Spots created before reverse geocoding have coordinates but no city/country
tag, so the map draws them while /stats can only show "Uncategorised". Counted
on the canonical relay: 54 raw -> 43 after admin-hide -> 13 tagged / 30
uncategorised, and 21 of the 30 are the two admin accounts' own pins.
useRegionBackfill finds the logged-in user's own untagged spots/shops, decodes
each pin's coords (lat/lon else geohash), reverse-geocodes via the shared
reverseGeocodeRegions(), and republishes the addressable event with the region
tags added - preserving every existing tag. Publishes to getWriteRelays()
(community + author) so the canonical stats see it, spaced 1.1s for Nominatim's
1 req/s policy. Surfaced as an admin-only RegionBackfillCard in Settings ->
Advanced.
EditSpotPage rebuilds eventTags from scratch, so saving a spot replaced the
addressable event WITHOUT its city/country/country_code — silently losing the
region tags it had. It also meant re-saving a legacy spot could never fix a
missing tag.
Both edit pages (and both create pages) now go through the shared
reverseGeocodeRegions()/pushRegionTags() helper, so region tags are re-derived
on every save. Re-saving a legacy spot now tags it.
Docs: AGENTS.md (any page that republishes an addressable event must
re-derive or preserve its region tags).
A spot the community has hidden as 'not a skate spot' no longer inflates the
community numbers.
- useGrowthStats calls useAdminModeration() and drops hidden content inside
dedup(), so it's excluded from totalSpots, the region breakdown and the
trusted counts (kind-agnostic: the coord is built from each event's own kind).
- The hidden-set sizes are part of the query key, so the stats recompute once
the labels load.
- Region footnote says 'community relay only, with moderation-hidden content
excluded'.
This narrows the map<->stats gap. The remaining deltas are legitimate and
documented in AGENTS.md (untagged pins; auto-hidden spots).
Answers 'why don't the map and stats match' by making the gap visible, and
brings shops into the region breakdown.
- src/lib/geocodeRegions.ts is now the single reverse-geocode helper (spot
create, shop create, shop edit), so the Nominatim call + Accept-Language + the
parse rules can't drift. Never throws.
- Shops now publish city/country/country_code (previously none), so they can
appear in the region breakdown.
- Stats: regions cover spots AND shops; a new 'Uncategorised (N)' row plus a
'{N} of {M}' footnote let the card reconcile with the location total. Section
retitled 'Spots & Shops by Region'.
- PERSIST_BUSTER v3 -> v4.
- /privacy: wording now 'spot or shop geocoding'.
The map deliberately still reads the community relay + the reader's relays
(forcing canonical would hide an opted-out user's own spots) - documented.
Answers 'how are these accounted?' and fixes the privacy-page gap it exposed.
- Spot creation now also publishes ['country_code', <ISO-3166-1 alpha-2>], kept
alongside the display 'country' name so existing search/display still work.
- Stats group countries by country_code when present, else a lowercased name;
the label is the most common raw name in the group, so a localised spelling
can't split one country across two rows.
- The tail beyond the top 6 is surfaced as an 'Other' row, and the section is
relabelled 'Spots by Region' (it counts spots, not people) with a footnote.
- PERSIST_BUSTER v2 -> v3 (GrowthStats shape changed again).
- /privacy: disclose OpenStreetMap Nominatim — spot geocoding sees your IP and
the spot's pin (never the device's GPS). It was missing from the IP list.
Docs: AGENTS.md.
/stats slammed with 'Cannot read properties of undefined' for anyone whose
IndexedDB held a growth-stats cache saved before the new fields existed —
stats.topCountries.length threw.
- Bump PERSIST_BUSTER v1 -> v2 (the documented way to invalidate a shape change;
it is wired into App.tsx).
- Read the new fields defensively (stats?.x ?? []) so a future field can't do
this again.
- AGENTS.md: note that the smoke test cannot catch this (clean profile, no
persisted cache).
Three improvements to /stats.
