Synchronous translation on read (Stage D) blocks the HTTP handler on
translator I/O. Stage E switches to "send moments-fast, deliver
when translated": recipients whose preferred_language differs from
the detected body_lang are inserted with available_at=NULL, and an
async worker turns them on once a LibreTranslate call materialises
the cache row (or fails terminally after 5 retries).
Schema delta on diplomail_recipients: available_at,
translation_attempts, next_translation_attempt_at, plus a snapshot
recipient_preferred_language so the worker queries do not need a
join. Read paths (ListInbox, GetMessage, UnreadCount) filter on
available_at IS NOT NULL. Push fan-out is moved from Service to the
worker so the recipient only sees the toast when the inbox row is
actually visible.
Translator backend is now a configurable choice: empty
BACKEND_DIPLOMAIL_TRANSLATOR_URL → noop (deliver original);
populated → LibreTranslate HTTP client. Per-attempt timeout, max
attempts, and worker interval all live in DiplomailConfig. The HTTP
client itself is unit-tested via httptest (happy path, BCP47
normalisation, unsupported pair, 5xx, identical src/dst, missing
URL); worker delivery + fallback paths are covered by the
testcontainers-backed e2e tests in diplomail_e2e_test.go.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces the LangUndetermined placeholder with whatlanggo-backed
body detection on every send path, then adds a translation cache
keyed on (message_id, target_lang) populated lazily on the
per-message read endpoint. The noop translator that ships with
Stage D returns engine="noop", which the service treats as
"translation unavailable" — wiring a real backend (LibreTranslate
HTTP client is the documented next step) is a one-file swap.
GetMessage and ListInbox now accept a targetLang argument; the HTTP
layer resolves the caller's accounts.preferred_language and
forwards it. Inbox uses the cache only (never calls the
translator) so bulk reads stay fast under future SaaS backends.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Closes out the producer-side of the diplomail surface. Paid-tier
players can fan out one personal message to the rest of the active
roster (gated on entitlement_snapshots.is_paid). Site admins gain a
multi-game broadcast (POST /admin/mail/broadcast with `selected` /
`all_running` scopes) and the bulk-purge endpoint that wipes
diplomail rows tied to games finished more than N years ago. An
admin listing (GET /admin/mail/messages) rounds out the
observability surface.
EntitlementReader and GameLookup are new narrow deps wired from
`*user.Service` and `*lobby.Service` in cmd/backend/main; the lobby
service grows a one-off `ListFinishedGamesBefore` helper for the
cleanup path (the cache evicts terminal-state games so the cache
walk is not enough). Stage D will swap LangUndetermined for an
actual body-language detector and add the translation cache.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Item 7 of the spec wants game-state and membership-state changes to
land as durable inbox entries the affected players can re-read after
the fact — push alone times out of the 5-minute ring buffer. Stage B
adds the admin-kind send matrix (owner-driven via /user, site-admin
driven via /admin) plus the lobby lifecycle hooks: paused / cancelled
emit a broadcast system mail to active members, kick / ban emit a
single-recipient system mail to the affected user (which they keep
read access to even after the membership row is revoked, per item 8).
Migration relaxes diplomail_messages_kind_sender_chk so an owner
sending kind=admin keeps sender_kind=player; the new
LifecyclePublisher dep on lobby.Service is wired through a thin
adapter in cmd/backend/main, mirroring how lobby's notification
publisher is plumbed today.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Phase 28 of ui/PLAN.md needs a persistent player-to-player mail
channel; the existing `mail` package is a transactional email
outbox and the `notification` catalog is one-way platform events.
Stage A lands the schema (diplomail_messages / _recipients /
_translations), a single-recipient personal send/read/delete
service path, a `diplomail.message.received` push kind plumbed
through the notification pipeline, and an unread-counts endpoint
that drives the lobby badge. Admin / system mail, lifecycle hooks,
paid-tier broadcast, multi-game broadcast, bulk purge and language
detection / translation cache come in stages B–D.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
After a host reboot macOS clears /private/tmp, so the per-game
bind-mount source under /tmp/galaxy-game-state/<uuid> vanishes and
Docker refuses to restart the long-lived engine container under
`restart: unless-stopped`. The container then sits in `exited` state
and the dev sandbox is unreachable until the developer manually rms
it and runs `make up` twice.
