- Tile store: the WebView's headers go along (a URL-restricted Mapbox token
is checked against the Referer), and anything but a tile is left to the
WebView, so a 401/403 never comes from the store.
- Widgets read their snapshot inside the composition, so every write reaches
a running Glance session; the trip picker updates only the widget the id
belongs to.
- Location: Approximate counts as allowed; the stream pauses with the app,
resumes with it, and stops when the page reloads or the activity goes.
- Billing: an already-owned subscription comes back as the purchase, so the
shell redeems it; a payment that clears outside the sheet is announced
(storeTransaction) and found again at start (storeUnfinished).
- getWindowLayout answers with the last posture and hinge, not the bare window.
- Share sheet: keeps the picked trip through rotation and folding, and takes
content:// only, never a file:// path read with TREK's own rights.
- The location shim runs at high accuracy only while a watcher wants it and
drops back once that watcher is gone, instead of keeping the 2 s GPS rate
for the rest of the session.
- The purchase client can hand over a Play subscription that was paid but
never delivered (a pending payment that cleared later), and the plans screen
redeems it at once, as a restore would, before Google Play refunds it after
three days.
The Android app asks the server once whether Google Wallet works there and
offers the button only after it says yes; a 404 from the pass route takes the
button away again. The wallet button's test mocks the new exports and covers
the Google case.
- GET /api/trips/:tripId/reservations/google-wallet answers whether this
server can make Google Wallet passes: passes on, and the TREK services'
signer with its Google route mounted (it is once liketrek has a Wallet
issuer). The probe is kept for ten minutes.
- The Google route is derived from TREK_WALLET_SIGNER_URL by its path, the
query kept; a signer URL that does not end in /pass has none, and the
booking is refused with a 404 instead of going to the Apple signer.
- End-to-end tests for the save link: owner, stranger, another trip, no
issuer yet, a link that is not Google Wallet, and the availability.
- Italian and Hungarian wording of the button.
Android lint counts every string without a translation once the app has
languages; the app name and the URL scheme are names, not text. Removing
location updates is guarded for a permission revoked meanwhile.
The new methods sit at the end of the interface and the restore's Android
answer has a type of its own, so the watch and large-screen branches, which
add their methods after storeRestore, merge without a conflict.
Everything the iOS shell gained, built natively for Android 16 with Material
You, and back to Android 7:
- Home-screen widgets in Jetpack Glance: trip, today, bookings, to-dos,
packing, world map and travel group, from the same snapshot the iOS widgets
read. Wallpaper colours on Android 12+, resizable in three layouts, a trip
picker per widget (reconfigurable), live previews in Android 15's picker
and preview images before it.
- The travel day as a Live Update: a promoted ongoing notification with the
next stop as the status-bar chip, a countdown and the day's stops as
segments (ProgressStyle), a quiet ongoing notification before Android 16.
- A share sheet of its own: a Material bottom sheet with the trip picker for
files, mails, pages and links shared from any app, and the coming trips as
launcher and direct-share shortcuts.
- Push: channels per kind of notification (chat, trip updates, invitations,
general) by the event the server now names, stacked per channel.
- Location for navigator.geolocation from the fused provider, a store for
Mapbox's map resources (30 days, 250 MB), Google Play Billing behind the
StoreKit methods, and Google Wallet save links.
- Large screens: every orientation and window size without a restart,
the window's size classes and a foldable's posture and hinge for the UI
(windowLayout), the UI's keyboard shortcuts as the app's (keyCommands),
predictive back, per-app language.
- The native strings in the 28 languages of the iOS app, generated from its
strings files (scripts/generate-android-strings.mjs).
Debug builds carry a widget gallery that draws every widget at every size.
The UI now turns these on by feature on Android as well as iOS:
- widgets, widgetTrips and the travel day (liveActivity) write the same
snapshot to the Android app's Glance widgets; the travel day also sends
the next stop's minutes and the stops' colours for the Live Update.
- geolocation answers navigator.geolocation from the app's location stream.
- tileCache: Android keeps Mapbox's resources in its WebView client, so the
map is offline-capable there too, without the iOS trek-tile:// rewrite.
- playBilling: the purchase client buys through Google Play on Android,
with the purchase token as the receipt the control plane verifies.
- googleWallet: a booking's button adds it to Google Wallet on Android,
through the save link the server asks the TREK services for.
The push payload carries the event it reports, so the Android app can file a
notification under a channel of its own (chat, trip updates, invitations).
A browser and the iOS extension ignore the field.
GET /api/trips/:tripId/reservations/:id/wallet-pass/google sends the same
request the Apple Wallet signer takes to the TREK services' Google route and
answers with the signed save-to-wallet link the Android app opens. Off with the
other TREK services, and a clear 404 while liketrek has no Wallet issuer yet.
- The claim screen says, right above the buy button, that the TREK Cloud
terms govern the instance and links them (liketrek.com/terms?app=1, the
view without the site's order buttons); the plan screen links them too,
next to Apple's EULA, which is now named as Apple's.
- On Android the payment note and the empty restore say Google account and
Google Play's own path, not Apple ID and App Store.
- The redeem sends the app's language, so the contract confirmation mail
arrives in it.
The wallet pass commit had reformatted the whole repository file, which put
it 200 lines over its size baseline. It is back to its old form, in the format
baseline again, and the pass-context query lives in a helper of its own.
It now matches the App Privacy answers in App Store Connect: user id, device id
(the push token for the relay) and emails or text messages (group chat, imported
booking mails), next to email, name, purchases, photos, other user content and
precise location; all for app functionality, linked, never for tracking.
- Orders closed at the cloud: the app shows the plans and that TREK Cloud
opens soon, and sells only in a test build (TestFlight, App Review), whose
sandbox purchases the cloud still sets up; a store purchase it would refuse
is never offered. The plans arrive without prices then, the store's own are
shown where it gives them.
- A purchase paid but never set up (Ask to Buy approved later, the app closed
mid-purchase) is picked up by Restore purchases and goes on to the claim with
that purchase, instead of failing; nothing is bought twice.
- The buy button names the period with the price (guideline 3.1.2).
- The privacy link opens the policy without the site's sales header.
- The privacy manifest declares what TREK Cloud keeps: email, name, the store
purchase, photos, other user content and saved places, all for the app to
work, linked to the account, never for tracking.
- The email hint promises messages about the contract, not receipts, which the
store sends.
- TREK_WALLET_PASS_URL read as a secret to the env check (the word PASS); it
is TREK_WALLET_SIGNER_URL now, everywhere it is documented
- findPassContext compared the TEXT accommodation_id with an integer id, which
Postgres refuses; it compares them as numbers now, as both databases accept
- the wallet-pass service test takes its subject from createTestModule instead
of building it by hand, and no longer names the DOM-only BodyInit type
- Reservations.repository.ts is formatted, so it leaves the format baseline
- the wallet button test reads the translated texts the render helper gives
"Add to Apple Wallet" on a booking in the iOS app turns it into a TREK pass:
a boarding pass from and to for flights, trains, buses and boats, check-in and
check-out for a stay, pickup and drop-off for a car, the time for anything
else, with the confirmation code as a QR code. Wallet brings it up on the lock
screen around its time and place: times carry the offset of the endpoint's zone
(looked up from the coordinates where the endpoint has none), stays and
restaurants their place's coordinates. The labels come from the booking forms'
own strings, so nothing new to translate on the pass.
Only liketrek holds the Pass Type ID certificate, so the server sends the pass's
texts to the TREK services' signer (TREK_WALLET_PASS_URL, push.liketrek.com)
once per tap; it stores nothing, and TREK_SERVICES=off turns it off with the
other services. Texts are cleaned of links and shortened here, the answer must
be a .pkpass within 512 KB, and a user gets 20 passes a minute.
- New wallet-pass domain: GET /api/trips/:tripId/reservations/:id/wallet-pass
behind the trip guard, the booking checked to belong to the trip; trip title
and place come from the reservations domain (findPassContext)
- MWalletButton in the transport and booking sheets, only where the app binary
has the wallet feature; undated bookings and failures are told apart
- TREK_WALLET_PASS_URL in .env.example, compose, Helm and the wiki
Opening a .pkpass in the iOS app brought up the generic share sheet. It now
shows Wallet's own sheet with the pass and "Add", and a pass that is already in
Wallet opens there instead of being offered twice (TrekShell.addToWallet,
feature wallet). A broken pass or a device without Wallet falls back to the
share sheet as before.
Passes carry an "Apple Wallet" badge in the trip's files and a wallet icon
where they hang on a booking, so it is clear what a tap does.
A share extension puts TREK in the share sheet of Mail, Safari, Files, Wallet
and any app that shares a document, a link, text or a photo. It shows what came
in (a Wallet pass, a PDF, a mail, the page as Safari rendered it, text, a photo)
and the user's trips with their covers, the next one already picked, and hands
both to the app, which opens that trip's booking import with the file in place.
"Choose later" leaves the trip question to the app's import sheet as before.
- ShareExtension target (add-extension.rb learns "share": JavaScript and the
localizations go in as resources), bundle com.liketrek.trek.share, signed in
the release lane, its plists linted in native.yml
- Safari hands over the rendered page (SharePage.js), so a confirmation behind
a sign-in arrives whole; other apps' links are fetched
- The item travels through the shared keychain group the widgets already use;
the app takes it on com.liketrek.trek://share or when it next becomes active
- Feature shareSheet; strings in the app's 28 languages
The plan picker sent buyers to the website, which the App Store does not allow.
It now sells through StoreKit 2 in the TrekShell plugin (feature storeKit):
App Store price strings in the buyer's currency, the purchase sheet, the signed
transaction redeemed at the control plane, and the transaction finished only
once the instance exists. A purchase whose redeem failed is redeemed again on
retry, never bought twice.
Restore purchases, the auto-renewal terms and the links to the terms of use and
the privacy policy sit under the plans. Where no store purchase is possible
(Android, an older binary) the picker shows the plans with no way to buy and no
pointer to the web.
vi.mock is hoisted above the file's imports, so the managed-tab mock could not
use the icons and JSX imported at the top (ReferenceError before initialization).
The trip PDF's print view linked Poppins from fonts.googleapis.com, so printing
a trip sent the reader's IP address to Google, the case German courts treated as
a GDPR breach (LG München I, 3 O 17493/20). It now declares the Poppins files
TREK already ships under /fonts, and the content security policy no longer
allows fonts.googleapis.com or fonts.gstatic.com, so no font can be fetched
from Google again by accident.
The mobile admin never read managedAdminTabs, so a tab an install attaches
there (managed/index.tsx) existed on the desktop only. The phone's section
picker now lists them after the built-in sections, as the desktop sidebar does,
and shows `mobileElement` when a tab brings a phone layout of its own, else its
desktop `element`. TREK itself attaches nothing; the seam stays empty.
MDemoWelcome used raw rem sizes, which ignore the text size setting and broke
theme:lint, and was missing from feature-pairs.json, which broke lint:pairs.
components/ may not import from mobile/ (lint:layers). The phone dashboard
mounts the sheet itself, as it already mounted the welcome, and the desktop
dialog no longer asks which shell it is in.
On a phone the head took half the screen, the list scrolled in a small window
and the start button broke over two lines next to the links. The phone now gets
a bottom sheet of the phone shell: a small head with the countdown, the same
content as a list, a full-width start button with the two links under it. The
photo credits open in its place and back, as on the desktop.
An object literal with airplaneSeat inside the legs() call failed the excess
property check of tsconfig.tests.json; the nodes are built as constants first,
as the other seat tests do.
The grouped flight wrote a seat only on the booking itself, from the first
segment. Each leg in metadata.legs now carries the seat of its own segment,
which the planner and the PDF already show per leg.
KItinerary writes a flight seat as airplaneSeat and a ticket seat as
reservedTicket.ticketedSeat.seatNumber, the schema.org fields, and it
removes seat and seatNumber from a script result. segmentDetail read only
seat and seatNumber, so an imported booking lost its seat.
segmentDetail now also reads airplaneSeat, and then the ticket seat when
no other key has a value. seat and seatNumber stay first, so a booking
that already has a seat gets the same result.
- The carousel and timeline tests step by the new 90 ms settle
- The journey search keeps its suggestion type through the fallback
- nearestCardIndex leaves useJourneyCarouselSync, which had grown past its
line limit; WidgetBag is not exported; the test baseline drops by the
non-null assertions a test no longer needs
The journal searched the full text only, which for a hotel in Osaka listed
the Japanese names the index holds. It now asks the planner's typed-ahead
search first (usePlaceSuggestions, the user's language, one billed session
per search), shows each place with where it is, and takes its point from the
details once picked; the full search still answers when there are no
suggestions. (#3003)
The track was w-10, which is rem and followed the text size setting, while
the knob and its travel are px; with a smaller text size the knob hung over
the end of the track. The track is 40 px now.
Widgets followed the dashboard's trip only, which is the one under way, so a
packing list or to-dos for a trip still being planned could not sit on the
home screen. The trip, today, bookings, to-dos, packing and group widgets now
take a trip in Edit Widget (an AppIntent with the trips running or ahead,
Automatic by default, which keeps following the dashboard); the world map
needs none.
- The app writes a snapshot per trip running or ahead (at most six, the
dashboard's order) and the list to pick from, under their own keys
(feature widgetTrips, so an older binary never mistakes one for the
dashboard's); bags, chat and polls are kept ten minutes, covers per trip
- The widget extension targets iOS 17, which configurable widgets need; the
apps are not released, so no installed widget goes away
- The choices and their texts in the 28 languages of the gallery
- The entry carousel draws the card at the centre larger while the swipe is
still moving (focusIndex, a transform), and the active card and the map
follow 90 ms after it stops instead of 150; the grow takes 200 ms
- The journey map's zoom buttons set their display inline, which beat the
class that hides them on touch screens; the display is a class now (#3001)
- The packing widget data no longer carries a colour literal
- Packing: a ring of what is in the bags, each bag against its weight limit,
the items still missing in their bag's colour, a nudge when departure is
two days or less away
- World map: the world as dots from the Atlas country shapes
(WorldGrid.swift, generated by WidgetPreviews/make-world-grid.py), the
visited countries glowing, the latest one ringed, continents as bars and
the latest countries with their flags; it needs no trip
- Travel group: who comes along as avatars, the latest chat messages as
bubbles and the open poll with its votes as bars
- The to-dos widget is about to-dos only; the packing ring moved out
The app sends the bags, the Atlas countries (an hour's cache) and the chat
and poll with the snapshot; all optional, as before.
- On a phone the shared journey's map stopped at the safe area, so Safari's
toolbar sat over a bare strip; it now runs to the edge (#3002)
- The shared journey on desktop gains the journal's day picker, every day and
its entries one click away (#2998)
With several trips running at once, the widgets took whichever came first
on the device, so an old test trip spanning today pushed out the journey
under way, and the today widget counted its days. The dashboard's choice
moves to utils/heroTrip.ts (running first, earliest start, else the next to
come) and the widgets ask the same function, so the home screen always
shows the trip the app opens on.
The active card grew its box from 136 to 164 px once the swipe had settled,
which moved every card after it and made the snap pull the strip a second
time. Every card now keeps the same box and the active one is drawn larger
with a transform, so nothing moves; the strips keep the room with a wider gap
and top padding. Where the browser reports scrollend, the card grows the
moment the snap lands instead of a beat later.
TestFlight builds now carry the release tag's message as What to Test.
- Names and descriptions of the four widgets in TREKWidgets/<lang>.lproj for
the 28 languages the app speaks, linted with the other property lists
- The preview renderer draws a cover photo state too, and the lock-screen
accessories as the system does, without a box of their own
- The large trip and today widgets put the stops right under the day and the
next booking at their foot instead of an empty middle; the large to-dos
widget names its two rings with their counts
- The trip and day for a travel day as a Live Activity: a progress ring, the
next stop large, the stops as segments, the Dynamic Island in the same green
- The place shows with a location arrow, and rows link through RowLink, which
the preview renderer draws without the Link it cannot render
Four widgets instead of one plain text block, drawn in the phone tokens
(warm paper by day, near black by night) with contour lines and a route arc
as the travel motif:
- Trip (small, medium, large, lock screen): the cover photo, a countdown ring
and a plane on its arc before the trip; the day of the trip as segments,
the weather and the next stops on frosted glass while it runs
- Today (medium, large): the day as a timeline in its category colours, with
a now line, the day's progress and the weather
- Bookings (small, medium, large, lock screen): flights, trains, hotels and
tables ahead as coloured tiles, the next one featured, status dots
- To-dos (small, medium, large, lock screen): to-dos and packing as nested
rings, the most pressing open to-dos with priority and due dates
The app now sends the destination, a small cover JPEG, the weather where the
trip is, the bookings ahead, the open to-dos and the packing count; every
new field is optional on both sides, so older binaries and servers keep
working. Timelines step every quarter hour so the now line moves without
the app. client/ios/WidgetPreviews renders every widget in every size, light
and dark, on the CI's Mac, uploaded as the widget-previews artifact.
The days are a scroll-snap pager since the native UX round: a page turns once
the finger has carried it past half its width. The spec swiped 180 px on a
412 px phone and read the chip right after; it now swipes 40 % either side of
the middle and polls the chip until the snap and the selection have landed.
- RTL: the measured pills and centred controls use logical forms or a marked line
- i18n: the two native widget numbers are positions, not counts (NOT_PLURAL)
- Types: styleOverride on the lazy GL map, the location and shared-file payloads,
the capabilities fallback, the transport sheet's stay metadata
- ESLint: latest-value refs move to layout effects, usePresence keeps the last
open state in state, the booking import opens clean with a shared document
kept in the list (the reset effect used to clear it again)
- Layers: the phone import sheet lives under mobile/ and reads through tripRepo;
the offline host check moves to utils/tileUrl
- Sizes: plugin layer geometry, hardNavigate and the exit animations get
modules of their own
- Tests follow the status radios, the stay chip overview, the dock + button
for the first list and the field token; store registry lists the new stores
Connecting the app to a server as someone not signed in there showed
"Server not reachable" with the server up. The UI's first API call answered
401, the client redirected with a full page load while the first page was
still loading, WebKit cancelled that load (-999), and Capacitor answers every
failed navigation, a replaced one included, with its error page: the shell's
unreachable screen.
The client now waits for the first page to finish loading before such a
redirect inside the apps, which helps the binaries already out; and the iOS
shell puts a small navigation delegate in front of Capacitor's that lets a
cancelled or interrupted navigation pass without the error page.
Docked sheets defaulted to the lighter bar glass, through which their header
buttons hardly stood out; every sheet now uses the same glass. The Vacay
people and legend sheets get the icon tile header.
Atlas bucket list, open route in, open in TREK, the place actions and the
MCP/OAuth client sheets now show the icon tile header (or, where a sheet
closes only by its buttons, the same title row without the close).
Admin sheets: their frame draws the icon tile header (an icon per sheet, the
plugin sheets' icons moved out of the title into the tile), the soft scroll
edge and 42 px footer buttons; admin fields use the sheets' eyebrow labels
and round inputs. The settings picker, the bulk category sheet and the save
to list sheet get the common header (TileHeader may now leave out the icon
for a plain picker). Every phone sheet now uses the glass material, as the
trip settings do; it stays 94 % opaque, so forms over the map remain
legible.
The shared confirm sheet (44 places) gets the icon tile with the title, a
warning icon for destructive questions, and the sheets' 42 px buttons with
the danger colour on a destructive confirm. The dashboard's and the
collections' own confirm dialogs now use it.
Vacay invite, shared calendars and settings get the icon tile header; the
settings' rows separated by hairlines become cards on the inner surface, as
the trip settings' sections.
Both get the icon tile header (the entry keeps its move buttons between
title and close), glass, the soft scroll edge and the sheets' 42 px footer
buttons; the journey name is a round field, delete in the entry's footer a
round icon button. The collection share sheet's invite button is round like
the trip settings' one.
The collections sheets had a header of their own with a hairline, white
square fields and short buttons. Their kit now draws the icon tile header
every phone sheet has (an icon per sheet), round 42 px fields on the field
surface and 42 px footer buttons; the sheets sit on glass with the soft
scroll edge. New list and add place lose their last own field styles: the
labels are the sheets' eyebrows, the search fields, the category picker and
the status choices are round fields of the same height.
The pending / confirmed toggle was shorter than the booking code field next
to it. It is now a 42 px framed field with the status colour as a dot in
each option, marked up as a radio group, and sits under the code at full
width, where both words fit.
New trip: the icon tile header, sections with labels above round fields
instead of floating labels in boxes, the round date fields, the shared
footer buttons, and no box within a box for the reminder and the cover
search.
The shared select, date, time and airport pickers style themselves inline
for the desktop and looked out of place in phone sheets, white and
squarer than the fields beside them. Inside the phone UI they now take the
sheets' 42 px round field on the field surface; the desktop is unchanged.
A tour tapped in the plan opened the desktop detail dialog in a hand-built
overlay: no slide, no drag to close, the desktop's frame on a phone. The
dialog now renders into the frame it is given; on the phone that is a sheet
like every other detail (MTourDetailShell), with the tour's name as its
title, its whole content and every action, the back gesture and an exit
animation.
Tapping the accommodation in the plan went straight to the stay's editor,
where travellers only wanted to look it up. It now opens the overview of the
booking behind it, as on the desktop, with check-in, check-out and nights as
tiles (one shared stayStats for both), and only a stay without a booking or
a place still leads to the editor.
Edit in that overview opened the transport editor for every booking, a stay
or a restaurant table included; non-transport bookings now open the booking
editor, as in the Bookings tab.
A stay with two linked bookings showed its chips twice in the day header,
because the list repeats a stay per booking; the header counts it once.
The shared form fields are round like the trip settings' fields (multi-line
ones keep a softer radius), and the shared footer's cancel, submit and
delete are the sheets' 42 px buttons. Every trip form built on these parts
(place, transport, booking, cost, note, packing, tasks and more) follows.
The journey settings sheet looked like a different app next to the trip
settings. It now uses the same parts: the glass card with the icon tile
header and the journey's name under the title, sections as inner cards with
a title, count badge and hint, 42 px round fields and buttons, the shared
link field, and switch rows for tracks, entry fields and the public link.