1. A relay outage must not read as real zeros. Every sub-query swallowed its
error, so a failed relay rendered a full board of 0s. The queryFn now counts
failures and returns 'partial'; the page shows 'Couldn't reach every relay —
these numbers may be incomplete'. Badge also gains 'updated Nm ago' + Refresh.
2. Surface more of the community: Marketplace (live listings, completed trades,
self-reported sats, cash/barter split) and Regions (top countries/cities from
each spot's own tags). Zap receipts are counted ONLY for the top contributors
and labelled as not-global-volume, because global zap volume isn't measurable
on a public relay.
3. Mobile/desktop: collapse the 13 Lab cards on mobile behind 'Show all' (a
display:contents wrapper keeps them grid items); widen max-w-6xl -> max-w-7xl;
StatCard gains a compact prop so big sats don't overflow.
Docs: AGENTS.md community-stats note.
The 'New here?' banner was a later sibling of the absolute toolbar inside the
same map container, so it became the first in-flow child at y=0 and the floating
toolbar painted over it — the search bar covered the banner.
Move the banner to be the container's first child (above the toolbar) and have
the toolbar drop into flow when the banner is shown, so it can't cover it while
still floating over the map otherwise. Desktop is unchanged (md:static).
Docs: AGENTS.md map-layout note.
Two Nostr-native auth upgrades (read-only npub sessions were declined).
NIP-49 (src/lib/ncryptsec.ts):
- Login: the Secret Key tab gains an 'Encrypted (NIP-49)' mode — paste
ncryptsec1… + password -> decrypt -> login.
- Signup: the done step's key card exports a password-protected backup, which
is safe to keep in cloud storage (unlike a raw nsec).
- Verified in Node: round-trips; wrong password, non-ncryptsec and empty
password all rejected.
NIP-46 nostrconnect (src/lib/nostrConnect.ts):
- Desktop -> phone pairing via QR. Nostrify implements only the bunker
direction; nostrconnect is signer-initiated, so the handshake lives here
(ephemeral client keypair + secret, subscribe kind 24133 #p:[clientPub],
reply {id,result:secret} signed with the ephemeral key), then hands the
derived bunker:// URI to login.bunker().
- Try NIP-44 then NIP-04 (the transport moved NIP-04 -> NIP-44; support varies).
- LoginDialog's Bunker tab leads with the QR; bunker:// paste stays a fallback.
- NOT verified against a real signer — needs an on-device Amber check.
Docs: AGENTS.md.
Owner feedback: the route dialog's map still dominated the phone screen, and
the 'nothing leaves to Google or Apple Maps' tagline cluttered the chrome.
- Map: 50vh -> 34vh on mobile (min-h 300 -> 220); desktop 440 -> 400. Follow
mode 60vh -> 54vh, so the bigger map is kept only when you need it.
- Drop the visible tagline from the route header (kept sr-only for a11y); the
promise now reads on /privacy -> 'What leaves your device' -> Directions
('routing stays in the app; start/destination are never handed off to Google
or Apple Maps').
Docs: AGENTS.md map-layout note.
Builds out the rest of the transit UX from data Transitous actually provides.
- Time picker: Leave now / Depart at / Arrive by (MOTIS time + arriveBy).
- Mode filter: Bus/Tram/Metro/Train/Ferry groups -> MOTIS transitModes.
- Stop departure boards: tap a stop name -> MOTIS stoptimes, with live and
cancelled badges and a feed-provided platform hint.
- Fares: MOTIS carries NO prices (verified live: debugOutput.fares=0,
agencyFareUrl empty), so a 'tickets' link is shown ONLY when the feed supplies
an operator fare URL - never an invented price.
New src/lib/transitUi.ts holds the mode vocabulary (emoji/colour/groups) so the
itinerary and the board cannot drift; StopDepartures.tsx is its own dialog.
Docs: AGENTS.md transit note.
Transit showed one cryptic summary line and threw away 3 of 4 itineraries.
Surface what the MOTIS/Transitous response already contains.
routing.ts: TransitLeg gains headsign, agencyName, from/to track, stop count,
realTime and distance; RouteResult gains transitOptions (all itineraries,
fastest first). fetchTransitRoute maps every itinerary (the API returns more
than numItineraries).
