* ci(i18n): fail on new English-identical translations
The drift check only warned about locale values identical to English, so
untranslated strings kept landing. It now fails on any such value that
tools/i18n/identical-en-baseline.json does not record for that locale and
key. The baseline captures today's 2,015 entries: legitimately identical
values (brand and technical names, language autonyms, PIN) and the
existing debt. An entry only covers the English text it recorded, so
copying reworded English into a locale fails too.
Baseline entries that are no longer identical are reported, not fatal.
`pnpm run i18n:baseline:update` rewrites the baseline deliberately; CI
runs `pnpm run i18n:validate` (node tests, then the check) and never
rewrites it. `--fail-on-identical` remains as a strict audit.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(i18n): keep the baseline intact on failed updates and strict audits
`--update-baseline` now writes nothing while any locale is unreadable or
has missing or extra keys, so an incomplete translation cannot reshape
the baseline. `--fail-on-identical` no longer reads the baseline it
ignores, so a damaged file cannot block a strict audit.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): measure the serial IPC depth before the J1 first card
Adds renderer.ipcSerialDepthToFirstCard to the launch journey: the length
of the longest chain of bridge calls in which each call started after the
previous one completed, among calls that completed before the first card.
The main IPC capture now records the ordered start/completion timeline;
the depth, its lower bound, the chain and the timeline are per-iteration
evidence, and the CI job summary prints the chain.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): keep the IPC timeline consistent around the J2 start marker
A call that started before the start marker no longer records its
completion in the timeline, and completions of a method with calls in
flight both inside and outside the timeline are attributed outside and
counted, instead of skipping the first marker-method completion.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: 4gray <fourgray@proton.me>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): add the J3 playback journey
J3 clicks a live channel of an Xtream portal and ends at the built-in
HTML5 player's first `playing` event, with `loadedmetadata` as a
secondary phase. It follows J2: every iteration is a fresh J1 launch on
a copy of a profile seeded through the app's dialogs, and the click
happens after the app has settled in the portal's first live category.
The portal is the mock's `live-fallback` account, whose `.ts` live URLs
serve the local H.264/AAC MPEG-TS fixture that mpegts.js plays through
MSE on every platform. The marketing accounts' local live bytes are
zero-filled and never reach `playing`. Seeding selects the HTML5 player
and the `ts` stream format; the catalog's picsum.photos logos are
cancelled from the test side so no request leaves the machine.
Counters: renderer.ipcCallsToPlaying, renderer.httpRequestsToPlaying,
renderer.domMutationsToPlaying, renderer.layoutShiftScore and
renderer.longTasks; wall-clock click->loadedmetadata and click->playing.
renderer.ipcSerialDepthToPlaying is listed as unavailable until the
serial-depth helper lands. No baseline yet.
The probe gains a media-event terminal; J2's pre-click settle moves to
journey-click-settle.ts so both journeys share it unchanged, and the
probe spec's jsdom fixtures move to a shared test helper.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): show J3's HTTP boundary margins and watch 1 s after playing
Review follow-up. renderer.httpRequestsToPlaying compares the ledger's
arrival stamps with the renderer's click and playing stamps, which come
from different processes on the same host clock. Each iteration now
records the distance of the nearest request on either side of both
boundaries, so a count a clock difference could flip is visible.
Requests after playing were a single snapshot taken right after the
probe; the test now watches the ledger for a fixed 1 s after playing.
A live stream never leaves the mock quiet, so J2's quiet wait does not
apply.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: 4gray <fourgray@proton.me>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): add the J2 open-source journey
Measure the click on the Xtream portal card until the category list and
the first page of the opened section are painted, in the same fresh
process as J1 after its counters are final and the app has settled.
- journey-renderer-probe: optional click start (capture-phase listener on
window, start sentinel before the app sees the click, entries before the
click dropped), companion selectors, recent-input layout shifts tallied
- journey-main-ipc-capture: optional start sentinel; counts calls between
the two sentinels
- journey-mock-request-ledger: loopback proxy that counts every request
the app sends to the mock without storing credentials
- open-source-journey-record: J2 counters and evidence
- journey-run / journey-summary: every journey spec of one perf:journeys
run adds its entry to the same summary.json
- docs: J2 contract in performance-journeys.md
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): stop echoing the request URL from the ledger spec's upstream
CodeQL flagged the fake upstream as reflected XSS. It now records what it
received server-side and answers with a fixed text/plain body.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): address J2 review findings
- Sentinels use cancelSourceProbe: the preload traces the call before
forwarding it and SOURCE_HEALTH_CANCEL is an in-memory map lookup, so a
marker no longer runs a SQLite query on the worker ahead of the measured
work (Codex P1). A spec pins that handler contract.
- A run is started only in the Playwright runner, replacing inherited
values, and carries a random harness.runId; summaries from another
invocation are never merged (Greptile P1, Codex P2).
- The mock ledger tracks in-flight requests; settling and the HTTP window
require none in flight (Codex P2).
- clickToFirstPagePaintMs reports click to the committed paint next to
the terminal-batch clickToFirstPageMs (Codex P1).
- The Playwright attachment carries the whole summary (Greptile P2).
