Runs the full catalog audit (runCatalogAudit) against the live hotel DB with
no repair/sql options — a pure read pass — and cross-checks the emitted
summary against independent DB + asset-dir measurements: row totals, duplicate
classname groups, orphaned catalog references and missing catalog / nitro /
icon counts must match exactly, with zero error events and a final
'batch_complete'.
Guarded by RUN_CATALOG_AUDIT_LIVE=1 so CI never runs it; loads the real .env
because vitest fakes DATABASE_URL.
The site-settings loader kept an in-process map forever after a Redis miss
and promoted DEFAULTS (no logo/theme) to Redis on any DB error, so a build
that started before the DB was reachable stuck the site on the preset logo
and default theme until a manual reload or full restart.
- Redis miss now reloads from the database instead of the stale in-process map
- a DB failure returns defaults only as an in-process last resort and never
writes them to Redis, so the shared cache can't be poisoned by a transient
error at startup
- regression tests: DB re-read on Redis miss after cache expiry, defaults never
promoted to Redis, recovery from transient DB failure
- verifyAndFixInteractionModesCount: paginated DB queries (500/batch)
instead of loading all items into memory at once
- withFurniDataLock: add 60s chain timeout to prevent deadlocks
when a lock holder stalls or crashes
- withGamedataLock: same timeout protection for gamedata locks
- conversion-pool: add 60s per-job timeout, fall back to main thread
- furni/batch: abort signal + post_import progress events + skip
post-import steps when client disconnects
- furni/batch-regen: abort signal + early exit when disconnected
- clone/sync-all: pre-fetch furnidata once per source instead of
per item (eliminates 2000+ redundant fetches)
- sse-client: add 60s idle timeout to prevent infinite hangs
Replace the pure-JS lzma decoder with lzma-wasm (Rust/WASM, base64-inlined,
zero-alloc decompress), giving an order-of-magnitude speedup on furni
imports. Remove the obsolete lzma type shim and the redundant top-level
vite dev dependency, which nothing imports directly.
- update-Nitrov3.sh: build CMS into .next-staging and swap atomically so a
failed build never takes the live site down; auto-merge new variables
from .env.example; validate env for duplicates/broken lines; restart the
emulator/CMS only when rebuilt or unhealthy; fix step renumbering
- next.config.ts: support NEXT_DIST_DIR for staged production builds
- fix all TS errors (unused imports, missing tryDownloadCandidates helper)
so tsc and the production build pass clean
- add Dockerfile/.dockerignore and switch docker-compose to a CMS container
- include prevailing UI/refactor changes (SurfaceCard, ticketing, tsconfig)
resolveHotelName() now returns env.HOTEL_NAME directly — the single source
of truth. The CMS hotel_name site setting and its DEFAULTS entry are removed
as dead code since they no longer influence the displayed name.
Call sites are unchanged (still await resolveHotelName()); only the lookup
behind it is gone, so the public site always shows the configured env name
with no DB round-trip and no preset.
Drop the hardcoded FALLBACK_HOTEL_NAME ("Atom") preset and the brand.ts
module. HOTEL_NAME is now a required env var: if it (and the CMS hotel_name
setting) is missing the site fails validation at startup/build with a clear
message instead of silently rendering a placeholder hotel name.
resolveHotelName() resolves CMS hotel_name -> required HOTEL_NAME only.
Callers that used the preset (api/home route catch branch, CMS settings form
default, mobile-nav/logo-generator prop defaults) now use the configured name
or an empty default; the real name is already passed in by server parents.
openhands
Simo