Category rows crossed the DB-worker IPC boundary with Drizzle's camelCase
property names while the renderer contracts declare snake_case, so backup
export dropped hidden-category IDs and restore degraded to a type-only
match that hid every category. Project category ops to the declared wire
shape, normalize restore state from untrusted sources (dropping entries
without a numeric xtreamId), reject entries with missing user-state
collections, and add full export→import round-trip coverage (unit
manifest-equality + Electron e2e) plus regression tests.
Closes#1017
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Adds a paused state to the Electron download manager with a full partial-file lifecycle:
- Pause keeps the .part and byte progress; cancel discards them; every lifecycle stage (queued, active, mid-transfer) is pausable.
- Resume continues via HTTP Range with If-Range entity validation (strong ETag / Last-Modified persisted in the new resume_validator column, idempotent migration incl. legacy-table rebuild). Non-206 answers restart from zero over the same .part; the 206 Content-Range offset is verified; responses that end before the advertised size are retained for a Range retry instead of being committed as completed.
- Crash recovery converts interrupted transfers to paused, keeps queued-with-partial rows resumable, and commits finalizations that crashed before the DB update.
- Destination collisions are non-destructive (retained partials finalize to the next numbered name); locked .part files never lose their DB owner across cancel/remove/restart; resume claims rows atomically and the queue dedupes ids.
- Stored request headers are re-filtered through the User-Agent/Origin/Referer allowlist on read, URL-derived extensions are sanitized, resume appends never follow symlinks, and transfer errors are logged by message only.
- UI: pause/resume/cancel/retry/remove surface failures in a snackbar; paused items show an active Resume button in VOD/episode detail views; translations for all 18 locales.
- Runtime split into download-runtime/transfer/finalize/broadcast modules; +30 unit tests and an Electron E2E covering pause -> retained .part -> Range/If-Range resume -> byte-exact assembly.
Co-authored-by: genrichh93-ui <genrichh93@users.noreply.github.com>
🤖 Generated with [Claude Code](https://claude.com/claude-code)
* feat(tmdb): clickable director/creator chips and directing credits on person pages
Directors were plain merged text — no photos, no navigation — while the
data was already sitting in the cached TMDB payloads (credits.crew and
created_by both carry id + profile_path; they just were not typed or
parsed).
- tmdb-merge: enrichedDirectors (crew, job === 'Director', deduped by
person id) and enrichedCreators (created_by) produce the same chip
shape as the cast (TmdbEnrichedCastMember) into a new tmdb_directors
field on all three merges (Xtream VOD, Xtream series, Stalker); types
widened (crew id/profile_path, created_by id/profile_path).
- Detail views (shared VodDetailsComponent, Xtream vod/serial routes,
Stalker series view) render the Director row as clickable avatar chips
when tmdb_directors is present — same markup and openActor handler as
the cast strip — falling back to the plain text otherwise. Stalker
re-normalization allowlist preserves the new field.
- Person pages: mapPersonFilmography now merges combined_credits.crew
(jobs Director/Creator) into the filmography — acting wins the
per-title dedup, directing-only titles show the job in the character
slot. Everything else (library matching, All-portals scope, filters,
search fallback, back button) works unchanged because the person page
is role-agnostic. Existing caches work as-is: crew/created_by were
always part of the stored payloads.
Tests: merge spec (director/creator chips + crew-row dedup ×3 merges),
person spec (crew credits, Producer excluded, acting-wins dedup),
stalker-vod.utils passthrough. Docs updated (CLAUDE.md + architecture).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(tmdb): address director-pages review — split oversized spec, stable track keys, translated crew roles
- tmdb-merge.spec.ts grew past the 400-line lint ceiling — the Stalker
merge suite moved to tmdb-merge-stalker.spec.ts (fixes the CI Lint job).
- All cast/director chip loops now track by TMDB person id with an
index fallback ('p<id>' / 'i<index>') instead of member.name — distinct
people can share a name and creator payloads carry no dedup (greptile).
- Directing-only filmography credits carry the role in a new crewJob
field ('Director' | 'Creator') instead of stuffing TMDB's raw English
job into character; ActorViewComponent renders it through translated
labels (XTREAM.CREW_JOB_DIRECTOR/CREATOR, added to all 18 locales via
the i18n patch workflow, matching each locale's existing glossary —
pt "Diretor", de "Regisseur") (Codex).
Tests: person spec asserts character/crewJob separation; merge suites
green after the split (15 + stalker file).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* fix(tmdb): purge obsolete search cache rows
* feat(playback): theater stage and opt-in ambient fill for the inline portal player
On wide-short windows the VOD/series inline player left a strip of app
surface next to the video: with `width: auto`, the viewport's `max-height`
transferred through `aspect-ratio` into a max-width (CSS transferred size
constraints), re-clamping the stage to 16:9 and leaving the leftover
outside it.
- Theater stage: give `.player-shell__viewport` a definite `width: 100%`
so it always fills the content row; the player renders as the largest
16:9 box that fits the stage height, centered — the leftover is always
the stage's black background, never app surface (YouTube-style
letterbox). Applies to every inline engine.
- Ambient fill: new `playerAmbientMode` setting (default off, Settings >
Playback, web players only) renders a blurred, dimmed copy of the
poster behind the player, filling the letterbox margins. Enforced at
runtime too: Embedded MPV never gets the extra DOM layer. Live channels
and non-http(s) poster URLs are excluded.
