Both server ratchets only logged an entry the tree had moved past. A
file that shrank could grow back to its old size entry unseen, a
deleted file left its allowance to the next file at its path, and a
boundary violation that was fixed stayed allowed until someone ran
--update. The client ratchets now fail on such entries, so the server
ones do the same.
lint:size fails on an entry above its file's size or for a file that is
gone, and lint:boundaries on an entry that no longer occurs; both tell
to run --update, which still only lowers or drops. Both baselines match
the tree today.
Each gate came with its npm script but ran nowhere. The server job now
runs lint:size, lint:boundaries and shared's contracts:open after the
shared build; the client job runs lint:warnings in place of lint:check
(it fails on errors the same way and also holds the warning count per
rule), then lint:format, lint:layers, lint:offline, lint:skips and
lint:i18n-keys. All of them are steps of jobs ci-ok already needs.
Both size baselines are taken on the merged branch. Four server files
grew while the other changes landed (the claimed reminder, the email
uniqueness error and the generator rule for an expression index) and
four client files by two or three lines with the plural forms; one
client file shrank and is lowered.
Nothing said which Node TREK runs on. The image is node:24, part of CI
ran 22 under comments saying 22 matched the server runtime, and the dev
guide said "22+". A bare-metal or dev install on an unsupported Node
installed without a word and then failed at runtime: isomorphic-dompurify
needs ^22.22.2 or ^24.15.0, the plugin jail's --permission flag is new
in 22.x.
.nvmrc now holds the major the Docker image runs, for nvm/fnm and for
every workflow to read. engines in the root and each workspace states
the range the runtime dependencies accept, the same one
isomorphic-dompurify declares. Without engine-strict npm only warns
outside it, so no existing install stops installing.
The image scan has been red since 10-06 on five advisories in four
runtime packages: proxy-addr 2.0.7 (critical, IP spoofing through an
IPv4-mapped IPv6 trust subnet), brace-expansion 2.1.4, source-map-js
1.2.1 and @modelcontextprotocol/sdk 1.30.0. brace-expansion had been
fixed once already; the ORM squash regenerated the lockfile and took it
back to 2.1.4 without anything noticing.
The overrides block in server/package.json was meant to hold such floors,
but npm reads overrides from the root package.json only, so none of them
ever applied: @nestjs/platform-ws kept its own ws 8.21.2 next to the
8.21.3 the block asked for. The floors move to the root, together with
new ones for every package this lifts, so the next lockfile regeneration
cannot quietly undo them. picomatch ^4 is dropped instead of moved: at
the root it would force micromatch's 2.x copy across a major. js-yaml is
lifted only inside its 5.x line, the one @nestjs/swagger pins.
The SDK goes to 1.31.0, the first fixed release; it stays hoisted, which
the CJS path aliases in server/tsconfig.json depend on. dompurify,
fast-uri, ip-address and js-yaml come along because npm audit flags the
same lockfile for them and each fix is inside the range already used.
@mikro-orm/postgresql (with pg), marked and topojson-client are dropped:
nothing in src, tests or scripts imports them, and the ORM config names
only the SQLite driver. The Postgres driver comes back with a Postgres
job that exercises it.
npm audit --omit=dev now reports only maplibre-gl, whose fix is the
6.x major tracked separately.
The open-request-schemas header pointed at scripts/open-request-schemas.spec.ts,
which does not exist; the spec lives in src/. The bracket-notation comment in
the server ESLint config kept a dash that the earlier cleanup on this branch
missed.
The foreign key index check took every index from PRAGMA index_list as a
usable prefix, partial ones included. An index like
CREATE INDEX ix ON c(p_id) WHERE state = 'open' would satisfy FKIDX-001,
yet the cascade lookup WHERE p_id = ? still scans the table, which is the
regression the check exists to catch. A partial index now counts only when
its predicate is IS NOT NULL on the key's own columns, the one form SQLite
can prove from p_id = ?. Nothing in today's schema depends on a partial
index alone, so the result on the migrated snapshot is unchanged.
contracts:open in shared only counts exports named *RequestSchema, but
the name is a convention: 37 body DTOs wrap a schema called *BodySchema,
*InputSchema or plain *Schema, and tags.dto.ts declares its schemas
locally. pluginInstanceConfigUpdateSchema (a record of unknown behind
PluginConfigDto), createAlbumLinkSchema with five open shapes and a dozen
more were never counted, and a new open body under one of those names
passed both checks.
tests/unit/contracts/dto-open-shapes.test.ts walks the other way round.