- jsdom probe specs wait for the post-paint cutoff instead of a fixed
40 ms, which flaked when the harness runs all files in parallel.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* docs(validation): describe perf:journeys as running J1 and J2
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): settle J2 on pending bridge calls and start HTTP at the click
- The journey IPC capture pairs every traced start with its success or
error and exposes the calls still in flight. J2 settles only when J1's
capture, installed before the document loaded, has none pending, so a
slow startup call cannot resolve after the click and count as J2.
- The mock HTTP window starts at the renderer's click stamp instead of
the test-side mark taken before Playwright's actionability checks.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): restart the J2 quiet period when pending work completes
Both waits in the open-source journey (settling before the click, closing
the mock window after the terminal) now use one waitForJourneyQuiet
helper that compares whole samples, in-flight counts included. The poll
that first sees a request or bridge call complete restarts the quiet
period, so the window is never measured from a poll at which work was
still pending. A fake-clock spec covers the in-flight to zero case.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): align J2 with the main-process counters from #1715
After rebasing on #1715, J1 measures main.sqlStatementsBeforeReadyToShow,
so J2's reason for listing main.sqlStatementsToFirstPage as unavailable
(no countable channel) was stale. State the actual limit: the running
total is read from the test process and cannot be bounded at the click
or the first-page batch. The performance-journeys CI job comment now
names both journeys.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): launch J2 without SQL counting and stamp mock requests in sub-ms
- runLaunchJourney takes the launch instrumentation; only J1 turns on the
main-process counters and IPTVNATOR_PERF_COUNT_SQL, so J2's click is not
measured under the hook that wraps every SQLite statement. The flags are
built in journey-launch-environment.ts, which the SQL opt-in guard now
expects, and a launch record without main counters is rejected.
- The mock ledger stamps arrivals with performance.timeOrigin +
performance.now(), the same sub-millisecond epoch as the renderer's
click, so a request later in the click's millisecond is not counted
before it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): reject J2 iterations with activity after settling
- The open-source record compares the settle snapshot with what the probe
and the IPC capture counted up to the click event, and with the mock
requests between the snapshot and the click stamp. Any change means
background work began during Playwright's actionability checks and
could land in J2, so the iteration is rejected.
- The SQL opt-in guard also checks who passes mainCounters: true: only
measureLaunchJourney may, and J2 must pass false.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): count the long task that dispatches the J2 click
A long task's startTime precedes the click event's timestamp when the
listener runs inside it, so the start-time filter dropped the task that
performs the interaction. Long tasks now count when their range overlaps
the window: on one main thread only the dispatching task can overlap the
click. Layout shifts keep the start-time filter. J1 is unchanged (its
window starts at -Infinity).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): bound J2's late-request check by the quiet sample's mark
The late-activity check compared requests against a fresh ledger mark
taken after waitForQuiet returned. A request that arrived while the final
quiet sample was still reading the IPC capture advanced that mark and
escaped the check. The boundary is now the ledger position read by the
accepted sample itself, like its DOM and IPC counts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): end J2's HTTP window at the accepted quiet sample
The post-terminal window read the ledger after waitForMockQuiet returned,
so a request arriving in between was counted although its completion was
never waited for. waitForMockQuiet now returns the ledger position its
accepted sample read; later requests are kept as evidence
(httpRequestsAfterSettledByRoute) instead of the counter.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): observe late mock requests before reading J2's ledger
httpRequestsAfterSettledByRoute read the ledger right after the accepted
quiet sample, so late requests had no chance to appear in it. The ledger
is now read after another quiet interval; the counter stays bounded by
the quiet sample's mark and late traffic shows up in the evidence.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): fail the J2 quiet wait when a sample stalls past its deadline
waitForJourneyQuiet accepted a sample that returned unchanged after a
stall longer than the timeout as the end of a quiet period, before the
deadline check ran. The deadline is now checked first, so a stalled
sample fails the wait instead of letting the click go ahead unobserved.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): detach J1's IPC capture before the J2 click
J2 used J1's capture to see pending launch bridge calls while settling,
but its ipcMain listener stayed attached and ran for every bridge call of
the measured click. The capture can now be detached; J2 detaches J1's
right after settling, before the click.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(perf): sample both J2 settle captures in one main-process snapshot
The settle sample read J1's capture (pending calls) and J2's capture
(call count) in two evaluate calls, so a call starting in between was
counted with a stale zero in flight and its completion went unseen. Both
states are now read in one synchronous pass, where no ipcMain event can
be handled in between.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: 4gray <fourgray@proton.me>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* ci(performance): give the initial-bytes ratchet slack and a labelled override
The exact renderer.initialBytes counter failed PRs for reasons outside
their diff: two concurrent merges left master 108 bytes over the baseline
for hours, and bundler identifier renaming moves the counter by hundreds of
bytes. PRs growing it by 243 and 302 bytes had no way to pass at all,
because the direction check refuses any raised baseline.
- Counter entries accept `slack` (integer, entry unit): the ratchet enforces
`value + slack`, reports how much slack a measurement uses, and still
prints the tighten hint below `value`. renderer.initialBytes gets 4096
bytes, so growth can accumulate at most 4 KiB past the last lowered
baseline while regressions such as +35 KB still fail.
- check-baseline-direction.mjs compares `value + slack`, treats widened
slack like a widened tolerance, and takes `--allow-increase`, which
reports weakened entries as ALLOWED instead of failing.