Verified live via CDP at 1720x760 (stage 1362x532, player 946x532 with
symmetric 208px margins) and 1280x950 (stage exactly 16:9, no bars).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(i18n): add ambient-mode setting keys to all remaining locales
The i18n drift gate requires SETTINGS.PLAYER_AMBIENT_MODE and its
description in every locale; the feature commit only covered en and ru.
Translated via the i18n-fill workflow (per-locale patch + mechanical
merge, glossary-matched against each existing file).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* test(settings): include playerAmbientMode in expected default settings
settings.component.spec asserts the persisted settings object with
toEqual; the new default-off field has to be part of the fixture.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* feat(matching): strip appended language/quality tags in title normalization
Real-world portal catalogs duplicate one show under dozens of tagged
variants ("|ALB| Fallout", "4K-DE - The Pitt (2025) (US)",
"Breaking Bad-eng", "Fallout_esp", "The Last of Us (2023) AF"). A third
of them normalized to polluted keys, silently skipping TMDB enrichment
and staying invisible to cross-portal title matching.
normalizeTitleKeys() now handles, conservatively:
- wrapped pipe tags: "|ALB| X", "|MULTI| X"
- longer/compound leads: "EXYU| X", "4K-DE - X", "AR-SUBS - X",
"4K-OSN+ - X" (dash/pipe only; colon stays
2-3 chars so "NCIS: LA" is untouched)
- underscore suffixes: "X_eng", "(US)_msub" (single-underscore only,
"The_Last_of_Us" stays intact)
- double-dash suffixes: "X--esp"
- joined dash tags: "X-DE", "X-eng" (vocabulary-gated and
case-uniform only; "Spider-Man", "Kick-It",
"Peut-être" are untouched)
- bare trailing tags: "X (2025) DE", "Breaking Bad ES" (UPPERCASE
vocabulary only, skipped for ALL-CAPS titles;
"Rocky II", "Made in USA", "Making It" are
untouched)
Every leading-tag segment must contain a letter, so numeric titles
("1917 - ...") are never treated as tags. The display-side
stripCountryPrefix() learns the same compound/plus-sign prefixes and the
numeric guard.
buildSearchLookupKey() gets a |v2 suffix so cached negative TMDB match
resolutions keyed on old polluted titles are invalidated.
Measured on 248 real catalog names from four shows (The Pitt, Fallout,
The Last of Us, Breaking Bad): clean matching keys 65% -> 99%, display
strip 91% -> 100%. The corpora are committed as spec fixtures.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(matching): use ES2015-safe trailing trim in title normalization
String.prototype.trimEnd is ES2019; the shared-interfaces lib compiles
against an older lib target (TS2550 in typecheck:web). Replace with a
regex-based trimRight helper. Jest uses its own tsconfig, so this only
surfaced in the CI typecheck, not local unit runs.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(matching): guard tag stripping against real-title false positives
Address code-review findings on the tag-stripping rules:
- underscore suffix is now vocabulary-gated, so "Mr_Robot",
"Cowboy_Bebop", "Mrs_Davis" keep their second word
- leading single-segment tags before a spaced dash stay 2-3 chars
(only hyphen-compounds like "4K-DE" and pipe-tags like "EXYU|" may be
wider), so "DUNE - Part Two" and "ALIEN - Covenant" are left intact
- "IN" is excluded from the weak joined-dash/underscore paths so
"drive-in" and "Plug-in" are not truncated (India still strips via
the strong "IN| " / "IN - " forms)
The display-side stripCountryPrefix() mirrors the narrowed dash rule.
Corpus coverage is unchanged at 99% (245/248); new counter-example
tests lock in the guards.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* feat(epg): manual EPG-to-channel mapping with mapping fallback in all EPG paths
* fix: epg mapping in live tv list
* fix(epg): harden manual EPG mapping — upgrade safety, perf, playlist-scoped keys
Follow-up fixes on top of the manual EPG-to-channel mapping feature:
- epg-database: dedupe existing epg_programs rows before creating the
unique (channel_id, start, title) index — a plain CREATE UNIQUE INDEX
crashed the EPG worker on upgrade when historical duplicates exist;
replace INSERT OR REPLACE with ON CONFLICT DO UPDATE so the
epg_programs_fts delete trigger is not bypassed (REPLACE skips delete
triggers unless recursive_triggers is on), with a plain-INSERT
fallback when the index cannot be created
- db: add idx_content_epg_channel — the mapping fallback scanned the
whole content table on every single-channel EPG lookup
- keys: scope Xtream mapping keys per playlist via shared
buildXtreamEpgMappingKey (xtream:{playlistId}:{id}) — bare stream ids
collide across portals; the backend fallback now joins categories to
resolve the playlist id
- pwa: hide "Map EPG channel" entries behind the supportsEpgMapping
capability — the menu item was a dead end in the PWA
- parser: parse the XMLTV offset sign from the string — Math.sign(0)
dropped the minutes of ±00:xx offsets
- cleanup: typed window.electron access instead of ad-hoc casts, drop
unused resolveChannelId and dialog data field, shared
EpgMappingDialogComponent.open() for all seven call sites
- dialog UX: minimum-characters search hint, save/remove snackbars,
current mapping shows the EPG channel display name,
takeUntilDestroyed on the search stream
- i18n: fill the new keys in all 17 locales
- tests: cover mapping CRUD/search escaping, the dedup-index guard,
offset parsing and playlist-scoped keys; update stale stream-resolver
specs for the new 50-item limit and 10s timeout
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(epg): escape backslashes in EPG channel search LIKE pattern
CodeQL js/incomplete-sanitization: a lone trailing backslash in the
search term paired with the closing wildcard under the ESCAPE clause
and corrupted the pattern.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* perf(epg): batch mapping lookups in the viewport preview queue
Expose the existing getEpgMappingsBatch operation over a new
EPG_MAPPING_GET_BATCH IPC channel and use it in resolveManualMappings —
the per-entry lookup issued one IPC round-trip per visible channel on
every scroll event.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* perf(epg): batch mapping prefetch in the collection preview loader
Resolve all candidate mapping keys for an Xtream preview batch with a
single getEpgMappingsBatch IPC call instead of per-channel lookups.