Those parts (SheetSection and the SHEET_* button and field classes) moved
into MTripSheetUi for every sheet to use.
Fields and buttons sitting directly in a column no longer shrink below
their height; name, subtitle and the archive button were squeezed flat.
Collections: the empty screen's own New list button is gone; the dock's +
starts the first list while there is none and adds places once there is one.
Trip settings: counts sit in a badge next to the section title instead of
reading as part of it (ACCESS 1), and every round button of the sheet,
invite and add guest as well as the member and guest actions, is one 42 px
circle; the icon-only buttons were wider than tall.
Journey settings: the public link is the same field as the trip's share and
invite links, through one shared MLinkField; the shared map's extras hang
under the map row with a guide line and smaller labels; removing the link is
a full-size button.
The transition waited for the new page to replace its route fallback and
polled for that with requestAnimationFrame. While a view transition waits
for its update the browser renders no frames, so the callback never fired
and every navigation to a page whose chunk was still loading sat frozen
until the browser gave up after four seconds. The wait now polls with a
timer: from Vacay to the journal takes 0.45 s instead of 4.1 s with an
empty cache.
They faded and scaled in but vanished at once. A small presence hook keeps
them for 170 ms after closing while the mirror of their entrance plays:
scaling down and fading on a desktop, sliding down on a phone where the
dialog is docked as a drawer, with the backdrop fading out. During the exit
they show their last content, since an owner often clears the subject in the
same step, and take no clicks. Reduced motion gets a short fade. Unit tests
keep the old instant close.
The card gets a header with its icon in a round badge, its section labels use
the eyebrow of the form's own fields below it, and the description sits on
the mobile field surface with a pill button to take its text over.
Add place on the phone: the place details showed the mascot half a screen
high while nothing was picked or nothing was found, and pushed the name field
below the fold. Both states are now one quiet line, as in the desktop dialog,
and the details sit in a card in the mobile tokens.
Sheets with a header over a scroll box (trip settings, day, export,
transport, files and more) cut their content off in a hard line right under
the header icon. The scroll box now keeps a little room under the header and
fades its content out at the edges.
Copy and share next to a share or invite link are round icon buttons of the
same size; the share button showed no icon because the copy button's padding
left it no room.
The previous commit carried, by mistake, shorter page transitions and a
route preloader that was never wired up. Both go back to how they were;
the warm-up and day rail fixes stay.
The iOS app pulls every file of a build into WebKit's cache so it can start
offline. It did so four at a time at normal priority, for all 1,900 files of
every new build including all 27 translations, and a page opened meanwhile
queued behind it: right after a server update, a first visit to Vacay took
five to seven seconds. The warm-up now runs two at a time at low priority,
takes only the user's language and English, and walks a build once.
The day rail in the trip header lit the new day only after the pager's snap
had landed and the plan had re-rendered, a second or two after the swipe.
The pager now reports the day a swipe heads for as soon as its page is the
nearer one, and the rail and its pill move with the finger; the plan still
switches when the snap lands.
The journal list (its own scroll box, so the hook now takes an optional
scroller) and the notifications screen reload by pulling, like the
dashboard. The list stays on screen while it reloads.
Page transitions: on the phone, pages move like a native navigation stack,
built on the View Transitions API (mobile/routeTransition.ts). A page opened
from another slides in from the trailing edge, back slides it away, and
switching between the dock's sections cross-fades while the dock stands
still. Lazy routes do not flash their skeleton: the swap waits, behind the
frozen snapshot, until the new page rendered. The edge-swipe back gesture
already moves the page and skips the transition. RTL mirrors the motion,
reduced motion and desktop get none.
Pull-to-refresh on the mobile dashboard: the page follows the finger with
resistance, ticks once past the threshold and re-reads trips and
notifications quietly, without a loading state.
Share: trip share and invite links and the journey link get a share button
next to copy, opening the system share sheet (Capacitor Share in the apps,
Web Share in browsers that have it).
Haptics: a selection click for switches and segmented controls, a tick when
a sheet is flung closed, and a firmer bump for the chat's long press.
Theme lint: this branch's new screens use the type and z-index tokens, the
Vacay company tint is a token, and the baseline is lowered to the new
counts.
The Android app now subscribes like the iOS one: TrekShell.registerPush asks
for the notification permission (Android 13+), creates the device's P-256
Web Push keys and returns the relay endpoint for its FCM token
(push.liketrek.com/v1/fcm/<token>). TrekPushService receives the relay's
data message, decrypts the record (RFC 8291, TrekWebPush) and shows the
notification. A tap opens the linked page, also when it starts the app.
The keys are sealed with an AES key kept in the Android Keystore. Firebase
is configured from res/values/firebase.xml, so no google-services plugin is
needed. The decryption is tested against the RFC 8291 example, and the
native workflow now runs the Android unit tests. On the web side push is no
longer limited to iOS: it follows the nativePush capability on both platforms.
A WidgetKit extension (TREKWidgets) with two parts: the next trip on the home
screen, with a countdown before it starts and today's day and next stops
while it runs, and a Live Activity for the travel day on the lock screen and
in the Dynamic Island showing the next timed stop and the day's progress.
The web app builds both from the trips stored offline on the device
(native/widgets.ts) and hands them over through TrekShell
(setWidgetSnapshot, syncTripDay, endTripDay). The texts are already
translated, so the widget needs no localisation of its own. The snapshot
lives in the keychain group the app shares with its extensions, so no App
Group is needed. Taps open com.liketrek.trek://navigate/<path>.
NativeWidgetsSync refreshes on start, on returning to the foreground and every
quarter hour. Both parts are gated behind the widgets / liveActivity
capabilities, so older binaries and the PWA are unaffected.
The relay hands APNs the encrypted record with a placeholder text and
mutable-content set; the new Notification Service Extension opens it with
the device's Web Push keys from the shared keychain group (RFC 8291 on
CryptoKit) and shows the real title, text and link. Without the keys the
placeholder stays, so a notification arrives either way and nothing on the
way reads it.
The target came in through ios/scripts/add-extension.rb (kept, idempotent),
its App ID is registered, and the release lane signs every target with its
own App Store profile. Web Push payloads now stop at 2800 bytes so the
encrypted, base64-encoded record still fits Apple's 4 KB.
The app subscribes to Web Push like a browser, only the app makes the
subscription: it asks for permission, registers with APNs and keeps its own
P-256 key pair and auth secret in a keychain group a notification extension
can share. The endpoint is liketrek's relay for this device's token, so the
server encrypts exactly as for a browser and nobody on the way can read the
notification. The settings card, the test push, sign-in resync and the logout
teardown all take this path where the binary has nativePush; a tapped
notification opens the place it points at, also from a cold start.
The App ID has push enabled, the app carries the aps-environment and
keychain-group entitlements.
The apps subscribe to Web Push like a browser, with their own key pair, but
with an endpoint on liketrek's relay: only liketrek holds the apps' APNs and
FCM keys. The server encrypts exactly as for a browser (RFC 8291), so the
relay hands on ciphertext it cannot read and the device decrypts it. Lookalike
hosts stay refused.
TREK now appears under "Open in" and in the share sheet for PDFs, calendar
entries, Wallet passes and mails (iOS document types, Android VIEW and SEND
filters). The native side reads the file (up to 15 MB) and hands it to the UI,
which asks which trip it belongs to, upcoming trips first, then opens that
trip's booking import with the file already chosen.
WKWebView asks for the location per server origin, under the server's name,
and again across launches. Binaries with the geolocation feature answer
navigator.geolocation from CoreLocation instead: one When-In-Use prompt as
the app, a stream that runs only while something listens, at the accuracy
the most demanding caller asked for. Every caller keeps the browser API.
A tap on the status bar now scrolls the list in the middle of the screen
back to the top, as in every iOS app.
A switch in the account settings, inside the app on a device that can unlock
that way (iOS LocalAuthentication, Android BiometricPrompt with the screen
lock as fallback). Turning it on asks once, so nobody locks themselves out.
With it on, TREK asks at launch and after half a minute in the background
before showing anything, and the app switcher shows a blurred cover (iOS) or
no snapshot (Android) instead of the screen. The setting stays on the
device, not the account.
WKWebView runs no service worker for a self-hosted server, so on iOS nothing
kept a map tile once it scrolled away. Binaries with the tileCache feature
answer a trek-tile:// scheme from a store on disk: a tile viewed once is
served from there for up to 30 days (Mapbox's limit for a device cache,
applied to every provider), then fetched anew; the store keeps under 250 MB,
dropping the least recently used tiles first, sends a unique app User-Agent,
and never prefetches. Mapbox's own resources (style, tiles, glyphs, sprites)
and Leaflet's raster tiles go through it, telemetry never does.
Offline, a Mapbox map in such an app now stays Mapbox where it was viewed,
and the pill offers the stored trip region instead. The offline settings
show what the store takes against its limit, with a way to empty it.
WebKit's swipe from the left edge slid away a screenshot of the whole page,
closed no sheet and took every drag near the edge from the UI. Binaries from
this one turn it off and report an edgeBack feature, and the UI draws its own:
the page follows the finger, a release past a third or a flick slides it off
and goes back, anything else springs it home, with a light detent.
Open sheets now register in one overlay stack, so a way back closes the one on
top first, through its own close path (unsaved changes still ask): the edge
swipe on iOS and Android's back gesture alike.
The UI ships with the server and the binary through the stores, so the two
drift. TrekShell.getCapabilities now reports the binary's shell API, version
and native features by name (iOS and Android), and the UI turns each native
feature on only where the binary has it; a binary from before the handshake
counts as shell API 1 with none. The other way round, the server names the
oldest binary its UI works with (minNativeShellApi), and an app below it is
asked to update with a link to its store.
The editor's twelve month chips were twelve grey boxes; they sit on one glass
surface now, the picked month under a pill that glides to it across both
rows. The month heads in the year overview are glass pills too.
In the editor the months of the leave year page like the trip's days, the
neighbour sliding in with the finger; in the overview the plan's years do
the same (only between years the plan has, the arrows still create one).
The day pager now pages anything with an id and can sit in a page's flow.
The row of person and legend chips is two compact pills now: People opens a
sheet to pick the member and show or hide shared calendars, Legend opens one
that explains the company days and holiday calendars, which were never a
choice.
Planned was a label with a switch in a pill of its own beside the full-width
wishlist button, and with Dawarich on the row grew cramped. Wishlist,
Dawarich and Planned are now buttons of one shape, equal width and the phone
shell's glass; Planned fills while the planned countries show.
The error and new-version fallback was a bordered box with a warning
triangle and the raw error message on show. It is now one centred card with
a round symbol (a friendly one for an update), a short title and line, and a
full-width reload button; the technical message stays for issue reports,
folded under Details. Phone, PWA and desktop alike.
The German Completed filter (Abgeschlossen) made the dashboard's filter pill
wider than its row, so it scrolled inside its box and showed cut off. It is
now Beendet, the help texts naming the filter follow, and the intrinsic
segments take a little less padding.
The day plan offered Google Maps and CoMaps as two tiles. One Maps button now
opens the day in whichever app the traveller picks: Google Maps, Apple Maps
with the stops between as waypoints, CoMaps, Waze and on Android any other
map app, the last two to the first stop, which they say. The preferred map
app from the settings opens straight away.
The day's actions sit together on one glass surface as round icons over
their labels, and the up-next card wears the same surface as the timeline
card under it.
- The rename and reorder controls beside the day's name, and the day and
weather chips on the timeline card, are now round or pill glass controls of
one height, so they read as tappable; the stay chips follow.
- A leg between two stops is one pill, as on the desktop: the mode, the time
in full weight and the distance beside it, larger and without a dot.
- The timeline flows under the up-next card at a fixed 12px, whatever its
height, instead of an offset that left a gap when the stop has no address.
- The places bar on the map wears the day rail's glass at its 36px height.
- A context menu stays inside all four edges of the screen, measured at its
real size rather than mid-animation, and scrolls when taller than it.
The phone map stacked four controls on the right and one on the left, in
three different looks. Now each side has two at most, all in one frosted
style, 20px off the edge and 8px apart: the whole-trip overview and locate on
the right, the base layer and the Dawarich route on the left, with the compass
appearing only while the map is turned off north.
The map's own places filter button and sheet are gone: the places tab carries
the same filter, and the map's pins follow it. The jump-to-today button beside
the day rail is gone too, since today's chip is already marked.
Also fixes a leftover of the segmented-control change that broke the parse.
Offline, a Mapbox map, another OpenFreeMap style or a raster template the
device cannot serve now switches to the stored OpenFreeMap basemap, light or
dark with the colour scheme, keeping every marker, route and track, with a
small Offline map pill. It switches back once the connection returns.
"Prepare for offline" stores that basemap's region for every trip, whatever
the provider, and both styles' sprites and glyphs, including the scripts a
trip in East or South Asia labels its places in. The vector prefetch keeps to
OpenFreeMap's download rules (four requests at a time, 50 MB per run), raster
tiles are no longer prefetched from OpenStreetMap's own servers, which their
policy forbids, nor inside the iOS app, which has no service worker to serve
them.
A pull on a sheet's content only answered touch, so with a mouse the sheet
could only be dragged by its top edge, and a pull that started on a button
read as a press. Mouse and pen now pull the content like a finger, a pull
that starts on a button swallows its click, and text fields keep their own
gesture.
The active day chip and the segmented controls now glide one shared pill to
the new item on the compositor, instead of repainting two buttons.
The docked sheet could only be pulled by an invisible 20px strip, re-rendered
React on every finger move, and a flick sprang back to the top before fading
out. A finger on the content now carries the sheet once the content is at its
top, the panel moves through direct style writes, and a dismissing release
throws it off screen from where it was let go, faster the harder the flick.
With unsaved changes it springs back and asks instead. Pulling up only
stretches a little.
The docked sheet now rises from below the screen edge and sinks back there,
and the floating card pops in place the way a native alert does.
The day swipe only rubber-banded the current day, then cut to the next one
after a fixed slide and a synchronous re-render of the whole planner. The
previous, current and next day now sit side by side in a horizontal
scroll-snap strip, so the finger drags the neighbour in, and momentum,
edge bounce and axis locking are the browser's own. The day changes once
the strip comes to rest, with a light haptic tick in the app.
The pager is one building block, MDayPager, used by the plan timeline and
the road trip stages alike. Each page renders its own day through the
timeline and stage hooks, which now take the day as a parameter, and keeps
its state when it slides into the middle.
The plan picker no longer shows a yearly badge, so its string reached nothing
and failed the i18n key check; the shell's start also initialised a value it
always overwrote.
The welcome headline and subtext, the demo button, the have-a-server link,
the unlimited-plan line and the claim title changed in en and de with the
redesign. All other locales now say the same, and the English headline
reads Your Trips to match the German plural.
The light scheme dropped the Kyoto band and drew the grid so faint it
vanished, leaving a bare white page. The photo now stays in both schemes
under a veil in the page colour (a dark dusk, or a bright haze in light
mode so the black wordmark still reads), and the light grid and wash are
strong enough to see.
Selecting a place no longer spiderfies a city-sized cluster: the reveal
only fans open bubbles of up to 12 stops, or stacks whose stops all share
one coordinate, so a zoomed-out trip with hundreds of places stays usable
(#2990).
The selected pin now carries a white border inside a pulsing blue ring on
both renderers (static under reduced motion), and the map measures the
place card instead of assuming a fixed 320px, so the selection pans the
pin above even a tall card on portrait tablets and eases it back out when
the card grows over it later (#2991).
Closes#2990Closes#2991
* chore(capacitor): base files for capacitor
* chore(capacitor): base plugins
* feat(auth): hand a browser sign-in over to the native app
The native apps show the server's own UI in a WebView, where neither SSO nor
passkeys can run. They sign in through the system browser instead and need a
way to carry that session back.
The app keeps a random verifier and sends only its sha256 into the browser.
Once signed in there, the browser trades that challenge for a one-time code
(cookie session only, so an API or MCP token cannot mint itself a browser
session), and the WebView redeems the code together with the verifier. The
code rides through a URL scheme any app could claim, so on its own it is
worthless; it is single use and good for a minute.
Also: /api/health/features reports nativeApp so the apps can refuse servers
that predate this, and hashed files under assets/ are served immutable,
which the iOS app needs to start offline without a service worker.
* feat(native): the apps load the server's own UI
The apps ship only an address screen now. Once a server is picked, the
native side points the WebView at it: TrekBridgeViewController on iOS and a
config copy with server.url on Android. The UI then comes from that server,
in the version it runs and on its own origin, so the session cookie,
uploads, /ws and plugins work exactly as in a browser and no app release
has to keep pace with the servers out there.
The address screen (native-shell/, built into native-dist/) probes
/api/health/features through native HTTP, tries https before http when no
scheme was typed, and refuses servers without the nativeApp flag. It also
serves as the error page when the remembered server does not load: retry,
change server, or on iOS the stored copy of the server's start page.
TrekShell is the app's own plugin for all of that: remember, change and
reload the server, sign in through ASWebAuthenticationSession or a Custom
Tab, and on iOS keep and open the offline copy of the start page, because
WKWebView runs no service worker for a domain the app does not list.
Plugins the frame does not use are gone (push, camera, geolocation,
calendar, file transfer, in-app browser and more): the WebView covers the
camera and location itself, and push needs a relay before it is worth
shipping. With them go the push entitlement, the calendar usage strings
and the Firebase bits. The iOS scheme falls back to capacitor://, since
Capacitor drops an iosScheme of https, and the arm64 capability replaces
armv7.
* feat(native): the UI knows when it runs inside the app
Everything the server's UI does differently inside the iOS and Android app
lives under src/native and is inert in a browser.
Sign-in: SSO and the passkey button go through the system browser
(signInThroughBrowser), because identity providers refuse embedded
browsers and WebAuthn needs a domain association no build can have for
every server. The browser end is the new /native-handoff page: it signs
the user in if needed and hands the session over only after an explicit
tap, so a link carrying somebody else's challenge cannot pass a session
on silently. The verifier is hashed with @noble/hashes because
crypto.subtle is missing on the plain-http servers many home networks
use. An OIDC-only install no longer bounces to the IdP inside the
WebView, and passkey management is hidden in the app.
Downloads: the WebView cannot save a blob, so downloadBlob, downloadFile
and openFile write to the app cache and open the share sheet. The eight
places that built their own download anchor now use downloadBlob, which
they should have anyway.
The rest: "Change server" on the login page and in the account settings,
the splash hides after the first frame, Android's back button walks
history before it leaves the app, the status bar follows the scheme, and
on iOS the start page is refreshed and every chunk of the build (listed
in the new asset-manifest.json) is pulled into WebKit's cache so the app
can start without network.
* ci(native): build the Android and iOS shells on every change to them
Compiles the Swift side on a macOS runner so nobody needs a Mac to know it
still builds, and leaves a debug APK on every run to install on a phone.
Unsigned: signing and store uploads need secrets fork PRs do not get.
* feat(native): offline vector maps in the iOS app
In a browser the service worker answers MapLibre from the gl-map-offline
cache that glPrefetcher fills for trips marked for offline. The iOS app has
no service worker, so nothing read that cache there and the map went blank
without network.
A trek-offline:// protocol takes the service worker's place. Inside the iOS
app the MapLibre style URL is rewritten to it, the handler rewrites every
https URL inside the style and its TileJSON the same way, and then serves
documents network first and tiles, glyphs and sprites cache first. Outside
the iOS app nothing changes. The default OpenFreeMap basemap is covered;
Mapbox GL has no protocol hook and stays online-only in the app, as does a
custom raster template, whose opaque responses only a service worker can
store.
* fix(native): keep the iOS offline copy inside the app
loadSimulatedRequest went to the server URL without its trailing slash, and
Capacitor keeps a navigation in the WebView only when it starts with the
bridge's serverURL, slash included. The offline start page would have
opened in Safari instead. Also covers the platform helpers and the
hand-off API module.
* feat(native): the small things that give a WebView away
- A short haptic tap on success, warning and error toasts, which is where
every save and delete already reports. Info toasts stay silent.
- No "‹ › Done" strip above the iOS keyboard.
- No Safari preview when a link is long-pressed.
- The swipe from the left edge goes back on iOS, through the same history
Android's back button walks.
* test(mobile): the backup-code download revokes its URL after the shared delay
The mobile twin of the AccountTab test still expected the object URL to be
revoked in the same tick. Since both screens download through downloadBlob,
the revoke happens after its 100 ms grace period.
* fix(native): remember the login by default inside the app
With the switch off the app got a session cookie, which dies with the app
process, so every time the phone unloaded TREK the user had to sign in
again. Found while running the Android build in the emulator.
* fix(native): no overscroll stretch on Android
The UI asks for no overscroll effect with overscroll-behavior: none on html,
and Chrome honours it. The Android WebView draws the stretch on its root
scroller natively, though, and it scales the whole view, so the top bar and
the bottom dock were pulled along with the page. Found in the emulator.
* fix(native): Android shows "server not reachable" instead of a blank page
Capacitor injects its bridge on Android only into the server's origin once
server.url is set, so the local error page it switches to when the server
does not load had no bridge. It could neither retry nor change the server
and stayed blank.
TrekWebViewClient now catches a failed page load (a network error, or a 5xx
from a proxy while TREK restarts) and the app starts over on the bundled
address screen, where the bridge is present and the screen offers exactly
those two choices. The server config no longer carries errorPath.
Two traps on the way, both found in the emulator: a restart during launch
has to be a fresh task, since recreate() leaves the system splash lying over
the new page; and a restart shows no launch splash at all, because the
splash plugin cannot take it down again when the page hides it early.
* feat(native): TREK's own app icon and launch screen, named just TREK
The shells still carried Capacitor's placeholder icon and splash, the blue
X on a grid. They now show the white TREK mark on a dark, slightly glassy
tile: a soft light from the top left and a sheen over the upper half, in
the app's launch colour #09090b. Android 12+ draws its launch screen on the
same colour instead of light grey.
scripts/generate-app-icons.mjs renders the sources into assets/ from
public/icons/icon-white.svg, and `npm run icons:app` turns them into every
size both platforms ask for through @capacitor/assets. The Android
foreground mark is sized for the adaptive icon's round mask.
The home screen label is TREK: the old "TREK - Plan your trips!" was cut
off under the icon, and the longer title belongs to the store listings.