- CI passes `--allow-increase` only when the pull request (or, for a master
push, the pull request merged as the pushed commit) carries the
perf-baseline-increase label, read from the API so a job re-run picks up
a label added later.
This change widens the slack itself, so its own PR needs the label.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0166PWobUjpoiWt8E9bBdsCs
* fix(performance): report slack usage only for entries that have slack
A wall-clock measurement above `value` but within `value × toleranceRatio`
fell into the slack branch and was reported as using "slack", conflating
timing tolerance with counter slack. Only entries with `slack` report it now.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0166PWobUjpoiWt8E9bBdsCs
* fix(performance): scope the push-time label and refuse raised counter values
- A master push compares the whole push, but the label was read from the
head commit's PR only, so one labelled PR could cover another commit's
increase in the same push. The label now counts only when the push added
exactly one first-parent commit (a squash or merge of one PR); any other
push that weakens a baseline fails.
- Raising a counter's `value` while narrowing its `slack` lowered the
enforced limit and was reported as "lowered". A counter's value is the
measured evidence and only moves down, so that raise is now a weakening
that needs the label.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0166PWobUjpoiWt8E9bBdsCs
* fix(performance): treat a counter/wall-clock type switch as a weakening
Turning `{ value: 100, slack: 10 }` into `{ value: 105, toleranceRatio: 1 }`
skipped the raised-counter-value rule and was reported as a lowered limit.
Switching an entry between counter and wall-clock now needs the
perf-baseline-increase label.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0166PWobUjpoiWt8E9bBdsCs
* ci(performance): paginate the label lookups of the direction check
The labels endpoint returns 30 entries per page by default, so a PR with
more labels could miss perf-baseline-increase. Both lookups now request
100 per page and paginate, like the release-note gate.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0166PWobUjpoiWt8E9bBdsCs
---------
Co-authored-by: Claude <noreply@anthropic.com>
* build(test): make spec tsconfigs resolve what Jest resolves
Lib spec tsconfigs used module: commonjs with node10 resolution, which cannot
see Angular's exports-only secondary entry points, and dropped global.d.ts, so
tsc reported thousands of resolution errors and no window.electron typing.
Switch them to module: preserve with bundler resolution (ts-jest still forces
CommonJS emit outside ESM mode), add global.d.ts to every spec program, type
jest.unstable_mockModule for the ESM workspace, include the ui-epg and
ui-playback specs that jest.web-esm.workspace.ts runs under the web spec
config, and drop the snack-bar stub that shadowed the real Material types.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* ci(test): gate spec type-checking with typecheck:spec
Add tools/typecheck/spec-typecheck.mjs, which runs tsc --noEmit over every
tsconfig.spec.json with a small pool and fails on any diagnostic, wire it into
the unit-and-typecheck job after typecheck:ci, and document the gate and the
spec tsconfig conventions in the validation map. Also bring the non-Tier-A
spec configs (remote-control-web, ui-remote-control, stalker-mock-server) to
the same conventions so the gate covers the whole workspace.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* test: fix the spec type errors surfaced by typecheck:spec
With the spec programs resolving modules and ambient typings correctly,
tsc reported 432 genuine errors across the Tier A projects: read-only
capability flags assigned on Partial<> doubles, signal-store values used as
types, fixtures missing required fields, index-signature property access,
partial bridge doubles cast through incompatible shapes, and deferred
resolvers narrowed to never. Type the doubles instead of casting to any:
writable mapped types for capability flags, InstanceType<typeof StalkerStore>,
typed jest.fn signatures, protectedState: false on test signal stores, and
completed fixtures. Production changes are limited to bracket access for
index-signature properties under the libs' noPropertyAccessFromIndexSignature
setting and two narrowing guards in the global favorites loader.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* test(playback): use the ESM setup's jest global in the controls fixtures
The fixture imported jest from @jest/globals, which is not a direct
dependency. Jest provides that module at runtime, so tests passed, but on a
clean pnpm install tsc cannot resolve it and typecheck:spec failed in CI.
The ESM test setup already installs import.meta.jest as the global, typed
by @types/jest, as the other ESM specs use it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test: type the parental lock doubles merged since the gate was written
The parental lock feature (#1601) and the Stalker actor route landed on master
with spec doubles declared as zero-argument jest.fn()s that the tests then
drive with the real arguments, plus a copy of the ResizableDirective override
imported from a library that does not export it. Give the doubles the lock
service's real signatures, drop the dead override as in the sibling layout
specs, use bracket access for the actor route's personId param, and keep the
Stalker layout spec within the 1200-line limit.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: 4gray <fourgray@proton.me>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
* ci(performance): run the performance journeys on the Linux runner
Adds a warn-only performance-journeys job to ci.yml that builds the
electron-performance configuration, runs the journey benchmarks under
xvfb and uploads dist/performance/journeys/ as evidence. Pull requests
run it only when they touch journey-relevant paths.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* docs(performance): document the journeys CI job and runner evidence
Describes the performance-journeys job and its path gate, and records why
the J1 runtime counters are not baselined yet: on the Linux runner the
launch journey is bimodal (13/576 vs 16/939 bridge calls/DOM mutations),
so the counters are not deterministic.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* ci(performance): run the journeys unless a PR changes only safe paths
The scope filter listed the paths that can move a journey, so a PR that
changed only a root build input (.nvmrc, nx.json, tsconfig.base.json)
skipped the measurement. List the paths that cannot instead: the E2E
workflow's ignore list plus release notes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: 4gray <fourgray@proton.me>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
PR CodeQL runs now share a per-PR concurrency group with cancel-in-progress; master pushes, the weekly schedule and manual dispatches get a unique group and are never cancelled. 69 superseded analyses ran to completion across 19 branches in the day before this change.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* chore(github): replace issue templates with issue forms
The Markdown bug and feature templates still asked "PWA or
Electron/Tauri application" with 0.16.0 as the example version, and
as free text they were mostly left unfilled: in the September backlog
triage only about a third of bug reports named the version or the
player, and 34 had to be sent back for a retest because they could
not be tied to a fix.