RouteDialog:
- Alternatives row — tap to switch; a dedicated effect redraws the map from the
selected itinerary, so switching never re-queries the service.
- Leg-by-leg itinerary: coloured line badge, headsign, board/alight stop + clock
times, platform track, N stops, operator, and a 'live' badge only when the
feed is realtime.
- Summary now reflects the selected option and shows the option count.
- 'Whole route' re-fit follows the selection.
Also clears the 2 pre-existing eslint warnings (unused StatusIcon/nip19) so the
documented zero-warning baseline holds.
Docs: AGENTS.md transit-detail note (incl. Transitous 403s a generic UA).
Two failures today were invisible: the ngit/Amber remote signer stalling
mid-push, and the deploy timer skipping while origin/main had moved.
- scripts/auto-deploy.sh records the reason for every skip (and the run time)
to ~/.local/state/skatespots-deploy/{last-skip,last-run}.
- scripts/ops-health.sh reads that back and prints repo/push sync, the skip
reason, the throttle window and timer health; exit 1 when action is needed.
--probe-push runs a bounded real push (the only true signer probe),
--notify desktop-notifies (deduped), --quiet is silent when healthy.
- Periodic watch: skatespots-ops-health.{service,timer} (every 15 min) plus
scripts/install-ops-health-timer.sh.
Docs: AGENTS.md Deploy section.
The stash is one replaceable NIP-44 blob, so a naive last-write-wins publish
silently dropped whatever another device added. Make it merge.
- StashItem.updatedAt (last edit, falls back to addedAt) and a top-level
Stash.deleted[] of {id, at} tombstones so deletes aren't resurrected.
- export mergeStashes(): union by id, newest stamp wins, tombstones win at/after
the last edit; tombstones pruned (180d / 500).
- fetchLatestStash() merges the live remote with the list already held (query
cache -> closure) and toasts when the other device contributed.
- Writes now require a confirmed relay read: assertConfirmed() refuses an
unconfirmed base so useNostrPublish can't queue a pre-signed blob that would
overwrite a newer edit later. Reads still work offline from cached ciphertext.
- StashPage surfaces the refusal via onStashError on every mutation.
Docs: NIP.md multi-device merge section; AGENTS.md updated.
Signup and onboarding are the first impression, so tighten both.
Signup (SignupDialog):
- Add a Choose -> Profile -> Save key step indicator.
- Make backup a confirmation, not a skip: an "I've saved my recovery code"
checkbox enables "Let's skate" (downloading the key also sets it); the old
co-equal "Skip for now" is now a quiet link.
- Fix copy that contradicted step 3 ("nothing to remember") and warn that a
signer may ask to approve the profile publish.
- Platform-aware signer nudge instead of implying a recommendation.
LoginArea: drop the dead SignupMethodDialog/MnemonicSignupDialog wiring (never
opened) and rename the misleading nsecSignupDialogOpen state.
Onboarding: add a "Where do you skate?" first step (Use my location caches the
fix so the map opens on the skater's area) and remove the dead disabled
"Log in first" button.
The Wishlist tab lived only in the encrypted stash, so 'I want an 8.25 deck'
was invisible to sellers. A public want (itemType wtb) now publishes a real
NIP-99 listing with listing_type=wtb (tag stash-want, price = what they'd pay),
which the Marketplace already surfaces under its WTB filter; the encrypted
Wishlist keeps the private copy.
- buildListingTags handles the want case (no showcase tag, listing_type=wtb,
no condition); syncItem no longer skips wtb.
- StashPage drops the item.itemType !== 'wtb' showcase guards (add/edit/bulk).
- AddItemDialog defaults wants to PUBLIC (a want is meant to be seen) and the
visibility hint + wishlist teaser explain where it goes.
The private stash was deny-listed from persistence with no local cache, so an
offline or relay-outage cold start showed an EMPTY stash even though decrypting
needs no network.