* fix(native): new lint warnings and the MapLibre protocol type after the dev merge
Awaits that sat in loops in the native-auth tests run together, the hand-off hook marks its navigations as ignored, the native tests drop their non-null assertions, and the offline protocol answers with the data type MapLibre 6.15 expects.
* fix(native): strict types in the address screen, Prettier on the native-auth e2e suite, file-local exports
The address screen's element helper returns the plain element instead of Object.assign's intersection, which strict mode could not represent; the haptics test reads an optional argument as optional; three exports only their own file used are local now.
* fix(native): Prettier on the native client files, rate-limit e2e requests in sequence
Formats the files the client formatting gate listed, and NATIVE-E2E-005 sends its twenty exchange attempts one after another: all at once they reset the test server's connections.
* fix(native): hide passkeys in the app through usePasskeys
The flag in useAccountSettings grew the identical destructuring in both account views past the duplication threshold; the shared passkeys hook already decides when its card is hidden, so the app check lives there now.
* feat(native): everything the App Store and Google Play submissions need
- iOS: microphone usage text for "Take Video", no-export-compliance flag,
ExportOptions.plist for App Store Connect uploads, the unused push
forwarders gone; CI builds with Xcode 26 and lints the property lists.
- Android: version and signing from Gradle properties (unsigned without
them), R8 with keep rules for TREK's and Capacitor's reflective classes,
no Auto Backup of the session cookie, no external file provider path,
template and Firebase leftovers removed, keystores and service accounts
ignored, full-size adaptive icon layers with a monochrome layer; CI also
builds the release APK.
- native-release.yml: signed bundle to a Play testing track and archive to
TestFlight, by hand or on a v* tag, behind the store environment.
- Version handshake: the server reports nativeApi 1, the app refuses a
server below its minimum and tolerates an app older than the offline copy.
- Address screen: privacy policy link, "Try the demo" and "What is a TREK
server?"; the privacy link also in the app's server settings. Three new
native strings in all 27 locales.
- Donation links and the notices asking for one are hidden in the app.
- Store listings for "TREK Travel Planner" in client/fastlane: texts, review
notes with placeholders, privacy and account deletion page drafts, the
feature graphic and icon, a length and image check in CI, and a Playwright
store project with the scripts that turn its captures into screenshots.
- Docs: wiki/Native-Apps.md, README no longer says there is no store app.
* fix(native): read a notice button's href without a weak type
An action button carries no href, so a parameter typed as only an optional href refused the notice DTO; the button type now names its kind as well.
* feat(native): a welcoming onboarding for the address screen
The app's own screen now opens on a welcome: the TREK mark, a line about what
TREK is, and two clear paths — connect the TREK you already open in the browser,
or get TREK Cloud. Both lead to polished screens: the address field with live
validation, a shake and haptic on a bad address and a success tick before the
server takes over; and a TREK Cloud intro with what it offers and a way in.
Screens slide and stagger in, buttons give a tap, and it all respects the safe
areas, light and dark, and reduced motion. Self-hosters are unaffected: cloud is
its own path, never in the way. New native strings are added across all locales
(English placeholders outside en/de, to translate).
* feat(native): a plan picker for TREK Cloud, with live prices
The Get TREK Cloud path now leads to a plan picker that reads the live plans
from the site (/api/plans), so the prices and quotas always match the website.
Monthly and yearly in a segmented toggle, the year marked with its saving, each
plan a card with its price and what it includes. Subscribe is wired to finish on
the web order for now; it becomes an Apple/Google in-app purchase once the store
products exist. New strings across all locales (en and de translated).
* feat(native): light status bar and a Go key on the address screen
The address screen sits on the near-black surface, so set the status bar to draw
its clock and icons in light while it is up; the server's own UI takes over the
style once it loads. The address field's on-screen return key now reads Go. Both
fail quietly where there is no native status bar (web, the simulator).
* i18n(native): translate the onboarding into es, fr, it, nl and pl
The welcome, TREK Cloud and plan-picker strings were English placeholders in
every locale but en and de. The five largest remaining European languages are
now translated in full; the others still carry the English text and are a
follow-up for native speakers.
* feat(native): the in-app TREK Cloud subscription flow
After picking a plan, the app now carries the whole purchase itself when the build
can take one: choose your address (checked for free as you type, against the same
endpoint the website uses), your email, and subscribe. The receipt goes to
/api/iap/redeem, and a Setting up your TREK screen waits on the new address and
loads it the moment it answers.
The purchase sits behind an IapClient seam (src/native/iap.ts): a fake drives the
whole flow in development and preview, the native client reports itself
unavailable until the store plugin and products exist, and until then the plan
picker still finishes on the web, so nothing half-wired faces a customer. New
claim and provisioning strings across all locales (en, de and the five largest
European languages translated).
* i18n(native): translate the app into the remaining 20 languages
The onboarding, plan picker, claim and provisioning strings were English
placeholders outside en, de and the five largest European languages. They are
now translated in full across the other 20 locales (ar, az, br, ca, cs, et, gr,
hu, id, ja, ko, ru, sk, sv, th, tr, uk, vi, zh, zh-TW), placeholders and brand
kept. The app now speaks every language TREK does.
* fix(i18n): add the Hebrew native strings, and correct the LoginPage size baseline
The he locale had every domain but native.ts, which broke i18n parity, the shared
coverage specs and the client parity test once the native strings existed. Add
he/native.ts with all native keys translated. Also correct the LoginPage size
baseline to its current line count.
* i18n(native): excuse the TREK Cloud brand, mark Email same-as-en, exempt the quota counts
The strict i18n checks, which the feature-branch CI does not run, catch three
things now that the native strings exist in every locale: TREK Cloud read as an
untranslated English phrase (it is a product name, like Vacay and Atlas, so add
it to the brand list); Email left identical in Italian, Indonesian and
Vietnamese, where it is the local word too (mark same-as-en) and in Greek (give
it a Greek term); and the GB and AI-document quota strings carrying {n} without
being plural groups (both are figures that never show a singular, like the other
unit-symbol counts in NOT_PLURAL).
* fix(native): typecheck, lint and knip on the native code; wire he's native bundle
The full suite, which the feature-branch CI does not run, catches four things the
native work still had: a non-narrowed union in the claim flow (tsc), a helper
named useFake that the react-hooks rule read as a hook (eslint), two iap types
exported but used only inside the file (knip), and the Hebrew native strings
present as a file but never spread into he's merged bundle (the aggregated key
parity test). All four fixed.
* fix(native): a flat redeem result, so the claim flow typechecks
The discriminated union did not narrow under the native-shell tsconfig, so the
failure code read as absent on the whole type. Use a flat result with optional
instanceUrl and code instead: both are always present on the type, no narrowing
needed, and the ok plus instanceUrl guard still routes success from refusal.
* style: prettier-format the native and i18n changes
The feature-branch CI does not run the formatting check, so the conflict-resolved
files (AboutTab, useAccountSettings, useTripExport), the native shell and iap
module, and the i18n files touched by the translation scripts were never run
through Prettier. Format them all; no behaviour changes.
* chore(native): re-trigger CI on the formatted tree
---------
Co-authored-by: Maurice <mauriceboe@icloud.com>
The field-to-column map of shared place details moves to @trek/shared,
where places and place-contrib both read it without depending on each
other, and the assistant reads a recorded call's arguments with a helper
of its own instead of reaching into the AI providers. The catch-up's
country from the address is taken back out: it made place-contrib depend
on atlas, which already depends on it, and it waits for a shared country
lookup both can use.
The packing row tests read the TREK tooltip and the accessible name instead
of the title the badges no longer carry, the journey delete suite takes its
service from the module over a snapshot db and seeds through the row
factories, the storage notifier suite builds its subject once, the POI
layer follows the view only while it has POIs, the collections repository
reads the list's places through its own query, and the baselines the moved
code left smaller come down with it.
Sending back every change called before its group was loaded cost a step
even when the model had the arguments right, and a valid change was no
longer held as the card the user decides on. A blind change is now held as
always when its arguments are of the shape the schema asks for, and sent
back with the schema loaded only when one is missing or of the wrong kind,
such as a number or a list written as text (#2967).
Eight files had grown past the line counts their ratchets hold, so code
moves to where it belongs instead of the baselines going up:
- a settle-up transfer is booked by applyTransfer in the money helpers
- whether an expense still carries its derived category is
isDerivedCostCategory in @trek/shared, next to the category map
- a list is deleted with the images it named by deleteWithImages
- the explore POIs and their world-copy tracking are PoiMarkers
- the journey delete tests get a file of their own, and two stale
migration-era comments in the journey service shrink to one line each
The bringers list moves next to the other bringer helpers so the card file
exports components only, three tests read without non-null assertions or
any, and three files take the formatting their checks expect.
The row's badges said who brings an item, who it is shared with and who
takes care of it in the browser's own tooltip. They now use TREK's, and
the bringers one lists everyone with their picture on a line of their
own. The others show as a +N circle tucked under the first picture.
Tooltip takes an optional content for tips that hold more than words.
Road trip day tiles still wrote numeric dates from the clock setting and
dates in the medium style kept the language's numbers. A poll deadline
whose time was cleared was read as UTC midnight.
A display-currency payment whose halves pointed opposite ways no longer
stayed settled (#1445), the plan sidebar and the PDF total still went
through the trip currency, and a zero-decimal display currency was not
rounded. Every booking type now maps to one Costs key on every path; a
cost the old map filed under 'other' still moves with a type change.
The waypoints switch showed in every Tours import dialog; it now shows in
the trip's import only. Confirming it when every point was already in the
trip closed the dialog without a word.
The catch-up read its state outside its transaction and could revive a
backfill a withdraw had just ended; switching a field off and on after a
withdrawal shared the withdrawn history again. The catch-up now resolves
the country the way a save does and reads one field map.
Deleting a journey left its uploaded photos behind for good, and deleting
a list queued neither its cover nor its place images. An object with a key
the store refuses no longer fails every night's pass, and the admin view
keeps the counts of a category skipped for a running backup.
The zoned flights traded clock times among themselves while stops kept
theirs, so a dinner timed after a flight landed could be drawn before it.
A timed stop near one of the day's zoned endpoints now joins the swap in
that endpoint's zone.
A write from a group the turn had not loaded was held as a change card with
schema-less arguments and failed on apply (#2967); it now goes back to the
model with the group loaded. Calls answered as not run are now listed on
the assistant turn too, so a thread replayed on another provider has no
orphan tool results. The blind-call hint only goes with the blind call.
The fitting ran in a layout effect and measured chips the row had already
squeezed, so a row that got wider kept its "+N". Every label now stays in
the row, the ones without room kept out of sight, and the observer that
sees the row change size fits them from each chip's full width. The name
takes the width it needs on a wide list and the labels fill the rest.
The first fix let a row wrap freely, which on a list beside the map put
the status on a line of its own and broke every row differently. A row is
now laid out the same way each time: on a list narrower than 760px the
name, address and status take the first line and the category and labels
one line below; a wider list keeps them on one line. The row shows as
many labels as it has room for, category first, and a "+N" chip names
the rest in its tooltip, so a wide screen shows every label. Chips that
cut a name short say it in full in their tooltip.
The label strip over the map hid the labels that did not fit behind a
scroll with no scrollbar. It now folds into a "Labels" button that opens
them as a list when the bar has no room for all of them, and unfolds
again when the room is back. The filter row stays on one line on a
desktop, a dropdown at its default keeping only its icon when narrow, and
the select bar's actions keep their icons with the words in a tooltip
when the bar is too narrow for them.
A place row was a three-column grid whose last column held the labels,
the category and the status, none of which could shrink. With two or
three long labels the name got what was left, often one letter or
nothing, and the status pill ran out of the card.
The row now wraps: the name keeps at least 11rem, and when the labels,
category and status do not fit beside it they go to a line of their own
under it and wrap there. A label or category too long for its chip ends
in an ellipsis instead of being cut mid-letter. Kept tours use the same
row and get the same layout.
Closes#2981
A booking dropped against another booking's time went back to the nearest
slot its own time allowed. With a 06:30 flight listed below the day's
untimed stops, every slot among them was against the flight, so a 13:00
transit dropped between two stops went back behind the flight on every
drop, and the flight's legs cannot be dragged out of the way.
The dropped row now stays where it was dropped and the other timed
bookings it puts out of order move next to it, the earlier ones ahead of
it and the later ones behind. Timed stops and notes never move: against
one of those the row is still seated by its time as before (#2631). The
phone's move applies the same rule before it refuses a move.
Closes#2983
Array.prototype.at is past the client's lib target, so the journey open-at
rule indexes the array. The packing row test expects the title that names
every bringer, a carousel test acts without a floating promise, and the
lint baseline drops to the counts this branch reaches.
A common item brought by several members showed the first bringer's avatar
and a bare "+2". The tooltip only named the first person, and on the phone
neither the row nor the item sheet said who else brings it (discussion
#2782).
packingBringerNames lists the bringer and every contributor; the desktop
row, the phone row and the phone item sheet use it for "brought by".
The public journey page on a phone shows the map with a card carousel, which
always opened on the first entry. The owner's "newest first" choice for the
link was ignored there, so readers following a running trip had to swipe
through every day to reach the latest one (discussion #2851).
The rule for where a journey opens moves out of the journal hook into
journeyOpenAtEntryId, which the public page now uses as well: a newest-first
link opens on the latest day, otherwise a running journey opens on today as
the journal already does.
The carousel also settles on the card at the end it is scrolled to. The last
card can never reach the centre, so opening on the latest day used to
activate the day before it, in the journal as well.
An assistant writing expenses through the MCP was told "e.g. Accommodation, Food,
Transport" and stored exactly that, so "Food" and "food" sat side by side in the
MCP totals while the Costs tab already read both as one. A typed category in any
case, or a common name like Hotel or Flight, is now stored as its fixed key on
every path (REST, MCP, linked booking costs), the totals the MCP reads group by
that key, and the tools name the real keys. The client reads categories through
the same shared mapping. Any other text is still kept as given.
The journal tools said that a private entry stays off the share link. It never
did: the share link shows every entry that is not a draft, whatever its
visibility, and the web editor has no visibility at all. The tools now say so
and point to is_draft, which is what keeps an entry to the contributors.
The wiki explains why the location button needs HTTPS and how to run TREK behind
an HTTPS tunnel while still using it over plain HTTP at home, and that Android's
photo picker drops a photo's location while Immich and Synology keep it.
With a room kept for the whole trip and a second stay for a night away, the day
after the side trip took the first stay in the list as the hotel it woke up in.
The Days view then drew a drive from the kept room, often hundreds of kilometres
away, to the side trip's first stop or airport, next to the flight back. The
morning now belongs to the stay checked into last, and the evening to the stay
that runs on past the day. The map also passes a booking's own point at the edge
of the day, as the sidebar already did, so a taxi leaving from the hotel adds no
zero-length road from the hotel to itself.
The Tour booking type read like the Tours tab, and a booked food tour left people
looking at an empty Tours tab. It is now called Guided tour in every language,
the empty Tours tab says that booked tours go under Bookings, and the help
center and wiki use the new name. The stored type stays 'tour'.
Moves the Leaflet map's event bridges (click, context menu, viewport, POI drop
and the longitude-frame follower) out of MapView into leafletMapBridges, and
shares one POI drop target between the Leaflet and GL maps, so both map files
shrink instead of growing. The repair migration does its arithmetic in one
UPDATE per table, and the new tests drop their non-null assertions and loops.
A right-click, click or drag on a repeated world copy reported the longitude it
has on that copy (274.9 for -85.1). The place was saved there, the geocoder
found nothing, and a route to a place on the real world ran across the whole
map (#2976). Longitudes are now brought back onto [-180, 180] where the map
hands them out (Leaflet and GL), where they are stored (places, booking
endpoints, collection places, so REST, MCP, plugin RPC and importers alike)
and before reverse geocoding. A migration moves the ones already stored.
Leaflet draws a layer at its literal longitude only, so a trip from Los
Angeles to Auckland had its pins at opposite ends of the world while the
flight ran off the edge to the next copy (#2975). A trip that crosses the
antimeridian is now drawn in one longitude frame, seamed in the widest
stretch it leaves empty, so pins, routes, flights and every overlay lie on
one copy and the map opens across the Pacific. Every other trip has no frame
and draws exactly as before. The GL day fit also takes the short way.
A new install now starts with Journey and Collections switched on, next to
the trip addons (all but Road trip), Atlas, Vacay and the Naver import that
already were. Existing installs keep their addons as they are.
Closes#2975Closes#2976
The admin reset link is a component of its own instead of one made during render, the version line of the plugin panels moved into small modules, the new settings rows hand their save on as void, the date picker takes an accessible name for fields without a visible label (the Vacay hire date), the Vacay and poll tests type through the shared picker, and the theme, size and lint baselines are lowered to what the files hold now.
Every date in TREK was written in the interface language's format, and
English is en-US, so an English interface always showed and read
mm/dd/yyyy with no way to change it. A few native date inputs followed
the browser instead, so one screen could mix two formats.
date_format is now a user setting with an admin default, next to the
time format: automatic (the language's format, as before), YYYY-MM-DD,
YYYY/MM/DD, DD.MM.YYYY, DD/MM/YYYY or MM/DD/YYYY. The date picker shows,
pre-fills, hints and reads typed dates in it, and every date written in
numbers across the app (lists, cards, admin panels, CSV and PDF exports)
goes through one formatter. Dates written with a month name keep the
language's wording. The remaining native date inputs are the shared
picker now. The display settings MCP tool reads and writes the key.
Closes#2971
The display-currency check is its own helper (inDisplayCurrency), used by the Costs model and the shared trip page. The tests that pinned #2525's reading of a display-currency bill through the trip currency now pin the as-typed reading, which the balances use too; read in the trip currency, such a bill and its repayment may stay a euro cent apart.
On a yen trip read in euros, an expense entered in euros was booked through the yen at the rate frozen on entry and shown back at today's rate, so 1,081.09 EUR read 1,078.86 EUR a day later, and repaying exactly what was typed left 2.23 EUR owed.
Rows in the display currency now net apart in whole display cents, as entered, and the lists show them as typed. A payment in the display currency first pays what is owed between the two in such rows; only the rest goes through the trip currency at its frozen rate, so a trip-currency debt settled in the display currency stays settled when rates move (#1445). Every other row still nets in trip cents and converts once (#1382), so the balances add up to zero in both halves.
Closes#2965
Refusing every call to a tool of a group not loaded in the turn also stopped calls whose arguments were fine. The call now runs as before; its group is loaded at once, so the next step offers the schema, and when the arguments fail the MCP check the error tells the model the schema is there now and to send the call again.
The model could call a tool it knew from an earlier turn without loading its group in this one. TREK ran it, but the model had never seen its schema in the turn, so numbers, arrays and objects came as strings and required arguments were missing, and the MCP check refused it with -32602 (mostly the Tours tools).
Such a call no longer runs: its group is loaded for the rest of the turn and the call is answered with a tool error asking for it again with the schema the next step now offers.
Closes#2967
The places toast is a plural group in every language, the points import goes through tourRepo instead of src/api, the hook and dialog keep their complexity, and the tests that checked the old confirm call or the GPX left-out count follow the new behaviour.
With the Tours addon on, a file picked or dropped on the places column goes
into the Tours import as soon as it holds one line, and that import reads
lines only. A Sygic or Trip-o-matic export (a route plus its stops) came in
as one tour and every place was dropped without a word.
The preview now counts a GPX file's waypoints as it already counted a KML
file's pins, and the dialog offers to add those points as places of the
trip, on by default. Confirming runs the Tours import for the lines and the
places import (points only) for the same files. A file whose tracks the
trip already holds can still bring its points.
Closes#2969
An iPhone or iPad running TREK from the Home Screen ignores print(), in
the page and in an iframe alike, so "Save as PDF" in the preview did
nothing there. Mobile Safari prints the same preview fine.
In the installed iOS app the preview now shows "Open in Safari" in
place of the dead button, with a line saying why. The link opens the
trip with ?export=pdf, and both shells open their export dialog or
sheet for it, so the PDF is one tap away once Safari is up. The
parameter is dropped once read and survives the login redirect.
Everywhere else the preview prints as before.
The preview overlay moves out of TripPDF.tsx into pdfPreview.ts, which
TripPDF had no room left for under lint:size, and its backdrop and
stacking now use the --overlay and --z-modal tokens.
- MAdminOidcCard is markup over the settings section's state, so it is
listed under mobileOnly with that reason.
- The default user settings views take their presets from
useDefaultUserSettings now and import only a type from mapSettingsModel,
so that pair names the map settings views alone.
- The phone's OIDC card moves into MAdminOidcCard, so the settings section
is back within its complexity limit after the managed-mode conditions.
- enabledPhotoProviderIds reads the enabled ids straight from the
repository and has a test of its own, which brings the addons coverage
back to its threshold.
- The owner sign-in and the photo provider the suite runs with move into
tests/helpers/journeyDetail.ts, so the page test is back under its line
baseline, which drops to 4028.
- Drop help/registry.base.ts, an unused copy of the help registry whose key
builders the i18n key check could not resolve.
- enabledPhotoProviderIds skips rows without an id instead of widening the type.
- The listed events read each addon once, in parallel, not in a loop.
- The narrowing test reuses the suite's service and swaps the addon answers.
- PlaceContribPanel keeps its complexity by naming the standalone check.
On a managed install the operator runs place search and routing, the plugin
runtime, the integrations and the documentation, so the admin and the users
no longer see switches and guides for them:
- The place search and routing card, the transit provider and the search log
are gone from the admin settings (desktop and mobile). places_provider is a
managed-locked setting, and the transit provider, search log and wanderer
routes refuse writes there. Share place details stays the admin's, as a card
of its own, hidden when TREK_PLACES_CONTRIBUTE switches it off.
- File types and trip reminders move to the right column, so the columns stay
balanced.
- Plugins disappear from the admin and the user settings, wanderer from Tours.
- The Help & Docs page and its links are not offered. The help center stays and
leaves out guides about what the operator runs (keys, OIDC, backups, storage,
releases, plugins, document sync) and about integrations that are off.
- Mapbox GL is offered only with the operator's token, Amap tiles only where
the operator set Amap up.
