Hotel reservations store their date range in day_accommodations rather
than on reservation_time, so the card date block never rendered. Pull
accommodation_start_day_id / accommodation_end_day_id from the SQL join
and surface them on the card.
Also apply Maurice's badge-pill pattern (day name + localized date pill)
to the day-range display, consistent with the modal day selectors.
fetchSynologyThumbnailBytes was still serving 240px while the
uncached streamSynologyAsset path had been bumped to 320px in
#761. Align the cached path with the streaming default.
Without these args, the Wikimedia fallback (used when no Google Maps key
is configured) silently skips the fetch because lat/lng are NaN. The plan
view (PlaceAvatar/photoService) already passes all three; this aligns the
PDF path with the same behaviour.
fetchPlacePhotos only checked google_place_id, skipping places that
only have osm_id (e.g. those added via MCP). Mirror PlaceAvatar logic
by falling back to osm_id in both the filter and the photo fetch call.
If a trip has no dates set but a day has a custom title, the
dropdown showed only the title with no context. Fall back to
'Day N' as the badge so users can still tell which day it is.
Day selectors in the Transport, Reservation and Hotel-Day-Range
modals only showed the renamed day title once a day had a custom
name — hiding the actual date. Added an optional badge prop to
CustomSelect, rendered as a pill next to the label, and wired the
date badge onto all affected dropdowns. FileManager day section
headers got the same pill for consistency.
Also translated transport.addTransport and transport.modalTitle.*
in all 13 non-English language files; the keys existed but still
carried the English source string.
MobileMapTimeline received the raw entries array, bypassing the
synthetic-container filter applied to timelineEntries. On screens
below the lg breakpoint (<1024px) the [Trip Photos] sync container
leaked back into the combined map+timeline view.
Fixes#773: isValidBackupFilename regex anchored to ^backup- rejected all
auto-backup-* filenames, causing 400 on download/restore/delete. Broadened
to ^(?:auto-)?backup-.
Fixes#774: three regressions in the trip Files tab —
- openFile import shadowed by a local function of the same name inside
FileManager; PDF preview modal was calling the local with a URL string,
corrupting state and crashing on the second click (mime_type read on
undefined). Fixed by aliasing the import as openFileUrl.
- GET /:id/download used a bespoke authenticateDownload that checked only
Bearer header and ?token= query param, ignoring the trek_session cookie.
After the JWT-to-cookie migration the client sends cookies only, so every
download silently 401-ed. Extended authenticateDownload to accept req and
check cookie → Bearer → query token in priority order.
- files.download and files.openError translation keys were missing from all
15 locale files; t() was returning the raw key as a truthy string,
defeating the || 'Download' fallback.
Migration 107 and the previous fix both wrote /api/maps/place-photo/<id>/bytes
into places.image_url without ever fetching the photo bytes. photoService
short-circuits on that URL prefix and hits /bytes directly, which 404s because
nothing is on disk.
- Add migration to null proxy image_url rows with no backing google_place_photo_meta
entry — restores the normal fetch-and-cache flow for affected rows
- Fix the previous legacy-URL migration to null instead of rewrite, so fresh
installs don't hit the same 404 path
Fixes#770 (follow-up)
Migration 107 only rewrote image_url rows matching /places/%/photos/%; URLs using
the /place-photos/ or /places/<opaque> paths survived the upgrade and were passed
verbatim to the Places API, producing a malformed request whose empty/HTML response
body threw SyntaxError before detailsRes.ok was checked. The resulting rejection was
leaked by placePhotoCache.setInFlight via an unhandled .finally() chain, triggering
Node 22's default unhandledRejection=throw and terminating the process.
