Add an env-driven `user:`/`group_add` to the filestash service plus
tooling/filestash-set-run-user.sh. Running the container as a real host
account (oxps: oier=1002) makes every file it writes into the share owned
by that account, so scp/ssh and the web UI share one owner — no chown dance.
Defaults FILES_UID/FILES_GID=1000, leaving cochi (image's own uid) untouched.
group_add 1000 keeps the 0730 image binary + 0770 /app/data usable under any
uid; the helper also chowns the filestash_state volume (1000:1000) so a
non-1000 uid doesn't crash-loop unable to write its own sqlite/config.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Two prod hosts now diverge on filestash auth: cochi stays htpasswd
(internet-facing), oxps runs FILES_AUTH=none/passthrough since it is
LAN-only, satisfying the internal-only gate. Document the per-host
status table and the flip-back rule for if oxps ever gains an internet
path.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
On a fresh mc_data volume the minecraft service crash-loops with
"AccessDeniedException: /data/kubejs/config" (ConsoleJS.STARTUP null NPE):
Docker auto-creates the bind-mount parent /data/kubejs as root:root, but
the itzg container runs as uid 1000, so KubeJS can't create its writable
siblings (config/cache/generated/logs). Only bites fresh worlds — cochi's
aged volume predates the ro submounts and is already 1000-owned.
Fix (durable): chown 1000:1000 /data/kubejs once, then restart.
- plan/22-migration.md: Troubleshooting section with symptom + fix.
- migrate-import.sh: manual tail now flags it with the exact commands.
Hit + resolved during the cochi->oxps migration.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Observed while staging generators. on cochi: with the nginx vhost installed
but the container not yet deployed, caddy answers an unmatched Host with an
EMPTY 200. Both / and the mirror path returned "HTTP 200, 0 bytes" while
nothing at all was running.
So the guidance to probe a mirror path instead of / was necessary but not
sufficient — a status-code-only check reports green on a completely absent
service. Assert the body.
Same failure family as the filestash green-monitor: the server answering
is not the app working.
Add generators.${BASE_DOMAIN}: a pinned fork of misode.github.io served
with its own mcmeta + vanilla-mcdoc mirror, so it works with zero internet
at the LAN party. Fork lives in ulicraft-generators (branch ulicraft),
pinned at upstream 2bfe5522.
The app is 100% client-side: every data fetch comes from the player's
browser, not from cochi. So offline capability is entirely about which
origins the bundle refers to, plus a mirror to serve them from. The vhost
sets Content-Security-Policy with connect-src 'self' as ENFORCEMENT: an
external origin that sneaks back in (or returns via an upstream merge)
then breaks loudly in prod, on an ordinary day, instead of silently at the
venue with no internet to debug it.
sync-generators-data.sh mirrors ~109M: the 1.21.1 summary/data/atlas tags,
assets-tiny (assets minus hundreds of MB of .ogg — hence no sounds
generator), vanilla-mcdoc symbols.json, and summary/versions/data.min.json,
which comes from a BRANCH head rather than a version tag and runs during
Spyglass init. It verifies a sentinel file per source instead of trusting a
clean clone exit, since a renamed upstream path would otherwise sync
"successfully" and only fail in the browser at the party.
The container serves the bundle and the mirror on one origin, so there is
no CORS, no second cert and no second DNS name.
DEPLOY ORDER MATTERS. GENERATORS_REPO must be set in cochi's .env BEFORE
pulling this: compose uses ${GENERATORS_REPO:?...}, and an unset required
variable makes docker compose fail to parse the file — taking down every
service including Minecraft, not just this one. Full runbook and the rest
of the prerequisites: plan/21-generators.md.
Verified offline in real Chromium with every non-origin request aborted:
zero off-origin attempts, zero same-origin failures, zero console errors,
generators rendering schemas and live JSON.
The pack grew 76 -> 164 jars (133 both / 31 client) without any doc
update. Create + ~18 addons filled the tech gap, ars_nouveau + ars_creo
the magic gap, and the full YUNG's/Moog's/TerraBlender set the worldgen
gap, which also resolved the ponderjs orphan.