Also fix a worker early-exit: a channel without tvgId/name returned out
of the shared-iterator loop and silently killed one of the three
concurrent preview workers.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: 4gray <serega05@gmail.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* fix(m3u): parse playlists with URLs longer than 2084 characters
Pluto TV style playlists (issue #1189) embed a session JWT in every
stream URL (~2200 chars). validator.isURL inside iptv-playlist-parser
rejected anything over its IE-era 2084-char default, and the parser's
stalled item index then collapsed the whole playlist into a single
channel.
Sync the 4gray/iptv-playlist-parser fork with upstream v0.15.2, which
removes URL validation entirely and adds an explicit branch so '#'
comments and unknown directives are never treated as URLs. Two fork
deltas are preserved on top: the radio attribute (radio player
detection) and pipe stripping (item.url is cut at the first '|' while
|User-Agent=/|Referer= params still land in item.http). The now-dead
validator/is-valid-path dependencies are dropped from the fork.
- pin iptv-playlist-parser to the fork commit SHA
- add a parser contract spec guarding long URLs, comment handling,
radio, pipe stripping, and header EPG attrs
- document the parser fork contract in the M3U architecture doc
Fixes#1189
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* chore(m3u): bump parser to optimized fork build
Pulls the fork's optimized parse() rewrite (2.5-3x faster: 100k
channels ~780ms -> ~285ms, 10k ~79ms -> ~25ms) and the README
documenting fork deltas. Output is differential-verified byte-identical
to the previous build; all parser-contract, unit, and import E2E suites
rerun green against the new pin.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* chore(m3u): bump parser for input robustness and library hygiene
Pulls the fork's real-world input tolerance: UTF-8 BOM, blank lines and
whitespace before the header, and case-insensitive #EXTM3U no longer
reject the playlist (all VLC-accepted forms); Node Buffers are decoded
as UTF-8 and other non-string input throws a clear TypeError. Also
brings truthful types (url?: string), fork metadata, an enforced 100%
coverage gate, and the fork CHANGELOG.
Extends the contract spec with a BOM regression test.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* chore(m3u): pin parser to the tagged fork release v0.15.2-iptvnator.1
Same commit as before (33f5e9c) — the readable tag replaces the raw
SHA in package.json while pnpm-lock still records the immutable
codeload tarball by commit. Fork release:
https://github.com/4gray/iptv-playlist-parser/releases/tag/v0.15.2-iptvnator.1
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(types): make ParsedPlaylistItem.url optional to match runtime
The parser fork's d.ts now truthfully declares url?: string (a trailing
#EXTINF without a stream URL yields url === undefined at runtime, and
always has). The local ParsedPlaylistItem mirrored the old type lie and
made the production typecheck reject the parser's Playlist type.
createChannel and createPlaylistObject already tolerate the absent url.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* feat: add strip country prefix setting
* feat: scope country-prefix stripping to live content and cover missing surfaces
- narrow the heuristic: pipes always strip, dash/colon separators only
when the prefix is a short uppercase tag ("UK - BBC One" strips,
"Sky - Sports F1" and "Mission: Impossible - Fallout" stay intact)
- fall back to the original name when stripping would leave nothing
- scope stripping to live content only: grid type (live/itv/radio),
playback isLive, external sessions without contentInfo, dashboard
cards with contentType 'live'
- cover previously missed surfaces: M3U player EPG timeline header,
M3U inline player title, radio player, dashboard live rails
- replace hardcoded settings strings with translate keys and add
SETTINGS.STRIP_COUNTRY_PREFIX(_DESCRIPTION) to all 18 locales
- add unit specs for the utility plus regression specs for
channel-list-item and external-playback-dock
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* test: cover strip-country-prefix call sites for codecov
- dashboard-rail: new spec for cardTitle live/movie/series scoping
- grid-list: strip enabled/disabled, VOD passthrough, 'No name' fallback
- portal-inline-player: live strip vs VOD passthrough
- unified-live-tab: timeline channel name strip + M3U name precedence
- video-player: timeline/radio/inline titles with the setting on and off
- settings-store: default false + persisted true round-trip
- settings-form.utils: new spec for form default and ?? false fallback
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: 4gray <serega05@gmail.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* spike(embedded-mpv): frame-copy pipeline prototype (helper + shm ring + Electron viewer)
Standalone macOS spike for the frame-copy unification direction from the
2026-07-10 analysis: a helper process renders mpv offscreen into a GL FBO,
reads frames back through an async PBO ring, and publishes BGRA frames into
a 3-slot POSIX shm seqlock ring; a minimal Electron viewer copies the newest
frame via a plain-C N-API addon and uploads it to a WebGL canvas per rAF.