It finds every class that extends createZodDto under src/, imports it,
and counts the open shapes in the schema the Zod pipe validates with
shared's walk. Each class is held to scripts/dto-open-shapes-baseline.json,
61 DTOs with 101 open shapes today, and a new DTO may hold none. It fails
closed: a createZodDto call that is not an exported class, two DTO
classes with one name, an empty tree or an unreadable baseline stop it.
It runs with the unit tests; npm run contracts:dto-open runs it on its
own and --update only lowers the baseline.
The five DTO classes declared inside roadtrip controllers are exported so
the walk can reach their schemas. The selector test picks up prettier's
wrapping of the cases added before.
MaintenanceRepository, DemoRepository and six seeders are exempt from the
SQL string selectors as whole files, so new SQL text added to any of them
passed lint unseen, and nothing noticed when a file no longer needed its
exemption. The comment promised a list that can only shrink, but nothing
held it to that.
The selector test now works out which files under src/db/repositories and
src/db/seeders the config really exempts and compares that set with a
pinned table, then runs the SQL selectors over each exempt file's source
and compares the hits with the count pinned for it (23 today). One more
statement fails, one fewer asks for the pin to come down, and a file at
zero has to leave the ignores. Per-line disables would have done the same
job but add suppressions the codebase is meant to lose.
The outbound-timeout selectors flagged safeFetch(url, undefined,
{ timeoutMs }) and safeFetch(url, { headers }, { timeoutMs }), although
timeoutMs is the deadline over the whole exchange that server/CLAUDE.md
tells callers to pass. At the same time an init spelled null or void 0
passed unseen, so a guarded call could still go without any timeout.
A call whose third argument is an object literal with a timeoutMs key is
now exempt, and undefined, null and void 0 (also behind an as cast) all
count as no init. The cast forms pin the inner node type, because esquery
compares String(value) and a missing name would otherwise read as
'undefined' and flag every cast init. The selector test covers both
sides.
The repository rule that keeps dialect SQL in sql-functions.ts only
matched raw() and the sql tag, so a SQL string handed straight to the
connection, em.getConnection().execute('INSERT OR IGNORE ...'), passed
unseen, and the seeders sat outside every repository rule. A new block
refuses a string or template whose text starts like a statement as the
first argument of execute, run, all or get in src/db/repositories and
src/db/seeders; a result mode such as qb.execute('run') stays legal.
The files that already hold such SQL are named one by one so the list
can only shrink: MaintenanceRepository (PRAGMA, VACUUM INTO),
DemoRepository and six of the seven seeders.
The selector groups are named once in eslint.config.mjs, and the
selector test now also checks the new block per path, its exceptions,
and that raw() stays forbidden in repositories.
journey-gallery-order.ts exported raw SQLite ORDER BY text from
src/nest that no production code read any more since the gallery reads
moved onto JourneyPhotosRepository. Two parity tests still compare the
Kysely rebuild against it, so it moves to tests/helpers as their oracle.
The ESLint rule from #2610 that requires a signal on outbound fetches
only matched the platform fetch. The guarded wrappers hand it a spread
init, which the rule trusts, so safeFetch(url, undefined, options) passed
lint with no timeout, and the dispatcher left undici's five-minute
default in place. Nothing capped the body either. Two callers did exactly
that on hosts TREK does not control: the Synology thumbnail read (a
user-configured NAS) and the place enrichment photo download (a URL out
of third-party place data), both buffering the whole answer.
The wrappers now carry defaults. A call without a signal waits 30 s for
the response headers; the body stays under undici's idle timeout, so a
proxied video that pauses with its player is not cut off. timeoutMs adds
a deadline over the whole exchange and joins the caller's signal, once
for the whole redirect chain. Every body is capped at 64 MB, refused up
front when the declared length is larger and failed mid-read otherwise;
callers pass a tighter maxBytes, or null where the body is streamed on
(pipeAsset, the doc-sync transfers that count their own bytes, the
Ollama pull).