Replace them with GitHub issue forms that make the version, install
method, OS, source type and selected player required fields, ask
whether the problem is a regression, and point to Copy diagnostics.
Add a dedicated form for playback problems, the largest class of
reports, and a config that disables blank issues and links questions
to Discussions, the Docker guide and the website.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* chore(github): address issue form review feedback
- Playback form: add an "Audio player (radio station)" choice, qualify
the Embedded MPV/VLC advice as desktop-only with a note on browser
limits for the self-hosted web app, and fold the last working version
into the description (12 inputs, the size immich ships).
- Feature form: add the credentials and private URL reminder.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* chore(github): narrow the self-hosted CORS note in the playback form
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* ci(e2e): skip Playwright browser installs in the Electron shards
The Electron suite drives Electron through Playwright's _electron API and
never launches a Playwright browser or records video, so the per-shard
`playwright install --with-deps` only downloaded unused browsers (about
3.5 minutes per Windows shard, nine shards per run). Linux shards now run a
quick check that xvfb-run and Electron's shared libraries are present on
the runner image instead. The web E2E job keeps its Chromium install.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* ci(e2e): fail the Linux Electron dependency check when ldd cannot run
A missing Electron binary or a failing ldd left the "not found" grep empty,
so the preflight passed and the launch failed later without a diagnostic.
Check the binary first and report missing libraries before an ldd failure.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* ci(e2e): resolve the Electron binary before checking its libraries
Electron 42+ downloads its binary on the first require('electron'), which
used to happen inside Playwright's _electron.launch(). The Linux preflight
ran before that and looked for node_modules/electron/dist/electron, which
does not exist yet on a fresh runner. Resolve the binary through
require('electron') so the download happens first and the check inspects
the same path Playwright launches.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: 4gray <fourgray@proton.me>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Pull requests whose changes cannot reach any Tier A test (allowlist checked against declared Tier A inputs and an AST scan of cross-project reads) skip the unit coverage suite; master pushes always run it. Jest's transform cache is persisted with actions/cache: PRs restore only, master pushes start empty and save. Paired CI runs: Tier A 9m04s cold -> 6m09s warm. Nx Cloud is intentionally not used.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
macOS runners are the scarcest on this account, so macOS now runs the
Electron E2E suite as two Playwright shards instead of three; Ubuntu and
Windows keep three. macOS shards get a 40-minute timeout. The summary job
needed no change because it reads each shard's total from its report.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Run the sequential Electron E2E suite as three Playwright shards per OS
(one runner each) and summarize all shards of an OS in one follow-up job.
The semantic summary script accepts a directory of shard reports, merges
them and refuses to write when a shard is missing, duplicated or
malformed, or when an explicit input does not exist.
Slowest shard per OS in the final run: ubuntu 12.5 min (was 26),
macOS 13.7 min (was 34), Windows 24.5 min (was 35).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Third step of the performance-journeys ratchet, stacked on #1693 (which is stacked on #1692; merge in order, GitHub retargets each to `master`).
- New `Initial bytes ratchet` job in `.github/workflows/ci.yml` (ubuntu-latest): install, `pnpm nx build web --skip-nx-cache` (production configuration, the one users download), then `pnpm run perf:initial-bytes:check`. The job fails when `renderer.initialBytes` exceeds `tools/performance/journey-baselines.json`.
- `dist/performance/` is uploaded as the `performance-journey-summary` artifact on every run, so a failing or tightenable run carries its evidence.
- After review: the job first runs the new `tools/performance/check-baseline-direction.mjs`, which compares `journey-baselines.json` with the revision the change is measured against (the target branch of a pull request, `github.event.before` for a `master` push, `master` for a manual dispatch) and fails on any raised enforced limit (`value × toleranceRatio`), any widened or newly added tolerance, or any removed entry, so a PR cannot grow the payload and raise the baseline to match (lowered limits and new entries pass; a target branch without the file has nothing to weaken). Node tests cover it.
- Docs: the performance-journeys contract and the validation map name the job, and the contract now states that this runner is the canonical measurer (take baseline values from its output, not from a local build).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* test(electron): reap Windows process trees and retain smoke diagnostics
* test(electron): require confirmed process exit before relaunch
* test(electron): retain process handle after application exit
* test(electron): bind captured processes to application instances
The `pull_request: closed` cleanup is the fast path, not a guarantee: GitHub
does not run that workflow when the head ref is already gone at event time,
which is what Dependabot does when it supersedes one of its own PRs. 15
`test-pr-<n>` drafts had been orphaned that way, 13 of them Dependabot's.