- The mobile admin no longer shows OIDC and the WebAuthn relying party fields,
matching the desktop tab.
For every install:
- Notification matrices list only events the install can raise: no plugin row
without the plugin runtime, no Vacay, collection, photo, Synology, wanderer or
packing rows while those are off. On a managed install the release and
replica alerts are left out.
- The journey entry editor offers External photos only where a photo provider
is switched on, as the phone already did.
- app-config is asked again after a failed answer. A page opened while the
server restarted kept every default for the session, which on a managed
install brought back the operator's settings.
- sharp 0.35.5 (librsvg CVE), with the musl pins that ship in the image.
- follow-redirects 1.16.2, postcss-selector-parser 6.1.5 and
serialize-javascript 7.1.3, from the open advisories.
- GitHub Actions: github-script 9 and the minor-and-patch group.
- plugin-sdk: @types/update-notifier 6.0.9.
The grouped npm update is left out: with it the server build runs out of
memory in tsc, and the culprit is not found yet.
- The npm minor-and-patch group (#2952) without mapbox-gl: 3.34 drops
setRTLTextPlugin from its typings, and the RTL plugin is what joins Arabic,
Hebrew and Persian labels. react and react-dom move to 19.3.1 together,
overrides included, so the workspace keeps one React.
- Security floors from the open alerts: sharp 0.35.5, vite 8.3.5,
follow-redirects 1.16.2 (#2927), postcss-selector-parser 6.1.5 (#2865) and
serialize-javascript 7.1.3.
- GitHub Actions: github-script 9 (#2951) and the minor-and-patch group (#2950).
- plugin-sdk: @types/update-notifier 6.0.9 (#2948).
An installation now asks for its TREK services key a few minutes after
it starts, and the daily check registers a missing key as well as
renewing a due one. Until now nothing happened before someone searched
for a place or opened the admin settings, so "Share place details"
stayed unavailable on a fresh or updated instance.
When an admin agrees to share place details, what users typed into
places since TREK 5.1.0 goes out too, not only what they type from then
on. The origin of every field is recorded since 5.1.0 whatever the
decision, so a catch-up walks those places page by page before each
send run: only values that are still the ones a user typed, only for
fields switched on, only for places the service can key, and only
while the queue holds fewer than 200 values. It goes on where it
stopped after a restart, never replaces a newer value a save queued,
and ends on switching off, declining or withdrawing everything. A field
switched on later catches up as well, under an agreement given with the
new prompt sentence that says so; an agreement from 5.1.0 does not.
Queuing moves into PlaceContribQueue so a save and the catch-up take
the same path.
Node gives each address of a dual-stack name 250 ms before it moves on.
From Australia that is shorter than one round trip to Europe, so on a
network without IPv6 the IPv4 handshake to places.liketrek.com was cut
off, the IPv6 attempt failed at once and every request ended in
ETIMEDOUT. The installation never got its services key, sharing place
details stayed unavailable and searches fell back to OpenStreetMap.
Raise the per-address attempt timeout to 2 s at startup, keeping a
larger value an operator already set, and log the reason behind
undici's bare "fetch failed" for each address tried.
Chrome on Android increasingly hands out push subscriptions on
jmt17.google.com, which the push service allowlist did not know, so
turning push on failed with a 400. The host is added as an exact entry,
next to the two FCM hosts.
Closes#2939
Replacing a cover or an avatar and deleting a note file now hand the old
bytes to the release queue instead of deleting them in the request; the
integration tests check the queue row and that the bytes are still
there. Release mocks answer with a count, the reclaim e2e sends its
refused bodies one after another, and a stale-count reload that fails is
told to the admin instead of swallowed.
While a restore's reset is still pending, or the database names another
installation, the nightly pass judges and reports but deletes nothing,
whatever the restored switch says.
A restore of this installation's own backup now resets it too, through the .trek-reclaim-reset file. Files dropped from the release queue after 30 days, or refused by a full queue, are only removed by the nightly cleanup once deletion is on.
- A restore (admin panel or RESTORE_FROM_BACKUP) now leaves
data/.trek-reclaim-reset before it swaps the database in. Boot reconcile
resets the reclaim state when it finds the file, whatever install id the
database names, and removes it afterwards; a successful reopen removes it
too. A restart after a failed reopen, or a boot restore of this
installation's own backup, no longer keeps the old switch, marks and
confirmations. The id mismatch reset stays.
- The release job skips a tick (install_mismatch) while the database names
another installation or a reset is still pending.
- An allowed owner verdict is cached for 15 minutes, a refusal for 26 hours.
The local driver reads a list prefix as a folder, so the category root is
still listed whole, at most once per category per window.
- Each tick reads every allowed category's due rows separately, an equal
share of the 200 each, so a category whose deletes keep failing cannot
starve the others.
- Rows dropped after 30 days overdue and releases refused by a full queue
are logged at warn with their counts.
Confirm sends the count the admin saw and waits for its answer; a stale
count reloads the state. A running backup is named as a skip reason, a
missing installation id and a stale count as errors, and the switch's
hint no longer promises removal right away: deleted files go a day
later, once no undo can bring them back. Strings in all 28 languages.
Covers what a clone must do (its own data dir, or delete .trek-install-id),
that an older TREK writes no owner marker and cannot be detected, and the
new backup_running reason.
A delete or a replace no longer deletes inside the request. The owner's call
queues the keys in storage_release_queue, due after RELEASE_DELAY_MS (26 h:
the assistant's 24 h approval window plus a margin), and StorageReleaseJob
takes the due rows every ten minutes, at most 200 a tick and 50 keys a
question, asks every reference source again and deletes what none names.
It refuses what the nightly pass refuses (StorageReclaimGuardService: the
ledger, the required sources, a shared location, another installation's
owner marker, cached per category), and waits while a restore or a backup
runs or another storage job holds the slot. An assistant undo or a late
write that names the old value again keeps its object.
The install id lives in data/.trek-install-id, outside any backup and never
written by a restore. At boot and after every reopen the database is
reconciled with it: a database naming another installation (a boot restore,
a restart after a failed restore, a copied travel.db) loses its switch,
confirmations, reports, marks and queue before anything reads them.
A confirmation now stores the count the admin was shown and when; it covers
up to one and a half times that count (at least 100 more) for 14 days, is
used up once the share is back under a fifth, and a count the last report
no longer holds is refused as reclaim_confirm_stale. The admin state folds
a live confirmation in, so the prompt goes at once.
A pass reads each source once for every category and drops each category's
keys once judged; right before a batch it reads the category's sources
again, whole and in memory, instead of a LIKE per key. A narrowed read holds
at most 50 LIKE terms. A pass held off by a backup reports backup_running,
and a category the ledger never reclaims reports not_reclaimed even after a
stop.
The confirm request now carries the unreferenced count the admin saw, so the
server can refuse a count its last report no longer holds. A pass that found
objects due while a backup read the uploads reports backup_running for them.
A trip cover is bounded to 2048 characters on REST, RPC and the MCP tool, as a
place image already is, since a replaced cover is queued for release by every
upload path it spells out.
The longest-run helper typed its reduce as never, reclaimView grew past
the complexity limit (one function per category line now), and a test
read a promise as a boolean.
The panels follow the reworked reclaim state: Check now starts a check
in the background and the panel polls until it is done, a category with
an unusually large share of unused files offers Confirm, Freed comes
from the last run that deleted, a stopped check says why and no longer
reads as nothing found, the switch waits for its answer and a poll that
was in flight cannot flip it back, and refusals name their reason in
the operator's language. Strings updated in all 28 languages.
Leaving every replaced cover, place image and deleted journey photo to the
nightly pass kept them publicly fetchable by URL for a week at least, and
for good while deletion was off. Owners now hand the values a delete or a
replace left behind to StorageReleaseService once their write committed.
It asks every registered reference source, with the same parser as the
nightly pass, and deletes at once what no row names any more: an object a
trip copy, a journey or a saved place still shows stays, which the old
inline deletes got wrong.
The release holds the pass's guards (reclaimable categories at their root
only, never an owner marker, never while a restore runs, never with a
required source missing or an overlapping location), never fails the
request and logs at debug. It runs for trip covers (update, upload, trip
delete), collection covers and saved-place images, place images (update,
bulk update, delete), journey covers and journey or gallery photos
(their thumbnail goes with a file that went), avatars (change, delete,
account deletion), trip files (permanent delete, empty trash) and collab
note and chat attachments. A trip file row now goes before its bytes, so a
failing backend no longer fails the delete. The objects of a deleted trip's
places and files are left to the nightly pass: their rows are other
domains'.
PlacesService drops the collection places repository nothing read any
more. Size, warning and boundary baselines follow the shrunk files.
The 20 % brake skipped a category and wiped its marks every night once a
fifth of it looked unreferenced. Any user could push a small install over
that line, a backlog from #2924 started braked, and a second install on the
same bucket below the line was not caught at all. It is gone:
- Each pass leaves a trek-owner-<install id> marker at the root of every
category it judges. A category holding another install's marker is
skipped as foreign_install, one where the marker cannot be written as
owner_unknown. Backups leave the markers out.
- The install id lives in app_settings outside the storage.* prefix. The
switch is stored as that id, so a database from another install reads as
off, and every reopen (a restore) writes this install's id back, turns
deletion off and drops the marks, confirmations and reports.
- More than a fifth unreferenced still marks and reports, but deletes only
after an admin confirmed the count shown (POST .../reclaim/confirm). A
larger count later waits again.
A skip no longer touches the marks, so a failed listing or lookup does not
restart the week. Check now runs in the background and answers 202; the
admin state says running and keeps the last deleting report apart, so a
report-only check never resets what was freed. A pass that deletes writes a
system audit entry, a pass that stops or fails keeps its partial counts and
marks the rest interrupted, a backend error stops deleting in its category,
and deleting stops for the rest of the pass once the switch is off or a
backup is reading the uploads. The usage scan waits for a pass still
running.
Sources read each category once per pass and answer the pre-delete
re-check with a query narrowed to the keys' plain runs. Bare names keep
their raw spelling (a '#', '?' or trailing blank is part of an upload's
name), and local roots are resolved through their symlinks before the
overlap check. A pass lists only a category's root.
The two admin routes go on the managed allow list, without which buildApp
refused to start. The reclaim calls move to their own api module so
client.ts stays at its size, the phone card sizes its text through the
type tiers, and the tests drop a hand-built service, a non-null assertion
and a promise read as a boolean.
Desktop and phone show what the last check found (can be freed, unused for
less than a week, freed), why a category was skipped, the switch that lets
the nightly run delete and a report-only check on demand, through the
shared storage panel hook. All 28 languages, and the Admin-Storage wiki
page explains the check, the week's wait and when to leave it off.
The last question to the sources now comes after the object's own check,
so nothing slower than the delete sits between the answer and the delete,
and it is asked once per batch instead of once per object: a pass read
every source's columns for each object it was about to remove. The
collection reads the removed inline reclaims used are gone with them.
A restore wipes and rehydrates the uploads under a swapped database, but it
never took the storage job slot, so a reclaim pass could start or carry on
beside it, and a delete already past the generation check could land on an
object the restore had just put back. BackupService now marks a restore as
running on StorageJobsService and waits for a pass in flight to let go; a
pass refuses to start while a restore runs and stops at its next check when
one begins. Backfills and migrations are not affected. The service test also
reads its fake source sets through a helper instead of non-null assertions.
The shared contract made reclaim a required field, so every client fixture
typed StorageAdminState stopped compiling. It is nullable now, as planned,
and the five base fixtures carry reclaim: null.
The reclaim e2e and the source-registration check mocked src/db/database,
which lint:test-mocks holds at zero for new files; both now boot buildApp
against the worker's snapshot db. The key parser drops its non-null
assertions (the indexed reads are typed string already), and the e2e sends
its refused bodies at once instead of awaiting inside a loop, so the eslint
ratchet stays where it was.
Replacing a trip or list cover, changing or deleting a place or saved-place
image and deleting a journey photo removed the bytes at once, without asking
whether a trip copy, a journey, a saved place or another photo row still
named them (#2924). Those paths now leave the object in place and the reclaim
job removes it once no row names it. A journey photo the user deletes takes
its local photo row with it when nothing else links it, so its bytes become
reclaimable instead of being kept by a row nobody sees.
A place image copied from a trip file is still dropped at once when the
update that would name it fails: that copy was never anyone else's.
Size and warning baselines follow the shrunk files.
Trips, collections, places, the contribution queue, journeys, photos, trip
files and users each register a reference source from their own domain. A
source reads its own columns and parses every spelling they hold through the
storage helpers, and the free-form columns claim every reclaimable category,
since a cover or an image can point into any of them. The contribution queue
joins the ledger as its own source, because place-contrib owns that table.
Domains whose rows name stored objects register a reference source with
StorageReclaimRegistryService; storage never imports them. A ledger in
storage decides per category whether it is reclaimed at all and which
sources must have answered first, and a category whose resolved location
overlaps another one is left alone.
A nightly pass at 04:05 lists the root level of files, journey, covers,
avatars and places, skips objects younger than a day or without a
modification time, asks every source in pages of 500 and marks the rest.
Only with deletion switched on, and only for marks a week old, it asks every
source about the single key again and deletes through StorageService. A
category is skipped and its marks start over when a source is missing or
fails, the listing fails, or more than a fifth of its objects look
unreferenced. A pass shares the one-job slot with backfills and migrations,
never runs on a demo instance and stops when the database is reopened.
Admins switch deletion and run a pass on demand under
/api/admin/storage/reclaim; both are audited, and GET /api/admin/storage
carries the switch and the last report. Deletion is off after install and
upgrade.
storage_reclaim_marks records since when the reclaim pass has seen an object
unreferenced, keyed by category and key. A mark is never moved forward, so
its age is how long the object has been unreferenced without a break; it goes
once the object is named again or gone. Owned by the storage domain.
The admin state now carries the reclaim switch and the last pass's report.
A report covers every served category, the legacy photos folder included,
with the objects a pass looked at, the unreferenced, eligible and deleted
counts and bytes, and the reason a category was skipped.
With "date first in day headings" a long day title squeezed the date in
the day list to "Th…" (#2920); the name now gives way and the date stays
whole. The day panel and the mobile day sheet kept the rename pencil on
the heading, which then is the date, and the input replaced it (#2921);
pencil and input now go with the title line under the date.
The ownership check still read every inode of the volumes on each start.
It now looks at the dirs and what sits directly in them, hands a volume
over once when it is not node's, and logs a warning instead of walking a
volume that refuses chown (NFS with root_squash).
The entrypoint no longer runs chown -R over data and uploads on every start;
it only hands over files that are not node's yet, and logs when it does.
Started as a non-root user, it skips chown and gosu and runs Node directly.
The chart exposes podSecurityContext and securityContext for that setup.
The checks behind a drag and the up/down buttons read each booking's raw
clock, so on a day whose flights were put in the order they happen every
move of a timed stop asked to change its time (desktop) or was refused
(mobile). They now read the same key the day is ordered by.
A booking's time is local to its own airport, so a flight leaving
Auckland at 20:20 and landing in Los Angeles at 12:25 the same date
was drawn below the 17:00 flight on from Los Angeles. Bookings whose
dated time has a known zone now take their clock slots in real order.
A step stopped while its calls ran (the user left the page) kept the
provider's whole record of it but answers only for the calls that ran, so
Anthropic and the Responses API refused that thread with HTTP 400 from then
on. Replaying the thread now answers every recorded call that never ran,
which also mends threads already stuck that way.
Claude Sonnet 5.5 at times ends a turn on suggest_follow_ups alone, without
writing its answer, and the turn failed as empty_answer. That step now goes
back to the model whole, its follow-ups answered that the user saw nothing
yet, and their chips come under the answer it then writes. The prompts no
longer say the call goes alone in its step.
A fix to a release workflow changes nothing that the tests cover, yet the
gate wanted a full new test run on it. With ci_from the release takes ci-ok
from the commit main's code comes from, provided that commit is an ancestor
of main, everything outside .github/ is identical, and one of its ci-ok runs
is green.
TypeScript 6 includes no @types unless a file imports from them, and the
issued dialect imports nothing from node. Type-checked on its own it did
not know Buffer, so the 5.1.0 release threw away a valid dialect as one
that does not type-check. The check config names node now.
The four cases added with the 5.1.0 popup patched the fixture through
n.release! and read the support card through card!, six warnings over the
tests baseline of the eslint ratchet. They go through a withRelease()
helper and optional chaining instead.
The release popup now shows 5.1.0: place search, shared place details
and faster routing as three cards side by side, each with its own
drawing. The support band is set as a card like the promise above it
(optional supportLabelKey; a release without it keeps the plain line),
and the two support buttons get the same glass look in their colours.
Settings > About gets a full-width panel for dataforest cloud, who
sponsor the servers behind place search, address search and routing.
Both logos ship in client/public/brands, so opening Settings reaches no
third party. Wherever a line names dataforest cloud it links to
https://cloud.dataforest.net/?mtm_campaign=trek, the link dataforest
gave us; the README moves to it too.
All 28 locales carry the new copy, and the old 5.0 card keys are gone.
The service reads a community place's details before the result cache and
every other place's after it, so the lookup has two call sites of which one
runs. It only fills the details row cache and turns a failure into no
details, so it is tagged @txStandalone and the enrich baseline drops to the
four writes it holds without it.
An OSM element can hold community values, so the service reads its details
before the cache to learn the version it is served under. The second request
reads them once and then answers from the row cache without another lookup.
The Hebrew locale arrived on dev while this branch was open, so its 80 new
admin.contribute, admin.servicesCard and places.refresh keys were missing
there.
Both answers were taken as typed. A value outside the published shape, such
as hours that are no text, reached the hours parser and failed the details
request. A place outside the shape is now left out of /v1/community, and an
element outside it sends the lookup to Overpass.
Like an OpenStreetMap summary, a summary credited to the community is read
straight off the request's details, so any caller could have it served to
every user of the instance for a week under the community's name.
Switching sharing off commits before the withdrawal asks the service. When
the service did not confirm, the decision had taken effect but left no audit
row. It is now audited with withdrawFailed, and the admin still gets the
error.
The upload and the trip-file routes vouched for every field that changed
between their read and their answer. A website the assistant or a
collaborator changed while a large picture uploaded was then recorded as
typed by the user and queued for the TREK Places API. Those writers now name
the one field they write, and other changes keep the origin their own writer
recorded.
A value saved while the request for its predecessor was out rewrote the same
outbox row with a new cid; the answer for the old value then deleted the row,
or charged it an attempt, and the newer value was never sent. The settle
writes now match the cid that was sent as well as the id, so such a row stays
queued for the next run. A replaced value also counts its age from the
replacement, so a long pause cannot expire a value queued days ago.
A saved place reopened for editing has no community version in its key, so
the answer the tab kept the first time stayed until the tab closed: a picture
another installation contributed since never showed. Kept answers now carry
when they were fetched and are asked again after ten minutes, the same span
the server keeps an empty answer. The key version moves to 5 so open tabs drop
what they kept in the old shape.
Two gaps kept it out. The service writes absolute picture URLs on its own
host, and the community service only took relative paths, so every picture
was dropped; it now takes the path of a URL on the configured service and
nothing from another host. And a saved place opened for editing sends no
details, so no community version, and the cached answer kept for a week; for
an index or OSM place the details are now read before the cache, so a new
version is a new answer.
The service delivers the pictures review approved for a place, but TREK never
read them. A new TrekCommunityService in the maps domain asks the TREK Places
API for them (the index row for a gers place, /v1/community for an OSM one),
accepts only paths of the service's picture store, and fetches the bytes with
the installation's key, capped at 4 MB. Enrichment puts them ahead of Google
and Commons, through the same photo cache, credited as TREK Community under
CC BY 4.0. The enrichment cache now follows the community version for OSM
places too, so a newly approved picture shows on the next open.
Also lowers the client's no-non-null-assertion baseline to 380: moving Use
this text into the description card removed two.
A busy log pushed the rest of the card down. It shows the newest five rows,
a Show more toggle opens the rest, and the page buttons appear once it is
open.
The pill compared website, phone, e-mail and hours only, so a description
the community published never showed up on a saved place. It now reads the
index description or the OSM summary, compares it as plain words against
the editor's markup, and shows it clamped to three lines.
The button was pinned to the foot of the details column, away from the text
it adopts. It now sits inside the description card, under the source line,
with the reason it is blocked directly beneath it.
A community description laid over OSM details or an index row showed up as
OpenStreetMap under ODbL, or as the place's own website. It now has a source of
its own, community, credited as TREK Community under CC BY 4.0, like the
service publishes it.
A checkout where npm install ran once carried shared/node_modules into the
image build, and its stale packages broke the client build with a vite that
lacked vite/internal.
The fresh lookup behind New place details goes through placeDetailsRepo
(online-only, like the service identity) instead of api/ directly. The form
loses the Translate type and the comment that moved with the head band's
pills, which takes it under its line limit and out of the size baseline; the
maps service suite's test-size entry follows its shorter count.
Editing a saved place looks it up again. Where the lookup has a website,
phone, e-mail or opening hours that differ from the place, a New place details
pill in the form's head band shows how many, and opens what the place holds
beside what the lookup has. Empty fields are ticked, filled ones are left to
the person and say they would replace; Take over writes only the ticked
fields, with the provider as their origin, and nothing the lookup does not
know is ever emptied.
The head band's pills move to PlaceFormHeaderPills.tsx, which brings the form
from 1157 to 1002 lines.
Community opening hours laid over OSM details now go through the same parser
as OSM's own, so they arrive as one line per day with open-now and the periods,
and a community description lands in the details' summary. A place copied from
a collection no longer records its values as an import: without an origin row a
value is never sent anyway, and a later edit marks only the fields it changes.
Opening an OSM place asked the public Overpass servers for its tags, which
took five seconds and more and answered nothing while they were down. Where
the TREK services are in use, the details now come from the Places API's
details layer (/v1/osm), with the community values in the same answer, and
Overpass is asked only for an element the layer does not hold.
The TREK API, the transit provider, the place search provider and the search
log move out of API Keys into a card of their own, which keeps them together
and saves as they change. API Keys holds only Google Maps, Unsplash and Amap,
and is not shown on a managed install, where the keys are the operator's.