- offlineCache v2 adds a 'stash' table storing the encrypted CIPHERTEXT (never
the decrypted list). useStash caches it on every successful fetch, and both
the read query and fetchLatestStash (the pre-write re-fetch) fall back to it --
so the stash loads at the park and an offline edit rebuilds from the real list
instead of throwing.
- /stash gains a 'Backup & transfer' card (desktop sidebar + mobile) with
Export (a JSON you control) and Import (one re-encrypt, skipping
title+category+size duplicates).
Offline banner: it keyed off a single relay probe (SKATE_RELAY) while the map
reads from several relays, so it reported 'showing saved spots' while fresh data
loaded fine - and its retry only refetched spots, never re-probed, so it appeared
impossible to clear. Now the saved-spots banner is driven by deviceOffline
(!useOnlineStatus) only, is dismissable (re-armed when the connection returns),
and both retries call refetchSpots() AND relay.recheck(). The relay probe still
gates the 'nothing at all' card.
Trade confirmations: usePendingTradeConfirmations now returns createdAt/eventId,
so 'someone confirmed your trade' becomes a trade-confirm notification - the
missing final step. Bell card + OS title cover it; confirming back is what makes
the trade mutual and builds both reputations.
Listings already supported price_type=trade, and attestations carry method=trade,
so swaps counted for reputation. Offers, though, could only name a sats amount.
- 30403 messages may now carry an optional 'item' tag: what the sender is
offering in kind. It combines with amount (e.g. 2,000 sats + an 8.25 deck) or
stands alone for a pure swap.
- OfferDialog: 'Or trading something' field + validation (amount, item or note).
- ListingOffersPanel: the thread shows the offered item, and the counter UI takes
an optional item alongside the amount.
- Bell/OS text reads it ('offered 2,000 SATS + an 8.25 deck').
- NIP.md documents the item tag; AGENTS notes barter is now first-class.
The bell had All/Social/News/Crews; trade steps were only inside All. Added a
5th tab (Handshake, orange accent) listing offers, counter-offers, outcomes and
trade confirmations, with its own count. All still shows them first.
Offers were one-way (buyer proposes, seller accepts/declines) and fixed-slot
addressable events, so the buyer could never counter back and the haggling left
no trace. Now it's a thread.
- useMarketplaceOffers: every message gets a UNIQUE d (append-only) with a
'role' tag (offer = buyer / response = seller), 'e' = the message it answers,
and status. New counterOffer() plus respondToOffer(previous, status). Parsing
still accepts the legacy fixed-slot d form when the role tag is absent.
- useListingOffers returns OfferThread[] (oldest first) with = whose
turn it is; useIncomingOffers returns the threads awaiting me (across listings).
- ListingOffersPanel: shows the volley, and accept/decline/counter (inline amount)
for whichever side's turn it is, with withdraw on your own open offer. Counter
notifications are nudged over DM too.
- New useTradeNotifications: incoming offers, counters and 'confirm our trade'
requests, derived from events already addressed to the user.
- Bell: trade items render first in All and count toward the badge (vs seenAt).
- OS notifications: new 'trades' category in notificationPrefs + a trades source
in OsNotificationBridge.
- NIP.md updated for the thread model (legacy form documented).
- Directions: the Distance/Time/Arrive tiles are now a compact one-line strip
(StatTile removed), and the dialog's DOM order matches its visual order
(header, map, vehicle row, From row, details, pinned Start) with redundant
order-* so a dev CSS-timing hiccup can't scramble it.
- Crews: the filter expand toggle now has a divider and always shows its
More/Less label, so it no longer reads as another crew category.
Three phone issues from a screenshot review:
- Map: the floating toolbar sat on top of Leaflet's top-left zoom control.
index.css now offsets .leaflet-top.leaflet-left by 4.25rem under max-width
767px, scoped to the map page via a new class
(desktop toolbar is in flow, so no offset). Verified at 390x844.
- Sasquatch: the cramped red 8px 'Go Pro' chip is now a proper lime pill with
the bolt, matching the panel.