- 17-mod-shortlist: rewritten around the current pack; gap doc marked
done, remaining gaps narrowed to map mod, Necronomicon orphan, C2ME.
- 04-mods: jar counts updated; the 24/52 cutover split is now labelled
historical alongside the current 31/133.
- CLAUDE.md: WS6 partly done, stale "65 default both" line dropped, and
the ~50-100 mod ceiling flagged as knowingly exceeded in both places.
MAX_MEMORY 10G was sized for that old assumption - noted for re-check.
- 18-landing-rework: mods.json count 76 -> 164, "50+" stat tile noted
stale.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
ars_nouveau, ars_creo and world_dimension_nexus all declare NeoForge
side BOTH and are content mods, so they stay on the server subset.
Create is already in the pack, so ars_creo's dependency is satisfied.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Propagate the bare-hostname fix (10b86ae) to the docs that still said "three
traps", and write down the meta-lesson it exposed.
The bug was invisible to every check we had: Filestash strips the scheme
before its own Host check, so `curl /` returned 200, the API answered, wrong
passwords were rejected, and the brand-new Uptime Kuma HTTP monitor stayed
green — while the SPA computed `http://https://files...` and no browser could
use the site. An HTTP monitor asserts "server returned 200", which is a much
weaker claim than "the app works".
- CLAUDE.md / README: four traps now, host-scheme first; add the post-change
check (`/api/config` → origin must be the real https URL).
- 19-routes: login form only renders at ?action=redirect (a bare 303 is
normal, not a bug); assert origin after config changes.
- 10-uptime-kuma: an HTTP monitor cannot see an SPA break — use a Keyword
monitor if you want teeth, and don't read green as "users can log in".
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The SPA computes its redirect origin as (force_ssl ? "https://" : "http://")
+ general.host. With host set to "https://files.${BASE_DOMAIN}" that produced
"http://https://files.ulicraft.net", so every client-side redirect broke and
the page never routed anywhere.
Server-side this was invisible: SecureOrigin strips the scheme before the
Host check, so `curl /` returned 200, the API answered, and the uptime
monitor stayed green — only real browsers failed. The tell is a
`POST /report?...msg=Redirecting to http://https://...` line in the log.
Set host to the bare hostname and force_ssl=true (which the origin needs to
build https://, and which only emits an HSTS header — it does not redirect,
so it can't loop behind the plain-http nginx→caddy hop).
Verified: origin is now "https://files.ulicraft.net"; login + ls still work.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Fold files./filestash into the docs that carry the service list, the vhost
map and the deploy procedure — README stack table, 00-overview, 02-caddy,
12-build-order, 14-deploy, and the deploy skill.
Two real gaps the filestash deploy exposed, now guardrails in the skill:
- update-all.sh never touches host nginx, so a NEW SUBDOMAIN silently 404s
after deploy. It needs issue-letsencrypt.sh + render-nginx.sh --install.
- a new service with ${FOO:?} guards fails the WHOLE compose file, so a
missing .env var means `up` starts nothing — and --hard has already run
`down`. Check the host's .env BEFORE tearing the stack down.
Also: filestash/config.json is a single-file bind mount re-rendered every
run, so a rolling update-all leaves it on a stale inode (same trap as the
caddy conf) — force-recreate after any FILES_* change.
Correct the cochi divergence section: `git diff HEAD` on the host is now
empty (verified this deploy — the ff-pull succeeded). The documented
docker-compose.yml/package.json divergence was converged; only untracked
dnsmasq/, caddy-root-ca.crt and a stale pack/ remain. The stash-pull-pop
procedure it prescribed is no longer needed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
One shared login (htpasswd) for all players, writable. Config is rendered
from a template and mounted read-only, so git stays authoritative — the
admin console loads but cannot save, by design.
