |
OpenFUT Core base |
OPENFUT_FIFA17_TABLES_DIR |
no | fifa17-recon/data/tables |
leagues/nations/teams.json for id⇄name |
Startup fails clearly if the catalog or identity store cannot be loaded — there is no placeholder fallback (exactly one production identity path).
Identity model (resolved)
FIFA renders an owned card by resolving resourceId & 0xffffff against the
client's own local players table; an invented id renders a blank generic
card (proven live — fut_cards.py:11-21). Two distinct identities, never
conflated, are resolved by [Fifa17IdentityResolver] (the single production path):
- Definition identity (
resourceId/assetId) — the card's real FIFA asset id, from the versionedOPENFUT_FIFA17_CATALOG. An unmapped definition is dropped and counted, never faked. - Instance identity (
id) — a stable, persistent, reversible wire integer from the genericopenfut-identitystore under the FIFA 17 wire-id policy (monotonic from100_000_001). The same owned instance keeps its id across restart and reverses exactly; two copies of one definition share aresourceIdbut get distinctids. The namespace is globally monotonic within(fifa17, owned-item)— no per-account column is needed because Core owned-instance ids are globally-unique UUIDs.
Remaining prerequisite for a rendering retail /club: Core inventory must
reference cards that exist in the catalog. The catalog + store + resolver are
built and tested; wiring a controlled real FIFA 17 dev-content inventory (the
curated per-game dev pack) is the next slice. rare=SP ("Special") stays
UNSUPPORTED (semantics unproven; parsed, reported, never guessed).
Retail A/B runbook (first /club gate)
Change only the UTAS routing layer; keep the validated Rust Redirector/Roster and the current Blaze path. Python remains the rollback oracle — do not modify it.
Preconditions (mirror the proven blaze/roster switch discipline):
cargo test -p openfut-utas-host -p openfut-adapter-fifa17green;clippy -D warningsclean;fmt --checkclean.- Built binary identity == HEAD (
scripts/verify-build-identity.sh); no dirty tree. - Python UTAS directly reachable; the Rust host directly probeable; no stale NAT/switch rules; FIFA fully closed.
Bring-up:
- Move Python UTAS to an alternate port (
FUT_PORT=8199in the container/openfut-fut.sh); it keeps serving there. - Start this host on the client-visible UTAS addr:
OPENFUT_UTAS_HOST_ADDR=<lan>:8099 OPENFUT_UTAS_PYTHON_URL=http://127.0.0.1:8199 OPENFUT_CORE_URL=http://127.0.0.1:8080 OPENFUT_FIFA17_CATALOG=<catalog.json> OPENFUT_IDENTITY_STORE=<store.json> OPENFUT_PERSONA_ID=33068179 openfut-utas-host - Launch FIFA → FUT → My Squad player picker and exercise: no-filter, position, nation, league, league+team, Gold+position, then scroll beyond page one.
Evidence to capture (all six):
- Switch: client traffic hits the Rust host.
- Rust positive: host log
owner=RUST route=club …for the client IP. - Python negative for /club: Python logs no
/clubrequest in the window. - Python positive for other UTAS: unimplemented routes still reach Python.
- Core positive: Core logs the
/collectionquery and returns the expected set. - Application + pagination: the UI shows filtered results; later pages differ from page one (no repeated-first-page amplification).
Rollback: point the UTAS addr back at Python directly; confirm FUT still usable;
then re-enable the host and confirm /club again (proves reversibility).
Logs are safe by construction: no auth/session/device/token material — only owner, route, filter summary, counts, status.