The two call sites get a signal and an 8 MB cap. The lint selectors now
cover safeFetch, safeFetchFollow and safeFetchAdminConfigured as well,
including an explicit undefined init, and are named once so the
repository block restates the same list. A new test asks ESLint for the
options it applies to a path and runs them over snippets, so a selector
that stops matching fails a test instead of passing silently.
coverage-thresholds.test.ts imported vitest.config.ts, which pulled the
config into tsconfig.tests.json's program; that program compiles as
CommonJS and rejects the config's import.meta, so typecheck:tests
failed. The test now loads the config at runtime. The other new gate
tests get prettier's formatting.
google-quota, managed, mcp-shared, mcp-transport, public-api,
receipt-scan, route-usage, scheduling, school-holidays, tokens and tours
had no entry in the per-domain ratchet, so they sat in the pooled
src/nest/** catch-all, where a domain can lose twenty points behind the
headroom of the files beside it. That includes the cron registrar, the
MCP HTTP transport, the public API and the new tours domain.
Each is pinned at floor(measured) minus 1 as scripts/coverage-thresholds.mjs
prints it, measured over the unit, integration and e2e suites that load
the domain. A full run can only measure higher. mcp-shared (59 branches)
and route-usage (66 branches, 77 functions) measure below the 80 floor
and are pinned where they are; managed holds an empty module.
A new unit test fails when a folder under src/nest has no entry, or an
entry names a folder that is gone, so the next domain cannot land in the
catch-all again.
Migration20200101042200 wrote 86 indexes by hand so that deleting a trip
or a user stops scanning every child table while it holds the single
connection. Nothing kept it that way: the next table with an unindexed
trip_id would bring the scans back, and the entity parity tests compare
index names, not whether a key is covered.
The test reads PRAGMA foreign_key_list off the migrated test database
and requires the columns of each foreign key to lead some index or the
primary key. Today every one does, so the allow-list for deliberate
exceptions starts empty, and a second case keeps that list honest. A
third builds a small schema to prove the check fails on an unindexed
key, a composite one, and an index that holds the column second.
Module boundaries were held by review alone. storage.module.ts documents
a cycle that only surfaced at runtime as "imports[1] is undefined", three
modules exist just to break cycles, and nothing stopped the next feature
from reaching into another domain's helpers.
scripts/import-boundaries.mjs parses every import under src/ with the
TypeScript parser and checks five rules against a baseline of today's
violations: load-time file cycles (none today once type-only and lazy
imports and the two-way MikroORM entity relations are left out), new
dependencies between domains that already depend on each other (262
pairs, one tangle around plugins and auth), shared-kernel files importing
a domain (8), imports of another domain's internals (69) and src/db
importing src/nest (5). A domain's public surface is what the code
already treats that way: modules, services, guards, decorators, types,
constants, the audit logger, the plugin RPC kit and the addon gates.
New violations fail, --update only drops entries, and an unreadable
baseline, a missing src/ or an import it cannot resolve stops the run.
The client has had lint:size since #2612, the server had nothing, and
the server is where the biggest files live: maps.service.ts holds 3231
lines and grew by 200 in a month, and nine more services are past 1100.
scripts/size-lint.mjs carries the client's semantics over to src/ and
scripts/: a file may hold 1000 lines, the 23 that already hold more are
pinned at today's length in scripts/size-baseline.json, and --update
only ever lowers an entry.
A line past prettier's 120 columns counts once per 120 columns it spans,
so joining lines does not make room under the limit. The check fails
closed: a missing or malformed baseline or a missing src/ stops it
instead of passing with nothing checked.
The backfill in Migration20200101042100 marks reminders as sent by UTC
date('now'), while the reworked job counts days in the app TZ, which read
like a seam that sends twice or skips on upgrade day. It is not: the
backfill stands in for the job before it, which matched start_date
against UTC date('now') at 9 AM app time, so UTC is the clock those
reminders went out by, and with one run per UTC day nothing it marks
was left unsent. The only effect is its strict "<": the upgrade day's
own reminders stay unmarked because the boot may precede that day's
run, so a boot after it reminds such a trip once more. One extra
reminder at most and never a lost one, so the shipped migration stays
untouched and the reasoning now sits on the job.
The todo claim's docstring writes its dedup bound as SQL.