Fix `make -C tools/local-dev up` to heal this in one cycle:
1. `prune-broken-engines` (new make target wired into `up`) walks
every container labelled `galaxy-game-engine` and removes the ones
not in `running` / `restarting` state. Healthy long-lived
containers survive normal up/down cycles untouched.
2. The backend now runs a single reconciliation pass before the
dev-sandbox bootstrap (`Reconciler().Tick(ctx)` in main.go).
Without it, bootstrap would reuse the soon-to-be-cancelled game
that the periodic ticker is about to mark `removed`. The pre-tick
cascades the orphan runtime row through markRemoved → lobby
cancel before bootstrap purges terminal sandbox games and creates
a fresh one — so a single `make up` lands a working sandbox with
a brand new state directory.
README troubleshooting section documents the symptom and the
recovery so the bind-mount-source error message is greppable.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds backend/internal/devsandbox: an idempotent boot-time hook that,
when BACKEND_DEV_SANDBOX_EMAIL is set, ensures (1) the configured
engine_version row, (2) the real dev user, (3) PlayerCount-1
deterministic dummy users, (4) a private "Dev Sandbox" game with a
year-out turn schedule, (5) memberships for every participant via
the new lobby.Service.InsertMembershipDirect helper, (6) a drive of
the lifecycle to running. Re-running on a populated DB is a no-op;
partial states from earlier crashes are recovered.
tools/local-dev gains the matching env vars in .env, surfaces them
in compose, and acquires a `make build-engine` target that builds
galaxy-engine:local-dev from game/Dockerfile (a prerequisite of
`up`/`rebuild`). The compose game-state mount is changed from a
named volume to a host bind on /tmp/galaxy-game-state so backend's
bind-mount source for spawned engine containers resolves on the
docker daemon.
After `make -C tools/local-dev up`, login as dev@local.test with
the dev code 123456 and the Dev Sandbox already shows up in My
Games. Per-user behaviour for the same email survives a backend
restart.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
`time.LoadLocation` is called from
backend/internal/server/handlers_public_auth.go:108 (confirm-email-code)
and backend/internal/user/account.go:218 (user.settings.update). Both
runtime images shipped today have no tzdata — production
backend/Dockerfile uses gcr.io/distroless/static-debian12:nonroot, and
local-dev tools/local-dev/backend.Dockerfile uses alpine:3.20 without
the optional tzdata apk — so the container-side binary resolves only
the no-data fallback (UTC and fixed offsets) and rejects every real
IANA zone with HTTP 400 `invalid_request: time_zone must be a valid
IANA zone`.
Adding `import _ "time/tzdata"` to backend's main is the idiomatic
Go fix: the binary embeds the IANA database, time.LoadLocation works
on every base image, no Dockerfile changes needed. Cost is ~800 KB
of binary growth — invisible next to the existing /usr/local/bin/backend
size and well below any container layer threshold.
The OpenAPI spec already documents the field as "IANA time-zone
identifier" (gateway/openapi.yaml:205, backend/openapi.yaml:2334)
and the UI sends Intl.DateTimeFormat().resolvedOptions().timeZone,
so neither the contract nor the client needs a change.
Why this slipped through: backend unit tests run as a host Go test
process (developer's tzdata covers them), Playwright tests mock the
gateway (backend never reached), and the integration suite — the only
layer that exercises the real backend container — uses
RegisterSession which hardcoded `time_zone="UTC"`. Switching that
default to "Europe/Berlin" makes every integration scenario that
enrols a pilot exercise the tzdata path, so the next regression
surfaces in the integration run instead of escaping into manual
smoke. (The integration suite is not in the per-PR workflow yet; that
gap is tracked separately.)
Verified end-to-end against `tools/local-dev`:
- Europe/Amsterdam, Asia/Tokyo, America/Los_Angeles → 200 +
device_session_id (was 400 before this patch).
- Mars/Olympus still → 400 (validation behaviour unchanged).
Host tests: backend/internal/{auth,user,config} green.
UI: pnpm test 14/14, CI=1 pnpm exec playwright test 44/44.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Ilia Denisov