- Directions: the map was 38vh and buried under the mode row. It's now 50vh
(440px on larger screens; 60vh/62vh while following) and sits directly under
the header, with the mode/vehicle row and 'From' row moved BELOW it. Done with
flex order-* utilities rather than moving JSX; Start bar stays pinned.
Follow mode was a live dot on a north-up map you had to keep watching. Now it
guides:
- Voice: device TTS via speechSynthesis (free, offline, no server, no
permission), preferring an on-device voice in the user's language. Speaks at
~150 m and ~40 m before each turn (once per band, per step) and on arrival.
Unlocked by the Start tap; mute toggle persisted in skatespots:route-voice.
- Wake lock: navigator.wakeLock keeps the screen on while following, released on
stop/unmount and re-acquired on visibilitychange.
- Arrival: within ~30 m -> "You've arrived", haptic, auto-stop, Done button.
- Haptics: navigator.vibrate on the 150 m / 40 m bands and arrival.
- Layout: while following the map grows to 58vh (60vh sm) and the planning
step-list steps out of the way; the NEXT banner is bigger.
- Copy updated: no longer claims 'no voice guidance'.
Still absent: auto-reroute if you leave the route. True screen-off navigation
needs a native foreground service (app must stay foregrounded).
The transport treats a crew's wss:// coordinator URL AS A RELAY, and an https://
one falls back to the public wss://relay.contextvm.org. So 'nothing crew-specific
hits a public relay' was false for https coordinators — it put ciphertext +
metadata (crew coord, timing) on a public relay.
- CrewCoordinatorCard: shows the actual relay in use, and states honestly
whether it's the crew's own or the shared public one.
- CrewForm: the coordinator field now says a wss:// endpoint is the crew's own
relay and an https:// one routes over the shared public relay.
- AGENTS.md: record the relay-vs-coordinator transport reality so this isn't
re-derived (and Cordn RPC is not called).
Carrier is the existing kind-0 'cordn' field (app convention, NIP.md) — no new
kind. It now holds a coordinator *pointer* rather than only a URL.
- EditProfileForm: accepts nprofile / cordn.net link / npub / hex / http URL,
validates via parseCoordinatorInput, normalises a bare pointer to Cordn's
canonical add-link on save, and adds 'Use my coordinator'. Label -> Coordinator.
- NIP19Page: when the profile's cordn value parses as a pointer, render a
Coordinator row with a one-tap Add (sets the viewer's coordinator, with toast)
and an 'Open in Cordn' link; otherwise keep the old plain-link row.
- NIP.md: document the broadened cordn field (pointer formats + one-tap add).
- smoke test: now boots a real profile route too.
Crew-level Cordn chat already existed (lib/cordn, useCordnChat, CrewChat,
CrewCoordinatorCard, the crew 'coordinator' tag). The real gap: the crew form
accepted a URL only, so a coordinator run from Cordn's phone app — which yields
an nprofile, not a URL — could not be entered at all.
- lib/coordinator.ts: parseCrewCoordinatorInput / crewCoordinatorInputValue —
the same pointer formats as Settings -> Coordinator (nprofile / cordn.net
link / npub / hex) plus a plain endpoint URL.
- CrewForm: parses the pointer, shows live valid/invalid feedback, sets
['coordinator', url, pubkey?] (schema unchanged), and gains a one-tap
'Use my coordinator' prefill from Settings.
- NIP.md: note the accepted pointer formats on the coordinator tag.
- AGENTS.md: record that crew-level Cordn chat already exists (the lesson:
grep src before building).
SkateSpots stays a Cordn *client*: it points at a coordinator the user
chooses or runs, and hands phone-hosting off to Cordn's own app. No embedded
coordinator, no default SkateSpots coordinator (deliberate: liability).
- src/lib/coordinator.ts: Cordn's own formats — an nprofile pointer, the
canonical add-link https://cordn.net/chat/coordinators?c=<nprofile> (from
Cordn packages/server/src/main.ts), default relay wss://relay.contextvm.org,
the docker one-liner. Parses nprofile/npub/hex/link; encodes an add-link.
