Kaj Schittecat
3171f5adea
Stacked on #585 and merged after it, as the PR asks. Two conflicts, both from the stack base moving: apps.json: unioned by id, so all fourteen apps survive. harness.lua: the additive blocks came from #586. Where both sides edited the same code, #585's is newer and wins: its ping scenarios pass cfg.gps as a FUNCTION, which is what the harness calls (cfg.gps()); #586 still passed the table and would have failed. Taking #586 there would have broken ping. Then restored the two dispatch branches my first resolution dropped. The breakout and tripodometer scenarios existed but nothing ran them, so both apps silently fell through to the gpscompass scenarios and "failed". All five new apps pass their own scenarios, and so do gpscompass and sdscan.
wadamesh.com infrastructure
Distribution stack for wadamesh: a VPS nginx origin behind Cloudflare.
tiles.wadamesh.com →CF (HTTP, edge-cached) → nginx → OpenStreetMap / OpenTopoMap
firmware.wadamesh.com →CF (cache bins) → nginx → /srv/wadamesh/firmware
flasher.wadamesh.com →CF (HTTPS) → 301 → wadamesh.com
wadamesh.com →CF (HTTPS) → nginx → /srv/wadamesh/site
Map tile styles. The default /{z}/{x}/{y}.jpg route serves OpenStreetMap
(the firmware default). An opt-in OpenTopoMap topographic style is served from
/opentopo/{z}/{x}/{y}.jpg (explicit OSM alias at /osm/...); the device requests
it only when the user enables Map → Options → Topographic map. Legal: OpenTopoMap
map tiles are © OpenTopoMap (CC-BY-SA) over © OpenStreetMap contributors
(ODbL) + SRTM — the touch UI shows that attribution when topo is active, and the
14-day disk cache keeps each tile hitting OpenTopoMap at most once per fortnight
(their tile-usage policy asks for a contactable UA + caching, both of which the
transcode service provides). Deploying the topo routes = update
tiles.wadamesh.com.conf + tile-transcode.py, then
systemctl restart wadamesh-tile-transcode && nginx -t && systemctl reload nginx
and purge the Cloudflare cache for tiles.wadamesh.com/opentopo/*.
The firmware fetches tiles + the update-check over plain HTTP (on-device HTTPS isn't viable — mbedTLS needs ~30 KB heap, only ~5 KB is free post-Wi-Fi), so the tile + firmware hosts must stay reachable over HTTP. Cloudflare provides the edge cache, HTTPS for the flasher, and hides the origin IP (so no IP lives in this repo or the firmware).
1. VPS (origin)
sudo apt install nginx
sudo mkdir -p /srv/wadamesh/firmware/releases/TOUCH /var/cache/nginx/wadamesh-tiles
sudo cp deploy/nginx/tiles.wadamesh.com.conf /etc/nginx/sites-available/
sudo cp deploy/nginx/firmware.wadamesh.com.conf /etc/nginx/sites-available/
sudo ln -s /etc/nginx/sites-available/tiles.wadamesh.com.conf /etc/nginx/sites-enabled/
sudo ln -s /etc/nginx/sites-available/firmware.wadamesh.com.conf /etc/nginx/sites-enabled/
sudo nginx -t && sudo systemctl reload nginx
2. Cloudflare
- DNS:
A/AAAArecords fortiles,firmware,flasher,@→ the VPS IP, all Proxied (orange cloud). - SSL/TLS: mode Flexible (CF↔origin HTTP) is enough since the origin is
HTTP-only. Do NOT enable "Always Use HTTPS" on
tiles.orfirmware.— the firmware needs plain HTTP there. - Cache Rules:
tiles.wadamesh.com/*→ Eligible for cache, Edge TTL ~14d.firmware.wadamesh.com/releases/*/*.bin→ cache, Edge TTL ~1d.firmware.wadamesh.com/releases/TOUCH(the listing) → short TTL (~60s) or Bypass, so new releases appear promptly.
3. Publishing a release
From a wadamesh checkout (builds both boards, refreshes the listing, rsyncs up):
WADAMESH_VPS=user@your-vps scripts/release.sh beta_2
The on-device check GETs http://firmware.wadamesh.com/releases/TOUCH, finds the
highest beta_<N>, and (once OTA-over-Wi-Fi is re-enabled) pulls
…/releases/TOUCH/beta_<N>/<board>.bin.
Done
- Web flasher ✅ — the guided install page at
wadamesh.com, served fromdeploy/site/and published byscripts/deploy-site.sh(esp-web-tools / Web Serial, per-board install buttons + .bin downloads, manifests generated per release byscripts/build/gen-flasher-meta.pyinto the/latest/and/latest-beta/feeds thatrelease.shrefreshes each publish). flasher.wadamesh.com301-redirects to the apex (seedeploy/nginx/flasher.wadamesh.com.conf) — it is an alias, not its own page.deploy/flasher/is the original standalone flasher page. No deploy script publishes it —deploy-site.shshipsdeploy/site/only — so it is effectively an offline/local copy kept in parity by hand. Retire it or wire it into a deploy target; until then, treatdeploy/site/index.htmlas the only install page users can reach.
TODO before public launch
- Re-enable OTA-over-Wi-Fi in the firmware (currently it version-checks then defers to manual flashing).
- Flip
wadameshrepo public = launch. - Decide tile-proxy sharing: dedicated
tiles.wadamesh.com(this config) vs reusing the meshcomod proxy.
Never commit the VPS IP, SSH keys, or
WADAMESH_VPS. Cloudflare fronts the origin; the deploy target is supplied via the environment at publish time.