Add a daily scheduled (and manually dispatchable) sweep to the same workflow.
It lists every draft tagged `^test-pr-[0-9]+$`, asks GitHub for that PR's live
state, and deletes only when the PR reports closed. It fails closed: a PR
lookup error leaves the draft untouched, a failed release listing fails the job
rather than sweeping a short list, and only a confirmed HTTP 404 excuses a
failed delete — `gh api` exits 1 for every failure alike, so the re-check reads
the response status instead of the exit code.
Workflow permissions drop to `contents: read`; the event job keeps
`actions: write` + `contents: write`, the sweep takes only `contents: write`.
No new actions. Docs: new "Rolling test drafts" section in
docs/architecture/release-pipeline.md.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(updater): nightly builds and a stable/nightly update channel
Every master push publishes its artifacts as a prerelease of
4gray/iptvnator-nightly instead of the rolling test-master draft, with a
version of <next patch>-nightly.<commit date>.<run number> applied in
every build job. Settings → About gains an Update channel switch;
AppUpdateService re-points electron-updater per check (feed repository,
allowPrerelease, channel name, allowDowngrade reset) and reads release
notes from the repository the requested version belongs to. Channel
switches are forward-only: a nightly build stays until a newer stable
release exists.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* fix(updater): compute the nightly version once and keep re-runs safe
Review follow-ups: the nightly version is resolved by a leading job and
handed to every build job, and the patch is bumped only when the base
tag already exists so the release-cut window stays below the imminent
release. A re-run never deletes a published nightly; only a draft left
by a failed run is replaced. Typed update-status literals in the
remaining specs carry the new channel fields.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* test(packaging): expect the nightly-version prerequisite in the build workflow graph
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
* chore(deps): upgrade Angular to 22.1 and Nx to 23.2
* fix(deps): complete Angular migrations after rebasing on master
* fix(ci): use the Node pin for Windows runtime refresh
* docs(deps): synchronize the workspace-shell Node requirements
`Build on macos arm64` started failing on master with
security set-key-partition-list -S apple-tool:,apple: -s -k *** <tmp>.keychain
SecKeychainUnlock: The user name or passphrase you entered is not correct.
Every failure ran the `macos-26-arm64` runner image 20260831.0337; the same
job on image 20260728.0273 kept passing, as did `macos-15-intel`. The image
exposed an electron-builder bug: `app-builder-lib` hands the certificate's
`.p12` import password to `set-key-partition-list -k`, which authenticates
against the temporary keychain and therefore needs the keychain's own
generated password. Older macOS builds accepted the wrong password once the
keychain was unlocked; the new one verifies it.
Upstream fixed this in electron-userland/electron-builder#10101 (master,
v27 alpha) and backported it in #10172 to release/v26 on 2026-09-03, but no
26.x containing it is published (26.16.0 predates the backport, #10167).
Apply the backport to the installed 26.15.7 as a pnpm patch, guarded by
`tools/dependencies/app-builder-lib-keychain-password.test.mjs`: it checks
that the patched version is the installed one, that the compiled source
passes `keychainPassword`, and — with `security` recorded — that
`createKeychain` unlocks the partition list with the password it gave
`create-keychain`, not the import password. The test fails 2/3 on the
unpatched code. CI runs it next to the Vite patch guard; CLAUDE.md and
AGENTS.md document when to retire the patch.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Add /download/ with Windows, macOS and Linux landing pages: OS-specific
install steps, requirements, feature list, FAQ, related posts and a
platform switcher, plus SoftwareApplication / FAQPage / BreadcrumbList
JSON-LD on every page.
Direct asset links resolve the latest published release from the GitHub
Releases API at build time (asset names, sizes, publish date) and fall
back to the root package.json version when the API is unreachable;
WEBSITE_SKIP_RELEASE_FETCH=1 forces the fallback. The deploy workflow
passes GITHUB_TOKEN to the build. The homepage download cards and the
header Download link now point at the new pages, the homepage schema
reads the resolved version instead of a hard-coded 0.20.0, and the
locale count is corrected to 19.
tools/testing/website-download-pages.test.mjs checks titles, canonicals,
direct asset links, structured data, cross-links and sitemap entries of
the built output; nx test website runs it beside the Giscus test.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The "Build pinned Linux Embedded MPV runtime" job failed twice on 2026-08-11
because www.freedesktop.org answered GitHub runners with HTTP 418 for the
fontconfig tarball. The Linux builder curled a single pinned URL with no
fallback, so upstream rate-limiting reddened the build.
Route downloadArchive() through the shared downloadPinnedSource() helper the
macOS builder already uses, and pin a mirror for each single-host source:
fontconfig and libdisplay-info (freedesktop-hosted) plus freetype, which the
macOS builder already mirrors. Each mirror was downloaded and verified to hash
to the existing pin. The curl hardening flags and assertArchiveMatchesPin are
unchanged, and the helper verifies every candidate against the same SHA-256,
so a mirror serving different bytes is rejected rather than used.
Unlike macOS, the Linux manifest keeps sourceUrl at the canonical pinned value
even when a mirror served the bytes: notice generation and the Snap publication
boundary compare that field against the immutable pin. A used mirror is logged
instead.
build-linux-runtime.mjs now imports the downloader, so download-pinned-source.mjs
joins the released source-archive tooling set (otherwise the archive would ship
a build script it cannot run) and the Linux runtime cache key.
