The committed FlatBuffers bindings were generated by flatc 25.x (the TS
runtime is flatbuffers@25.9.23), but nothing pinned the compiler, so a
regen on a box with an older flatc (Debian apt ships 23.5.26) silently
churns output and flips nullable-scalar builder defaults. PR #82 hit this
and shipped 5 report files from the wrong compiler.
Unify the whole toolchain on 25.9.23 (the only version available as an
npm package, a prebuilt flatc binary, and a Go tag) and make the bindings
reproducible:
- Downgrade the flatbuffers Go module 25.12.19 -> 25.9.23 (schema,
transcoder, gateway, integration) so compiler and both runtimes match.
- Regenerate every schema with flatc 25.9.23. The only resulting change
is order/command-item.ts: the lone straggler still on the old
optional-scalar builder default (cmd_applied/cmd_error_code: 0 -> null).
Inert in practice — the TS side never builds those response-only fields
(the engine sets them in Go); the reader is unchanged.
- Pin the version in tooling: a flatc-check guard in ui/Makefile (fbs-ts)
and a new pkg/schema/fbs/Makefile (fbs-go); both refuse a mismatched
flatc and point at the release binary. Fix the stale apt install hint.
- Add a path-filtered CI guard (.gitea/workflows/fbs-codegen.yaml) that
regenerates with the pinned flatc and fails on any diff.
- Document the pinned version and the regen commands in the schema README.
No wire-format change: Go build/vet, transcoder roundtrip + engine tests,
pnpm check and the full vitest suite (888) stay green.
Three-stage refactor of the game-engine plumbing (game logic untouched):
Stage 1 — lock-free persistence + admin serialisation. Remove the file
lock from repo/fs (the .lock file, the Read/Write-vs-*Safe duality and the
dead ReadSafe polling) and replace the two-step rename with a single atomic
rename so concurrent reads are torn-free without a lock. Serialise the
state-mutating admin writers (init/turn/banish) with one shared router
LimitMiddleware, rewritten to block on the request context instead of a
racy shared 100ms timer.
Stage 2 — remove the obsolete immediate-command path end to end. Players
submit through PUT /api/v1/order; the legacy PUT /api/v1/command path is
deleted across game (route, handler, 24 command factories, Ctrl), backend
(Commands handler/route, engineclient.ExecuteCommands), gateway (dispatch +
executeUserGamesCommand + routing entry), the FlatBuffers/model contract
(UserGamesCommand[Response]) and transcoder, plus every affected
OpenAPI/README/FUNCTIONAL/ARCHITECTURE doc. The integration proxy test is
converted to the order path.
Stage 3 — flatten the REST->engine wrapper. Replace the executor adapter,
the controller package functions and RepoController with one concrete
controller.Service; drop the single-implementation Repo and Storage
interfaces (repo.Repo / fs.FS are now concrete). Handlers depend on a thin
handler.Engine seam and own the domain->REST projection; storage is
resolved once at startup instead of per request.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
ARCHITECTURE.md §15 "Verification order" specifies signature verification
(step 4) before payload_hash (step 5), but the authenticated-edge
decorator chain wrapped the payload-hash gate outside the signature gate,
so the hash was checked first. gateway/README.md and gateway/docs/flows.md
had drifted to match the code (hash-first), leaving ARCHITECTURE.md as the
lone source describing the intended order.
Swap the two decorators in server.go so the signature gate runs first, and
align README + flows.md to ARCHITECTURE.md. Signature-first is the
cryptographically sound order: the signature covers the payload_hash field,
so the request is authenticated before any of its content is processed.
Observable side effect: a request carrying a tampered payload_hash whose
signature was computed over the original hash is now rejected at the
signature gate (UNAUTHENTICATED "invalid request signature") instead of the
hash gate (INVALID_ARGUMENT). Security is unchanged — both refusals happen
before the payload is handled. The four payload-hash unit tests re-sign
over the tampered hash so they keep exercising the hash gate; the
cross-service integration test signs over the overridden hash and already
accepts both codes.