Allowed File Types moves to the left column, which had room for it.
Both blocks in the TREK API card now use SettingsCard like Single Sign-On and
two-factor: an icon tile, the title with its state as a badge, the note under
it and the switch or badges on the right, with a flush variant for a card
inside a card. A figure badge no longer cuts the last pixels of its value, and
the withdrawal toast names how many values it took back.
A place found through Photon or Overpass, a relation or a way, is no row of the
TREK index, so the community values the index lays over its own rows never
reached it. Opening its details now asks /v1/community for the element and
lays what is published over the record, with the version, where the TREK index
is in use. No answer leaves the details as they were.
The state and its badges sit in a head band, the new-identity button in a foot
band, each with a rule and the card's own surface, so neither reads as part of
the allowance tiles between them.
A value is due the moment it is queued, and the save starts a send run right
away; the two-minute tick stays for the retries. One run goes at a time, and a
run asked for meanwhile follows right after it. Only where the schedules run, so
tests and replicas send nothing.
A place copied from a collection and a venue a booking import creates now
record their details as an import, so a later edit compares against what was
really written. Dawarich and the background enrichment write none of the
shareable fields, so they need nothing.
The admin panel asks again for a few seconds while a fresh installation is
still fetching its key, so the prompt appears without a reload.
- The plugin RPC fallback tests build PlacesRpc with its origin bookkeeping.
- The e2e guard and contract cases send their requests one after another; the
test server reset connections opened all at once.
- The job's two cron registrations are tagged @txIndependent: neither writes.
- The panel draws its fold head itself instead of importing it from pages/.
- The pause setting's type comes from its schema, so nullable fields read the
same under the server's compiler settings.
- The place form passes the origin of a field through to the hook.
- Serial steps run through inOrder (src/nest/common/in-order.ts) and
independent ones through Promise.all, instead of an await inside a loop. The
bulk update in places.mcp.ts goes the same way, so its baseline drops by one.
- The status a call without a key gets is a named constant, not an envelope
literal.
- A picture test no longer inherits the previous test's mock answer.
- The entity counts include the three new tables.
- The log keeps its failure per query instead of resetting it in the effect,
and usePlaceForm sets its picture through one helper.
What is sent and when, what never is, how values are checked before anyone
sees them, and where an admin switches it off or takes it back. The new
TREK_PLACES_CONTRIBUTE variable is listed with the compose file and the chart.
The first time an admin opens the settings on an installation that can share,
a dialog asks for the whole installation, with every field switched on and a
plain list of what never goes along. Nothing is sent before Agree; Not now
asks again in 14 days, No keeps the question away until the terms change, and
closing it decides nothing.
Inside the TREK Places API card, Share place details then holds the switch for
the installation and one per field, the figures of the last 30 days, the
transmission log and the way to withdraw everything sent. The place form
reports for each field whether the user typed it or a provider filled it.
Every shareable field of a place now records where its value came from: typed
by a user, filled by a provider, written by the assistant, a plugin or an
import. Only values a user typed are ever sent, and only after an admin agreed
to the current terms for the whole installation, field by field.
The new place-contrib domain keeps the decision in app settings, queues values
in an outbox a cron job sends from, logs what the service made of each one and
can withdraw everything this installation sent. Development, test and demo
instances, an own TREK_PLACES_URL, TREK_SERVICES=off and
TREK_PLACES_CONTRIBUTE=off never send anything. Index results carry the
community version, so the details caches refresh when it changes.
The Atlas server no longer keeps the admin-1 regions of every country in memory. Each country's regions are now a pre-gzipped file of their own, sent to the map as they are.
**Why:** `GET /addons/atlas/regions/geo?countries=…` built `admin1Store` (every country's region text, kept until restart), then joined the requested countries, `JSON.parse`d them, and Nest stringified them again. On the 512 MB – 1 GB boxes TREK runs on, the first z≥5 view could OOM the server (same shape as #1576).
**What changes:**
- `server/assets/atlas/admin1/<ISO_A2>.geojson.gz`: 198 files, 14.85 MB, the same 3,447 features as `admin1.geojson.gz`.
- New `GET /addons/atlas/regions/geo/:code` with a Zod contract (`regionGeoParamsSchema` in `@trek/shared`: two letters, any case). It sends the country's file as it is, with `Content-Encoding: gzip` and the current 1-day `Cache-Control`, as `countries/geo` does. A country without regions gets an empty FeatureCollection. Anything that is not two letters gets a 400.
- `useAtlas.ts` fetches one URL per uncached country and still rebuilds the layer once. Each country has a stable URL, so a country evicted by `REGION_CACHE_MAX` comes back from the browser cache.
- `getRegionFeatures` reads only the country it needs. `admin1Store` is removed.
- `?countries=` keeps working for cached PWA clients. It joins the per-country files as bytes, without a parse.
- The migrations are untouched. `admin1.geojson.gz` still ships, only for `Migration20200101021500` and `Migration20200101031500`. This costs ~15 MB of disk in the image and no RAM. `build-atlas-geo.mjs` writes both, and ATLAS-BUNDLE-008 keeps them equal.
**Memory** (real `atlas-geo` module, first z≥5 view over Europe = 49 country codes, JS heap + Buffers):
| | peak | kept after |
|---|---|---|
| before, first view (builds the store) | +372 MB | +61 MB |
| before, later views | +115 MB | 0 |
| after, `/regions/geo/:code` | +4.7 MB | 0 |
| after, legacy `?countries=` | +59 MB | 0 |
**Tests:**
- New: `atlas-region-files.test.ts` (moved from `atlas.service.test.ts`, plus join and dedupe cases), the `regionGeoParamsSchema` spec, ATLAS-BUNDLE-007 (the file format the byte join depends on), ATLAS-BUNDLE-008 (bundle and per-country files are equal), ATLAS-016/017/018 (the new route over HTTP, including the 400).
- The `getVisitedRegions` tests (ATLAS-UNIT-019..024) used to pass only when `getRegionGeo`'s `beforeAll` had built the store first. They now also pass on their own.
- The shared, server and client gates pass. Baselines go down only where this change removed code (`no-console`, `no-explicit-any`, `no-swallowed-catch`, file sizes).
**Trade-offs:**
- A first zoom over Europe sends ~49 small parallel requests instead of 1. Over HTTP/1.1 the browser queues them 6 at a time. The total bytes are the same (4.4 MB gz).
- The per-country URLs are unversioned and cached for a day, like `countries/geo`. After an upgrade that changes a country's regions, a browser can show the old shapes for up to 24 h.
Approved on Discord in #github-pr: <paste the thread link here>
Related: #1576, #2559
- [x] Bug fix
- [ ] New feature
- [ ] Breaking change
- [ ] Documentation update
- [ ] I have read the [Contributing Guidelines](https://github.com/liketrek/TREK/wiki/Contributing)
- [ ] My branch is [up to date with `dev`](https://github.com/liketrek/TREK/wiki/Development-environment#3-keep-your-fork-up-to-date)
- [x] This PR targets the `dev` branch, not `main` *(wiki-only PRs are exempt)*
- [x] I have tested my changes locally
- [x] I have added/updated tests that prove my fix is effective or that my feature works
- [x] I have updated documentation if needed
Translate the 55 keys dev added since the branch was cut (TREK services
allowance, Valhalla routing service, PDF routing progress, overview
distances, Unsplash key hint, tour day limit) and drop the two reservation
toasts dev removed, so he matches en key for key again. Refresh the two
admin strings and the Valhalla hint whose English changed.
Check he in the placeholder spec and the RTL test, and count 28 languages
with Hebrew in the README and the Languages, Environment Variables,
Contributing and Help Center wiki pages.
Fill in every domain added since the first Hebrew pass (help, tours,
assistant*, studio*, wanderer*, journey, demo, ferratas and the rest),
retranslate strings whose English changed, and give every count group
its Hebrew one/two forms. Register he as a non-Latin locale for the
untranslated ratchet, add the Hebrew 429 message, and count 28 locales
in the tours locale spec.
Full Hebrew translation of every UI domain plus email/notification
templates. Registered in SUPPORTED_LANGUAGES, marked right-to-left,
wired into the client locale loader and the rate-limit message table.
transit is not one of the fixed Costs categories, so typeToCostCategory
would put an uncategorised fare under other. Default to transport, the
category every rail and bus booking maps to.
- create_transit_journey takes price / budget_category, as create_transport
does: a fare > 0 files the linked expense in the booking's own transaction
and rides along in metadata.price. The tool now writes through
ReservationWritesService.createBooking, so its events and booking notice are
the ones every other booking write sends
- create_budget_item and create_budget_item_with_members take reservation_id,
the field POST /budget already accepts; create_budget_item now runs
BudgetWritesService.createItem like its sibling, so a linked booking's
mirrored price is worked out again before budget:created goes out
- bookingPriceSchema / bookingBudgetCategorySchema in @trek/shared back the
price fields of all three booking tools (mcp-inline-zod baseline lowered)
- wiki: MCP-Tools-and-Resources describes the new params
The trimming if swallowed the blank line, so the HTTPRoute note ran
straight on from the Persistence section. Move the blank line inside
the conditional so it only appears with the section.
Add disabled-by-default HTTPRoute support with configurable parent references,
hostnames, paths, annotations and optional timeouts. Route to the existing TREK
Service and keep Ingress support unchanged, including simultaneous use.
Document configuration and extend the existing Helm Chart & Compose job with
HTTPRoute render variants and strict validation using a pinned CRDs catalog.
Approved on Discord:
https://discord.com/channels/1488298068427411591/1553537637455761500
Validation: strict Helm lint, 37 manifests validated without errors or skipped
resources, and both Compose configuration checks passed locally.
Live Gateway WebSocket and streamed assistant response checks were not performed
because no Kubernetes test cluster was configured.
The install guide lists extraEnv and extraEnvFrom but had no word on the
new volume values. Add a Further volumes section with the KItinerary
extractor example and point the Reservations-and-Bookings page at it.
The chart mounts only the data and uploads volumes, and no value adds
another one. Add two lists that go into the Deployment as written:
extraVolumes for the pod, and extraVolumeMounts for the trek container,
next to the data and uploads mounts. Both lists are empty by default,
so the default render does not change.
The values file shows one use: a ConfigMap of KItinerary extractor
scripts, mounted at /usr/local/share/kitinerary/extractors. The chart
README and the Reservations-and-Bookings wiki page tell Helm and Compose
users how to mount extractors there.
The variants render in the Helm Chart & Compose job sets both values,
so kubeconform checks them.
The empty-result fallback can call Google after Transitous, so lint:tx now sees two googleQuota.record() calls in TransitService.plan. Only one of them runs per request, and each is a call counter written after its own upstream request, so it cannot share a transaction with it.
Admin settings, transport, environment and MCP pages still said Transitous
is the only fallback and that Google is billed only when picked. Document
that an empty plan asks the other backend, that the Google fallback uses
only an instance key within the daily limit, and mention it in the
search_transit_routes tool description.
Logroño-Agoncillo (RJL) and fifteen other airports gained their IATA code in
OurAirports after the bundled list was built, so the airport search and the
code lookup of create_transport could not find them. The list is rebuilt from
the current data. The build now keeps every code the previous list carried
that upstream no longer lists (a scheduled-service flag flipped back, PBI moved
to a new record), since bookings and MCP calls still name them.
Refs #2897
A car or other transport booking carried its route only in the description, so
a calendar app had no place to navigate to. Without a location of its own, a
transport event now takes the endpoint the traveller has to get to: the
destination of a drive of their own (car, bicycle) and the departure of a ride
they board, the side the per-leg events already use. A car's pickup and
drop-off events are located at their own endpoints. A location the booking
states keeps winning.
Closes#2909
Claude Sonnet 5.5, Opus 5.5 and Fable 5.1 reject tool_choice type "tool" and
"any" with a 400, so every booking import through the Anthropic provider failed
on them. The request now sends tool_choice auto with parallel tool use off,
which every Anthropic model and Messages-compatible gateway accepts, and the
system prompt tells the model to answer through the emit tool exactly once. A
model that still answers in text is reported as before.
Closes#2902
Opening an entry on the share page showed the hero at no height until the
original arrived, then pushed the story down, and the lightbox stayed black
until the full photo had loaded. The hero now reserves the photo's proportions
(4:3 when none were recorded) with its thumbnail behind it, and the lightbox
shows the thumbnail and a spinner until the original is in.
The phone entry sheet probed the photo providers again on every open and added
the External photos button once the answer came, squeezing the icons of the
other two buttons. The journey screen already knows the connected providers, so
it hands them to the sheet, and the icons keep their size.
The picked neutral mood was a grey tint on a grey sheet; a picked mood is now
framed in its full colour and announced as pressed.
Closes#2907Closes#2908Closes#2905
The live trail arrived in Dawarich's own day palette, counted from the first
recorded day, while a journey colours its days by the entries. A day's line and
its pins therefore wore different colours. The trail of a day that has entries
now takes that day's colour, the way the kept recordings already did.
Closes#2906
The journey map's photo thumbnails set position:relative inline on the marker
element. That overrides the engine's .mapboxgl-marker/.maplibregl-marker
position:absolute, so every thumbnail entered the normal flow below the ones
drawn before it and the engine's translate was added on top: the further down
the list, the further the photo drifted from its spot. The location dot had the
same inline position. Both now leave position to the engine's class.
Closes#2910
The search turned an airport code into its airport by reaching into the
airports domain's data module. airports.contract.ts publishes that lookup
instead, as the import boundaries ask.
maps.helpers.ts and osm.client.ts had grown past the 1000-line limit. The
merge behind the search and suggestion lists, with its lead rules, moves to
maps-merge.helpers.ts; the search records and Photon's half of the
OpenStreetMap geocoding (search, suggestions, airport codes, identity,
reverse, element details) move to osm-geocoding.ts. OsmClient keeps the
same methods and hands them on, so no caller changes.
- The lead is Photon's own row first; a row of its name inside the lead
only stands in where Photon's row went as the same place. Rows without a
line below the name are never folded as reading the same.
- A miss of Photon where it is the geocoder no longer asks it a second time
on the way to an empty answer.
- The transport words of the identity match use a Unicode word boundary
(straße, estación) and match 駅 anywhere.
- Opening an OSM place reads the coordinate it was placed at explicitly.
- The ferry spans of a Valhalla leg move to valhallaFerries.ts, and a day
cut into pieces is asked with one options object.
- "Limit" on the allowance tiles is "Daily limit", which no locale left
as an English copy.
- Tests follow the new behaviour: no Nominatim behind Photon, zoom 13,
and the expectations of the merge and edit-distance cases.
The card names place search, address search and routing with what each
answers and what is left of today's allowance. Each service is a tile with
what is left as its figure, the day's use as a bar and used and limit as
badges; the state, the kind of installation, a key from the environment
and the time the count starts again sit as badges in the header. The
sources add the terrain tiles the routes take their heights from.
A request to the TREK Valhalla that runs out its deadline (a long bike
route) is answered 504 and no longer rests the relay for every user or a
tab for ten minutes; that one day goes to OSRM. Only an unreachable host
or a server error rests it.
A day goes to the relay with up to twenty stops per request and no pause
between pieces; a request that ends up at the public FOSSGIS Valhalla is
cut to ten and paced, as before. The 1.1 s spacing of the server applies
to the public hosts only, not to an admin's own OSRM or Valhalla. The
relay keeps its connections open and asks for maneuvers only, which the
relay contract now accepts.
The suggestions and the full search now share one merge. Photon's hits are
ranked with places that have a Wikidata link before their unlinked
namesakes, and its context bias is narrower (zoom 13). The list opens on a
lead where Photon's answer has one: a country or region always; a town,
island or sight when it is called what was typed (or a near typo of it),
lies within 1000 km of where the traveller is looking and no place of that
name ranks above it. Rows of the lead's name inside it, the same element
under two ids and rows that read the same are shown once. A three-letter
code in capitals is asked for as the airport of the bundled list.
Where Photon is the geocoder it answers alone: an empty answer and a miss
are the answer for search and imports, the reverse lookup asks again
within 50 km, a Google place is matched inside a box through Photon, and
an opened OpenStreetMap place is placed by Overpass and named by Photon.
Nominatim stays the search of installations without the TREK services or
with an operator's own Nominatim.
Measured on 5000 places from Wikidata: the right place first while typing
48 % before, 65 % now; full search 53 % to 68 %; no repeated rows.
A car booking without an end day is a one-day booking everywhere else in
the plan, but the road trip read it as a car never handed back: only its
pick-up desk stood on the road. Two one-way car bookings on one day
(Bologna to Villach, then Villach to Klagenfurt) therefore drove from the
first pick-up to the second and on to whatever came before, counting the
road between the two cities twice and never reaching Klagenfurt. This
applies to the planner, the phone and calculate_roadtrip alike, since all
three seat the desks through the shared carrierSeam.
Such a car now has its return desk on the pick-up day. A return without a
time goes back before the next car of that day is collected rather than
closing the day behind it, so each one-way car drives its own road once.
Bookings with a return day and cars without a located return desk are
unchanged.
Closes#2894
A rail, bus or ferry ticket with changes arrived as one booking per
segment, each with a linked cost for the full fare, so the trip total
was the fare times the number of segments.
Import: the KItinerary mapper now groups consecutive train segments
under one booking reference that connect (same station by name or
within 500 m, no return, within a day) into one train booking with
stops and metadata.legs, the way connecting flights already were. The
connection rule is shared between flights and ground transport, and a
stop carries the onward departure for flights as for trains, as the
planner form and the MCP tools write it. The three copied ground
mappers are folded into one.
Within one such booking a fare repeated on every segment is booked
once, and fares that differ per segment are added in whole cents. That
summing now applies to connecting flights too, which used to keep the
first leg's price only.
Segments of one order that do not join into one booking (an outward
and a return trip, a gap, a change of station named two ways without
coordinates, bus and boat tickets, Rail&Fly) are compared across the
document: when every priced transport booking of a reference carries
the same amount in the same currency, the first to depart keeps it and
the others lose it. Amounts that differ stay. This runs on what
KItinerary extracted only.
AI import: the repeated-total safety net from #2651 now covers train,
bus and ferry tickets and any mix of them with flights, still only
when the document prints that amount on a total line.
MCP and assistant: create_transport says that a ticket with several
segments is one booking with endpoints[] and legs[], and the price
fields say they hold the whole booking's cost. When a new booking
repeats the price an earlier booking with the same confirmation number
already carries in the trip's currency (its own currency, or the
trip's when it names none), create_transport and create_reservation
still save it and return a warning naming the earlier booking. REST is
unchanged.
Existing duplicated costs are left alone: they cannot be told apart
from two tickets that really cost the same, and deleting expenses
automatically would change settled balances. They can be removed in
the Costs tab.
Closes#2895
A stable release without a dialect of its own now ships the one on main only while that one has 90 days left; dialect 0 or less stops the run before anything is committed, unless allow_stale_dialect is ticked. Fetching and checking moved into registration-dialect.yml, which never fails its caller and also type-checks the module under the strict config and lints it with --max-warnings 0, since the bump commit skips CI. The bump and everything after it run whatever became of the dialect job. Prereleases ask the issuer for a dialect of their -pre.N version and build it into the image without committing it. The release check reads the module token by token against an allowlist of keywords, globals and member names, refuses any assignment to a property, computed access and regular expressions, and names Object, JSON and defineProperty among the forbidden. The issuer's kitchen-sink sample is gone; the tests use a hand-written dialect instead.
The stable release asks the issuer for the registration dialect of the
version it cuts and commits it in the bump commit. Fetching and checking run
in a job of their own that holds no write token and keeps no git
credentials: scripts/ci/check-registration-dialect.mjs refuses a file that
imports or names anything beyond hashing, then runs it on its self-check
inputs. Only a file that passes reaches the bump as an artifact. When the
issuer is unreachable or its answer fails the check, the release goes ahead
with the dialect already on main and says so in the log and the summary.
Prereleases get no dialect of their own; dev takes the stable one through
the merge of main into dev after a release, and docker-dev.yml warns while
dev still speaks the original registration or a dialect older than 150 days.
Registration now goes through a registration dialect module: the paths of
the challenge and the registration, the bytes scrypt runs over and the body
and headers of the request. Each stable release will carry the dialect the
TREK services issuer writes for it; until the first such release the module
speaks the original registration, so nothing changes on the wire.
An issuer that no longer takes the release's dialect answers 410. TREK then
says so once per process, waits a day (or what Retry-After asks) and keeps
searching through the shared pool. The issuer client only lets a dialect
register at the issuer's own paths, with X-TREK-* headers that are not the
gate's and a body of registration size.
Renewal needs no dialect, and an expired key may now be renewed for a year
instead of 30 days, the same period the issuer allows, so an installation
that was offline for long comes back without an update. The key format
ignores the new optional dialect claim; the shared vectors gain one key that
carries it.
The marker effect reads the card handler through the component's own ref
again, so exhaustive-deps sees a ref it knows, and the card hook takes it
alongside the map's other refs. The pin event assertion types the mock
calls it maps under strict mode.
The phone card logic moves out of JourneyMap into useEntryCard, which brings
the component back under the line limit. The new card tests reach the
card element through a guard instead of non-null assertions, the public
hook test reads the hook through renderHook, the GL card tests use
mock.lastCall where Array#at is outside the client's lib, and the
complexity baseline drops to the count the branch now holds.
listPublicEntries changed for #2886, and the repository rules move a
touched QueryBuilder call to the entity methods or Kysely. Kysely keeps
the method's answer what it was, the stored row as SELECT * reads it,
which the entity methods would hand back as entities with a journey
reference in place of journey_id. The query-api baseline drops to 8.
ConnectionStop, EntryConnections and JourneyMapStop are only used in the
signatures of their own modules, and knip counts an export nothing
imports. The connection tests take the FE- prefix and header the other
journey tests carry.
The card a phone host opens its entry from was still a tooltip bound to
the pin, so Leaflet opened and closed it on the pin's hover and focus. A
tap on the card reaches a phone browser as the pointer leaving the pin
first: the pin's mouseout closed the card between the tap and its click,
and on iOS a content change in that phase can drop the click entirely.