First numbers on M1 Pro (see spike README): 4K60 HEVC hwdec sustained at
60 fps end to end, ~1.2 ms shm copy + ~3.5 ms texture upload, ~10 ms
produce-to-upload age, zero torn frames. Remaining gates: weak hardware,
long-run pacing, HDR, latency flash test.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* spike(embedded-mpv): add RESULTS.md measurement log with M1 Pro baseline
Structured per-machine table with repro commands so the pending Intel Mac
and Windows iGPU runs can be appended and compared one-to-one.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* spike(embedded-mpv): pacing/judder instrumentation + 50/25 fps and HDR gate results
Viewer now measures inter-frame intervals on both clocks (present side and
producer side): stddev/p99/max, late-frame counters vs the producer's median
interval, and a cumulative LONGRUN summary every 30 s. The addon exposes the
producer timestamp (produceMs) for this.
Measured on M1 Pro: 50 fps and 25 fps cadences are clean (late frames only
at startup; residual jitter is 120 Hz rAF grid quantization, bounded by one
display tick), and 4K25 HDR10 PQ/BT.2020 is tonemapped to SDR by mpv before
readback at full rate with unchanged copy costs. RESULTS.md carries the
tables and HDR-clip repro commands.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* spike(embedded-mpv): record 10-minute 4K60 HEVC long-run results
Zero dropped frames and zero torn reads after the first-minute warmup over
~8.5 minutes; steady-state late frames (~0.4%) track the 12 s test clip's
--loop restarts, not the copy pipeline.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* spike(embedded-mpv): viewport-scaling measurement + integration design draft
Confirms the render-at-viewport-size claim (4K source in a 720p FBO costs
720p: 0.17 ms readback / 0.16 ms copy / 0.17 ms upload at 60 fps) and adds
DESIGN.md — the draft integration architecture: per-session helper process
linking bundled libmpv on all platforms (finally full-featured + Wayland-
agnostic Linux), JSON-over-stdio control evolving the Linux wid protocol,
unchanged EmbeddedMpvSession renderer contract, shm generations for resize,
packaging via the existing vendored-runtime tooling, rollout behind its own
flag with the docked path as default.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* spike(embedded-mpv): auto-detect Homebrew prefix and Node headers for Intel Macs
BREW_PREFIX was hardcoded to /opt/homebrew (Apple Silicon) and NODE_INC to
one nvm version; both now resolve via brew --prefix and the PATH node's
execPath, so the pending Intel Mac run needs no Makefile edits. README gets
a fresh-machine checklist.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* spike(embedded-mpv): self-contained measurement bundle for machines without Node/pnpm
make-bundle.sh assembles a tarball with the spike sources, vendored N-API
headers (Makefile prefers them when present, so no Node install is needed),
pre-generated 4K HEVC/HDR10 test clips, and an official Electron dist
download for the target arch. collect-results.sh builds and runs the full
RESULTS.md scenario suite automatically (plus an optional --long 10-minute
run) and writes one results-<host>-<date>.txt to send back. Target-machine
prerequisites shrink to Xcode CLT + brew mpv — built for the pending Intel
Mac baseline run.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* spike(embedded-mpv): support MacPorts libmpv and legacy-macOS bundles
Makefile and collect-results.sh now detect libmpv in the Homebrew prefix or
MacPorts /opt/local (Homebrew is unsupported on legacy macOS like High
Sierra; 'sudo port install mpv +libmpv' provides libmpv there). make-bundle
takes ELECTRON_VERSION/BUNDLE_SUFFIX overrides — Electron 27+ needs macOS
10.15, so High Sierra bundles ship Electron 26.6.10 (LSMinimumSystemVersion
10.13).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* spike(embedded-mpv): scope macOS frame-copy engine to Apple Silicon only
Owner decision 2026-07-10: skip Intel Mac measurements and gate the future
frame-copy engine on arm64. Intel Macs able to run the app at all are a
shrinking 2015-2020 cohort and keep the docked/external/web player paths;
the macOS hardware gate closes with the M1 Pro numbers, and remaining
hardware risk moves to the Windows/Linux ports.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(embedded-mpv): frame-copy helper process and shm frame reader (native layer)
iptvnator_mpv_helper: one-process-per-session libmpv host that renders
offscreen at viewport size (headless CGL + async PBO ring, validated in
spikes/mpv-frame-copy), publishes BGRA frames into a seqlock shm ring with
resize generations, plays audio directly, and speaks a stdio protocol —
tab-separated commands in, JSON events out. The snapshot event mirrors
NativeEmbeddedMpvSessionSnapshot; status semantics (END_FILE reasons,
eof-reached with keep-open, pause gated on loaded path, fatal-only status
flips) are ported from embedded_mpv.mm.
embedded_mpv_frame_reader.node: plain-C N-API reader the preload script
uses to memcpy the newest complete frame into a V8 ArrayBuffer (Electron's
memory cage forbids zero-copy). Stub exports off macOS.