Filestash's OIDC/SAML middlewares are enterprise-only, so Keycloak SSO is
out; FILES_AUTH keeps htpasswd/passthrough switchable for later. Auth-off
(passthrough) is only valid once the host is off the public internet —
render-config.sh warns loudly when rendering it.
Three traps, all found by testing against the pinned image:
- the `local` backend is admin-gated: without LOCAL_BACKEND_SECRET in the
connection params every login fails with "backend error - Not Allowed"
- the htpasswd plugin only accepts $2a$ bcrypt, so a $2y$ hash from
`htpasswd -B` fails every login silently — use `openssl passwd -6`.
render-config.sh rejects the wrong format rather than shipping it
- APPLICATION_URL/ADMIN_PASSWORD env make filestash rewrite config.json at
boot, which fails on the :ro mount — both live in the rendered config
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Distribution added 11 jars. classify-mods.py seeds all as `both`
(none declare CLIENT in neoforge.mods.toml). Hand-flip the two
pure-client cosmetic ones so sync-server-mods.sh drops them:
- ExtremeSoundMuffler: client-side sound muffling UI
- ToastControl: client-side toast popup suppression
Rest stay `both` — Placebo/moonlight/caelus are libs, and
amendments/etched/fishingoverhaul/immersive_paintings/
createornithopterglider/emotecraft add server-side content or sync.
Also map iris as `client`. It is referenced by distribution.json but
absent from the local (gitignored) dist jar dir, so classify-mods.py
never sees it; mapping it keeps mods-sides.json aligned with the
manifest. Client-only, so the sync selection is unchanged.
server-mods/ now 130 jars (was 123).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
voice_host lives in the mc_data volume and regenerates empty on world wipe,
silently breaking remote voice (secret sent, UDP handshake never completes →
red icon). Document the set/restore steps and the VPN-client false-positive
(Cloudflare One/WARP, FortiClient drop voice UDP). Add a known-good prod
snapshot at server-configs/voicechat-server.properties (voice_host=ulicraft.net).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
online-mode=false made the server skip the online-auth handshake entirely, so
authlib-injector never validated sessions against Drasl: players got offline-hash
UUIDs and no skin textures (everyone Steve). authlib-injector's whole purpose is
to redirect the *online* handshake from Mojang to Drasl, so online-mode must be
true; only the 1.21+ secure profile stays off (ENFORCE_SECURE_PROFILE=false +
Drasl SignPublicKeys=false). Correct the conflation in CLAUDE.md and
plan/05-minecraft.md.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Mounting the whole kubejs dir read-only made KubeJS's BaseProperties.save()
throw NPE on boot (it writes config/cache/generated/logs under kubejs/),
so KubeJS never initialised. Bind only the source subdirs (server_scripts,
startup_scripts, data, assets) read-only and let the kubejs root live
writable in mc_data.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The minecraft container only received forgemods jars via ./server-mods.
The distribution's files/kubejs (server_scripts, data, startup_scripts)
and CustomSkinLoader never reached the server, so server-side kubejs
recipes/scripts were silently absent.
Bind-mount both from DISTRIBUTION_WEB_ROOT (read-only, to avoid the
server polluting the client distribution with kubejs-generated files).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Distribution replaced the wrong-version jars: rechiseled now
mc1.21 (gate [1.21,1.21.2) — valid for 1.21.1, no longer the broken
mc1.21.11 build) so back to `both`; fusion bumped 1.3.2a→1.3.2b, stays
`client` (resource-pack texture feature). Dropped the two stale entries.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
rechiseled-1.2.5-neoforge-mc1.21.11.jar targets Minecraft 1.21.11 /
NeoForge 21.11, not this server's 1.21.1 / 21.1.233. As `both` it crashed
the server at pre-load (FATAL ModLoader: requires minecraft 1.21.11). Mark
`client` to exclude it from server-mods so the server boots.