Where a host wires onMarkerCardClick, the pins now bind no tooltip. One
card per map stands over the active pin, the way the GL popup does: it
follows highlightMarker, leaves when no entry is active or the entry
pins are switched off, and only a change of entry moves it. A press on
it never reaches the map. Desktop hosts keep the hover label on each pin.
The card tests run against the handlers the pins really get and the
card the map really adds, and the file parses again: two test names held
an unescaped apostrophe.
On the phone journey screen and the public share page on a phone, a tap on
a numbered pin brings up the entry's card above it, but the card could not
be tapped: the GL popup and the Leaflet tooltip both let the pointer fall
through to the map, and no phone host gave the card anything to do.
The map now takes onMarkerCardClick, and both renderers make the card open
that entry once a host wires it. The GL popup handles it in its delegated
listener next to the photo strip, which keeps opening the photos where the
desktop asks for that. Leaflet handles it on the tooltip element itself:
a press that reached the map container let Leaflet focus the map, which
blurred the pin and closed the card before its click landed. The phone
screen and the timeline open the entry through the carousel hook, the
same way a tap on the active card does. The desktop asides keep the card
as a hover label.
The unused hideMarkerTooltip prop goes, together with the docs that still
said the phone hides the card.
Closes#2885
The share link has counted its map stops through the whole journey since
#1962, while the owner's journal still restarted the count every day, so
the same stop read 66 on the link and 6 in the journal.
Both views now take their stops from one helper, journeyMapStops: located,
non-skeleton entries in the server's reading order (date, sort_order, id),
numbered across the journey. A draft keeps its pin in the journal but gets
no number, so the published stops after it carry exactly the numbers the
share link shows. The public entry query gains the id tiebreak the owner
query already had, so entries sharing a sort_order come back in the same
order on both sides.
Closes#2886
The journey map joins every located entry with a faint dashed line in the
order of its pins. It drew that line through everything, also across days
whose real route is on the map as a kept recording or the live Dawarich
trail, so straight guesses fanned out over the recorded track. The phone
has no layer switches, so the line could not even be hidden there.
A recorded track now carries the day it belongs to (an optional `date` on
the track contract, set by every builder of recorded lines: kept days, the
shared page's recordings, the live Dawarich trail and the day import
preview). A pure model splits the connection into the stretches still open
and those whose days were all recorded, the night between two recorded days
included. Both renderers draw only the open stretches with the entries; the
covered ones sit on a layer of their own that shows while the reader hides
what was recorded, so the guess comes back exactly where nothing else
explains the way.
The live trail is now matched to a kept day by that date instead of reading
the track name as one.
Closes#2888
The journey map grouped its photo thumbnails and its plan and stay marks
on a grid laid over container pixels. Every pan shifted that grid against
the map, so photos near a cell edge changed groups after each move: the
count badges jumped and the thumbnail switched to another photo, although
the zoom never changed.
Both renderers now bucket in world pixels at the current zoom, Leaflet
through its own CRS (so GCJ-02 basemaps stay right) and GL through a
plain Web Mercator projection that ignores centre, bearing and pitch.
Panning keeps every group as it is; only a zoom regroups.
Closes#2889
The booking and transport dialogs each spelled out the notes and the link
as four value and change props, and with the saving flag the two calls
became one copied block. The shared lower part now takes the form and its
setter, as the dialogs hold them, and writes the two fields itself.
The editor follow-up always hands the planner an afterSave, and its type now
says so, so the hook test calls it under strictNullChecks. The travel stats
test reads a country without a status as visited, as the standing helper
does for the hand-marked list. The client baseline holds the one non-null
assertion the count missed.
The booking save helper answers a Reservation the way the dialogs' onSave
does, and takes the payload the sheets read back. The export tests resolve
their routing promise without non-null assertions, the autocomplete test
reads its own console spy, and the Photon tally test asks its five dropped
questions together instead of awaiting in a loop. The Photon address line
falls back to empty when a feature has none.
The repair test now expects the dashboard to place the cleared beach in
Germany through the borders, as the Atlas map does since #2875. The client
lint baseline follows the counts this branch lowered.
The resolver's comment claimed the places of a city trip share their
coordinates, which they do not. It now says what an uncached place costs
per count and why the dashboard does not queue the Atlas region lookup:
that lookup falls back to Nominatim, and a dashboard load must not send a
user's places to a third party.
The #2875 cases continue the file's AUTH-DB numbering.
The suggestions waited for Photon beside the index for up to its full
2.5 s deadline on every keystroke, and their misses and spent allowance
paused Photon for the search and the reverse lookup as well.
Once the index has answered, Photon now gets 300 ms more and is then
called off; with nothing from the index it is still waited for. The
keystroke question has a shorter deadline and its own tally of misses and
pauses, a question called off by the list counts as no miss, and a pause
of the search still holds the suggestions back.
Refs #2877
A PDF of a long trip routes for minutes, and only the skip button could
stop it. Closing the export dialog, hiding the phone sheet or leaving the
trip left the legs going out and opened the document later on whatever
page the user was on. Closing now stops the routing and the document never
opens; skipping still prints without the remaining legs.
The plan filter moves from TripPDF to pdfScope beside the other plan
predicate, which keeps TripPDF under its size entry. The routing test name
gets its quotes back, and the export dialog and sheet tests cover the leg
count, the skip button and closing while routing.
The follow-up kept the booking a save stored until the editor session
changed, and on the phone the session was only the open flag. The import
review swaps the draft without closing the sheet, so the second booking of
the same kind went out as an update of the first and lost its linked cost.
The stored booking now lives only as long as a failed try needs it, and
the phone sheets name their session the way the desktop dialogs do, by the
booking or draft they hold. The sheets also seed their form from that
session, so the next draft of the review shows its own fields.
The follow-up hook gets its own test, the notes block test passes the
saving flag the component now takes, and the swallowed-catch baseline
follows the one catch the save merge removed.
The passport card and the last/next trip countries read place countries
from the place_regions cache alone, while the Atlas map falls back to the
bundled borders for a place without a cached region. A place in Vatican
City never gets a region row (the region bundle has no VA features and
Nominatim names no region there), so the map counted VA and the card did
not. The same gap hid any country whose place was still waiting for its
background lookup, or whose lookup had failed.
Both views now resolve each place through one helper: the cached country
first, the borders otherwise, once per location. The dashboard reads every
place of a started trip with its cached country instead of the distinct
cached codes, and the trip countries count resolved places the same way.
The MCP travel stats tool and the public stats API use the same service
methods and are fixed with it.
Closes#2875
The PDF export routed the whole trip under an 8 s deadline while the router
is asked one leg at a time, a second apart. A cold month-long trip got about
seven legs routed per export; the rest printed as straight lines at "0 m"
under a total that looked final, and every further export routed a few more
days from the device cache.
The export now waits for every leg, bounded by the request pacing, the
rate-limit retries and a per-leg timeout. The export dialog and the phone
sheet count the legs while they route and offer to print without the rest.
Any leg left without a road is marked in the document: the day reads "not
routed" or "at least ...", the total says it is a floor, and a note names
how many legs are missing. The planner overview shares the same wording.
The PDF also routes what the planner map routes now: the road trip's vias,
the multi-day tour stages and the reader's distance unit.
Closes#2876
The opening day is seeded while the loading splash still covers the shell,
so the rail's auto-scroll ran before any chip existed and never ran again:
its dependencies did not change when the rail first mounted. A trip opened
mid-way showed today's day below a rail stuck at day 1.
The rail element is now a dependency of the alignment, so the active chip
is brought into view when the rail appears, including after re-entering
the trip. The first alignment jumps instead of gliding across a rail
nobody has seen yet, and runs before paint; later day changes still glide.
Closes#2880
Saving a booking or a transport closed the editor and reported success
before the travellers and picked files were written. A failed upload left
a booking without some of its PDFs, no list to retry from, and a create
form that would add a duplicate when saved again. On desktop a file picked
on an existing booking went up at once and vanished when that failed.
The planner's save now takes the editor's follow-up as a step it waits
for: it stores the booking, runs the traveller write and the uploads, and
only then says so and closes. Each file that lands leaves the list, so a
failure leaves exactly the files still to send, and the editor keeps the
stored booking so the retry updates it instead of creating another one,
without a second linked cost or stay. All four editors (desktop booking
and transport dialogs, both phone sheets) share this through
useBookingFollowUp, an import review only moves on once an item's files
are up, and desktop files picked on a saved booking wait for Update like
on the phone.
Closes#2883
The suggestion list asked only the TREK index and its OpenStreetMap layer,
because Nominatim's usage policy rules out autocomplete. The layer is a slice
of OpenStreetMap, so many places showed up only once the full search was sent
(#2877). Photon is built for this keystroke: the suggestions now ask it in
parallel with the index and interleave its rows, dropping any the list
already holds, by OSM id or by the search list's same-place rule.
Without Photon (services off, an operator's own Nominatim, a paused or failed
answer) the list reads exactly as before. A failing index no longer throws
away what Photon found, and the Google-only switch skips Photon as it skips
the index.
Refs #2877
MCP-SCHEMA-005 finds tuple keywords, it does not validate against the 2020-12 metaschema, so the name now says that. A restored map point lost its range and order to the tuple fix; the description keeps the order visible to a model that builds a page itself.
With the Journey addon on, ChatGPT refused every TREK tool: restore_book_page
took a map's path and roads as [lat, lng] tuples, which the SDK writes as the
draft-07 array form of items, and JSON Schema 2020-12 refuses that form.
The restore now takes a map point as two numbers and parses each page as a
spread before it is put back, so lat and lng stay checked and a point out of
range is an invalid_op instead of a thrown parse. MCP-SCHEMA-005 holds every
tool's input schema free of tuple keywords.
Closes#2879
The built-in profile check returned the open profile key and the run request
returned an unnarrowed answer, both of which the client typecheck refused. The
new road trip tests asserted with non-null assertions where optional calls do.
The graceful stop step piped docker logs into grep -q under pipefail, so grep
leaving early after its match could fail the step with SIGPIPE although the
server had logged SIGTERM.
A road trip day OSRM drew while the relay could not answer was cached for five
minutes, and a key added or removed kept serving the old engine's answer. The
stand-in is no longer cached and the key's availability is part of the cache
key. A walk kept off the boats may still go to FOSSGIS, which is now said and
tested.
The per-user 429 carried no Retry-After, so the planner left the relay alone for
its full ten minutes and routed every day with OSRM meanwhile. It now answers
like the installation limit does, with the minute the window takes.
The route cache key followed the relay's rest, so after one failed relay request
(always the case offline) every day the device held only from Valhalla missed
the cache for ten minutes. The key now follows the configured provider, and a
resting relay only stops the question from going out.
A route now names the engine that drew it. A day OSRM drew because Valhalla
could not answer is OSRM's on the rail, and its other ways are asked of OSRM,
instead of being labelled Valhalla's and compared with Valhalla offers.
An own Valhalla gets the relay's twelve second deadline, and a day that runs it
out goes to OSRM without a second wait. A day that avoids something keeps its
old cache key, so the days a device already holds stay found. The leg segment
and the offer types move to modules both engines import.
The hint under Routing service now explains that the TREK services and an
own Valhalla calculate every route, that OSRM calculates a day they cannot
answer, and that FOSSGIS keeps Valhalla to what OSRM cannot do. In all 27
languages.
Where the TREK services or an own Valhalla route the days, the MCP road trip
router asks Valhalla for a day that avoids nothing, through the relay alone,
and OSRM routes the whole day when it cannot answer. Requests carry twenty
locations unless they may end up at FOSSGIS, which keeps ten.
With the TREK services or an own Valhalla, the day plan, the road trip and
the other ways of a leg are routed by Valhalla on foot, by bike and by car,
twenty locations a request. A day it cannot answer (relay refusal, time out,
no way for a leg) goes to OSRM whole and without a word; it is never handed
to FOSSGIS unless it avoids something only Valhalla can avoid. With FOSSGIS
nothing changes: OSRM routes the days and Valhalla the avoidances.
A Valhalla answer is cached under its own key beside the OSRM one, so
switching engines never serves a day from the other engine. The Valhalla
day routing moves out of RouteCalculator into valhallaDayRoute.ts.
VALHALLA_COSTING maps driving, walking and cycling to auto, pedestrian and
bicycle once for the planner and the server. valhallaRoutesEveryDay says
whether the resolved provider carries every day: the TREK services and an
own Valhalla do, FOSSGIS and no Valhalla leave the days to OSRM.
An empty Photon search, or one without a hit that has an OpenStreetMap id, is asked again at Nominatim as the reverse lookup already was. The background lane goes to Nominatim alone, a reverse lookup leaves Nominatim half of the caller's deadline, and a 429 with Retry-After pauses Photon for that long.
A route or height question through the relay now gives up after 12 s. A 400 without Valhalla's error_code and a 404 or 405 count as the relay unable to answer instead of no way or no Valhalla, and a Retry-After sets how long the tab leaves the relay alone.
A hung TREK Valhalla cost every leg, height chunk and trace the full 30 s before FOSSGIS answered. The relay now waits 10 s for a route or heights, rests a minute after an error or timeout and answers 503 at once meanwhile, follows a 429's Retry-After (an hour at most), passes that header on to the browser, and holds the installation to 100 questions a minute below the gate's 120 (60 per user).
The status asked a fixed Photon origin and compared NOMINATIM_URL with the public address, so a moved Photon or an explicitly set public Nominatim showed the wrong service as in use. It now reads the same photonUrl and nominatimOwn the Photon client decides by. TREK_VALHALLA_URL joins the commented compose variables.
GET /api/admin/service-identity lists places, photon and valhalla, each with
today's allowance read fresh from /v1/quota of its own host (cached for 20
seconds per host) and whether this instance asks it right now: Photon only
beside the public Nominatim, the TREK Valhalla only where the routing choice
comes to it. Photon and Valhalla are asked only with a key and only while
TREK_SERVICES_HOSTS lists them. The last allowance an answer stated is kept
per host for when a host does not answer.
The card lists the three services under each other in the look it had,
a service this instance does not ask dimmed and named as such.
routing_valhalla_provider (fossgis, trek or custom) sits beside the routing
URLs under Admin, Default user settings, admin only and locked on a managed
install. Without a choice an own Valhalla is custom, an own OSRM alone asks
no Valhalla, and otherwise the TREK services answer.
The planner asks them through the same-origin relay, so no connect-src
changes. A relay that answers 401 (a shared page without a sign-in), 403,
413, 429 or 5xx sends the question to FOSSGIS without a word and is left
alone for ten minutes. A long line to describe goes as a polyline.
POST /api/routing/valhalla/route, /trace_attributes and /height ask the
Valhalla at TREK_VALHALLA_URL (default https://valhalla.liketrek.com) for a
signed-in user. The body is held to the shared contract (64 KB, at most 20
locations), every question counts against 120 a minute per user, and the
installation key goes along only to a host TREK_SERVICES_HOSTS lists.
Valhalla's answer is passed on as it came, refusals included.
The road trip router asks the same relay when the routing choice lands on
the TREK services, so the MCP road trip, city walk and hike tools share its
limits, and hands a question it cannot answer to FOSSGIS.
An installation that holds a key for the TREK services and has no
NOMINATIM_URL of its own now sends the OpenStreetMap search and the
reverse lookups to photon.liketrek.com first. The answers are read into
Nominatim's shape, so the place records stay the same, and every miss
(refused key, allowance spent, timeout, error, foreign body) falls back
to Nominatim, with a one minute pause after four misses in a row.
TREK_PHOTON_URL points at another copy, and the default of
TREK_SERVICES_HOSTS now lists photon.liketrek.com and
valhalla.liketrek.com beside places.liketrek.com.
Regenerating clears the stored key at once and writes the new one only after the issuer answered in the background, so the two writes cannot share a transaction.
The proof of work and the bounded read of a 403 body recurse instead of looping, the background-work checks compare against null instead of testing a promise, and the tests repeat their requests through oneByOne. The renewal test loads the stored key before its first request, and the panel test picks the last confirm button without Array.prototype.at.
An expired key is no longer sent; the request goes through the shared pool
while a due key, or an expired one within its grace, is renewed in the
background. Refusals in one burst count once, a refused renewal counts like
a refused search, and a Places request whose key is refused is sent once
more without it.
docker-compose.yml and the Helm chart name the switch and the key (the key in the Secret, never the ConfigMap), and the wiki explains the installation key, what it carries, where it goes and the admin view of the allowance.
A calm panel under the card's facts: a dot with the state, the tier, a bar of today's use against the limit, what is left and the local time the count starts again. It reads when the card comes into view and every 30 seconds while it stays there and the tab is in front, says why when the services are off, revoked or cannot be read, and offers a new identity behind a confirmation unless the key comes from the environment.
A new service-identity domain registers the installation with places.liketrek.com on its first Places request, in the background and behind a light scrypt proof of work, keeps the key encrypted in app_settings, renews it from day 60 through a daily job, drops it after three 401s in a row with up to six hours of jitter, and freezes it on a revocation until an admin starts over. TREK_SERVICES_KEY wins and turns registration off; TREK_SERVICES=off sends no key; a mirror set with TREK_PLACES_URL never registers and never sees the key, which goes only to the hosts in TREK_SERVICES_HOSTS over https.
The Places client takes its headers from the identity port, sends a TREK/<version> User-Agent and keeps the instance token. A 401 or 403 does not open its breaker. GET /api/admin/service-identity reports the state, tier and today's allowance, asking the issuer at most every 20 seconds; POST .../regenerate starts a new identity.
Zod schemas for what the key issuer answers (challenge, key grant, quota) and for GET /api/admin/service-identity. The card no longer claims there is no quota: every locale now says there is no key to enter, a fair daily allowance per installation, and that only the number of requests is counted.
Without an Access Key the image search goes through Unsplash's web endpoint,
which refuses datacenter addresses with a 401 or 403 or answers with a
challenge page. The search then said only "Unsplash search unavailable",
although an admin can already enter a free key in the settings. Such a
refusal now answers 502 with that hint and the code unsplash_key_needed, and
the cover search shows it in the user's language.
Closes#1449
The candidate a GPX or KML row is checked as is built in lineCandidate, so
places.service.ts stays within its size baseline, and the atomic import test
imports the same Ridge line it later expects to be skipped.
Two GPX tracks a watch named alike (two "Afternoon Hike"s) were taken for one:
the import compared a track only by its name, so the second was skipped as
already in the trip. A track now matches a track of that name only when it
also starts, ends and reaches as far as that one, to about a hundred metres,
so the same file imported again is still skipped. The GPX, KML and tour
imports and the tour preview all read the same rule.
Closes#2870
A tour whose walking time and breaks came to more than 24 hours a stage
only greyed out Save, with nothing saying why. The planner now marks the
planned total and says what to do: make it a multi-day tour by ending a
stage at a stop, or enter a shorter planned total. The details section
opens by itself so the reason is in sight.
Closes#2869
The assistant sent an own server's chat to the address as entered plus
/chat/completions. Ollama, entered as http://host:11434 the way its own docs
write it, answers that with a 404, so every message ended in "something went
wrong" while connecting had worked: the model list is found under the root
at /api/tags. A bare server address now chats under /v1, where Ollama, LM
Studio, vLLM, llama.cpp and LiteLLM all answer; an address with a path, such
as https://api.groq.com/openai/v1, is used as entered.
A book held at most 150 spreads. The auto layout cut everything past that
without a word, so a long journey's book ended about three quarters of the
way through, with its last page and back cover gone, and adding, copying or
applying a designed page to a full book was refused.
The spread cap is gone from the book schema, the page operations, the
design composer and its apply, the assistant's book card and the design tool
contracts. A design's own page budget (pages.max) still bounds what it lays
out. The save route's 8 MB body still bounds a book's size, which at two to
five kilobytes an auto laid page is a few thousand pages.
The templated design digests move because their counter ids shift by two;
the pages, their order and their elements are unchanged.
A wider gutter left the button floating in the corner over the first card and the scrollbar. It now sits in its own row above the scrolling content, level with the Help title, and the content keeps symmetric padding.
The close button floats over the scrolling content, which only kept a 40px gutter on the end side while the button reaches 52px in, so the first card ran right up to it. The content now keeps a wider gutter on that side.
The cluster source, its layers, the expansion zoom and the DOM bubbles
now live in placeClusterLayers.ts with their own tests, which brings
MapViewGL back under its size baseline instead of growing it. Lowers the
size and eslint baselines that this change made smaller.
The MapLibre and Mapbox maps painted their place clusters as a canvas
circle plus a symbol layer, so the count used whatever glyphs the map
style ships and the bubble had no hover. They are now DOM pins built by
the same markup and CSS as the Leaflet bubbles, with the same sizes,
font and hover scale. The canvas layers stay, invisible, as the hit
area the track and route layers check a click against.
The date picker of the external photo picker hung absolutely under its
trigger, inside a dialog that clips its content. The last week of a
six-week month fell below the dialog, and for the end date in the
right-hand column the calendar opened past the dialog's right edge, so
the last weekday and the next-month arrow could not be reached.
The calendar now opens in a portal, pinned to its trigger through
useAnchoredPosition: it flips above the trigger when below is too short,
scrolls when neither side holds it, and is kept inside the viewport
sideways.
Closes#2858
With Google as the transit provider, the result list still credited
Transitous, and a route added from the panel or through
create_transit_journey was saved with provider 'transitous'. The panel
now credits "Google Maps" (untranslated, as Google's terms ask) or
Transitous after the backend that answered and saves that one; the MCP
journey records the backend that plans routes for the caller, and
search_transit_routes reports it.
Gemini 3 signs a tool call on its OpenAI-compatible endpoint
(extra_content.google.thought_signature) and refuses the next request
with a 400 when the call comes back without it, so every turn in which
the model called a tool failed. The Chat Completions client now keeps
the signature per call as the step's provider state and sends it back
with the call, to the model that signed it only.
An own server entered with Gemini's endpoint now also gets Gemini's
narrower tool schemas, which it refused before.
With connected days, every join between two days kept its one-request-a-second
slot even when the route came back from the device's cache, so a long road
trip still waited about a second per join on every open. A join answered from
the cache now gives its slot back; requests that go out stay spaced as before.
Refs #2794
The Tours import takes over a same-named track a trip kept from 4.x, but a
route file attached in the assistant chat was read with the plain duplicate
rule only, so that track was reported as already in the trip and never
proposed. The read now leaves such a track to the import, which takes it
over as a tour like the Tours dialog does.