- src/hooks/useCoordinator.ts: device-local store + presence probe. Health is
NOT Cordn RPC — it queries the coordinator's relays for authors:[pubkey] (its
ContextVM announcement/heartbeat), so a private/unannounced coordinator
honestly reads 'Not seen' rather than failing.
- src/components/CoordinatorSettingsCard.tsx + Settings -> Coordinator:
empty/connected states, add by paste or QR scan, add-link copy + QR, and a
'Run your own' hand-off (open Cordn, scan its QR, copy the docker command).
- smoke test now boots /settings?section=coordinator.
Verified against Capacitor v8 docs: Background Runner exposes fetch but not
WebSocket, and Nostr relays are WS-only — so on-device background relay polling
cannot work. Recorded the two real alternatives (push gateway / foreground
service) and the owner decision to hold at the zero-server design.
An OS alarm an hour before each upcoming event the user RSVP'd going or
interested to. Because the OS holds the alarm it fires with the app fully
closed — the only closed-app notification achievable without a server.
- src/lib/reminders.ts: schedule/cancel via @capacitor/local-notifications,
tracking handed-out ids in localStorage so withdrawn RSVPs and passed events
are cancelled without disturbing one-off notifications. Web is a no-op.
- src/hooks/useLocalReminders.ts: reuses the existing upcoming-events and
my-RSVPs queries; no new relay destination, no signing.
- Bridge mounts the hook only on native and only when the new 'reminders'
pref is on; unmount cancels everything (opt-out/logout).
- Skipped the Pro-expiry reminder on purpose: it would drag the heavy
useSasquatchPro query into a nudge.
Bumps to v1.5.7 (versionCode 10507); APK rebuilt.
Mirrors the activity the app already computes (bell, crews, DMs, newsletter)
to the OS notification shade, with no new relay traffic and no new network
destination.
- src/lib/osNotify.ts: one delivery API. Web uses the Notification API; native
uses @capacitor/local-notifications (lazily imported, so the web bundle never
pays for it). Tap-through navigates in-app.
- src/components/OsNotificationBridge.tsx: mounted once in Layout, inert until
opted in. Per-source silent priming + a seen-id set means history is never
replayed as a flood, and it stays quiet while the app is visible (the bell
already covers that).
- DMs notify from the gift-wrap ENVELOPE only (src/hooks/useDmEnvelopes.ts,
no decryption) — so Amber is never prompted just to be told you have a
message. Preview is honestly generic.
- src/lib/notificationPrefs.ts + useNotificationPrefs: device-local, opt-in,
default off, per-category toggles. Settings -> Notifications UI.
Deliberately local-only: a fully closed app needs a server push, which is not
built (Phase 2 = on-device background runner; Phase 3 = content-free FCM).
Both are production-only (Netlify Function + the CVM-host proxy), so on
localhost the dev server answered them with index.html and the NIP-05
availability check showed 'Could not check right now' (and dev AI silently fell
back to Pollinations). Proxying them to the real endpoints (as /api/mints
already is) makes dev behave like prod. Verified: the check returns a real
availability result and /api/ai/status returns the live Routstr balance.
Dev-only config — the built output is unchanged, so the deploy guard skips it.
- lib/badges.ts + ProfileBadges: a profile shows (1) the built-in SkateSpots
badge, derived from a nip05 ending @skate-spots.org (app logo linking to
https://skate-spots.org/ — no storage, works for existing users), and (2) user
badges from a kind-0 badges array ({label,url?}) rendered as clickable chips.
- NIP19Page: app badge in the header, user badges in a 'Badges' card (replaces
the plain Sponsors section).
- EditProfileForm: Sponsors editor becomes a Badges editor (label + optional
link); saves the badges array and retires the legacy newline sponsors field.
Existing sponsors are read back as label-only badges, so nothing disappears.
The 'test of secret island airport' case: a no-photo, one-line, reported pin sat
on the map looking legit. Four fixes:
- A) Report has a visible effect: a 'Reported as not a skate spot by N — under
review' banner on the spot page, for everyone.
