4grayandClaude Fable 5.1
01c423ac43
* fix(portals): apply the IPv6->IPv4 fallback budget in the Electron process The 2500 ms happy-eyeballs attempt timeout from #1404 only ever ran in the web backend. The Electron main process and its playlist-refresh and EPG workers kept Node's 250 ms default, so a dual-stack panel hostname behind a VPN or a slow link failed every connection attempt in a row and tripped the host connectivity guard. The module now lives in `@iptvnator/shared/host-health` and every Node isolate that opens connections applies it. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix(portals): stop treating a slow panel as a dead one in the host guard axios raises the same ECONNABORTED whether the SYN went unanswered or the panel accepted the connection and then thought for longer than the request budget. Two such timeouts opened the breaker and every request to the panel was refused for 30 s with "portal is not responding" — the shape behind the "connection keeps dropping" reports on 0.23 and nightly. Both transports now report whether the TCP connection was established (`onConnect`: Electron through a per-request observed agent instead of the shared keep-alive globalAgent, the web backend through the transport that owns the ClientRequest), and `classifyHostRequestFailure(error, { connected })` downgrades a host-level code observed after the handshake to inconclusive. Redirect attribution keeps precedence. A host that never accepts the connection trips the guard exactly as before. The Xtream mock gains a `silent:silent` scenario whose detail actions accept and never answer, plus a real-socket regression spec for the guard. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix(portals): let an accepted connection clear the host-failure streak Review finding: an unanswered SYN, then an accepted-but-slow timeout, then another unanswered SYN still reached the two-failure threshold, because the middle request was merely not counted. An accepted TCP connection is the reachability the guard measures, so it now reads as `responded` and clears the streak like an HTTP response would. Regression coverage for the mixed sequence on one flapping loopback origin (Electron) and through the proxy route (web backend). Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix(portals): credit an accepted connection when it happens, not when the request settles Review findings. A request that connected and then hung for 30 s cleared, on its eventual timeout, the failures later requests had recorded while it waited — reopening a host that had just died on evidence older than theirs. The connect hook now reports the connection the moment it fires through a new `HostConnectivityGuard.reportConnected`, which clears the failure streak but closes no open or half-open breaker (the trial keeps its slot until it settles), and the settled timeout is inconclusive. Electron also skips the socket observer while an environment proxy (`http_proxy` / `https_proxy` / `all_proxy`) applies to the request: through a proxy the socket connects to the proxy, whose handshake proves nothing about the portal, so those requests keep the pre-observer behaviour. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * fix(portals): decide the proxy exemption with axios' own resolution Review findings. The hand-rolled environment check ignored `no_proxy`, so a LAN portal exempted from the proxy lost its connect observer and slow requests to it still tripped the breaker; it also read the variables with `??`, letting an empty lowercase one mask a populated uppercase one that axios would honour. The decision now calls `proxy-from-env`'s `getProxyForUrl`, the same pinned package axios' http adapter uses, declared as a direct dependency so the packaged app carries it. The validated-axios spec clears and restores every proxy variable around each case, so a runner that exports a proxy cannot change what the cases prove. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> --------- Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Release notes (.changes/)
Every PR with a user-visible change drops one file here describing that change
in plain language. At release time
tools/release/build-release-notes.mjs turns the accumulated files into the
GitHub release body, the CHANGELOG.md section, a blog-post scaffold for
the website, and Telegram/Reddit announcement drafts — then deletes them.
The point is to write the note while the context is still fresh, instead of reconstructing three months of work from commit titles at release time.
File
Name it <area>-<short-slug>.md, e.g. .changes/playback-up-next-rail.md.
---
type: feature
area: playback
issues: [1187]
screenshot: up-next-rail
highlight: Up Next rail
---
Series now show an "Up Next" rail beside the player on wide windows: the rest
of the current season, watch progress, and click-to-play inline.
| Field | Required | Value |
|---|---|---|
type |
yes | breaking, feature, fix, perf, or internal |
area |
yes | lowercase slug, same as the conventional-commit scope |
issues |
no | issue numbers this closes — [1187] or 1187 |
screenshot |
no | slug from tools/release/screenshots.manifest.json |
highlight |
no | short headline (max 60 chars) marking a release highlight |
There is no version field. The release version is chosen deliberately at release time, not derived from these files.
You never write a PR number: the generator resolves it from the commit that added the file.
Writing the body
One to three sentences, present tense, written for a user, not a reviewer. The body is capped at 400 characters — depth belongs in the blog post.
-
❌ "Refactor
WebVideoControlsAdapterto hoist volume state into the session" -
✅ "The player now remembers volume between episodes"
-
❌ "Fix off-by-one in
resolveEnrichmentSeasonNumber" -
✅ "Series whose title carries a season marker no longer show the wrong season"
highlight marks the change as one of the release's headline features and
gives it a short, poster-worthy name. The 60-character cap keeps it roughly to
one line on the hero card; card text wraps by estimated width, so a headline of
unusually wide glyphs may still wrap or ellipsize rather than overflow. Highlights lead the Telegram/Reddit
announcements (everything else collapses into a "+N more" counter) and open the
blog scaffold: a row in its "What changed" table and a ## section ahead of
the themed sections, while fixes without a highlight collapse under a spoiler
there. Set it on the two or three
changes worth announcing — a release where everything is a highlight has none.