Login, password reset and the sign-up checks all match emails
case-insensitively, but the only constraint was the column's own
UNIQUE, which SQLite compares case-sensitively. Two accounts whose
addresses differ only in case (legacy rows, two sign-ups at once) left a
login to resolve to whichever row came first, and LOWER(email) lookups
had no index to use.
A migration adds a unique index on lower(email). An install that
already holds such a pair must still start, so the migration then
leaves the index out, logs how many addresses are shared and by which
user ids (not the addresses), and records email_case_index_pending.
Since a migration runs once, db/email-case-index.ts runs on every boot
while that flag is set: it repeats the warning until the accounts are
told apart, then creates the index and clears the flag. The app-level
checks keep guarding meanwhile.
A write that passes the up-front check and still loses the race is now
refused by the database. The users repository turns that unique
violation into UserIdentityTakenError, and sign-up, admin create and
edit, and the profile update answer it with the same 409 their checks
give, instead of a 500 or an unhandled rejection (ADMIN-SVC-094 pinned
the old behaviour and now pins the 409).
The entity generator could not see an expression index (PRAGMA
index_info names its column NULL), so PARITY-009c found a db index with
no entity declaration. Rule 14 carries such an index into the entity as
a named entry with its own statement; Users.entity.ts is regenerated.
The trip and todo reminder ticks sent with `.catch(() => {})` and then
marked the reminder sent whatever happened. A send that threw (a DB error
resolving the recipients) or whose every channel failed (SMTP down, no
in-app) was recorded as delivered and never retried, without a log line.
And because the mark came after the send, two processes running the same
tick (a rolling update overlap, a stray container on the same volume)
both sent.
Each reminder is now claimed first with one conditional UPDATE through
its repository: reminder_sent_for is set only while it differs from the
start date (and the start date is still the one read), reminded_at only
while it is empty or older than the 20 hour dedup window. Only the
process whose update lands sends. If the send throws, or it attempted
in-app and channels and none went out, the claim is given back, the
failure is logged, and the next tick tries again until the trip starts
or the todo falls out of its window. A send nobody wanted (everyone
opted out) stays marked, since a retry would reach no one.
For that, NotificationsService.send() now reports what it attempted and
what was delivered; a channel that rejects or answers false counts as
not delivered. Existing callers ignore the result.
The legacy-upgrade suites stopped at legacy-v242, while v4.3.3, the
release every 4.x install will come from, left schema_version 244. The
most common upgrade path only ever ran against synthetic step maps.
legacy-v244.sql is that release's own runner booted once on an empty
file, plus synthetic rows across the main tables: users (one with a
mixed-case email), trips with a member, days, places, assignments with
the canary times, a booked stay, a flight with endpoints, budget items
with splits, packing, todos, notes, file metadata, tags, a journey with
an entry and a notification. scripts/build-legacy-fixture-v244.mjs
rebuilds it byte for byte from the tag: it archives v4.3.3 into a temp
dir outside the repo, boots that runner, seeds, pins every timestamp,
VACUUMs and dumps.
The suite runs the shared upgrade assertions and checks that every row
of those tables comes through with its values, that the boot copied the
database first and the copy still is the 4.3.3 schema, and that the
reminder backfill marks only the reminders whose day has passed.
Migrations only go forward, and since #2610 an older image refuses a
database a newer one migrated. Going back after a bad update therefore
needs a copy from before it, and the only copies were backups the admin
takes by hand or auto-backup, which is off by default. The first boot
on 4.4 migrates every 4.3.x install through the legacy baseline and the
new migrations without either.
Before applying pending migrations to a file database that holds
anything, the boot now writes a VACUUM INTO copy next to it, named
pre-migrate-<schema>-<time>.db (legacy-<schema_version> for an install
the positional runner left, otherwise the last recorded migration), and
keeps the newest TREK_DB_PRE_MIGRATE_SNAPSHOT_KEEP (3). The copy is
written under a temporary name and renamed, so a crash never leaves a
truncated file that looks good. If it cannot be written the boot refuses
before anything is migrated; TREK_DB_PRE_MIGRATE_SNAPSHOT=false is the
documented way past that. The newer-database refusal names the copy
taken right before the newer release migrated past this one.