Both build as extra binding.gyp targets through build-embedded-mpv.js; the
helper gets the same libmpv dependency-path rewrite + ad-hoc re-sign as the
addon and is validated by the forbidden-link check.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(embedded-mpv): frame-copy engine wiring in main process and preload
EmbeddedMpvFrameCopyAdapter implements the NativeEmbeddedMpvAddon surface
over a per-session helper process (spawn, stdio protocol, snapshot cache,
graceful quit->SIGTERM->SIGKILL teardown), so EmbeddedMpvNativeService
reuses its polling/diff/power-blocker/recording logic unchanged. The
IPTVNATOR_ENABLE_EMBEDDED_MPV_FRAME_COPY flag (darwin/arm64 only) routes
getAddon() to the adapter and reports engine: 'frame-copy' in support.
The preload frame pump loads the shm reader addon, copies the newest frame
once per rAF into a reused buffer, and uploads it to WebGL2 on the
renderer's canvas — no frame data crosses the contextBridge; the bridge
only gains attachEmbeddedMpvFrameView/detachEmbeddedMpvFrameView. The
experiment flag relaxes the window sandbox for that native require;
contextIsolation and nodeIntegration:false stay on.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(embedded-mpv): frame-copy canvas mode in the player component + docs
EmbeddedMpvPlayerComponent renders <canvas data-embedded-mpv-frame> when
support reports engine 'frame-copy' and the session controller starts/stops
the preload frame pump around the session lifecycle. The bounds provider
skips HIDDEN_BOUNDS and the popover cutout for this engine — the canvas is
ordinary DOM, dialogs and popovers stack above it natively; bounds sync
still drives the helper's render size. Adapter unit tests cover spawn args,
snapshot caching, shm generations, protocol encoding, unexpected-exit
mapping, and dispose escalation. Architecture doc and CLAUDE.md describe
the engine, its flag, and the sandbox trade-off.
Verified end to end in the built app (M1 Pro): engine detection, helper
spawn, lavfi playback onto the canvas via CDP-injected smoke — including an
orientation fix (helper FLIP_Y already yields texture-order rows; the pump
shader must not flip uv again).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(embedded-mpv): close helper stdin on dispose + lifecycle logging
Live testing surfaced a stray idle helper that survived a session switch;
until the root cause is pinned down, dispose now also closes the child's
stdin (the helper exits on EOF) as a second kill path besides quit ->
SIGTERM -> SIGKILL, and spawn/dispose/exit are logged with the session id
so leaks are attributable.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(embedded-mpv): reap sessions when the renderer reloads or crashes
Root cause of the stray idle helper found during live testing: session
teardown lives in the renderer's Angular lifecycle, which never runs on a
renderer crash or hard reload — the main process kept the session (and its
frame-copy helper process / native mpv handle) alive until app shutdown.
EmbeddedMpvNativeService now watches the main window's webContents for
render-process-gone and did-navigate (full reloads only; in-app Angular
routing emits did-navigate-in-page) and disposes every session. Applies to
both engines. Verified live: location.reload() during frame-copy playback
logs 'Disposing 1 session(s): renderer reloaded' and the helper exits
cleanly. Regression test drives both events against the service.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(embedded-mpv): Settings toggle for the frame-copy engine
New 'Embedded MPV: frame-copy engine' checkbox in Settings > Playback,
shown only when the machine can run it (macOS arm64 with the helper binary
present — support now reports frameCopyAvailable). The choice persists to
the main-process config store because the engine relaxes the window sandbox
for the preload frame pump, which is fixed at window creation: main.ts
reads the store before creating the window and sets the engine env var; an
explicitly set env var (including '0') always wins, and the UI shows a
restart hint while the saved choice differs from the active engine.
Localized in all 18 locales.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(embedded-mpv): aspect-fit rendering in the frame-copy helper
The helper now observes dwidth/dheight and renders its FBO at the
aspect-fit size of the video inside the requested viewport, bumping a shm
generation on change — letterbox bars are never baked into frames (the VOD
watch shell's ~2:1 box no longer shows black side bars; the canvas
background is transparent so the sides show the app surface, while
fullscreen keeps its black backdrop). Frames also get smaller than the
viewport when aspects differ, trimming copy cost. Aspect override changes
refit automatically. Snapshots now carry videoWidth/videoHeight, and the
adapter forwards IPTVNATOR_EMBEDDED_MPV_AUDIO_DELAY to mpv's audio-delay
for lip-sync tuning until proper calibration lands.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(embedded-mpv): require frame-copy artifacts in macOS package validation + docs
macOS packages that ship embedded_mpv.node must also ship the
iptvnator_mpv_helper binary and the embedded_mpv_frame_reader.node addon —
they come out of the same binding.gyp run, and a package missing them would
silently lose the frame-copy engine. Covered in the package-identity test.
Architecture doc and CLAUDE.md document the Settings toggle, aspect-fit
rendering, audio-delay passthrough, and the renderer-reload session reaping.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(electron): inline TS helpers so the sandboxed preload keeps working
The frame pump's async/await (target es2015 + importHelpers) made webpack
externalize tslib in main.preload.js. Sandboxed preloads can only require
Electron's built-in module whitelist, so the entire preload script failed
to load and window.electron disappeared for every run without the
frame-copy flag. importHelpers:false for electron-backend keeps the preload
bundle self-contained — and future async code in preload can no longer
silently reintroduce the breakage. Verified live: sandboxed run now has the
bridge, reports engine 'native' and frameCopyAvailable true.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* docs(embedded-mpv): commit the frame-copy analysis handoff + source inventory
The 2026-07-10 analysis that led to this branch now lives next to the spike
(spikes/mpv-frame-copy/ANALYSIS.md), and the architecture doc's What To
Commit section lists the frame-copy engine sources.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(embedded-mpv): address review findings on the frame-copy engine
- Stale pump attach can no longer win over a newer session: attach/detach
bump a shared epoch and async attach waits re-check it after every await,
so an attach for a replaced session aborts instead of installing itself
(greptile P1).