NOTE: the jar is still wrong-version for clients too (same gate) — it must
be removed from / replaced in the distribution to be usable at all.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Two new distribution mods. rechiseled adds decorative blocks → both
(server needs the block registry). fusion is a client-only resource-pack
texture/model feature (no blocks/logic) → client, so sync drops it from
server-mods.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
New custom mod ulicraftmod-1.21.1-1.0.0-6.0.10 added to the distribution;
content mod (server + client), so both.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The pack./packwiz service was removed operationally long ago, but several
plan docs still presented pack. as a live caddy vhost / DNS subdomain and
render-pack.sh as a current tool. Remove those references from the
current-architecture docs (overview, caddy, tooling, uptime-kuma,
letsencrypt, mc-backup); leave the migration log (04-mods.md), superseded
banner (04-packwiz.md), commit history (12-build-order.md), and planned
landing rework (18) intact as deliberate history.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Remove the stat-grid from the status section (server card only); prune the
now-orphaned site.stats, MOD_COUNT_KEY, statLabels (3 locales), and
.stat-grid/.tile CSS.
- Empty-roster message -> "Nobody online", now localized (status.empty in
en/es/eu) and wired to both ServerCard instances.
- Flatten the server-card icon (drop the inset emboss bevel on .sc-icon).
- Drop "Solo LAN" (hero metaLan) from the hero meta row; prune metaLan (3 locales).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
YUNG's Cave Biomes hard-requires geckolib>=4.7.6 and terrablender>=4.1.0.8;
both now in the distribution. Classified both (server worldgen libs).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
OPS env resolves names via Mojang lookup, but the server is offline-mode
with Drasl auth (own UUIDs) → wrong UUID, inert op entry. Document the
RCON procedure that uses the cached Drasl UUID from usercache.json.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The home page header already used ulicraft-logo-mini.svg; apply the same
brand mark to the register, account, and fjord page headers (and footers),
replacing the leftover Creeper pixel-art.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- landing header + footer brand: swap the Creeper pixel-art for the
ulicraft-logo-mini.svg goat mark.
- ServerCard fallback icon: same logo (real SLP favicon still overrides on
a successful ping).
- minecraft: set ICON=/extras/server-icon.png + OVERRIDE_ICON so the
in-game server-list entry and the SLP favicon use the goat logo. PNG
ships in ./runtime (mounted /extras).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The pack service was already gone from compose/caddy and the landing is
off packwiz. Remove the remaining live references: the pack. TLS vhost in
the host nginx template, pack from LE_SUBDOMAINS, and stale docs in README
and CLAUDE.md.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Describe the self-hosted mc-status pinger (mcstatus.io v2 JSON, same-origin via
caddy /api/mcstatus/* → mc-status:8080) that the landing ServerCard reads, with
the apex endpoint and a direct-to-caddy curl example.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Document the gotcha hit fixing distribution CORS: caddy/conf.d/*.caddy are
single-file bind mounts pinned to the host inode at container creation. git pull
rewrites tracked files via atomic rename (new inode), so a running caddy keeps
the old config; `restart`/`caddy reload` reuse the stale inode. Apply conf
changes with `docker compose up -d --force-recreate caddy` (or the --hard
down/up). Added to plan/14-deploy.md and the deploy skill guardrails.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The landing page (apex origin) fetches launcher.json from the distribution
vhost cross-origin to render download buttons; add Access-Control-Allow-Origin
on /launcher/* so the browser doesn't block it. Public assets, no credentials.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The build-time download buttons baked stale/placeholder URLs (the host has no
synced launcher.json, so the build fell back to launcher.example.json). Render
a skeleton at build time instead and fetch the live launcher.json client-side
once on page load (no polling) from
https://distribution.${BASE_DOMAIN}/launcher/launcher.json, building per-OS
buttons from files[]. On fetch error the skeleton clears.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Rename all "crew" mentions (en/es/eu) to the untranslated proper name
"El valle" across hero, members, and features copy in i18n/ui.ts.