A fresh install, a restart with nothing pending, an in-memory database
and a restore's re-bootstrap (the archive is the copy) take none, and
NODE_ENV=test leaves it off unless a suite sets the switch.
Migration20200101042300_every_tour_type_the_contract_names runs after
the legacy steps like the other post-legacy migrations, but was missing
from the test's list of migrations without a step, so LEGACYMAP-001
failed.
Three gaps in the notice from the previous commit, found in review.
A page served from the network and claimed later by its own build's
worker counted as already taken over, so Reload went to that worker,
got the old build again and spent the session's reload: no notice and
no new build until the next start. Only a worker that replaced the one
that served the page skips the handover now, as at launch.
A reload that came before the new worker could take over (a slow
precache, a failed update, an install gone redundant) also spent that
guard, so the launch behind it could no longer finish the handover with
its own reload. That reload leaves the guard alone now.
An install whose server never reports the bundle's version (a hand-set
APP_VERSION) showed the notice in every session, and Reload could not
clear it. The store keeps the first version a page hears of, from the
app config at launch, and the notice only asks when a later report
differs: a different build at launch is the launch handover's to fix.
The server's wire-envelope test pins the welcome frame's fields and now
expects the version beside the socket id. The wiki's update page says
what an open tab shows after an update.
A tab kept open across a deploy ran the previous build until its next
start. The server now names its version in the realtime welcome frame,
which every open client reads again when it reconnects after the
restart a deploy brings. When that version is not the build the page
runs, a notice offers to reload. The reload goes through the same
service worker handover a launch uses: it waits for the new worker to
take over, or for a short timeout, and records the version only when
the takeover happened.
The notice reads the error boundary's texts for the same situation, so
every language has it already. It is one component for both shells,
sits above the offline pill, stays away offline and after a reload for
that version that did not bring it, and can be closed for the session.
Each zone's VTIMEZONE carried one fixed offset, so a client that reads it
literally showed every event after a change to or from summer time an hour
off. It now lists every offset change in the calendar years its events
fall in, each as its own observance after the offset in force before the
first one. A zone without changes, or a span with no readable date, keeps
the single fixed offset.
A tourRepo replaces the direct toursApi calls: reads fall back to a new
Dexie tours table, and a tour saved or edited offline is written there and
queued like a place, under one temporary id for the place and its facet.
The queue moves both to the real id on replay and drops both when the
server refuses the create. Trip sync caches each trip's tours while the
addon is on, so a tour no longer reads as a plain place offline.
The tour metrics move to shared so the offline copy shows the server's
numbers, and useTourPlaceIds keeps one Set per change as intended.
tour_types now holds every key the shared contract accepts, all but hike
disabled, and create and edit turn a disabled type away with a 400 instead
of failing on the foreign key. tours.tour_group_id is dropped: nothing ever
wrote or read it. difficulty and wanderer_ref stay for the Wanderer import.
Updates the wiki for the changes in this branch: TREK refuses an
ENCRYPTION_KEY that disagrees with the key file, the rotation script
covers the AI Parsing key and moves the file along, a restore refuses a
newer backup and keeps the replaced database, the chart's startup and
readiness probes, PORT and extraEnv, retrying or discarding failed
offline changes, trip reminders catching up, and why a Portainer
recreate used to keep reporting the old version.
One try around the holiday fetch and the deletes that follow it caught
both, so a database error while removing vacation days that fall on a
public holiday was logged as nothing and left the days standing. Only
the fetch is tolerated now. Also drops two files long gone from the
Sonar coverage exclusions.
86 foreign-key columns had no index. SQLite does not add one by itself,
so every cascading delete of a trip or a user scanned each such child
table in full, inside the transaction, while every other request waited
on the one connection. One migration adds them all, and the entities are
regenerated to match.
The list read marked tour places with tour_place_id, but the single-place
read behind place:created and place:updated did not, and the phone's
places browser kept a second tour list of its own that no realtime event
updated. A tour another member made, or one imported from GPX, showed as
an ordinary place on the phone until the screen remounted, and every
tour did offline. The single read carries the mark now, the browser asks
the planner, and the tour id set is memoised instead of rebuilt on every
render.