- B) Publish floor: a spot needs a photo OR a non-empty description. Photo stays
optional (per the earlier decision); title + type alone no longer publishes.
- C) Auto-hide from the map when >=3 DISTINCT accounts report it AND it has no
community signals (no community photo, no status report). The 'no signals'
clause is the anti-brigade guard; reversible. useSpotReports now exposes
distinct .
- D) The flattering 'Conditions ... NN/100' block is hidden until the spot is
confirmed (community or staff) — on an unconfirmed pin it read as the app
endorsing the spot itself.
- #3 usePersonalContext decrypts the Stash ON DEMAND (unlockStash, called by
getMyStash) instead of never — 'what's in my stash?' is now correct, while the
globally-mounted guide still never decrypts on mount. stashRef avoids the
stale-closure bug.
- #4 new self-contained action draft wtb_gear (NIP-99 listing_type=wtb) — a
'wanted' gear ad. SYSTEM_PROMPT + tool handler + publish branch.
- #5 panel header shows 'Pro · Nd' days left (title = exact expiry).
- #6 tier picker computes live USD via useExchangeRate; removed the hardcoded
~$60k/BTC usd field from PRO_TIERS.
- #7 copy: 'skate status' not 'spot status'; trip planning stated as free.
#1 Extend now actually extends. useSasquatchPro collected receipts and took the
max of (receipt.created_at + tierDays), so buying again before expiry
under-delivered. It now stacks every valid proof with renewal semantics
(expiry = max(expiry, proofTime) + days), client and server alike.
#2 Owner-grant fallback. Pro can now also be granted by an owner-authored NIP-32
kind-1985 label (org.skatespots.pro, p = user, expiry) — so Pro is not solely
dependent on the Lightning provider publishing zap receipts. New owner-only
GrantProCard (Settings -> Advanced) publishes it to getWriteRelays(); the hook
reads via querySkateContent.
Protocol: the x-sasquatch-pro payload is now {receipts,grants}; BOTH ai-proxy and
the Netlify function accept the new payload and a bare receipt (old APKs). Proxy
was deployed first and verified healthy before this commit.
Removed from App.tsx defaultConfig, NostrSync NEW_USER_RELAYS, NostrProvider
DEFAULT_RELAYS, TestApp, and the CVM DEPLOY.md example. It has been failing for
a while and just adds a dead socket + latency to every default relay set.
Existing accounts keep it in their stored config until they remove it in
Settings -> Relays (no migration applied). APK rebuilt (1.5.5 / 10505) so the
release matches.
backup-relay.sh now pushes the archive to restic.env AND, if present,
restic-2.env — so adding an R2/B2 bucket gives true 3-2-1 (local + CVM +
object storage) instead of replacing one offsite with another. Each repo is
sourced in its own subshell so RESTIC_* don't leak between them.
Verified with a throwaway local second repo: both pushes ran, both snapshots
exist. Also documents the umask trap (mode-600 cache dirs break restic).
All four DM read hooks used config.relayMetadata.dmRelays, which is only set by
the login-time sync. When it is stale/empty the app reads only NIP-65 relays —
where gift wraps almost never are (real case: 94 addressed wraps on nostr.mom,
0-1 on the user's NIP-65 relays), so Messages showed 'No messages yet' with a
full inbox.
- useResolvedDmRelays(): discovers the user's own kind-10050 inbox relays at
query time (per-relay, 6s each, over read relays + aggregators) and unions
them in. Used by useConversation/useConversations/useGroupConversations/
useGroupConversation.
- useConversations now returns { conversations, undecryptable, wraps } and the
DM page surfaces 'N encrypted messages couldn't be opened', so a decrypt
failure stops looking identical to an empty inbox. Updated DirectMessagesPage
and ContactsPage.
- DNS: A record ai.skate-spots.org -> 185.18.221.2 (Netlify DNS).
- TLS: Caddy on the CVM host, automatic Let's Encrypt (verified live).
- Client: production builds (web + APK) now default to
https://ai.skate-spots.org/api/ai; dev uses local /api/ai. Override with
VITE_SASQUATCH_AI_BASE to revert to the Netlify Function, which stays
deployed as a fallback.
