This reverts commit 153fc21. With no fallback in code, the deploy
crash-looped: Coolify keeps an entry for every variable compose.yml has
referenced, and empty entries appear to reach the containers as empty
values instead of the compose defaults.
compose.yml now holds each default as ${VAR:-default}, and the code keeps
no fallback values. The bot stops at startup on a missing or invalid
PORT or LOG_LEVEL, /addsticker refuses without STICKER_PACK_NAME, and the
renderer refuses to start until every RENDERER_* setting is a valid
value, listing each problem. Settings whose empty value means none or
all (MODULES, OWNER_ID, ADMIN_IDS, the API tokens) use ${VAR:-}.
The renderer's npm start and dev load .env when present, so a local run
works from a copy of .env.example.
BREAKING CHANGE: running outside compose now requires LOG_LEVEL and PORT
for the bot, STICKER_PACK_NAME for /addsticker, and every RENDERER_*
variable for the renderer.
Each service's environment is grouped into required, optional, and
stack-fixed values. Optional variables are referenced without a default,
so Coolify lists them as settings that can be filled in or left empty;
an empty value falls back to the code's default.
Coolify lists every ${VAR} that compose.yml references as an app setting,
so the four renderer tuning values showed up as entries to fill in. They
are no longer referenced: the renderer's own defaults apply, and an
override goes into compose.yml as a literal.
The renderer's settings now read RENDERER_HOST, RENDERER_PORT,
RENDERER_MAX_CONCURRENT_RENDERS, RENDERER_RENDER_TIMEOUT_MS,
RENDERER_MAX_OPTIONS, and RENDERER_MAX_OPTION_CHARS, so they read as
renderer settings and cannot clash with the bot's variables, which
Coolify injects into every service. NODE_ENV keeps its standard name.
BREAKING CHANGE: the unprefixed HOST, PORT, MAX_CONCURRENT_RENDERS,
RENDER_TIMEOUT_MS, MAX_OPTIONS, and MAX_OPTION_CHARS are no longer read
by the renderer.
The bot now takes the renderer's base URL and appends each /api/<name>
route itself, instead of taking the full /api/gif endpoint and swapping
its last path segment for /gacha and /genshin. The client, its files, and
its errors are named after the renderer rather than wheelofnames; the
/wheelofnames command keeps its name.
BREAKING CHANGE: WHEELOFNAMES_API_URL is replaced by RENDERER_URL, which
holds the base URL (e.g. http://renderer:3000) rather than the /api/gif
endpoint. compose.yml sets it, so Coolify needs no value.
The renderer now runs only on the compose network with no published port
or domain, so a shared bearer token adds nothing. The renderer no longer
checks Authorization or requires API_TOKEN in production, and the bot no
longer sends a token.
BREAKING CHANGE: WHEELOFNAMES_API_TOKEN and the renderer's API_TOKEN are
removed. Never publish the renderer's port: its API is unauthenticated.
The bot now reaches the renderer at http://renderer:3000/api/gif on the
compose network instead of a separately deployed service. Both share
WHEELOFNAMES_API_TOKEN. CI runs the renderer's lint, typecheck, tests,
render smoke, and image build.
Brings back the Genshin-style meteor wish that /api/gacha rendered before pack-cards, under genshin names. It takes the same request as /api/gacha and returns the 7-second landscape MP4.
A meteor coloured by rarity falls across a night sky, lands in a white
flash, and the label is revealed with its stars popping in. The caller
picks the result and rarity; the route only draws it. Renders as silent
H.264 so Telegram plays it as an animation without GIF banding, and
shares the render slots with /api/gif.
Replace pnpm-lock.yaml with package-lock.json. allowBuilds for esbuild becomes
package.json#allowScripts. The Dockerfile installs with npm ci --omit=dev and
no longer prepares pnpm through corepack; .npmrc keeps engine-strict, which npm
honours natively.
minimumReleaseAgeExclude is dropped rather than translated: pnpm's
minimumReleaseAge was never set, so the 30-entry exclusion list had nothing to
exclude from. Direct dependency versions resolve identically to the pnpm
lockfile.
README described confetti as reusing theme slice colors and made no mention of
the winner callout or the reveal timing.
Fill in the benchmark rows that were TBD, and record where the GIF growth
actually came from. Isolating each change shows the celebration rework, pill,
spin retune, palette reorder and pointer together account for +3.3%, while
grayscale label antialiasing accounts for +22.4% on its own because it affects
every spin frame rather than only the celebration. That trade is deliberate:
it removes visible color fringing and stops the palette being exhausted.
Also correct the size levers. Particle count is not one - cutting 43% of the
particles changes the file by 1.1%. Note that render times on this machine vary
more than 2x across repeat runs of identical input, so single timings are
indicative only.
tiennm99