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* fix(build): include shared UI stylesheets in Nx cache inputs
`libs/ui/styles` held shared SCSS partials but had no `project.json`, so its
files belonged to no Nx project and were absent from every task hash. Editing
a partial and running `pnpm nx build web` reported 4 of 4 tasks cached and
shipped the previous CSS — a silent wrong build rather than a failure.
Nx derives its project graph from TypeScript imports only, so a relative Sass
`@use` that crosses a project root creates no edge. Verified directly: after
adding the project but before declaring anything, `ui-styles` still had zero
dependents in the graph.
Make it the `ui-styles` project (no targets — it exists to be hashed) and
declare `implicitDependencies` on the 8 consumers. Chosen over adding the path
to `sharedGlobals`, which would put shared styles into every project's hash and
make a one-line SCSS tweak mark the whole workspace affected. A styles edit now
marks 15 projects affected and leaves electron-backend, website, the mock
servers and the shared libs alone.
`libs/ui/styles` was the only projectless directory holding files under `libs/`
or `apps/`.
Add `pnpm run styles:inputs:validate` to keep it closed: it resolves every
relative stylesheet import against Nx's real project graph and fails when one
escapes the input closure of a build that compiles it, naming the project to
declare. It exits 1 with 21 diagnostics on the pre-fix tree. Imports of
`apps/web/src/nav-list.scss` are deliberately accepted — `web` already hashes
that file, and a lib -> app edge would make the graph cyclic.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(build): spawn git without a shell in the stylesheet check
`execSync("git ls-files '*.scss'")` runs through `cmd.exe` on Windows, where
single quotes are literal characters rather than quoting. Git received the
pathspec with the quotes intact, matched nothing and exited 0, so
`styles:inputs:validate` reported success after checking zero stylesheets —
silently disabling the check for Windows developers while staying green.
Spawn with `execFileSync` so no shell is involved and git expands its own
pathspec; verified to return the identical 133 files.
Both this and the eslint glob trap next to it in the docs report success while
covering nothing, so also make an empty scan fail rather than pass: the
workspace always contains SCSS, and a listing that returns none means the scan
broke.
Reported by Codex review on #1360.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* refactor(styles): move nav-list partial into ui-styles (#1361)
* fix(build): count every target of a comma-separated Sass @import
`@import` is the only rule that takes a list, and the scan read just its
first target. A later entry crossing an Nx project boundary escaped the
cache key while the check still reported success — the same silent-pass
failure the tool exists to prevent.
Parse every target of an `@import` list. The obvious "read all quoted
strings" fix trades one silent gap for a phantom one, so the rule decides:
`@use`/`@forward` load exactly one module and a quoted string after it is
`with (...)` configuration, and `url(...)` stays a plain CSS import the
browser resolves at runtime. Neither is a module Sass compiles.
The workspace has no relative `@import` at all today, so the scan still
finds the same 42 imports across 133 files; this closes the gap before
someone writes one.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* test(stalker): enforce portal auth in the mock and cover the full-portal flow
The mock server implemented neither get_profile nor get_events and
validated no auth at all, and the e2e suite imported the portal through
/portal.php — which the app classifies as a *simple* portal. The entire
authenticated branch (handshake, token, watchdog, re-auth) therefore had
zero coverage, right before a series of PRs that reworks exactly that.
Mock server:
- serve the canonical /stalker_portal/server/load.php endpoint, which
enforces the Bearer token and the Infomir MAC format like the real
middleware; /portal.php stays tolerant so the existing suite keeps
covering the simple-portal branch
- auth-store.ts models the parts of Stalker 4.9.35 a client can get
wrong: plain-text auth failures with HTTP 200, a handshake that is not
yet a session, idempotent token re-presentation, and permanent
device_id pinning (including the blank-after-pinned lockout)
- add get_profile (status 0/1/2, device conflict, block_msg) and the
get_events watchdog; profile advertises watchdog_timeout/timeslot
- new login-required scenario MAC and POST /invalidate-session so tests
can force a mid-session token loss
- the /stalker proxy route now forwards the token as a Bearer header and
wraps auth failures in the { payload } envelope, matching web-backend
Also moves extractMac into request-mac.ts: importing it from the
categories handler dragged the whole data generator into any consumer,
which broke unit tests on the workspace alias.
E2E: new stalker-auth.e2e.ts asserts handshake precedes get_profile
precedes content, that content requests carry the token while the
handshake does not, that the plain-text failure body is never rendered,
and that the client re-authenticates after the portal drops the session.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(mock): address CodeQL findings in the new portal auth code
Two genuine defects in the code this PR added:
- readBearerToken's /Bearer\s+(.*)$/ backtracks polynomially on
"bearer" followed by a long run of spaces; require the token to start
with a non-space character instead
- the /stalker proxy route read query params as strings without
narrowing, so a repeated key (?url=a&url=b) arrives as an array and
String.prototype.includes silently changes meaning
The remaining three alerts (missing rate limiting x2, sensitive data in
a GET query) are web-service hygiene rules aimed at internet-facing
services. The mock servers bind to localhost, serve fabricated data,
ship in no artifact, and deliberately mirror the real backend proxy's
token-in-query contract; a rate limiter would break the E2E suite that
hammers them. Exclude only those two apps from analysis via a documented
CodeQL config; every shipped path keeps full coverage.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(mock): tighten portal-auth fidelity per review
Review follow-up on #1324 (Greptile 2xP1, Codex 3xP2), all valid:
- adoptToken only accepts tokens the mock actually issued (or the
already-bound one). The stock server pins any presented Bearer —
handshake is stateless there — but a fixture that does the same
cannot catch a client with a broken token pipeline; documented as a
deliberate strictness divergence.