Refs #39
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Serve the whole stack behind one host: site at /, game UI at /game/,
gateway REST at /api + /healthz, Connect at /rpc (prefix stripped by the
edge Caddy). The built artifact is domain-agnostic — the UI talks to the
gateway same-origin via relative URLs, so the same bundle runs under any
host with no rebuild and with CORS disabled.
- Rename the Connect proto service galaxy.gateway.v1.EdgeGateway ->
edge.v1.Gateway; regenerate Go + TS; public path /rpc/edge.v1.Gateway.
- Move the game UI under base path /game (env BASE_PATH); make the
manifest, service-worker scope, WASM loader, and all navigation
base-aware via a withBase helper.
- Relative API + /rpc Connect prefix; Vite dev proxy mirrors the strip.
- Rewrite the edge Caddy (dev + prod) for path-based routing; empty CORS
allow-lists (same-origin); single host.
- New VitePress project site (site/): i18n en/ru with switcher, LaTeX
math, minimal monospace theme; built and served at /.
- dev-deploy compose/Makefile + CI (dev-deploy, prod-build, new
site-build) build and seed the site; probes hit /, /game/, /healthz.
- Sync docs (ARCHITECTURE, gateway README/openapi, dev-deploy &
local-dev READMEs, CLAUDE.md, ui/PLAN).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Browser fetch-streaming layers close response bodies they consider
idle after roughly 15-30 s without incoming bytes. Safari is the
most aggressive, but the symptom matters everywhere: a quiet
SubscribeEvents stream (lobby, between turns, mailbox empty) gets
torn down by the browser, the EventStream singleton reconnects with
backoff, and any push event that fires inside the reconnect window
is lost because `push.Hub` queues are not persisted across
subscription closes. The user-visible failure mode is the
intermittent "Fetch API cannot load … due to access control checks"
console error (a misleading WebKit symptom — CORS headers are
actually present) plus missed turn-ready / mail-received toasts.
Server-side fix: a silence-based heartbeat at the
`authenticatedPushStreamService` wrapper layer. After the signed
`gateway.server_time` bootstrap event, gateway wraps the bound
stream with `heartbeatingStream`. Every tail Send (fan-out, future
variants) resets the silence timer; when the timer elapses, a
goroutine emits `gateway.heartbeat` with only `EventType` set —
everything else stays at proto3 defaults, so the wire frame is
~45 bytes amortised. A `sendMu` serialises the heartbeat goroutine
with tail Sends because grpc.ServerStream.Send is not goroutine-safe.
The heartbeat is intentionally UNSIGNED: heartbeats carry no
payload, dispatch to no handler on the client, and an injected
heartbeat trivially causes no user-visible state change. TLS still
protects the wire and real events keep the signed envelope
unchanged. Documented in `docs/ARCHITECTURE.md` § 15 alongside the
per-scale bandwidth projection (100…100 000 clients × 15…60 s).
Config: new `GATEWAY_PUSH_HEARTBEAT_INTERVAL` (default `15s`,
`0s` disables). Telemetry: new
`gateway.push.heartbeats_sent{outcome}` counter so operators can
budget bandwidth and spot a sudden `outcome=error` bump as an
upstream-failing-before-flush signal.
Client (`ui/frontend/src/api/events.svelte.ts`): early `continue`
on `event.eventType === "gateway.heartbeat"` before `verifyEvent`,
`verifyPayloadHash`, or dispatch — empty signature would otherwise
trip SignatureError and reconnect. A leading heartbeat still flips
`connectionStatus` to `connected` and resets backoff, because
receiving one is proof the stream is healthy.
Tests:
- `push_heartbeat_test.go`: unit tests for the wrapper — zero
interval returns nil, heartbeat fires after silence, real Send
resets the timer, Stop / context-cancel halt the goroutine,
Send errors propagate.