- placePhotoCache: add .catch() after .finally() to prevent unhandled rejection crash
- mapsService: reject URL-shaped placeIds early; read response as text before JSON.parse
- migrations: add migration to rewrite remaining googleusercontent/places.googleapis URLs
- MapView/MapViewGL: prefer stable proxy URL form of image_url before google_place_id
Fixes#770
- Remove NOT NULL constraint on day_accommodations.place_id (migration)
and change ON DELETE CASCADE → SET NULL so deleting a place no longer
cascades to the accommodation row
- Switch listAccommodations / getAccommodationWithPlace to LEFT JOIN so
accommodations without a linked place are visible to the modal
- Relax create/update guards in reservationService to only require
start_day_id + end_day_id, not place_id; place_id remains optional
- Client save guard now sends create_accommodation whenever FROM/TO days
are set, regardless of whether a hotel place was selected
- Add re-hydration useEffect in ReservationModal to back-fill hotel
fields from the accommodations prop when it arrives after modal opens
(race between isOpen and the tripAccommodations fetch)
- Fix demo-seed TDZ crash: move db Proxy declaration before DEMO_MODE
block so circular require in demo-reset resolves correctly
- Sidebar accommodation badge falls back to reservation title when
place_name is null; click/cursor disabled for placeless accommodations
- listAccommodations now joins reservations to expose reservation_title
ReservationOverlay was Leaflet-only: react-leaflet components, L.divIcon,
panes, useMap/useMapEvents. When the user switched the planner map to
Mapbox GL, the entire feature disappeared — no polylines, no endpoint
badges, no clickable IATA labels.
Add a matching overlay for the Mapbox renderer:
- New reservationsMapbox.ts with an imperative `ReservationMapboxOverlay`
class — mapbox-gl is imperative, so a React component wrapper would
fight its own lifecycle every render. The manager owns one GeoJSON
source + line layer for the arcs, one HTML `mapboxgl.Marker` per
endpoint badge, and one per flight stats label. It cleans itself up
when the map is rebuilt (style/token/3d toggle) or unmounted.
- Geometry helpers (great-circle arc, antimeridian split, haversine,
tz-aware duration math, label formatting) are copied from the Leaflet
overlay so both renderers produce the same lines. Great-circle is
useful even on the Mapbox globe because the mercator projection mode
still draws the short-way line, and the antimeridian split prevents
a NYC↔Tokyo flight from wrapping halfway around the planet.
- Flights / cruises get geodesic arcs; trains / cars get straight
lines. All four types get clickable endpoint badges with the
matching lucide icon; only flights render the rotating mid-arc stats
label (IATA → IATA · distance · duration) — same rule as the Leaflet
overlay.
- The stats label's rotation is recomputed on every `render` event by
projecting two points straddling the arc midpoint, which keeps it
parallel to the arc as the camera rotates/zooms on the globe.
- Visibility thresholds mirror the Leaflet overlay (per-type min pixel
distance before a line / endpoint label is worth drawing).
- MapViewGL now accepts the `reservations`, `visibleConnectionIds`,
`showReservationStats`, `onReservationClick` props that the Leaflet
MapView already took. `visibleConnectionIds` is honoured the same way
— the per-booking toggle in DayPlanSidebar controls which routes
appear, so switching the renderer doesn't lose that UX.
- Added a `mapReady` gate so the overlay can only add its source/layer
once the map's `load` handler has attached the other trip sources;
the gate resets on every style rebuild.
Replaced loose includes()/startsWith() path checks with exact equality
for static routes and strict prefix matching for dynamic-token routes.
Added /forgot-password and /reset-password to the allowlist so the
password-reset flow is usable without auth. Extracted isAuthPublicPath
as a pure testable function with 14 unit tests covering regressions.
- Budget table column alignment: the NAME data cell had
`display: flex` directly on the <td>, which pulled it out of the
table-layout and desynced the column widths between data rows and the
AddItemRow. Moved the flex wrapper into a <div> inside the cell.
Closes#759
- Packing list: template-apply and bulk-import handlers called
`window.location.reload()` to refresh the list, which re-rendered the
whole trip loading screen. Both flows now merge the returned items
into the trip store instead. Closes#760
- Journey timeline: move-up / move-down arrows were rendered on
skeleton suggestions — skeletons are places from the linked trip and
don't participate in sort order. Skip canReorder when
entry.type === 'skeleton'. Closes#763
- Journey public view: the synthetic `[Trip Photos]` and `Gallery`
entries produced by syncTripPhotos were leaking into the public
timeline and map. The owner view already strips these in
JourneyDetailPage — apply the same filter on JourneyPublicPage.