- A failed frame-view attach (no canvas, no WebGL2, reader missing) now
disposes the session and surfaces the error UI instead of leaving audio
playing behind a black canvas (codex P2).
- A stale frame-copy opt-in without the helper binary falls back to the
native engine instead of reporting embedded MPV unsupported, and the
Settings checkbox stays visible while a saved opt-in exists so it can
always be cleared (codex P2). Regression test covers the fallback.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(packaging): make the darwin frame-copy packaging test host-agnostic
On non-macOS CI hosts validatePackagedEmbeddedMpv also reports that macOS
link validation needs a macOS host, so the success-path assertion now
checks only the frame-copy artifact requirement instead of expecting an
empty error list.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* docs(embedded-mpv): Windows/Linux porting handoff for the frame-copy engine
Self-contained entry point for porting sessions on other machines: current
state and coordination constraints, per-OS task lists (Linux EGL first,
then Windows WGL + named shm — the decisive iGPU perf gate), the
hard-won gotchas from the macOS integration (preload/tslib sandbox
breakage, V8 memory cage, frame orientation, stale-attach epoch, dispose
escalation, node-gyp naming, snapshot protocol semantics), testing
recipes, and the suggested milestone order.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* docs(embedded-mpv): branching and merge strategy in the porting handoff
Port work goes to stacked branches off the frame-copy branch (PR base =
frame-copy branch, sequential merges, stack depth one), never into the
frozen PR #1169 branch itself.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* docs(embedded-mpv): drop stale uncommitted note from porting handoff
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(embedded-mpv): harden frame-copy helper startup
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* fix(tmdb): fall back to original language for missing trailers too
TMDB language-filters videos the same way it filters overviews: a
Russian-only title has its trailer tagged iso_639_1=ru, so an en-US
request returns no trailer — the trailer only appeared when the user
switched the app language. The original-language fallback previously
triggered on an empty overview only, so titles that had a translated
overview but a language-only trailer never got one.
detailsFallbackLanguage now also triggers when the payload has no usable
YouTube trailer, and fillDetailsFromFallback fills the overview and the
trailer independently — a present app-language overview is kept while the
trailer is pulled from the original-language payload. The fallback fetch
stays best-effort and cached per language.
Tests: extended tmdb-language-fallback.spec.ts (trailer-missing trigger,
independent overview/trailer fill, YouTube-only guard). Docs updated.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* test(tmdb): cover combined overview+trailer fallback fill
Adds the missing case where the primary payload lacks both overview and
trailer and the fallback supplies both — the two fill branches are
independent (greptile).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* feat(dashboard): TMDB trending rail and hero enrichment (backdrop, badges, S/E)
Two dashboard additions, both async after first paint so the page renders
exactly as fast as before:
Trending rail ("Trending this week", dashboardRails.tmdbTrending toggle,
default on, rendered only when TMDB is opted in AND the Electron DB
worker is available):
- TmdbTrendingService fetches /trending/{movie,tv}/week (one request
each, cached one day per language in tmdb_metadata under
trending:week), merges by popularity; exposed via the enrichment
facade (getTrendingWeek)
- DashboardTrendingService matches the titles against imported Xtream
playlists with ONE batched DB_MATCH_TITLES request, applying the same
two-tier + year-compatibility rule as actor pages; matched cards show
the playlist name and navigate straight to the detail view, unmatched
cards open the global search prefilled (?q=)
- The load fires only after the dashboard's own recent/favorites data
is in (never competes for the worker at startup) and once per session
- DashboardRailCard gained optional queryParams for the search links
Hero enrichment:
- DashboardHeroTmdbService patches the hero with a TMDB backdrop (only
when the item has none), a rating badge and up to two genre chips —
via the enrichment facade, so previously opened items resolve from
the SQLite cache without network; memoized per title per session,
staleness-guarded against hero changes in flight
- Series heroes show the tracked "S{n}·E{n}" badge from the playback
position; the watch-progress bar no longer applies to live heroes
Settings: new dashboardRails.tmdbTrending toggle in Settings > Dashboard.
i18n: 3 new keys translated into all 17 locales via tools/i18n patches.
Tests: dashboard-trending.service.spec.ts (gating, matching, year guard,
single-flight); settings fixtures updated. Docs updated
(tmdb-metadata-enrichment.md Dashboard Integration section, CLAUDE.md).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(dashboard): resolve hero TMDB extras for Stalker embedded-series items
Stalker vclub items carry type 'movie' in activity rows but are TV shows
on TMDB, so the hero's movie lookup found no confident match and the
backdrop/badges never appeared — while the detail view (which resolves
via is_series) showed them. When a movie-typed hero item has no movie
match, retry the lookup as TV: the detail view has usually already
cached that resolution, and misses are negative-cached.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(tmdb): cross-portal "Similar" rail — Stalker gets it, Xtream gains other-portal matches
The Similar rail only existed on Xtream because it matched against the
locally loaded catalog; Stalker catalogs are server-paginated, so its
detail views had no rail despite tmdb_recommendations being cached.