- `server_test.go`: integration tests through the full gateway
pipeline — heartbeat fires after the configured silence window,
zero interval keeps the stream silent.
- `config_test.go`: default applied, env-override parsed,
negative value rejected.
- `events.test.ts`: heartbeat skipped before verification + not
dispatched to handlers; leading heartbeat still flips
`connectionStatus` to `connected`.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The public REST listener already exposes
`GATEWAY_PUBLIC_HTTP_CORS_ALLOWED_ORIGINS`; the authenticated
Connect-Web listener on the separate gRPC port had no equivalent.
That worked in `tools/local-dev` (Vite proxy makes everything
same-origin) and would work in production once UI and gateway share
a single hostname, but the long-lived dev environment serves the
UI from `https://www.galaxy.lan` and the gateway from
`https://api.galaxy.lan` — every `/galaxy.gateway.v1.EdgeGateway/*`
fetch failed in the browser with the WebKit "Load failed" generic
message because the response carried no `Access-Control-Allow-Origin`
header. Lobby rendered as "[unknown] Load failed" with no game.
Mirror the public-REST CORS surface for the authenticated handler:
- new env `GATEWAY_AUTHENTICATED_GRPC_CORS_ALLOWED_ORIGINS`;
- new `AuthenticatedGRPCConfig.CORSAllowedOrigins` field;
- new `grpcapi.withCORS` middleware wrapping the Connect mux;
- dev-deploy stack sets the env to `https://www.galaxy.lan`.
The middleware speaks plain net/http (the Connect handler is mounted
on a ServeMux, not gin), handles preflight 204 immediately, and
exposes the Connect-Web header set the browser needs to read the
response (`Grpc-Status`, `Grpc-Message`, `Connect-Protocol-Version`).
Empty allow-list disables the middleware — production stays at
"single hostname" by default.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The Phase 27 BattleViewer was the last UI surface still issuing raw
fetch() against the backend REST contract (`/api/v1/user/games/...
/battles/...`). The dev-deploy gateway never proxied that path, so
the viewer worked only in tools/local-dev/. Move it onto the signed
ConnectRPC channel every other authenticated surface already uses.
Wire pieces:
- FBS GameBattleRequest in pkg/schema/fbs/battle.fbs, regenerated
Go + TS bindings.
- MessageTypeUserGamesBattle constant + GameBattleRequest struct in
pkg/model/report/messages.go.
- pkg/transcoder/battle.go gains GameBattleRequestToPayload and
PayloadToGameBattleRequest helpers.
- gateway games_commands.go switches on the new message type and
GETs /api/v1/user/games/{id}/battles/{turn}/{battle_id}; the JSON
response is re-encoded as a FlatBuffers BattleReport before being
returned. 404 from backend surfaces as the canonical `not_found`
gateway error.
- ui/frontend/src/api/battle-fetch.ts now builds the FBS request,
calls GalaxyClient.executeCommand, and decodes the FBS response
into the existing UI shape (Record<string,string> race/ship maps,
string-form UUID). BattleFetchError carries an HTTP-style status
derived from the result code so the active-view's not_found branch
keeps working.
- battle.svelte pulls the GalaxyClient from the in-game shell
context. While the layout's boot Promise.all is in flight the
effect stays in `loading` until the client handle becomes
non-null.
- ui/Makefile FBS_INPUTS gains battle.fbs.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds the gateway-side translation layer that maps the eight new
ConnectRPC mail commands onto backend's
`/api/v1/user/games/{game_id}/mail/*` REST endpoints.
- `gateway/internal/backendclient/mail_commands.go` defines
`ExecuteMailCommand` and one helper per command (inbox, sent,
message.get, send, broadcast, admin, read, delete). Each helper
decodes the FlatBuffers request envelope, issues the REST call
via the existing `*RESTClient.do`, decodes the JSON body, and
re-encodes a typed FlatBuffers response. Recipient identifiers
travel through unchanged so the new `recipient_race_name`
shortcut introduced in Step 1 reaches backend untouched.