Gallery photos still come from every entry so a shared gallery keeps
showing the trip-synced photos. Closes#764
- Journey thumbnails: public gallery grid was loading the original
asset for every tile. `photoUrl()` now takes an optional kind and the
grid requests `thumbnail`; the lightbox still opens the original.
Synology thumbnail default bumped from `sm` (240px) to `m` (320px)
because `sm` looked pixelated on retina. Closes#761
- SEC-H6: remove conditional audience check in mcp/index.ts — audience is
now always enforced against the mcpResource URL. Add migration to revoke
pre-existing oauth_tokens with audience=NULL so dead rows don't linger.
- SEC-H4: validate doc.issuer against config.issuer inside discover() to
prevent a MITM'd discovery doc from supplying a crafted expected issuer.
verifyIdToken caller now passes config.issuer as ground truth, not
doc.issuer.
- tests: cover three new OIDC callback failure paths (no_id_token,
id_token_invalid, subject_mismatch) and two idempotency caps (key length
>128 chars returns 400, body >256 KiB skips caching).
- CI-C1 false positive: actions/{checkout,setup-node,upload-artifact}
@v6 do exist (v6.0.0 releases published Oct-Dec 2025). Restore the
@v6 refs — the earlier batch-1 commit downgraded them unnecessarily.
- Widen idempotency_keys primary key to (key, user_id, method, path)
via new migration. Batch 1 widened the middleware lookup but left
the table PK at (key, user_id), so `INSERT OR IGNORE` silently
skipped the second endpoint that reused a key — the cache was
never populated for it and a replay re-ran the handler. The
migration rebuilds the table preserving existing rows (the old
narrower PK guarantees no conflicts against the new looser key).
- HSTS: keep `includeSubDomains` OFF by default. Enabling it for
every NODE_ENV=production install would break apex-domain setups
where siblings still serve HTTP. Operators who want the stricter
policy opt in with HSTS_INCLUDE_SUBDOMAINS=true.
- Extend the idempotency unit tests to cover the (method, path)
dimension — same user+key on different path no longer replays.
Second-pass fixes caught by a self-review after the initial commit — each
one would have undermined a fix from the previous commit.
- mfaPolicy now goes through `verifyJwtAndLoadUser` too. Without this,
a JWT stolen before a password reset still satisfied `require_mfa`
until its natural 24h expiry, defeating the whole point of the
password_version bump.
- Drop the `?? keys[0]` fallback in OIDC JWKS key selection. When the
token carries a `kid` that is not in the current JWKS, refuse
outright instead of picking an arbitrary key and letting the
signature check produce a generic failure — the real failure mode
deserves a specific error code.
- Tighten OAuth DCR custom-scheme rule so `javascript:`, `data:`,
`vbscript:`, `file:`, `blob:`, `about:`, `chrome:` are all rejected.
Previously the catch-all "not http/https" check admitted them; the
authorize flow later 302s the browser to whatever is registered,
which with a `javascript:` URI would execute attacker script on
redirect. Also require the private-use scheme body to be reverse-DNS
(contain a dot), matching RFC 8252 §7.1.
- permanentDeleteFile / emptyTrash only delete the trip_files row when
the on-disk unlink actually succeeded. Previously Promise.all
swallowed individual unlink failures and DELETE ran unconditionally,
so a permission / ENOSPC failure would orphan bytes on disk.
- restoreFromZip also invalidates the permissions cache in the outer
catch. If extraction threw before the DB swap even started, the
cache wasn't stale, but belt-and-braces is cheap and guarantees no
failed-restore path leaves stale cache behind.
Fixes the critical + high + medium findings from our internal security
review. Bundled into one PR because the changes overlap heavily (JWT
verification unifies across three call sites; backup-code hashing and
demo-email handling cross-cut several services); splitting them out
would mean redundant reviews of the same files.