Refs #2810
remember_for_trip runs at once during a turn, and read_attached_file and
keep_attached_file are answered by the server, but all three are calls the
model made. Rebuilding a thread from its stored steps left them unanswered,
and ChatGPT, the OpenAI reasoning models and Anthropic replay their own record
of each step, call included, so the provider refused the conversation with a
400: the continuation after applied changes failed, and every later message of
the thread with it. A step that kept a note also sent an empty user message.
Every call the model made is now answered as in the live turn, the attached
file's outcome as the answer of its call; threads that broke this way work
again on their next message. A refusal's log line names the refused field's
path, never its message.
Closes#2845
Once a store was connected for a trip, its address and login could only be
entered again after a sync run had been refused, and the stored connection
could not be removed at all: the server's DELETE route had no way in from the
app. A binding's settings now open the connection at any time, the folder
picker leads back to it before a folder is chosen, and the connection form
removes a stored connection after a confirmation, on the desktop panel and the
phone sheet alike. Removing it drops the trip's bindings through it; documents
stay on both sides.
Closes#2842
create_day_note and update_day_note described `time` as a clock label and capped
it at 250 characters, while the planner shows it as the Markdown body under the
title and allows 2000. The assistant put long text in the title and had to cut
notes it was asked to edit. Both tools now take text and time from the REST
contract with descriptions naming title and body, and an update still refuses an
empty title. The assistant's change card labels the two fields title and content
instead of text and time.
Closes#2840
A new list starts with a '...' placeholder row. Typing the first item there
and pressing Enter saved it but left the cursor nowhere, while the "Add item"
field puts it straight back for the next item. Enter on a filled placeholder
row now opens the category's add field and focuses it. A blur still just
saves, so clicking elsewhere does not pull the focus back.
Closes#2825
The airport picker never offered Little Cayman (LYB): OurAirports marks it as
having no scheduled service although it has daily flights, and the build
dropped every airport on that flag. The same flag also hid Vestmannaeyjar,
Western Sydney and around 1,280 other large and medium airports with an IATA
code. The build now applies the flag to small airports only, where it keeps
private airstrips out.
The bundled list gains exactly those airports. Every entry it already carried
stays unchanged, because a fresh OurAirports export currently renames PBI to
DJT and drops nine other codes that saved bookings point at.
Closes#2820
A GPX or KML import skipped every row whose name was already a place of the
trip. A watch names a hike after where it went, so "Hokitika Gorge" was
refused as already in the trip because the gorge itself was a planned
place, with no track to show for it (#2810). A track now matches only the
tracks of that name; a pin still matches every place of its name.
The Places filter also hid its "Tracks" choice whenever Tours was on, on the
grounds that tracks are tours then. A track imported before Tours is a
plain place and stays in the list, so the choice now counts the tracks that
are no tour.
Auto layout gathers journal entries with no photo and no text, one after the
other, onto one list page instead of giving each an empty page. Its heading
said "Stations" (German "Stationen"), which read as train stations and gave
no hint why those places were on it. It now says "Along the way" in every
language; each stop still carries its date.
A GPX track imported in 4.x is a plain place: a line, no tours row. It never
showed under Tours, and importing the same file there was refused as a
duplicate by its name, so there was no way to turn it into a tour (#2810).
The tour import now takes over a track of the trip that has the same name
and is no tour yet: the place keeps its name and line and gets the tours
row, with the recorded times the file adds. The preview says so instead of
"Already in the trip". New and taken-over tracks commit in one transaction.
Connecting an own server asked Ollama's /api/tags and gave up on any answer
that was not a 404, and every failure, a refused key included, read as "not
responding". The model list is now looked for at /api/tags, at the address as
entered and at /v1/models, an answer only counts when it is a list, and a
refused key or a server that answered without a list say so (#2808).
AI parsing through an own server always called Ollama's /api/chat, which an
OpenAI-compatible server such as Groq answers with 404. A server without it
now gets the same extraction over /chat/completions at the address as
entered (#2806).
TREK 5.0 left most registry plugins blocked because their authors still declare
"<5.0.0", and the only way past that was TREK_PLUGINS_IGNORE_TREK_RANGE for every
plugin at once. An admin can now install or switch on one such plugin after a
warning; the consent is kept on its row for that plugin version and TREK major,
so an update or a new major asks again. The environment switch works as before.
Discover marks a plugin not confirmed for this TREK on its card, and Install opens
the warning first. Admin > Plugins reads cleaner: one row of alike facts per plugin
with each state that needs attention on its own line, type, status and order behind
one Filter button, and the containment notes in a dialog behind the shield in the
header. The row, the filter menu and the dialog moved into their own files.
PocketBase answers a back-relation with a single record as the object itself,
not a list, so the user lookup failed as an answer TREK did not understand and
every search with it.
Closes#2799
TREK's code for Brazilian Portuguese is 'br', which Intl reads as Breton.
The assistant passed it straight to Intl.DisplayNames for the system
prompt's answer language and to the card formatters, so the model was told
to answer in Breton. Convert TREK codes to their BCP-47 tag with
getIntlLanguage before handing them to Intl.
Closes#2800
The return address already carries a query, so a code appended with ? instead
of & ended up inside the ai marker and the page did nothing, without a word.
An imported hotel was looked up once, by its name and its address together.
A booking site's romanized address (Agoda for a stay in Japan) often matches
nothing and sank the hotel's name with it, so the place was created without
coordinates and silently never showed on the map.
The review now finds the stay right away (name and address, then the name,
then the address) and says under the address whether it is on the map and
where, or that it is not yet, with a way to search again. Picking an address
suggestion sets the spot by hand. The save uses that spot, falls back the same
way when there is none, and a stay saved without a spot says so in a toast.
Desktop dialog and phone sheet share the lookup through useReservationForm.
With "Show whole trip" on, every leg of every day went to the router one at a
time, a second apart, and the route cache lived in memory only, so each visit
started over: a trip of a month took 30 to 45 seconds before its route was
drawn, on every open.
Routes the router answered are now also kept on the device (Dexie v14,
routeCache by RouteCalculator's cache key, a month at most, the oldest give way
past 3000). A leg found there needs no request and no pause, so a trip whose
legs did not change opens with its roads at once; only new or changed legs are
routed. Plugin routes stay in memory only, since they depend on trip and day.
The cache moves out of RouteCalculator into routeStore.ts.
Closes#2794
A token whose account has no local ActivityPub author was reported as
invalid_response, which the settings show as 'The answer did not come from
wanderer', although wanderer had answered fine. It now has its own code,
no_author, with a message in every language that says what is missing.
A failed wanderer call kept its reason in WandererError.detail (the field that
did not match, or what came instead of JSON), and the client comment said it
stays in the log, but nothing wrote it anywhere. The log only had the generic
'wanderer answered with something TREK does not understand', so neither an
admin nor we could tell what wanderer had sent. Every refusal now logs its
code, message and detail.
gpt-5.6-sol, gpt-6.1-sol and the other reasoning models refuse function tools
on Chat Completions, so every assistant message on such a model came back as
HTTP 400. An OpenAI connection whose model is from the o-series or GPT-5 and
later now goes through the Responses API, the client a ChatGPT plan already
uses, with the key as its bearer token and no token refresh on a 401. Older
models and the -chat variants stay on Chat Completions.
Closes#2792
Sonnet 5.5 and Opus 5.5 think by default and write what they say between tool
calls into thinking blocks, which they want back unchanged with the rest of
the turn. The client dropped those blocks (and their signatures) and rebuilt
the turn from its text and calls, so after a tool result the next step came
back empty and the assistant gave up with empty_answer. A turn that thought
is now replayed block by block like one that searched the web; the cache mark
before the question skips a thinking block, and replayed blocks are copies so
a mark never sticks to the stored answer.
Closes#2789
The compose file drops every capability and adds back CHOWN, SETUID and
SETGID. dumb-init runs as root and node as the node user, and signalling a
process of another user needs KILL, so docker stop never reached node and
always ended in SIGKILL (exit 137). KILL only reaches processes inside the
container's own PID namespace. CI starts the image with the same set and now
checks that docker stop exits 0 after a logged SIGTERM.
Closes#2790
The server takes a sign-in only once. The clipboard read on return and a paste
of one's own could send the same address twice within one render, and an
address left over from an earlier attempt used up the one in progress, so the
real address came back as 'took too long'.
The model picker in Admin > Addons did nothing for a ChatGPT connection: the
only way to change the instance's model was saving the connection again, which
needs an API key a plan does not have. A PATCH /api/admin/ai/connection/model
now switches it on its own, to a model the provider offers.
Closes#2783
docker logs trek found nothing because the run command gave no --name, and
over plain HTTP the browser dropped the Secure session cookie, so the printed
admin login did not stick.
The marks module exports components only, so fast refresh keeps it whole and react-refresh no longer counts it; the map by provider id moves to documentProviderIconMap.ts, which every caller now imports.
The 4.3.3 release card that imported the five marks one by one is gone, so they stay inside the module and knip no longer counts them as unused exports.
The overview of a trip's tours decides for itself when it shows, which keeps the planner page hook inside its complexity and length limits, and adds nothing to the camera's key while there is no tour to frame. The elevation helper names its empty case for the strict type check, the self-host wording exemption for the retired Dawarich card goes, the day plan test reads the wrapping pills, and the new tests drop their non-null assertions and loops.
The update dialog now shows the 5.0.0 headliners, Tours, TREK Studio, the assistant and help on every screen in 27 languages, each with a drawing of its own, translated into every language TREK ships. The 4.3.3 cards and their drawings are gone.
Dismissals were compared with the prerelease part dropped, so a notice closed on 5.0.0-pre.2 counted as closed on 5.0.0 and the final release showed nothing to anyone who had tried a prerelease. Versions now compare in full.
The map stayed empty until a tour was opened. It now draws every tour shown in the list, line and marker as on the plan, frames them, and opens a tour from its marker.
Ascent and descent were the raw sum of every height difference, so a flat walk across Venice (highest point 16 m) read 121 m of climbing. Heights now count once they move 5 m from the last one that counted; the server's stored figures and the planner's live analysis share the one function.
Each member's amount was converted and rounded to a cent on its own, so four 34,000 ISK quarters of a 136,000 ISK bill came to 997.52 EUR against the 997.51 the payer was credited. The balances ended a few cents off zero and settle-up had no payment for them. A split whose amounts match the payers in the bill's own currency is now converted as one set by largest remainder, in the settlement and in the per-person summary alike.
Every page of an install with the assistant and AI parsing off asked /api/ai/status and logged the 404 the addon guard answers. The store asks only while an addon opens the route and the install is not managed, and one hook on the desktop window and both phone screens asks again when the assistant or AI parsing is switched.
The rail gave its labels 48px and let them break anywhere, so Elements read Element and s, and nine languages broke a word at the default text size. Labels wrap only between words now, and the rail grows to its longest word from 64px.
The map framed its first view for the browser window instead of its own box and then skipped the fit, so a list of cafes across Europe opened on northern France. It now fits the places in its own box with room for the label bar and the controls, fits again when the set of places changes, keeps a view the user moved, opens a single place at town level and keeps a list across the date line on its short arc.
Tours, TREK Studio, Road Trip, Trip Settings, the notifications inbox, the phone layout, Help Center, Passkeys and the admin pages for settings, storage and notifications get pictures where the reader needs them, and the captions of the recaptured pictures say what they now show.
The 74 screenshots of TREK itself show the 5.0 look and the three demo trips, journeys and collections, all at 1600x1000 in the light theme. The Portainer pictures stay as they were.
Six upload routes refused the shared demo account in demo mode; journey photos, videos, gallery media and covers, collab note files and chat images did not, so any visitor could put images in front of every other one for the hour. They answer the same 403 now, checked in the handler after the upload is read, and a text-only chat message still goes through. The Demo Mode wiki page lists them, and says document sync is not covered.
The hourly reset puts only the database back, so a cover, photo, place image or trip file a visitor deleted in between was missing from the restored rows for good. On a demo instance StorageService.delete now leaves the content categories alone; the caches, the backups and the assistant's attachments are still deleted, and the orphan thumb sweep no longer counts a kept thumb as removed.
A Photo credits link in the dialog's footer opens a second view with every photo: its title linking to its page on Wikimedia Commons, its author and its license linking to the deed. The list is its own chunk, loaded when a visitor opens it. The countdown test reads the clock the interval moved.
The trip Photos addon of old releases is still seeded, switched off, and nothing reads it any more. The policy test now fails on no row, and its checks run without awaiting in a loop.
One UPDATE per shelf and state and one upsert for the switches, inside the same transaction, and the demo account's settings in one upsert. The two new log lines go through the logger.
The seed put the example trips in whenever the admin owned none, so a demo
whose content belongs to the demo account got them back on every reopen,
the hourly reset included, and the next baseline saved them along.
The demo account can look at its Account tab, API keys, MCP configuration,
own AI connection and language, but not change them, and the server already
refuses each of those writes. DemoLock renders such a section dimmed and out
of reach under diagonal stripes in theme tokens, with a small Demo mode badge
and one line saying why; useDemoAccountLock decides, on the desktop settings
and the phone screens alike (a new pair in feature-pairs.json).
Locked: the whole Account tab (profile, password, 2FA, passkeys, deletion),
API keys, the MCP card with its tokens and connected apps, connecting an own
AI model and the language picker, which shows English. The admin of a demo
instance sees everything unlocked, 2FA included, so the old instance wide
"Not available in demo mode" line under 2FA and its key are gone from all
locales.
The old popup was a narrow white card with its own copy of the texts in
eight languages, hex colours and a minute counter on the browser clock. The
new DemoWelcome is a wide dialog on DialogShell with theme tokens only: what
TREK is in one line, the hourly reset with a live countdown in the server's
time zone, the 5.0 highlights to try right away (trip planner with bookings
and costs, Tours with 3D terrain, TREK Studio photo books, live
collaboration, Journey, Atlas, Vacay, Collections, offline use), what the
demo leaves out and what the full version adds, and the link to the source.
Its strings are real i18n keys in a new demo domain file, translated into
all 27 locales. Both dashboards open it instead of the old DemoBanner, which
leaves the theme and format baselines with its file and the test non-null
assertions it held.
The client is about to lock settings sections for the shared demo account,
and it has to ask the same question the server does. DEMO_ACCOUNT_EMAILS and
isDemoAccountEmail now sit in @trek/shared, and the server's DEMO_EMAILS is
built from them instead of spelling the two addresses a second time.
A demo instance now runs with one explicit addon policy (demo-addon-policy.ts)
instead of whatever its admin last switched: every trip, global and
integration addon on, and off only what reaches a server of the visitor's
own, the five document sync providers, wanderer under Tours, the Immich and
Synology Photos providers, AirTrail and Dawarich. The switches under the
addons (Tours maps, piste routing, bag tracking, the collab features) are on.
The demo seed applies it, together with the demo account's English
language, on every boot and after every reopen, and the hourly reset applies
it once more after the baseline swap, in one transaction each. A test pins
that every seeded addon and provider is named in exactly one list, so a new
addon has to be classified before it ships.
Every visitor of a demo instance signs in as the same account, so one
visitor's API key, MCP token, OAuth client, passkey or new email address
became everybody's. These writes now answer 403 for the demo account through
refuseForDemoAccount in demo-write.ts: creating API keys, creating and
deleting MCP tokens, registering, rotating and deleting OAuth clients,
approving an app and ending a connected app, changing the name, address or
picture, and enrolling, renaming or deleting passkeys. Password, 2FA,
account deletion, sessions and the own AI connection were refused already.
The demo account's language is fixed to English: a single settings write of
the language answers unchanged, a bulk save drops it and keeps the rest. The
MCP settings tool is held back for the demo account by the tool gate, as
every write tool is. The admin of a demo instance and every real instance
are untouched.
The sidebar now follows the way a user meets TREK: account, everyday use and
settings before planning, Tours beside Road trip and wanderer, Studio beside
Journey, and an AI and MCP section led by the assistant pages. Admin pages
follow the real tab order, with Backups and the Audit log at its end.
Each of these is a screen or a connection users meet in 5.0 that no page
explained: the trip's People, Sharing and Journal areas, the inbox page and
bell, what the phone layout offers and leaves to the desktop, and connecting
and watching a wanderer instance.
Add the paddle region alert, the to-do reminders, the new audit keys for Tours, paddle water, wanderer and AI, the release history paging and the backup hour hint. Dashes are replaced with plain punctuation.
Packing template deletes ask for confirmation, plugin versions can be changed and updates paused, and the toolbar buttons are icon buttons. Dashes are replaced with plain punctuation.
The addons are searchable and filterable in cards per type, Tours is on by default and carries the wanderer, hiking map, cycling map and piste router switches, MCP, AI Parsing and the AI Assistant share one AI tile with a setup dialog, and the paddle water card loads regions automatically or on request. The overview lists the AI Assistant addon.
Add the Reset 2FA action of the edit user dialog, the invite delete confirmation, the Link a journal permission and the phone admin sections. Replace dashes with plain punctuation.
Document the demo welcome popup, name wanderer and the AI assistant hosts in the SSRF guard lists, and list the paths a restricted reverse proxy must forward for MCP sign-in.
Describe the world map login panel, the first-run admin form, the plain HTTP cookie warning, the SSO error messages, the reset page states and the Reset 2FA action in the admin user dialog. Disabling 2FA accepts a backup code.
The Account tab is a stack of separate cards, the Integrations tab gained the wanderer connection and the AI connection, and the notification forms have test buttons. Dashes in the text are replaced with plain punctuation.
The offscreen container is built inside buildGlMap's closure, where the
let lost its narrowing; a const of its own keeps it an HTMLDivElement.
The stage tests hand the stage a t that returns the key, so the toast
reads map.webgl2.terrain there. One floating promise fewer lowers the
eslint baseline to 164.
MapLibre 6 and Mapbox GL 3 both throw out of their Map constructor without
WebGL2. Every place that builds a GL map now asks one helper first,
components/Map/glSupport.ts, and a browser without WebGL2 gets the Leaflet
map with the user's tiles instead of a map that fails to start:
- the trip map (MapViewAuto) and the journey maps (JourneyMapAuto) pick
their renderer through useGlRenderer, which folds in the Mapbox token
rule they each had;
- the settings preview, desktop and phone, shows the Leaflet preview
through useMapSettingsForm and says why under it;
- the tour planner keeps the 2D map when 3D is asked for and says that
3D needs WebGL2;
- the PDF route map and the Leaflet vector basemap already fell back and
now report a GPU failure too.
Behind the probe, buildGlMap wraps each constructor (MapViewGL,
JourneyMapGL, the settings preview, the terrain map, the PDF map). A GPU
that claims WebGL2 and still refuses the context is reported through
utils/webgl, which turns the answer to no and tells every mounted map,
so they switch to Leaflet without a reload; any other error is still
thrown. Two new strings, map.webgl2.fallback and map.webgl2.terrain, in
all 27 locales.
GHSA-jrc7-96c5-q579 covers every maplibre-gl up to 6.4.0 and has no 5.x
release, so npm audit kept flagging the patched 5.24.0. Moving to 6.15.0
retires the local backport: patch-maplibre.mjs and its postinstall,
predev and prebuild hooks go, along with the Dockerfile copy and its
format-baseline entry. 6.x sanitizes attributions against an allow list.
What the major changes for TREK:
- ESM only, no default export: the engine module imports the namespace
and hands that out as its default, so every caller stays as it was.
- The worker is no longer inlined, and inside a bundle MapLibre cannot
find it next to its own module. The engine sets it from Vite's
?worker&url build before any map, the Leaflet bridge's included.
- WebGL2 only: the probe asks for webgl2 alone, so a WebGL1-only browser
gets the raster basemap instead of a GPUInitializationError.
- 6.9 shapes right-to-left text itself and deprecated setRTLTextPlugin;
the old plugin would replace that and could only run through eval in
the module worker, so MapLibre no longer registers it. Mapbox does.
The lockfile moves maplibre-gl and its own dependencies only; the
Leaflet bridge already allows ^6.0.0 as its peer.
The quote inside a reply sat on one line with an ellipsis, and that unbroken
line set the bubble's minimum width, so a long quote pushed the bubble past
the edge of the chat and cut its words off. It wraps now and is cut after two
lines, on desktop and on the phone.
A stay taken into a journal now has a card of its own, so ASTTOOL-001
looks for its tool; Dawarich's tools are listed only with its addon on,
which the test now turns on beside the others.
The welcome chose the journal's words whenever the Studio had not
registered as the chat's host, which also hit the Studio's own panel
before (or without) that registration. A journey's chat now speaks of the
journal only where the window or the phone's chat screen say they show it
for its journal; everywhere else, the Studio included, it stays the book's.
AssistantWindow had gone past the complexity limit choosing between the
planner's trip, the journal on screen, a picked chat and the home chat;
that choice now lives in windowChat.
The journal header's tool list was a value read only for its type, so it is
a union now. A journey's session reads the entries a call names side by side,
the photo announcements go out together, and a photo move links its photos
one after another through a promise chain, so each still takes the next
place among the target's photos.
isHeldInJourney took every tool of the journey's controllers from the Studio
side, delete_journey among them. The Studio never offered it (its writes are
the book's), so leaving it unattached changes nothing the model sees and
keeps the denial in one place.
A Dawarich stay's row read fields one side of a union did not have, the
photo cards narrowed an entry's day through an optional chain, and the
settings service cast a journey row straight to a record. Helpers only
their own file reads are no longer exported, and the photo tests seed
their rows with insertRow like the other journey suites.
The four new photo link tools join the journal's tools: the journey chat
offers them to editors, the trip chat loads them with the journey group and
the home chat keeps none. A journeyPhotos card shows the photos as
thumbnails with the entries they leave and reach, the first one marked as
the cover in a new order, and a caption old to new; a library add shows the
photos once they landed. A linked photo comes off again, an unlinked one
goes back, moved photos go back and a caption is written back, each while
nothing changed since. A new order of an entry's photos and a library add
have no way back yet. The journal prompt names the photo tools.
update_journey takes show_trip_tracks and cover_gradient and answers
previous, update_journey_entry answers previous, reorder_journey_entries
answers previous_order, and the four photo link tools are listed.