- `routes.go` exposes a `MailRoutes` constructor and a matching
`mailCommandClient` implementing `downstream.Client`.
- `cmd/gateway/main.go` registers the new routes alongside the
existing user / lobby / game-engine routes.
- `mail_commands_test.go` covers the inbox, send-by-race-name, and
read-state paths end-to-end against an `httptest.Server`,
asserting request shapes (path, body, X-User-ID) and the
decoded FlatBuffers response.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a `GATEWAY_PUBLIC_HTTP_CORS_ALLOWED_ORIGINS` env-driven allow-list
on the public REST server so the dev UI on https://www.galaxy.lan can
call https://api.galaxy.lan without the browser blocking the
cross-origin response. Defaults to empty (no CORS) so the production
posture stays closed.
The middleware mounts before route classification and anti-abuse, so
OPTIONS preflights never charge against per-class rate-limit buckets.
`tools/dev-deploy/docker-compose.yml` opts the dev gateway into a
single allowed origin (`https://www.galaxy.lan`); local-dev keeps the
defaults because Vite proxies through the same origin.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Backend now owns the turn-cutoff and pause guards the order tab
relies on: the scheduler flips runtime_status between
generation_in_progress and running around every engine tick, a
failed tick auto-pauses the game through OnRuntimeSnapshot, and a
new game.paused notification kind fans out alongside
game.turn.ready. The user-games handlers reject submits with
HTTP 409 turn_already_closed or game_paused depending on the
runtime state.
UI delegates auto-sync to a new OrderQueue: offline detection,
single retry on reconnect, conflict / paused classification.
OrderDraftStore surfaces conflictBanner / pausedBanner runes,
clears them on local mutation or on a game.turn.ready push via
resetForNewTurn. The order tab renders the matching banners and
the new conflict per-row badge; i18n bundles cover en + ru.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The owner reported two symptoms after pulling the Phase 14 stack:
1. user.games.order.get answered with `unimplemented: message_type
is not routed`. The gateway/backend code was correct, but the
local-dev compose images were stale — `make rebuild` picked up
the new routes table and the symptom went away. To prevent this
class of regression from depending on docker-image freshness,
gateway/internal/backendclient/routes_test.go now asserts that
every authenticated MessageType constant declared in
pkg/model/{user,lobby,order,report} is registered, and verifies
that user.games.order.get specifically resolves to the game
command client.
2. The inspector kept the un-renamed name after a successful submit.
ui/frontend/tests/inspector-overlay.test.ts mounts the inspector
tab against a real OrderDraftStore + a stubbed GameStateStore
and walks the full happy path (add planetRename → markSubmitting
→ applied → simulate refresh) plus the integration scenario
driven through the order-tab Submit button. Both cases pass —
the underlying overlay path is reactive and resilient to a
refresh that returns the un-renamed snapshot. The original
in-browser symptom was the rebuilt-image freshness issue from
point 1; this test pins the reactive contract for future
refactors.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Wires the first end-to-end command through the full pipeline:
inspector rename action → local order draft → user.games.order
submit → optimistic overlay on map / inspector → server hydration
on cache miss via the new user.games.order.get message type.
Backend: GET /api/v1/user/games/{id}/orders forwards to engine
GET /api/v1/order. Gateway parses the engine PUT response into the
extended UserGamesOrderResponse FBS envelope and adds
executeUserGamesOrderGet for the read-back path. Frontend ports
ValidateTypeName to TS, lands the inline rename editor + Submit
button, and exposes a renderedReport context so consumers see the
overlay-applied snapshot.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Replaces the Phase 10 map stub with live planet rendering driven by
`user.games.report`, and wires the header turn counter to the same
data. Phase 11's frontend sits on a per-game `GameStateStore` that
lives in `lib/game-state.svelte.ts`: the in-game shell layout
instantiates one per game, exposes it through Svelte context, and
disposes it on remount. The store discovers the game's current turn
through `lobby.my.games.list`, fetches the matching report, and
exposes a TS-friendly snapshot to the header turn counter, the map
view, and the inspector / order / calculator tabs that later phases
will plug onto the same instance.