New CrossPortalSimilarService (libs/services) matches recommendations
against ALL imported Xtream playlists with one batched DB_MATCH_TITLES
worker request — the same two-tier normalized-title + year-compatibility
rule as actor pages and the trending rail. Electron-only; resolves to []
in the PWA.
- Stalker: the shared VodDetailsComponent (movies; covers catalog and
inline detail hosts) and stalker-series-view (series) now render a
"Similar" rail from cross-portal matches, each card badged with the
source playlist and navigating into that portal's detail view.
- Xtream: vod/serial detail rails keep instant local-catalog matches and
append cross-portal matches (current playlist excluded, deduplicated
against local hits by normalized exact title), also playlist-badged.
- Loads async after the detail view renders, staleness-guarded; the
section only appears when there is something to show.
Tests: cross-portal-similar.service.spec.ts (PWA gate, navigation
targets, playlist exclusion, type/year guards). Docs updated.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(tmdb): drop TestBed from cross-portal similar spec
The services Jest target has no @angular/core/testing (same CI failure
as the cache spec earlier) — construct the service via Injector.create +
runInInjectionContext instead.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(dashboard): address PR review — reactive opt-out gates, retry after empty trending load
- Trending rail and hero TMDB extras now vanish immediately when the
TMDB opt-in is switched off mid-session: the render computeds read the
settings signal through isAvailable/isEnabled instead of trusting data
loaded earlier (Codex P2 ×2).
- loadedOnce latches only after a successful non-empty load, so a
transient TMDB outage on first visit no longer suppresses the rail for
the whole session — the next dashboard visit retries (greptile P2).
- Unified the duplicated heroTmdbExtras() read in the hero computed
(greptile P2).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* fix(tmdb): restore trailers in packaged builds, fall back to original language for missing text
Two production-only gaps reported against the packaged 0.22 build:
1. Trailers showed YouTube "Error 153 — Video player configuration
error". YouTube requires a Referer on /embed requests; the packaged
app loads the renderer from file://, which never sends one (dev works
because the origin is localhost). The Electron main process now
injects the project site as Referer for YouTube embed hosts when the
header is absent (request-header-overrides.service.ts, registered at
startup via registerStaticHeaderShims). Existing Referers are never
overwritten, other hosts untouched.
2. Descriptions vanished for content whose original language differs
from the app language: TMDB does not fall back on missing
translations, so a Russian-only series has empty overviews in en-US.
When the app-language payload carries no text, the enrichment
refetches once in the content's original_language and fills ONLY the
missing text fields (tmdb-language-fallback.ts): details overview,
season overview and per-episode names/overviews. Genres, credits and
artwork stay in the app language; both language rows share the cache.
Tests: 3 new shim cases in request-header-overrides.service.spec.ts,
new tmdb-language-fallback.spec.ts (9 cases). Docs updated
(tmdb-metadata-enrichment.md).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(tmdb): address PR review — best-effort fallback fetch, /embed/ scope, names-only trigger
- A failing original-language fallback fetch no longer costs the
already-fetched primary details payload (Codex P2) — it is wrapped
best-effort and the primary payload is returned on error.
- The YouTube Referer shim now scopes to /embed/ paths, so regular
www.youtube.com requests keep their real (missing) Referer (greptile).
- seasonNeedsTextFallback also triggers when episode overviews exist but
ALL episode names are empty (partially translated seasons; greptile).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* feat(ui): slim ContentHero to non-scrolling hero block
The page scroll container moves out of the hero into the upcoming
PortalDetailShellComponent; the hero no longer hosts a default content
slot for player/seasons/extras.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(ui): add PortalDetailShellComponent with browse/watch states
Two-state detail layout shell: owns page scroll, hero collapse animation
(~300ms, reduced-motion aware), Escape-to-close-player handling, and an
About block that re-stamps host-provided *detailTags/*detailMeta
templates in watch state. The [detail-player] slot is never wrapped in a
shell conditional so the host-owned player subtree survives state
changes.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(playback): restyle inline player header as now-playing bar
Adds a back button (emits the existing closed output — in watch state
back returns to browse, not route navigation) and replaces the
hardcoded header strings with i18n keys (PORTALS.NOW_PLAYING,
CLOSE_PLAYER, BACK_TO_BROWSE).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* refactor(xtream): migrate VOD and serial detail templates to detail shell
Hero chips/meta/actions become *detailTags/*detailMeta/*detailActions
templates; the inline player moves to the full-width [detail-player]
slot; trailer and similar rail move to [detail-extras]. Escape and the
now-playing back button close the inline player back to browse.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* refactor(stalker): migrate shared VodDetails and series view to detail shell
The shared VodDetailsComponent (Stalker VOD, collections, search) and
StalkerSeriesViewComponent move their hero content into
*detailTags/*detailMeta/*detailActions templates and host the inline
player in the full-width [detail-player] slot.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(ui): season tabs replace two-level seasons navigation
SeasonTabsComponent renders a pill row (dropdown beyond 6 seasons), an
optional season description and a back-to-playing chip. The season
container auto-selects a season (inline-playing episode's season →
most recent in-progress → first), emits seasonSelected for
auto-selections so lazy-load/enrichment hooks keep firing, highlights
the inline-playing episode, and loses the seasons grid + 'Back to
seasons' button. Download/progress helpers move to pure utils; season
strings are now translatable.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(portals): season descriptions under the season tabs