Rewire the homepage launcher download list to the custom-launcher build
output launcher.json (productName/version/files[]) instead of the
never-produced launcher-manifest.json shape. site.ts now reads launcher.json
and normalizes files[] -> internal builds[]; launcher.example.json is the
committed fallback. Drop the stale launcher-manifest.example/schema, update
.gitignore and docs (CLAUDE.md, plan/18).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Added picaxe (both). Pruned 5 stale keys for removed jars: LittleFrames,
waterframes, watermedia, watervision (WaterMedia stack dropped from the
pack) and the orphan iris key. sides keys 135→131 = current dist count.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
watermedia throws IllegalEnvironmentException on server-side ("keep it
on client"). Its natives (wm_binaries) and dependents (watervision,
waterframes) are client-render media mods too — flip all four to client
so sync drops them.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
cwb declares @Mod(dist=CLIENT); classify-mods.py read its toml side as
BOTH (the dist annotation isn't in the manifest), so it synced to the
server and RuntimeDistCleaner FATAL'd on CubesWithoutBordersImpl. Flip
to client so sync drops it.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Distribution grew 76→135 forgemods; classify-mods.py seeded the 59
new jars into mods-sides.json (57 both, 2 client). Unblocks the server
mod sync — server was stuck on the old 52-jar subset.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Rewrite the routine-redeploy section around update-all.sh (rolling vs --hard),
list the steps it runs, note the python>=3.11 requirement for build-modlist on
cochi (default python3 is 3.10), and that www/ + mods.json are generated every
run. The deploy skill = pull + fetch-authlib + ./update-all.sh --hard.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The host's default python3 may predate stdlib tomllib (cochi/Ubuntu 22.04 ships
3.10), so the shebang-resolved python3 failed the version guard and strict-halt
aborted the run. Pick the first python3.11+ interpreter instead.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
update-all.sh is the single source of truth for the on-host post-pull build +
restart steps: render drasl/nmsr configs, sync server mods, regen the landing
mod list, rebuild www/, then apply via docker compose. Two modes:
- default (rolling): up -d --build, then restart ONLY services whose mounted
inputs changed since a pre-run snapshot — drasl/nmsr on a config re-render,
minecraft on a mod change. Minimal downtime. (Cannot detect caddy static-conf
changes — use --hard for those.)
- --hard: down --remove-orphans && up -d --build (full restart, MC downtime).
Strict halt on any error; landing build sources nvm + pnpm (never npm) and bakes
.env. fetch-authlib stays out (rarely changes) — the deploy skill runs it before
update-all. Refactor the deploy skill to `pull && fetch-authlib && update-all.sh
--hard` so the step list can't drift.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replace Pixelify Sans with Block Blueprint as the heading font (theme.headFont).
Block Blueprint isn't on Google Fonts, so it's vendored straight from 1001fonts
(License: 1001Fonts Free For Personal Use — fine for this private friends'
server) as a TTF, with a hand-written @font-face appended to fonts.css. The
vendor step is added to fetch-fonts.sh (runs after the Google rewrite so it
survives re-runs). Pixelify stays available as a HEAD_FONTS option + the
hard-coded fallback.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Members roster avatars render 230px tall (full body); NMSR size bumped to 256
and tiles widened to fit the portrait.
- Features section is now two columns: the 4 feature cards stacked on the left, a
Minecraft goat image on the right (public/Goat_JE1_BE1.webp). Stacks on narrow
screens.
- Header nav links trimmed to How to Join + Status; the auth CTA cluster is now
Register / Account / Play (the old "Log in" button relabeled Account →
Cuenta/Kontua).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Default language Spanish at root /, English moves to /en/ (eu unchanged):
getStaticPaths {lang:undefined}->es + LANG_*_PATH maps swapped in ui.ts.
- Section order: hero -> join -> status -> members -> mods -> features.
- Join flow 3 -> 2 steps (drop the Connect/IP step; launcher handles the
address). Remove the hero IP CopyChip. Header CTA renamed Play/Jugar/Jolastu
and gains Register + Login links on all four pages.
- Account page: hide the whole login block once authenticated (not just the
form), add an avatar Refresh button (cache-busted NMSR re-fetch), flatten the
avatar (no bevel). New i18n key account.refresh.