Linking a gallery photo to an entry and taking it off told nobody, so other
tabs, and the journey page after an assistant change, stayed as they were.
Both now go through a photo links service that broadcasts
journey:entry:updated after the write, with the asking socket left out on
REST; a caption change through PATCH photos/:id broadcasts journey:updated
the same way. The service also moves photos from one entry to another of
the same journey in one transaction, as POST
/api/journeys/entries/:entryId/photos/move, with a strict contract. The MCP
tools link_journey_photo, unlink_journey_photo, move_journey_photos and
set_journey_photo_caption call the same methods; the caption tool answers
the caption it replaced.
The update card silently left out the tags and the pros and cons it was
about to write, so the user approved a change the card did not show. An
entry now shows its first tags and how its verdict leans, and an update
names the tags, each side of the verdict and the coordinates it changes.
A new order of the entries is drawn as the entries that move, in their new
places. A changed entry is written back from what the tool answered it
replaced, and a reorder is taken back to the order before, each while
nothing changed since. A Dawarich stay taken in as an entry gets the
entry card; it has no way back, since deleting the entry would leave the
stay marked as taken in.
update_journey_entry now answers previous, the entry's earlier values of
exactly the fields it names (a suggestion written up goes back to being
one), and reorder_journey_entries answers previous_order, the ids in the
order they stood in before. Both come from the service the REST routes
use, which now builds an entry's patch in one pure helper; the routes'
bodies are unchanged. The tools' inline zod moves into strict schemas in
@trek/shared.
Starting a journal, changing its settings, linking or unlinking a trip,
bringing back its suggestions, the user's own view of it and deleting it
fell back to the generic card, which names ids and offers no way back.
They now share a journeyHeader card: the journal with its trips, the one
the change links or unlinks marked, what a new journal brings in and what
a deletion or an unlink takes, and each setting old to new in the
journal's own words. A new journal is deleted again while nothing was
written, added or laid out in it, changed settings are written back from
what the tool answered, a linked trip is unlinked by its owner while none
of its suggestions was written up, and the own switch flips back.
The MCP tool lacked show_trip_tracks and cover_gradient, which PATCH
/api/journeys/:id has always taken. Both now read the input through one
helper, and the tool goes through a small settings service that writes the
same patch in a transaction with the stored values it replaces, answered
as previous, so a change can be written back. The subtitle and the cover
gradient take null to clear them. The tool's inline zod moves into strict
schemas in @trek/shared; the REST route and its body are unchanged.
On /journey/:id the assistant window and the phone's assistant screen now
open the journey's conversation, the one TREK Studio holds too, instead of
the home chat; the Studio route keeps its own column. A turn sent from a
chat the Studio does not host names the journal page, a hosted one the
Studio, so the server offers the right tools. A journal card from the home
chat opens that conversation in place on the desktop and the journal page
on the phone, and the empty chat asks about the journal rather than the
book.
Asked on /journey to create "Japan 2026", the home chat had nothing to
offer: it keeps no write that names a trip or a journey. It now keeps
create_journey, whose trips the tool checks one by one, and delete_journey,
which the journey's own chat leaves out because the conversation goes
with it. For anything inside a journal it offers open_journey_chat, the
twin of open_trip_chat: the journal is read from the user's own journals
and shown as a card whose button goes on in the journal's chat with the
question waiting there. The trip and journal cards share their question
and button.
Asked on a journey page to write an entry, the assistant said it could not:
the only journey chat was the Studio's, which keeps the journey's writes
out. A journey session now holds what either page offers and shows the
model only the asking page's tools. The journal page offers a viewer the
reads, an editor the entry, photo, recorded day and day import writes, and
the owner also the journey's own settings and trips; delete_journey stays
out of the journey's own chat. A few reads that name no journey come along
(place search, the photo libraries, Dawarich stays), the Dawarich journal
entry is pinned through its journalId, and an entry an argument names must
belong to the conversation's journey before the tool runs. The turn reads a
journal card and instructions of its own; the Studio keeps exactly what it
offered before.
The journey's assistant lives in TREK Studio today, and the journal page is
about to get one of its own over the same threads. The chat and decision
bodies take an optional surface, studio or journal, and the journey scope
carries it to the service; left out it reads as the Studio, so today's
client is unchanged. The scope key ignores it, so both pages keep one
thread list per journey.
The refusal test proposed two keeps in one thread under the same call id, so
the second decision met the first call, already decided, and no refusal came
back. Each proposal now has its own id, and the refused right is set after the
proposal so it is the apply that refuses.
Saving the custom connection lists the server's models first, and the fake
provider answered that with a 404, so the setup got 502. It answers /api/tags
the way the other assistant suites' fake provider does.
lint:test-sql counts every .prepare( under tests/ as a raw statement, and the
attachments service's prepare() was called five times in its unit test. The
test holds no SQL; the method takes a name that does not read as one.
The server built the answer through words.text(), words.count() and maps of
plain strings, so the check read all 13 assistant.attach keys as unused. The
helper's lookup is t() now (a numeric count picks the plural form, as the
client's does) and the chip and reason maps hold their keys under *Key
properties.
The chips offered to keep an attached file in Files, but nothing could do it:
the assistant's upload_trip_file stays denied, and the file never reached
Files. A new server write, keep_attached_file, is offered to the model where
the thread holds an attachment and held as a change card naming the file and
the booking, expense or place it goes with.
Applied, it runs the checks of a Files upload with the user's rights (the
demo block, the trip, file_upload and file_edit for a link, the instance's
file types, every linked id on the same trip), copies the bytes to a trip
file of their own, and writes the trip file, its expense link and the
attachment's pointer to it in one transaction; the members' planners get
file:created after it. A second keep of the same attachment answers the file
kept the first time. FilesService gains storeBytes and discardBytes so the
storage write stays out of that transaction.
The receipt opens the Files tab; the card's verb, its "Attached to" row and
the receipt's open label are new keys in all 27 locales. Undo is not offered:
no tool takes a trip file back yet.
The assistant could not read a file the user attached: the bytes were gone
after the request, the thread only held the file's name, and read_trip_file
stays denied because it would send any document of the trip to the provider.
A new server tool, read_attached_file, reads only the files attached in this
conversation by this user; any other id is answered "No such attachment in
this conversation.", whoever it belongs to. It is offered in a trip turn whose
thread holds a file with text. The first read of a file is held as a change
card that names the file and the model and host its text would go to; once
the user agrees, the agreement is kept on the attachment and later reads run
at once. The text reaches the model through the history and, for a later
read, as a tool answer only the model gets (AgentToolRun.modelText): no message
part, stream chunk, shared answer or client store ever carries it.
The model is told each kept file's attachment id, and the standing
instructions say to read the file when the user wants to work with it, to
propose a booking found in it through create_reservation and then to offer
keeping the file with it. The card's verb, its "Sent to" row and the step's
name are new keys in all 27 locales.
A booking file without a booking ended the answer on the error "No booking
found in this file." with no action, and the user was left to type. Now the
server answers every booking file itself, without a model: one line that
names the file and what TREK found in it, or that it found nothing, why, and
asks what to do with the file. Under it come up to three chips, picked from
what the user may do on the trip and what the file allows: keep it in Files,
make a booking from its text, attach it to a booking or an expense, ask what
is in it, or keep it with the booking just found. A chip sends its words as
an ordinary chat message.
The line and the chips are written in the language of the request from new
assistant.attach.* keys in all 27 locales, through a small helper the sports
proposals now share. The empty answer finishes as stop; no_bookings_found
stays in the contract for the threads older builds wrote. Route files keep
their answer as it was.
The chat's booking reader kept only the bookings of the import preview and
dropped its warnings, so "AI reading is not set up", "the AI failed",
"a scan without text" and "nothing in it" all came out as the same
no_bookings_found. The reader now reads the reason from the preview (the mode
it ran in and the AI fallback's warnings) and keeps it in the stored read
step, and the note the model is sent names it.
A file attached in the trip chat lived only as a buffer for the one request
that read it, so the assistant could not look at it again and nothing could
keep it in Files. Now its bytes go to storage before the turn writes anything
(under the trip files' category, which is never served by path) and a row in
the new assistant_attachments table keeps its name, type, size, hash and the
text TREK read out of it, written in the turn's opening transaction. The
message's attachment part carries the row's id; old messages without one still
read.
The text is read the way the AI import reads it, through a new public
LlmParseService.readText, and a file whose text cannot be read is kept all the
same. A storage that refuses the bytes refuses the turn before any row exists.
Deleting a thread takes its files along, and the nightly retention sweep drops
files older than seven days and bytes no row names any more.
AssistantService sat three lines under the size limit, and the attachment
work ahead grows exactly this answer. Moving it into AssistantFileAnswers
leaves the rows and their transactions with the service, which hands the
answer the two ways it writes its first row. Nothing the user sees changes.
2026-10-24 22:32:47 +02:00
3098 changed files with 142248 additions and 13596 deletions
# The stable release. Dispatched by hand on main, it:
# 1. refuses to run unless ci-ok passed on the main commit it releases,
# 2. works out the version from the tags and the bump input, commits the
# bump to package.json, the lockfile and Chart.yaml, tags it and pushes,
# 3. builds both platforms from exactly that bump commit (never from
# 2. works out the version from the tags and the bump input,
# 3. asks the TREK services issuer for the registration dialect of that
# version and checks it, in a job without write access
# (registration-dialect.yml); without one the release ships the dialect
# already on main, but only while that one still has 90 days to run,
# 4. commits the bump to package.json, the lockfile, Chart.yaml and the
# dialect, tags it and pushes,
# 5. builds both platforms from exactly that bump commit (never from
# whatever main points to by then) and publishes :VERSION,
# :MAJOR.MINOR, :MAJOR and :latest,
# 4. publishes the Helm chart from the same commit.
# 6. publishes the Helm chart from the same commit.
#
# Big or risky changes go out as a prerelease first (docker-dev.yml, tagged
# :latest-pre and :<next>-pre.N); bump=auto here then finalizes that line.
# Without a prerelease in flight there is no default: the run stops until
# someone picks patch, minor or major.
#
# Each stable release registers new installations in a dialect of its own,
# which the issuer accepts for 180 days, and for as long as it is the newest
# release's; renewing a key needs none. dev gets the module when main is merged
# back into dev after the release; a prerelease asks the issuer for one of its
# own (docker-dev.yml). The issuer's side, and what to set up before the first
# release with a dialect, is in the operations guide of the TREK services.
on:
workflow_dispatch:
@@ -26,6 +38,14 @@ on:
description:"Type MAJOR (all caps) to confirm a major release"
type:string
default:''
ci_from:
description:'Take ci-ok from this earlier commit instead of the one on main, when main differs from it only under .github/ (a fix to a release workflow needs no second test run)'
type:string
default:''
allow_stale_dialect:
description:'Ship without a dialect of its own even if the one on main is the original one or has under 90 days left'
type:boolean
default:false
permissions:
contents:write
@@ -36,13 +56,16 @@ concurrency:
cancel-in-progress:false
jobs:
version-bump:
version:
if:github.repository == 'liketrek/TREK'
runs-on:ubuntu-24.04
timeout-minutes:120
timeout-minutes:30
permissions:
contents:read
checks:read
outputs:
version:${{ steps.bump.outputs.VERSION }}
sha:${{ steps.bump.outputs.SHA }}
base:${{ steps.base.outputs.SHA }}
steps:
- name:Release from main only
if:github.ref != 'refs/heads/main'
@@ -50,17 +73,17 @@ jobs:
echo "::error::Dispatch this workflow on main, not on $GITHUB_REF."
report "$VERSION ships without a dialect of its own and registers in dialect $id from $created, which the issuer takes for new installations for $left more days at least. Its operator can write one by hand for the next patch." "::warning::"
exit 0
fi
if [ "$id" = 0 ]; then
what="the original registration, which the issuer accepts only until the end of its transition period"
elif [ -n "$left" ]; then
what="dialect $id from $created, which has $left days left"
else
what="a dialect whose id or date cannot be read from $target"
fi
if [ "$ALLOW_STALE" = true ]; then
report "$VERSION ships without a dialect of its own and registers in $what, as allow_stale_dialect asked." "::warning::"
exit 0
fi
report "$VERSION got no dialect of its own and would register in $what. Nothing was committed or pushed. Make sure the issuer runs with its dialect seed and release token and that TREK_DIALECT_TOKEN is set, then run the release again, or tick allow_stale_dialect to ship anyway." "::error::"
exit 1
- name:Update the version, commit, tag and push
id:commit
env:
NEW_VERSION:${{ needs.version.outputs.version }}
run:|
set -eu
# Update all workspace + root package.json files and the root lockfile in one shot
npm version "$NEW_VERSION" --workspaces --include-workspace-root --no-git-tag-version
sed -i "s/^version: .*/version: $NEW_VERSION/" charts/trek/Chart.yaml
<img src="docs/subtitle-dark.png" alt="your trip. your plan." height="28" />
</picture>
### Every part of your trip. One app you own.
A self-hosted, real-time collaborative travel planner — with maps, budgets, packing lists, a journal, and AI built in.
TREK has grown from a travel planner into a whole ecosystem for trips: plan together in real time on a live map, build hikes, tours and road trips, keep bookings and shared costs in order, pack, write the journal, lay out the photo book and hand the busy work to the built-in AI. Self-hosted, on the web, on your phone and offline.
<br />
@@ -32,11 +27,16 @@ A self-hosted, real-time collaborative travel planner — with maps, budgets, pa
@@ -182,7 +182,7 @@ Most of what follows is an addon an admin switches on or off. Lists, Costs, Docu
#### 📱 Mobile and offline
- **Installable**: iOS and Android straight from the browser, no App Store, standalone with a themed status bar
- **Apps**: TREK Travel Planner for iPhone, iPad and Android connects to any TREK server, and the browser installs TREK as a standalone app with a themed status bar too
- **Phone layouts**: their own screens for every part of the app, with a bottom bar and dashboard widgets the user orders
- **Offline reads**: the app shell is precached and trips, places, tours, files and nearby places live in a per-user IndexedDB store, so a trip opens with no network
- **Offline writes**: changes queue and replay with an `X-Idempotency-Key`, and a stale edit waits for a keep-mine or keep-theirs choice
@@ -200,7 +200,7 @@ Most of what follows is an addon an admin switches on or off. Lists, Costs, Docu
- **Admin panel**: users and the permission matrix, defaults for new users, categories and templates, addons, plugins, storage, notifications, MCP tokens, backups, audit log and releases
- **Backups and storage**: scheduled backups with retention that restore their own secrets, and local or S3-compatible storage per content category with replicas
- **Appearance**: light, dark or follow the OS, seven colour schemes plus a custom accent, transparency, density, reduced motion and text size
- **27 languages**: en, de, es, fr, it, nl, hu, ru, zh, zh-TW, pl, cs, ar (right to left), br, id, tr, ja, ko, uk, gr, sv, vi, ca, az, et, sk, th
- **28 languages**: en, de, es, fr, it, nl, hu, ru, zh, zh-TW, pl, cs, ar (right to left), he (right to left), br, id, tr, ja, ko, uk, gr, sv, vi, ca, az, et, sk, th
- **Help center**: a ? in the top bar with 381 step-by-step guides with screenshots, searched together with the wiki that ships in the image
</td>
@@ -214,7 +214,39 @@ Most of what follows is an addon an admin switches on or off. Lists, Costs, Docu
**Every TREK installation in the world searches through servers that [dataforest cloud](https://cloud.dataforest.net/en?mtm_campaign=trek) sponsors.**
The place index with over 88 million places, the address search in every TREK language and the planet-wide routing all run on [dataforest cloud](https://cloud.dataforest.net/en?mtm_campaign=trek) in Frankfurt, Germany, free of charge for everyone who self-hosts TREK. Without this sponsorship, these services would not exist in this form.
Thank you, [dataforest](https://cloud.dataforest.net/en?mtm_campaign=trek), for backing open source this generously. 💚
Open `http://localhost:3000`. On first boot TREK seeds an admin account — if you set `ADMIN_EMAIL`/`ADMIN_PASSWORD` those are used, otherwise the credentials are printed to the container log (`docker logs trek`).
Open `http://localhost:3000`, or `http://<server-ip>:3000` from another machine. On first boot TREK seeds an admin account — if you set `ADMIN_EMAIL`/`ADMIN_PASSWORD` those are used, otherwise the credentials are printed to the container log (`docker logs trek`).
`COOKIE_SECURE=false` is only for trying TREK over plain HTTP: without it the browser drops the session cookie and the login does not stick. Once TREK sits behind HTTPS (see [Reverse Proxy](#reverse-proxy)), remove it.
> [!TIP]
> **No server yet?** TREK's own public services run on [**dataforest cloud**](https://cloud.dataforest.net/en?mtm_campaign=trek), our infrastructure sponsor. A small seed is plenty for a TREK instance and is ready within a minute.
<div align="center">
@@ -297,7 +335,7 @@ See [`charts/README.md`](https://github.com/liketrek/TREK/blob/main/charts/READM
<h2 id="install-as-app-pwa">Install as App (PWA)</h2>
TREK works as a Progressive Web App — no App Store needed.
TREK works as a Progressive Web App, and the TREK Travel Planner app for iPhone, iPad and Android connects to the same server ([Native Apps](https://github.com/liketrek/TREK/wiki/Native-Apps)). To install the web app instead:
1. Open TREK in the browser (HTTPS required)
2.**iOS**: Share ▸ *Add to Home Screen*
@@ -395,16 +433,6 @@ Caddy handles TLS and WebSockets automatically.
@@ -5,9 +5,10 @@ This is a minimal Helm chart for deploying the TREK app.
## Features
- Deploys the TREK container
- Exposes port 3000 via Service
- Optional persistent storage for `/app/data` and `/app/uploads`
- Optional persistent storage for `/app/data` and `/app/uploads`, plus extra volumes and mounts
- Configurable environment variables and secrets
- Optional generic Ingress support
- Optional Gateway API HTTPRoute support
- Configurable health checks (liveness/readiness probes) on `/api/health`
## Helm Repository
@@ -27,7 +28,7 @@ helm install trek trek/trek
Or install directly from the local chart:
```sh
helm install trek ./chart \
helm install trek ./charts/trek\
--set ingress.enabled=true\
--set ingress.hosts[0].host=yourdomain.com
```
@@ -47,7 +48,7 @@ See `values.yaml` for more options.
-`JWT_SECRET` is managed entirely by the server — auto-generated into the data PVC on first start and rotatable via the admin panel (Settings → Danger Zone). No Helm configuration needed.
-`ENCRYPTION_KEY` encrypts stored secrets (API keys, MFA, SMTP, OIDC) at rest. Recommended: set via `secretEnv.ENCRYPTION_KEY` or `existingSecret`. If left empty, the server falls back automatically: existing installs use `data/.jwt_secret` (no action needed on upgrade); fresh installs auto-generate a key persisted to the data PVC.
- Liveness and readiness probes are configurable under `probes`. Overrides are merged over the defaults, so `--set probes.liveness.initialDelaySeconds=60` keeps the `httpGet` handler. When switching to `exec` or `tcpSocket`, clear `httpGet` in the same override. Set `probes.liveness: null` to remove a probe.
- If using ingress, you must manually keep `env.ALLOWED_ORIGINS` and `ingress.hosts` in sync to ensure CORS works correctly. The chart does not sync these automatically.
- If exposing TREK through an Ingress or HTTPRoute, set `env.ALLOWED_ORIGINS` and `env.APP_URL` to the public URL. The chart does not sync these automatically.
- Set `env.ALLOW_INTERNAL_NETWORK: "true"` if Immich or other integrated services are hosted on a private/RFC-1918 address (e.g. a pod on the same cluster or a NAS on your LAN). Loopback (`127.x`) and link-local/metadata addresses (`169.254.x`) remain blocked regardless.
- Set `env.ALLOW_LINK_LOCAL_IPS` to a single link-local address (e.g. `"169.254.1.2"`) only when a service TREK needs sits behind one, such as a host gateway. Cloud metadata addresses (`169.254.169.x`, `169.254.170.x`) cannot be listed.
-`FORCE_HTTPS` is optional. Set `env.FORCE_HTTPS: "true"` only when ingress (or another proxy) terminates TLS. It enables HTTPS redirects, HSTS, CSP `upgrade-insecure-requests`, and forces the session cookie `secure` flag. Requires `TRUST_PROXY` to be set.
@@ -55,3 +56,5 @@ See `values.yaml` for more options.
-`COOKIE_SECURE` is auto-derived (on when `NODE_ENV=production` or `FORCE_HTTPS=true`). Set `env.COOKIE_SECURE: "false"` only during local testing without TLS. **Not recommended for production.**
- Web Push needs no chart settings: TREK generates its key pair on first start and keeps it in the database on the data PVC. To bring your own pair, set `env.VAPID_PUBLIC_KEY` and `secretEnv.VAPID_PRIVATE_KEY` together (and optionally `env.VAPID_SUBJECT`); changing the pair later means every device has to turn push on again.
- Set `env.OIDC_DISCOVERY_URL` to override the auto-constructed OIDC discovery endpoint. Required for providers (e.g. Authentik) that expose it at a non-standard path.
-`podSecurityContext` defaults to `fsGroup: 1000` with `fsGroupChangePolicy: OnRootMismatch`, and `securityContext` (the container's) is empty. The container then starts as root, hands `/app/data` and `/app/uploads` to `node` (uid 1000) if the directory or anything directly inside it is not owned by `node` yet, and drops to `node`. Deeper files are not checked. Only the owning user is changed (plus a `root` group), so a different `fsGroup` stays in place. A volume that refuses `chown` (NFS with `root_squash`) logs a warning instead; set `runAsUser: 1000` then. To run as `node` from the start, for example under the `restricted` Pod Security Standard, set `podSecurityContext.runAsUser: 1000` and `runAsGroup: 1000`; the image then skips that step and the privilege drop. See the example in `values.yaml`.
- Use `extraVolumes` and `extraVolumeMounts` to add your own volumes to the pod. To give the booking import more KItinerary extractors, mount a ConfigMap at `/usr/local/share/kitinerary/extractors`, with one `.json` and one `.js` key for each extractor; see the example in `values.yaml`.
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.