Critical
- CI-C1 — .github/workflows/test.yml: restore actions/{checkout,setup-
node,upload-artifact} to @v4. The @v6 refs don't exist, so the test
workflow was errorring before a single test ran.
- SEC-C1 — mfaPolicy now extracts the token via extractToken() (cookie-
first, Bearer fallback). Previously it only read Authorization, so
every cookie-authenticated SPA session bypassed require_mfa entirely.
- SEC-C2/C4/C6 — all JWT verification paths (MCP bearer, file download,
photo route) now go through the shared verifyJwtAndLoadUser that
checks password_version. resetPassword additionally deletes every
mcp_tokens row and marks outstanding oauth_tokens revoked, so a
password reset invalidates ALL credential classes — not just the
cookie JWT.
High
- SEC-H2 — reset email URL is built from server-side APP_URL /
ALLOWED_ORIGINS (via existing getAppUrl()), not request headers.
Closes the host-header-injection vector into reset links.
- SEC-H3 — OIDC findOrCreateUser wraps the invite-redemption UPDATE +
user INSERT in a transaction. The UPDATE is the capacity check; if
a concurrent callback takes the last slot, the whole transaction
aborts with registration_disabled instead of double-creating users.
- SEC-H4 — new verifyIdToken() performs full JWT signature
verification via the provider's JWKS (Node's crypto.createPublicKey
accepts JWK directly — no extra dependency), plus iss/aud/exp
checks. The callback also rejects the login when userinfo.sub does
not match id_token.sub.
- SEC-H5 — OAuth DCR now validates redirect_uris against an allowlist
of schemes: https, http-loopback, or a private custom scheme. Plain
http://non-loopback is rejected.
- SEC-H6 — oauthService audience defaults to mcpResource when the
`resource` parameter is missing, so tokens are always audience-bound
to /mcp instead of being issued with audience=null.
- SEC-H7 — HSTS is enabled any time NODE_ENV=production (previously
required FORCE_HTTPS=true), includeSubDomains defaults on and can
be disabled with HSTS_INCLUDE_SUBDOMAINS=false.
- SEC-H8 — trek_session cookie Secure flag is also driven by
req.secure (which Express resolves from X-Forwarded-Proto once
trust proxy is set), so instances behind a TLS-terminating proxy
get Secure cookies without needing FORCE_HTTPS.
Medium
- SEC-M1 — permanentDeleteFile / emptyTrash / avatar unlink now use
fs.promises.rm with { force: true } (one async op vs the previous
existsSync + unlinkSync pair per file).
- SEC-M2 — invalidatePermissionsCache() is called inside restoreFromZip
so a restored DB with different permission rows is honoured
immediately.
- SEC-M3 + C1 — idempotency store bounds the key at 128 chars, caches
only responses ≤ 256 KiB, and scopes the lookup by (key, user_id,
method, path) rather than (key, user_id). Same key replayed against
a different endpoint no longer returns a stale unrelated body.
- SEC-M4 — share_tokens gets an expires_at column; new tokens default
to 90-day TTL, expired tokens are denied at lookup. Existing tokens
stay NULL = no expiry so already-published links don't break.
- SEC-M5 — /uploads/photos/:filename now resolves the photo to its
trip_id and requires the share token to cover THAT trip. Previously
any share token for any trip would unlock any photo filename.
- SEC-M6 — BLOCKED_EXTENSIONS is the single source of truth shared
between fileService and collab uploads. The '*' allowed_file_types
wildcard now still rejects executables/scripts.
- SEC-M7 — single DEMO_EMAILS constant (services/demo.ts) used by
demoUploadBlock, mfaPolicy, and every demo-mode guard in
authService. The old demoUploadBlock only matched 'demo@nomad.app'
so the seed 'demo@trek.app' could in fact upload in demo mode.