Not allowed on type: internal.
type: internal records invisible maintenance. Internal notes stay collapsed in
CHANGELOG.md, are omitted from the blog scaffold, and are removed from the
authored public GitHub body by
extract-changelog-section.mjs --public. GitHub's generated commit list remains
separate. An internal-only release can therefore have an empty authored body.
When a note is not needed
The gate auto-exempts website, E2E and mock-server apps, *.spec.{js,ts},
*.e2e.{js,ts}, snapshots, any /testing/ path, and Markdown. For other
test-only, documentation, CI/workflow, or pure-refactor changes under
apps//libs/, apply no-release-note when no user-visible note is warranted.
Commands
pnpm run release:notes:validate
pnpm run release:notes:github
pnpm run release:notes:changelog
pnpm run release:notes:blog
pnpm --silent run release:notes:telegram
pnpm --silent run release:notes:reddit
node tools/release/build-release-notes.mjs --consume
The release version comes from the root package.json — bump it first, then
generate. --version 0.24.0 overrides it to preview a release before the bump:
pnpm run release:notes:github --version 0.24.0
A bare -- separator is accepted and ignored, so the npm habit of
pnpm run release:notes:github -- --version 0.24.0 works too: pnpm forwards
that separator to the script rather than consuming it the way npm does.
--validate, --format github, --format telegram, and --format reddit
only read and print. --format changelog and --format blog write their
target file (rerunning changelog for the same version replaces that section
rather than duplicating it). Only --consume deletes anything.
The announcement formats print paste-ready posts to stdout, each guaranteed
to fit its platform's limit: Telegram plain text within 4096 characters,
Reddit markdown within 40,000, with a suggested post title on the first line.
Whatever does not fit collapses into a counter; a breaking change is never
collapsed, and if the highlights alone will not fit, the render fails with an
actionable error rather than shipping a post that cannot be submitted. Use pnpm --silent run for these two —
plain pnpm run prints its lifecycle banner to the same stdout, so a
redirected post starts with two lines of build noise. Render and save them
before
--consume — the changelog keeps the entries, but the highlight: metadata
lives only in the note files. Publishing is manual; nothing posts anywhere.
An internal-only release has nothing to announce: both formats then print an
explanation on stderr, leave stdout empty, and exit 0 — the same shape
extract-changelog-section.mjs --public uses for its empty public body.
The release sequence is: bump the version → release:notes:changelog →
release:notes:blog → release:screenshots → release:notes:telegram /
release:notes:reddit → release:cards:generate → --consume → commit → tag
→ push → release:verify:draft. Everything reading highlight: comes before
--consume, because that step deletes the only copy of it. Full contract:
docs/architecture/release-pipeline.md. The tag build then
extracts the new CHANGELOG.md section into the GitHub release body
(tools/release/extract-changelog-section.mjs) and fails the release if
the section is missing — a tag cut without the changelog step cannot silently
ship PR-title-only notes.
The website publishes one post per minor version (v0-18 … v0-22), and
release screenshots live under the matching blog/v0-24/ directory. A patch
release therefore edits the existing post rather than generating a new one, so
--format blog refuses to overwrite unless you pass --force.
Screenshots
pnpm run release:screenshots captures every manifest shot in dark and light
against the built app plus the Xtream mock server — never a real account.
The run is fail-closed: it proves the real ~/.iptvnator/databases directory
(including the SQLite WAL sidecars, checked after Electron exits) was not
touched, launches the app with an allowlisted environment, records and blocks
all non-localhost traffic, scans every frame for external resources and
credential-shaped text, and asserts TMDB enrichment stays disabled. Frames are
staged outside the repository and published only once every shot and every
guard has passed.
Adding a shot for a new feature = one entry in
tools/release/screenshots.manifest.json (plus, if navigation is new, one
named action in tools/release/capture-navigation.ts).
Highlight cards
pnpm run release:cards:generate renders one branded 1200×630 card per
highlight: note (headline, body, and a framed screenshot strip when the note
names one) plus a release hero card — for Telegram/Reddit previews and the
blog hero.jpg. It reads screenshots from the published blog directory, so it
runs after release:screenshots and, like the announcement formats, before
--consume. Output goes to dist/release-highlight-cards/v<version>/, and a
rerun replaces the cards it previously wrote there; copying a card into the
website tree is a deliberate manual act. release:cards:dry-run lists what
would be rendered. A release without highlights still gets its hero card, and
an internal-only release writes nothing — neither is an error.
pnpm nx run electron-backend:build-e2e # once, before capturing
pnpm run release:screenshots # all shots, both themes
pnpm run release:screenshots -- --only dashboard --theme dark
pnpm run release:screenshots -- --release v0-24