- /invalidate-session clears tokens but keeps pinned device identity:
losing a token never unpins device_id on a real portal, so changed
identity after re-auth must still hit the device-conflict branch.
- The login-required scenario gates on actual do_auth completion
instead of auth_second_step: the app sends auth_second_step=1 on its
very first get_profile, so the parameter check was trivially
bypassed and the status-2 flow never exercised. do_auth is now the
faithful boolean step (non-empty credentials -> {js:true}, recorded;
empty -> {js:false}).
- /server/load.php — the second URL shape isFullStalkerPortal
recognizes — is now served and enforced, directly and through the
/stalker proxy predicate, so full-portal tests cannot silently fall
into the tolerant branch.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* test(stalker): prove content actually reloads after re-authentication
Review follow-up on #1324 (Codex P2, valid — the "passes for the wrong
reason" class): the re-auth test only polled for a fresh handshake and
a negative body-text assertion, both of which pass even if the original
content request is never replayed or stays unauthorized. Capture the
content token from the initial import, then assert a post-invalidation
CONTENT request goes out under a DIFFERENT token and that the ITV
categories actually render — the mock only answers content for an
adopted token, so this proves the new token round-tripped through
get_profile. Verified against a live mock that the token genuinely
rotates (old token -> "Authorization failed.", new token -> content).
Also documents the second Codex P2: the mock is deliberately strict on
/server/load.php (a real portal enforces auth there); the import dialog
vs session predicate divergence is a separate app bug the strict
endpoint will let a later PR cover.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* test(stalker): serialize the portal specs and bind mocks to loopback
Review follow-up on #1324 (Codex, 4xP2):
- Parallel-reset race: under the workspace `fullyParallel` preset the new
auth file ran concurrently with stalker.e2e.ts against one shared mock
process, and each `beforeEach` wiped global state (sessions, favorites)
mid-assertion in the other. Reproduced locally: both suites green in
isolation, two failures when run together. Merged the auth tests into
stalker.e2e.ts and pinned the file to `mode: 'serial'`, which also
removes the pre-existing race between that file's own tests. 19/19
green across three consecutive runs.
- Watchdog was recorded but never asserted, so the suite would stay green
if the full-portal workflow stopped pinging or dropped its token —
`sendWatchdogPing` swallows failures. Now polls for an authenticated
`get_events`.
- Both mock servers listened on every interface (stalker: `listen(PORT)`
with no host; xtream: an explicit `0.0.0.0` default), which made the
CodeQL exclusion's "binds to localhost" rationale untrue. Both now
default to `127.0.0.1` with a `HOST` opt-in, and the config comment
states plainly what the directory-wide ignore trades away.
- Documented that the login-required scenario is HTTP-level only for now:
the client's `do_auth` path is dormant and sends empty credentials, so
the fixture is waiting on that client-side work rather than claiming
end-to-end coverage.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* test(stalker): force a real auth failure before asserting it stays hidden
Review follow-up on #1324 (Codex 2xP2 + 1xP3), all valid:
- The "never surfaces the plain-text auth failure" test only performed a
successful import, so its negative body assertions were vacuous. It now
imports with a MAC outside the Infomir OUI: the strict endpoint answers
get_profile with a bare {status:1}, no token is ever adopted, and every
content request keeps returning "Authorization failed." Unlike an
invalidated session this cannot be repaired by the client retry, so the
failure is genuinely observed (asserted directly against the proxy) and
only then checked for not leaking into the UI.
- docs/architecture/xtream-mock-server.md still documented the wildcard
bind that 4b31f7167 replaced with a loopback default; it now states the
new default and the HOST=0.0.0.0 opt-in needed for phone/STB/container.
- Removed a dangling "Known app-side gap: the" fragment left in the
stalker mock README.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(mock): scope /reset by MAC so parallel specs stop wiping each other
The re-authentication test passed locally but failed all three CI
attempts: no request carried a token, because self-hosted.e2e.ts issues
a GLOBAL `POST /reset` against the same mock from a parallel Playwright
worker, destroying the session mid-import. Running only stalker.e2e.ts
locally never triggered it.
Serializing within one file (4b31f7167) could not fix this — the
interference is between files. Mock state is per-MAC, so `/reset` now
accepts `?macAddress=` and clears only that MAC's data, favorites,
session and watchdog counters; the unscoped form is kept for callers
that own the whole server. Both spec files now reset only the MACs they
own, so no worker can disturb another.