- SEC-M8 — MFA backup codes are now bcrypt-hashed at rest
(hashBackupCodeBcrypt). matchBackupCode accepts both bcrypt and
legacy SHA-256 hex hashes, so existing installs keep working until
the user regenerates codes via enableMfa.
- SEC-M9 — document the "security via UUID v4 filename" model for
/uploads/avatars|covers|journey. Requires no code change but
captures the decision so future reviewers don't re-flag it.
- SEC-M10 — already covered by the resetPassword revocation logic
above: mcp_tokens DELETE + oauth_tokens UPDATE … SET revoked_at.
Performance
- PERF-H1 — new migration adds the indexes flagged in the audit:
trips(user_id), trips(created_at DESC), photos(day_id),
photos(place_id), reservations(day_id), share_tokens(token), plus
conditional day_accommodations and notifications indexes depending
on which columns are present.
Tests
- tests/integration/oidc.test.ts now mocks verifyIdToken and passes
an id_token in the exchangeCodeForToken stub for the three flows
that exercise a successful callback. The three remaining failures
tests pointed out were all pre-existing (file-upload flakes +
notificationPreferences event_types count drift), none introduced
by this PR.
Todos already support a due_date field but nothing notifies the user
when a deadline is approaching — you'd only remember if you happened
to look at the Lists tab. This wires a reminder into the existing
notification pipeline so due-date todos behave like trip-start
reminders.
Details:
- New `todo_due` event type alongside trip_reminder; all four channels
(in-app, email, webhook, ntfy) supported and toggleable per user in
Settings > Notifications.
- New daily scheduler task (9 AM local TZ) queries unchecked todos
whose due_date is within the next 3 days. Each todo gets at most
one reminder per 24 hours, tracked via a new todo_items.reminded_at
column (migration 116).
- If the todo has an assigned user, only that user is reminded; if
not, every member of the trip gets the notification.
- Strings added in all 15 UI languages and for all notification
carriers.
- Gated by app_settings.notify_todo_due (default on) so admins can
disable it globally.
- Suppress trek-stagger animation on the day list while a drag is active
so nth-child delays (0–320 ms) no longer re-fire on every hover change
- Replace sibling drop-indicator <div> injections with borderTop/borderBottom
on the target row to prevent nth-child index shifts during drag
- Dedup setDragOverDayId calls in onDragOver handlers so setState is only
invoked when the active day actually changes
- Move initTransportPositions out of getMergedItems (render path) into a
useEffect to stop mid-drag setState cascades
Moves the 91 MB product-tour GIF out of the repo entirely. Standard
clones and CI checkouts no longer pull it — even LFS-aware clients
previously downloaded the blob on checkout, which made `git pull`
noticeably slower for everyone.
The file now lives as a release asset on a separate repo and is
referenced from README via its GitHub Fastly-backed download URL.
Removes the LFS tracking entry from .gitattributes.
Update Password-Reset.md and Login-and-Registration.md to reflect the
email-based forgot-password flow added in feat(auth): 51387b0, including
the SMTP-less console fallback, MFA gate, session invalidation, rate
limits, and security properties.
The 'From Gallery' picker on the journey entry editor used `aspect-square`
on grid items inside an overflow-scrolling container. Safari (desktop and
iOS) collapses the computed height of aspect-ratio boxes in this layout,
which stacked every thumbnail at y=0 — making selection impossible.
Swap to the classic padding-top spacer pattern (`paddingTop: '100%'` on
the cell + absolutely positioned image) which is bulletproof across
browsers and preserves the 5/6-column grid on mobile/desktop.
- Animated TREK logo (light + dark variants) via <picture> + prefers-color-scheme
- 60-second product tour GIF (91MB, 1100x619, 10fps) stored via Git LFS so
standard clones don't pull it by default
- 9 feature tiles as composite SVG grids: 3x3 on desktop, 2x4 on mobile
- 8 fresh screenshots captured from dev.pakulat.org
- Feature details folded into a collapsible 2-column table
- Environment variables moved behind a collapsible
- Roadmap badge added next to Live Demo / Docker / Discord
- Removed redundant Community section and footer
jubnl