A note's attachment moved to the trash from the file manager kept
showing on the note; the hydration read now skips it, as the chat one
already did. The MCP session sweep logged a line every minute while any
session was open; it logs only when it cleaned something. The compose
file caps the container log, and three test mocks that pointed at
modules long gone are removed.
Every refresh writes a new oauth_tokens row and nothing ever deleted the
old ones, so an MCP client refreshing hourly grew the table, and every
backup with it, without end. A nightly job now deletes tokens whose
refresh token expired more than 30 days ago, leaves first, and keeps any
row a live chain still points at for replay detection. Also types the
backup's per-entry store flag.
Changing PORT left the Dockerfile and compose healthchecks, the chart's
containerPort, Service target and probes on 3000, so the container went
unhealthy and Kubernetes restarted it in a loop. All of them follow PORT
now. The chart gains a startup probe (ten minutes for a first boot, a
long migration or a restore at boot, which liveness used to kill after
about 105 s), a readiness probe on the new /api/health/ready that asks
the database, extraEnv and extraEnvFrom for the settings the chart does
not list, slots for the four secrets it had none for, and the
DEFAULT_LANGUAGE and MCP_MAX_SESSION_PER_USER values it documented but
never set. The Dockerfile no longer hides a failed npm ci behind a ';',
pins gosu, and the language lists name all 27 locales.
Resetting a locked-out or compromised admin left every session on that
account valid, because the script did not move password_version on the
way every password change in the app does. It also looked the address
up case-sensitively and created a second account when the stored one
had capitals.
CronJob ran a tick on schedule even when the previous one had not
finished, so an hourly auto-backup of a large library could overlap
itself, and a rejected tick only reached cron's own console line. Jobs
now wait for completion and log failures through the app logger. The
backup also stores photos, videos and PDFs instead of deflating
already-compressed files at level 9.
Overpass, the Wikipedia photo lookup (which holds one of five shared
photo slots), the plugin download and registry, and the two API key
checks had no signal, so a provider that took the connection and went
quiet held the request for undici's 300 s default. ESLint now refuses a
fetch() whose init carries no signal.
The reminder crons run at 09:00 in the instance's TZ but computed today
in UTC, so east of Greenwich a trip reminder arrived a day late. A day
the job did not run also lost that day's trip reminders for good: the
match was 'starts exactly N days from today'. Trips now record the start
date a reminder went out for, and the job sends every reminder that is
due and not yet sent. The active-trip redirect and the upcoming
reservations use the instance's date and clock as well.
Filling in a missing end date and keeping a trip's length when its start
moved both shifted a local-time Date, so a trip crossing a DST change
lost or gained a day. Both now use addIsoDays. The one-date-makes-a-week
default moves from the REST controller into TripsService.create, so the
MCP create_trip tool no longer stores a half-dated trip, and the trip and
its days are written in one transaction.
That branch numbered its own steps after the upstream step it was cut
from, so a database at schema_version 243 to 246 (or 216 to 218) with the
Tours tables had four upstream migrations marked as run without ever
running them. The baseline now cuts such a version back to the last
upstream step, and the Tours migration adds the difficulty cap to a tours
table an earlier branch stage created without it.
Login and the OIDC fallback match emails case-insensitively, but the
password reset lookup and the admin create and edit checks compared
exactly. An account stored as Foo@x.com never got a reset mail, and an
admin could create bob@x next to Bob@x, after which login picked
whichever row came first. The signup collision check and the first-user
count also move inside the signup transaction.
budget_settlements.created_by_user_id has no ON DELETE action and the
account cleanup never cleared it, so deleting such a user failed on the
foreign key and rolled the whole deletion back. The cleanup now clears
the recorder; the payment itself belongs to the two members and stays.
Rolling an image back, or restoring a backup a newer release made, ran
older code against a schema it does not know and failed later, far from
the cause. Boot now refuses a database that records migrations this
build does not ship, and the restore check refuses such a backup while
the current database is still in place. A restore also keeps the
database it replaces as data/pre-restore-<time>.db.
An env value different from data/.encryption_key used to overwrite the
file on boot, destroying the only copy of the key the stored secrets were
encrypted with. The server now refuses to start and names the rotation
script, which re-encrypts the AI Parsing addon key as well and moves the
key file to the new key. Under test the suite's fixed key no longer
overwrites the developer's key file.
Maurice