Same proxy as netlify/functions/ai.mjs, rebuilt as a standalone Bun service on
the CVM host so AI calls stop spending Netlify compute/bandwidth credits and
the sats-paid Routstr key lives on our box.
- ai-proxy/server.ts: Bun.serve on 127.0.0.1:8787 — NIP-98 + Pro gate, model
allowlist, per-IP throttle, /api/ai + /api/ai/status. Own tsconfig
(typecheck:ai in npm test); eslint-ignored like cvm-server.
- ai-proxy/{package.json,.env.example,skatespots-ai.service,Caddyfile,DEPLOY.md}.
- scripts/deploy-ai-proxy.sh (npm run deploy:ai): scp + bun install + unit +
restart, mirroring deploy-cvm.sh.
- Client: one place decides the AI base, src/lib/sasquatchAi.ts (also the NIP-98
u tag). Defaults unchanged (/api/ai on Netlify); set VITE_SASQUATCH_AI_BASE to
move it. Fully reversible.
Deployed and verified on the host: /api/ai/status returns the live Routstr
balance. Public TLS (DNS + Caddy) is the remaining step before switching the app
over, so no client behaviour changes yet.
The timer deployed on every site-affecting commit at 15 Netlify credits each
(~68 deploys / ~1,020 credits in a week). auto-deploy.sh now enforces a minimum
gap between successful deploys, default 24h, overridable with
SKATESPOTS_DEPLOY_MIN_INTERVAL=<seconds> (0 disables). A throttled tick records
nothing so the change still deploys later; failures retry immediately.
Deploy on demand with: scripts/auto-deploy.sh --force
scripts/docs only — skipped by the timer, so this costs 0 credits.
- Offsite backup is now live: restic over SFTP to the CVM host via a scoped
~/.ssh/config alias (no host/key in tracked files); restic.env (600, outside
the repo) holds the repo URL + generated password. Verified: first push
dedupes to ~32 MiB and 'restic check --read-data-subset=10%' is clean.
- scripts/drill-relay.mjs: throwaway in-memory NIP-01 relay (ws) for restore
drills. Ran the loop end-to-end against the real archive: 114/114 events
restored, 0 missing.
Scripts/docs only — the deploy timer skips this (no Netlify credits).
Map tiles: new lib/tileCache.ts (Dexie SkateSpotsTilesDB, bounded to ~3000
tiles) and lib/cachedTileLayer.ts, a Leaflet TileLayer that reads through it
(cached -> fetch+cache -> plain URL fallback). Wired into both maps (SpotMap,
RouteDialog). Chosen over a service worker so it works in the Capacitor
WebView too, where a SW never sees tile requests.
Photos: useUploadFile drops from three Blossom hosts to two and records
['fallback', '<mirror>/<sha256>']. BlossomUploader signs once and Promise.any
PUTs to every server in parallel, so the old three-host list was triple-
uploading every photo (the 'tries them in order' comment was wrong). Read
side: lib/imageFallback.ts derives same-hash mirrors and ImageWithFallback
retries them before showing the placeholder - one choke point for ~20 photo
call sites, so a dead host no longer blanks a spot.
scripts/archive-relay.mjs snapshots the app's own content off
goskatespots.nostr1.com to gzipped JSONL of the raw signed events, so a
restore can republish them verbatim to any relay. scripts/restore-relay.mjs
re-verifies signatures and publishes to an explicit --to (no default, to
avoid replaying a snapshot onto the live relay). scripts/backup-relay.sh
wraps it and pushes with restic when a repo is configured; a weekly systemd
user timer runs it.
Scoped on purpose: that relay is PUBLIC and carries general traffic (15.7k
kind-0 profiles, 2k kind-30078 app-data from other clients, 408 foreign
NIP-99 listings), so the archive keeps only the app's own kinds/tags plus
contributor profiles, SkateSpots labels, and reports that reference archived
content. ~114 events / 63 KiB.
Not configured yet: offsite (restic.env) and a full restore drill.