The pipeline forced one cross-stage decision: the user surface needs
the current turn number to know which report to fetch, but
`GameSummary` did not expose it. Phase 11 extends the lobby
catalogue (FB schema, transcoder, Go model, backend
gameSummaryWire, gateway decoders, openapi, TS bindings,
api/lobby.ts) with `current_turn:int32`. The data was already
tracked in backend's `RuntimeSnapshot.CurrentTurn`; surfacing it is
a wire change only. Two alternatives were rejected: a brand-new
`user.games.state` message (full wire-flow for one field) and
hard-coding `turn=0` (works for the dev sandbox, which never
advances past zero, but renders the initial state for any real
game). The change crosses Phase 8's already-shipped catalogue per
the project's "decisions baked back into the live plan" rule —
existing tests and fixtures are updated in the same patch.
The state binding lives in `map/state-binding.ts::reportToWorld`:
one Point primitive per planet across all four kinds (local /
other / uninhabited / unidentified) with distinct fill colours,
fill alphas, and point radii so the user can tell them apart at a
glance. The planet engine number is reused as the primitive id so
a hit-test result resolves directly to a planet without an extra
lookup table. Zero-planet reports yield a well-formed empty world;
malformed dimensions fall back to 1×1 so a bad report cannot crash
the renderer.
The map view's mount effect creates the renderer once and skips
re-mount on no-op refreshes (same turn, same wrap mode); a turn
change or wrap-mode flip disposes and recreates it. The renderer's
external API does not yet expose `setWorld`; Phase 24 / 34 will
extract it once high-frequency updates land. The store installs a
`visibilitychange` listener that calls `refresh()` when the tab
regains focus.
Wrap-mode preference uses `Cache` namespace `game-prefs`, key
`<gameId>/wrap-mode`, default `torus`. Phase 11 reads through
`store.wrapMode`; Phase 29 wires the toggle UI on top of
`setWrapMode`.
Tests: Vitest unit coverage for `reportToWorld` (every kind,
ids, styling, empty / zero-dimension edges, priority order) and
for the store lifecycle (init success, missing-membership error,
forbidden-result error, `setTurn`, wrap-mode persistence across
instances, `failBootstrap`). Playwright e2e mocks the gateway for
`lobby.my.games.list` and `user.games.report` and asserts the
live data path: turn counter shows the reported turn,
`active-view-map` flips to `data-status="ready"`, and
`data-planet-count` matches the fixture count. The zero-planet
regression and the missing-membership error path are covered.
Phase 11 status stays `pending` in `ui/PLAN.md` until the local-ci
run lands green; flipping to `done` follows in the next commit per
the per-stage CI gate in `CLAUDE.md`.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
vite.config.ts now proxies `/api` and `/galaxy.gateway.v1.EdgeGateway`
to the gateway, so the browser sees only `localhost:5173` and never
trips a cross-origin preflight. `.env.development` accordingly points
`VITE_GATEWAY_BASE_URL` at the Vite origin. The proxy target is
overridable via `VITE_DEV_PROXY_TARGET=...` for non-default gateways
without touching the compose file.
`gateway/Dockerfile` previously failed to build because gateway
imports `galaxy/core` (replaced to `../ui/core` in `gateway/go.mod`)
but the Dockerfile did not copy `ui/core/` into the build context
nor declare the replace in the synthesised `go.work`. Adding both
makes `docker build -f gateway/Dockerfile .` succeed; this is the
same fix already shipped in `tools/local-dev/gateway.Dockerfile`,
back-ported to upstream.