- Register page: new REGISTRATION_SHOW_INVITE env (default false/hidden) toggles
the invite-code field independently of the REGISTRATION_MODE backend gate.
- Members roster: drop the `admin` account, render full-body via a new
ROSTER_AVATAR_MODE (default fullbodyiso); portrait tiles.
- Style: flatten emboss on feature icons (.picon) + player tiles (.roster-tile).
PixelIcon gains optional image support (/icons/<name>.png|svg).
- Favicon now /favicon.png (committed) on all pages.
- Docs: plan/09-landing.md Assets section (logo/favicon/features-icons drop
points) + default-language note; .env.example documents the two new vars.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Capture what this session learned the hard way, in plan/03-drasl.md and CLAUDE.md:
- Correct admin key is DefaultAdmins (not DefaultAdminUsernames); wrong/unknown
keys are silently ignored by drasl (logged as "unknown config option").
- CORSAllowOrigins must be top-level, before any [Section], or it's ignored.
- First-admin bootstrap under invite mode is a chicken-egg → temp-flip
RequireInvite=false, register, revert.
- NMSR needs the auth. host-gateway pin + https mojank endpoints or avatars
render the default skin (cross-ref plan/19-routes.md).
- Fix the stale example config and mark the admin-bootstrap task done.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Every avatar rendered the default skin. NMSR fetched the player profile from
Drasl fine, but the profile embeds an absolute HTTPS skin URL (Drasl BaseURL is
https), and NMSR could not connect to auth.${BASE_DOMAIN}:443: the mcnet alias
resolves auth. -> caddy, which only listens on :80 internally. Logs showed
`fetch_texture_from_mojang … client error (Connect)` for the skin URL, so NMSR
fell back to the default skin.
Mirror what the minecraft service already does: pin auth.${BASE_DOMAIN} to the
host via extra_hosts so NMSR reaches the host nginx on :443 (real LE cert,
publicly trusted — no private CA needed), and switch the nmsr mojank endpoints
from http:// to https:// so the profile/texture scheme matches the embedded
skin URLs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
NMSR has no per-request avatar refresh (the `?skin=` override is ignored by this
build), so the only lever for a changed skin to appear is the resolve cache.
Drop resolve_cache_duration 15m -> 60s so account-page skin changes propagate
within ~1 min; texture cache stays 48h (Drasl skin URLs are content-hashed, so a
new skin is always a new URL — never stale). Cost is one extra internal Drasl
profile lookup per avatar per minute, trivial for 4-10 players.
Update the now-stale "~15 min" copy (skinLead + skinPreviewNote, en/es/eu) and
code comments to "~1 min".
Add plan/19-routes.md: a reference for the nmsr / landing / auth HTTP routes
(public via caddy + browser->Drasl + internal service-to-service), indexed in
plan/00-overview.md.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Two config keys were silently ignored by drasl (logged as "unknown config
option"), so neither took effect in production:
- `DefaultAdminUsernames` is not a drasl option; the correct top-level key is
`DefaultAdmins`. With the wrong key, the `admin` user never got promoted —
admin-only API routes (e.g. GET /players for the registered-players roster)
returned 403.
- `CORSAllowOrigins` sat after the `[RegistrationNewPlayer]` table header, so
TOML parsed it as `RegistrationNewPlayer.CORSAllowOrigins` (ignored → no CORS
headers → landing register/account browser calls blocked). Moved it top-level,
before any [Section], with a comment warning against re-nesting.
Verified against drasl master configuration.md.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Host cochi carried an untracked landing/pnpm-workspace.yaml with a bogus
`allowBuilds:` key and no `packages:` field, which made pnpm 11 abort every
command with "packages field missing or empty" — breaking the deploy landing
rebuild. Track a correct file: `packages: ['.']` plus `onlyBuiltDependencies`
for esbuild/sharp. Verified install + astro build clean locally.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
oierandClaude Opus 4.8