Verified: a scoped reset of one MAC leaves another MAC's session intact
(and its own dies), and stalker.e2e.ts + self-hosted.e2e.ts run together
23/23 green — the combination that reproduced the CI failure.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* test(e2e): scope the last global Stalker reset in sources-pwa helpers
Completes 3a93fef0f: that commit scoped self-hosted.e2e.ts but missed
resetPwaMockServers, which still wiped the whole Stalker fixture from a
third spec file. Scope it to the two MACs this suite owns.
The auth tests use dedicated MACs no sibling touches, so portal sessions
— the fragile state — can no longer be cleared by a parallel worker.
Content MACs still overlap between files, which is harmless: that data is
regenerated deterministically from the same seed.
Verified with the full interfering set running together:
stalker.e2e.ts + self-hosted.e2e.ts + sources-pwa.e2e.ts, 26/26 green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* test(stalker): await the first authenticated content request
The re-auth test kept failing on CI (3/3 attempts) with an undefined
token while passing locally. My earlier diagnosis — a sibling spec's
global /reset — was wrong: the failure survived the scoped-reset fix.
Real cause is a race in the test itself. `addFullStalkerPortal` only
awaits the route change, so on a slower runner the first authenticated
content request has not been recorded yet when the token is read; the
sibling test that passes happens to await `.category-item` first. Poll
for a content request carrying a token before capturing it.
The scoped-reset work stands on its own merits (cross-file resets were
a real hazard), it just was not what broke this test.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* test(stalker): drop serial mode, batch resets, cover the auth handlers
Review round on a44f8135f plus a stability regression I introduced.
Codex, both valid:
- The proxy route stripped `token` from the forwarded query, so
`handshake` never saw a presented token and the idempotent-handshake
behaviour I documented was unreachable through the PWA path. The real
backend forwards every param except `targetId` *and* sets the header;
match it. Verified through the proxy: re-handshake now returns the
same token with not_valid 0.
- The login-required scenario had no committed test, so the README claim
was unbacked. Added auth-handlers.spec.ts (status 2 -> do_auth ->
profile, MAC-format rejection, device conflict, idempotent handshake,
watchdog). Handlers are called directly because the dispatcher pulls in
the faker-based generator, which this project's Jest cannot transform.
- Sibling suites now own disjoint MACs (00:1A:79:5F:*) instead of
sharing the Stalker suite's, so no reset can reach another suite's
state at all.
Stability: a baseline run of master passed 23/23 first try while this
branch failed a different test each run, so the flakiness was mine.
`mode: 'serial'` was a stand-in for isolation that per-MAC scoping now
provides properly, and it amplified every flake by aborting the rest of
the file; removed. `beforeEach` also fired seven sequential resets — the
endpoint now accepts repeated `macAddress` params so a suite clears all
of its MACs in one request. Added a retrying POST helper after an
ECONNRESET on a control call.
Verified: three consecutive runs of stalker + self-hosted + sources-pwa,
26/26 each; 28 mock unit tests; lint clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* test(stalker): restore serial mode for the shared-scenario file
Review follow-up (Codex P2), valid: the previous commit removed
`mode: 'serial'` while every `beforeEach` still resets all OWNED_MACS,
so under fullyParallel one test in this file could clear another's data
or session mid-run.
Of the two suggested fixes, serialize rather than give each test its own
MAC: the tests here are written against scenario fixtures (default,
minimal, embedded-series) whose shapes the assertions encode, so a MAC
per test would mean inventing a scenario per test and rewriting
pre-existing assertions. Cross-file isolation stays with the disjoint
sibling MAC range, which is what serial was wrongly standing in for
before.
The header now states both levels explicitly so the next reader does not
undo one of them.
Verified: three consecutive runs of stalker + self-hosted + sources-pwa,
26/26 each.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* chore(deps): bump actions/setup-node from 4 to 7
Bumps [actions/setup-node](https://github.com/actions/setup-node) from 4 to 7.
- [Release notes](https://github.com/actions/setup-node/releases)
- [Commits](https://github.com/actions/setup-node/compare/v4...v7)
---
updated-dependencies:
- dependency-name: actions/setup-node
dependency-version: '7'
dependency-type: direct:production
update-type: version-update:semver-major
...
Signed-off-by: dependabot[bot] <support@github.com>
* chore(ci): allow actions/setup-node v7 in the Snap workflow policy
The Snap supply-chain policy test pins the exact major of every action
the build workflow may use, so bumping actions/setup-node in the
workflow without updating BUILD_ACTION_ALLOWLIST fails
publish-snap-workflow.test.mjs.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: 4gray <serega05@gmail.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
* chore(deps): bump actions/cache from 4 to 6
Bumps [actions/cache](https://github.com/actions/cache) from 4 to 6.
- [Release notes](https://github.com/actions/cache/releases)
- [Changelog](https://github.com/actions/cache/blob/main/RELEASES.md)
- [Commits](https://github.com/actions/cache/compare/v4...v6)
---
updated-dependencies:
- dependency-name: actions/cache
dependency-version: '6'
dependency-type: direct:production
update-type: version-update:semver-major
...
Signed-off-by: dependabot[bot] <support@github.com>
* chore(ci): allow actions/cache v6 in the Snap workflow policy
The Snap supply-chain policy test pins the exact major of every action
the build workflow may use, so bumping actions/cache in the workflow
without updating BUILD_ACTION_ALLOWLIST fails
publish-snap-workflow.test.mjs.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---------
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: 4gray <serega05@gmail.com>
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