Verified:
- docker build -f gateway/Dockerfile . — builds cleanly
- pnpm test 14/14, pnpm exec playwright test 44/44 (with CI=1 to
force a fresh dev server; reuse keeps the previous startup env)
- curl POST through localhost:5173/api/* and /galaxy.gateway.v1.* —
reach the gateway, no CORS preflight on the browser side
tools/local-dev/README.md updated with the new network map and the
`VITE_DEV_PROXY_TARGET` override.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- Extend pkg/model/lobby and pkg/schema/fbs/lobby.fbs with public-games
list, my-applications/invites lists, game-create, application-submit,
invite-redeem/decline. Mirror the matching transcoder pairs and Go
fixture round-trip tests.
- Wire the seven new lobby message types through
gateway/internal/backendclient/{routes,lobby_commands}.go with
per-command REST helpers, JSON-tolerant decoding of backend wire
shapes, and httptest-based unit coverage for success / 4xx / 5xx /
503 across each command.
- Introduce TS-side FlatBuffers via the `flatbuffers` runtime dep, a
`make fbs-ts` target driving flatc, and the generated bindings under
ui/frontend/src/proto/galaxy/fbs. Phase 7's `user.account.get` decode
now uses these bindings as well, closing the JSON.parse vs
FlatBuffers gap that would have failed against a real local stack.
- Replace the placeholder lobby with five sections (my games, pending
invitations, my applications, public games, create new game) and the
/lobby/create form. Submit-application uses an inline race-name
form on the public-game card; create-game keeps name / description /
turn_schedule / enrollment_ends_at always visible and the rest under
an Advanced toggle with TS-side defaults.
- Update lobby/+page.svelte to throw LobbyError on non-ok result codes;
GalaxyClient.executeCommand now returns { resultCode, payloadBytes }.
- Vitest binding round-trips, lobby.ts wrapper unit tests, lobby-page
+ lobby-create component tests, Playwright lobby-flow.spec covering
create / submit / accept across all four projects. Phase 7 e2e was
migrated to the FlatBuffers fixtures and to click+fill against the
Safari-autofill readonly inputs.
- Mark Phase 8 done in ui/PLAN.md, mirror the wire-format note into
Phase 7, append the new lobby commands to gateway/README.md and
docs/ARCHITECTURE.md, add ui/docs/lobby.md.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds a minimal Svelte 5 i18n primitive (`src/lib/i18n/`) backing the
login form, the layout blocker page, and the lobby placeholder.
SUPPORTED_LOCALES drives both the picker and the runtime lookup;
adding a language is a two-step change inside `src/lib/i18n/`.
Login form gains a globe-icon language dropdown (English / Русский
in their native names), defaulting to navigator.languages with `en`
as the fallback. Switching the locale re-renders the form in place;
on submit, the locale rides in the JSON body of `send-email-code`
because Safari/WebKit silently drops JS-set Accept-Language. Gateway
gains a body `locale` field that takes priority over the request
header for preferred-language resolution.
Email and code inputs disable browser autofill / suggestions
(`autocomplete=off` + `autocorrect=off` + `autocapitalize=off` +
`spellcheck=false`) so Keychain / address-book pickers and
remembered-value dropdowns no longer fire on focus.
Cross-cuts:
- backend & gateway openapi: clarify that body `locale` is honored.
- docs/FUNCTIONAL{,_ru}.md §1.2: document body-vs-header priority.
- gateway tests: body `locale` overrides Accept-Language; blank
body `locale` falls back to header.
- new ui/docs/i18n.md; cross-links from auth-flow.md and ui/README.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Replace the native-gRPC server bootstrap with a single
`connectrpc.com/connect` HTTP/h2c listener. Connect-Go natively
serves Connect, gRPC, and gRPC-Web on the same port, so browsers can
now reach the authenticated surface without giving up the gRPC
framing native and desktop clients may use later. The decorator
stack (envelope → session → payload-hash → signature →
freshness/replay → rate-limit → routing/push) is reused unchanged
behind a small Connect → gRPC adapter and a `grpc.ServerStream`
shim around `*connect.ServerStream`.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Ilia Denisov