Xtream reads season overviews from get_series_info; Stalker exposes the
TMDB season overview through TmdbEnrichmentService.getSeason (same
cache rows as the episode enrichment) and keys it per tmdbId so views
reused across detail navigations cannot leak descriptions.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* docs: document two-state detail layout and season tabs
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(ui): app-prefix detail template directive selectors, build fixes
detailTags/detailMeta/detailActions → appDetailTags/appDetailMeta/
appDetailActions per @angular-eslint/directive-selector; type the
Escape host listener as Event; drop a redundant ?? in season tabs.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* chore(i18n): translate new detail-page keys across all 17 locales
13 PORTALS.* keys (now-playing bar, About block, season tabs, episode
empty states) translated via the i18n-fill workflow.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(playback): keep now-playing bar actions inside the viewport
The nowrap title made the header grid track grow to min-content and
pushed Copy/Close out of the page; min-width: 0 on the header/title row
lets the title truncate instead.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(ui): expandable season description and episode info dialog
Season descriptions get the same show more/less toggle as the hero
(new shared ExpandableTextComponent); episode cards and list rows get
an info button that opens a dialog with the full plot, duration, air
date and a Play action — card heights stay fixed and click-to-play
semantics stay intact. Also: drop the per-frame backdrop blur on the
player card and make the watch-enter scroll instant so the hero morph
stops competing for frame budget.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(portals): re-run TMDB season enrichment when the show match arrives
With season tabs the first seasonSelected fires as soon as seasons
load — usually before the async show-level TMDB enrichment has written
tmdb_id — so the season fetch silently no-oped and was never retried,
leaving episode stills/plots and the season description empty. Both
detail hosts now key the fetch on (tmdb_id, selected season) in an
effect, so it runs whichever arrives last. Also adds breathing room
above the About divider in watch state.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(ui): move episode info button into the card body, compact overlay actions
Three overlay buttons crowded the thumbnail on hover and the new info
button had no styling at all. Info belongs with the content text: it
now sits as a quiet ghost button at the end of the title row (revealed
on card hover, MDC touch target clamped so it cannot swallow the
play-on-click card area). Download + watched-toggle stay on the
artwork, shrunk to 28px with a subtle edge border. List view already
used the shared action style and is unchanged.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(ui): single watched indicator-toggle on episode cards
Watched state was signalled three times on one card: the green badge,
the hover toggle in the top-right cluster, and a checkmark appended to
the title. Now one control does both jobs — a toggle at the top-left
of the artwork that looks like the old green badge when watched
(always visible) and appears as a ghost circle on hover when not.
The right cluster keeps only the download action; the title checkmark
is gone.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(playback): back arrow in watch mode navigates straight to the list
Leaving the theater took two clicks: back to browse, then back to the
list. Watch state already shows everything browse offers (episodes,
About, extras), so the bar's back arrow is now route-level back
(new backClicked output wired to each host's goBack), while Close
player and Escape keep exiting to browse without navigating. Replaces
PORTALS.BACK_TO_BROWSE with a generic top-level BACK key (all locales).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(ui): compact episode stills in the list view
When a season has genuinely distinct per-episode art (TMDB stills or
provider images), list rows show a 96x54 thumbnail with the episode
number riding on it instead of the number square. When every episode
repeats the same image (providers often send the series poster), the
plain number square stays — a column of identical posters is worse
than none.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* test(ui): compact list-thumbnail spec under the max-lines limit
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(ui): address Greptile review — real fallback asset, untracked overflow measure
default-episode.png never existed (pre-existing in the grid card, copied
into the list thumbs) — both onerror handlers now fall back to the real
default-poster.png. The expandable-text and content-hero measuring
effects read the expanded signal via untracked so toggling show
more/less no longer rebuilds their ResizeObservers.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* feat(tmdb): opt-in TMDB metadata enrichment for Xtream and Stalker portals
Adds an opt-in TMDB integration (Settings > Metadata) that enriches
detail views with a field-level merge — the provider stays authoritative
for stream data, TMDB fills editorial fields when the match is confident.
Enrichment:
- Movie/series details: plot, cast (avatar chips), director, genres,
rating, poster/backdrop, official YouTube trailers
- Confidence-gated matching: provider tmdb_id trusted; otherwise
normalized-title search with year gate (±1; series accept earlier
premieres), season-suffix stripping, Cyrillic search-language override,
and language-prefix fallback variants
- Lazy season/episode enrichment: real episode names, overviews, stills
- "Similar" rail (Xtream): TMDB recommendations matched to the catalog
- Actor pages per portal with full filmography, availability filter and
an Electron-only "All portals" scope backed by a batched DB_MATCH_TITLES
worker op over the trigram FTS index
Infrastructure:
- SQLite cache table tmdb_metadata (details, search verdicts, seasons,
persons; per-language, TTL-guarded), in-memory fallback for the PWA
- Settings: enable toggle, own-API-key override with a live "check key"
button; TMDB attribution in Settings and About
- Embedded key stays an empty placeholder; CI injects TMDB_API_KEY via
tools/tmdb/inject-tmdb-key.mjs when the secret is configured
- normalizeTitle shared between renderer and DB worker
- CSP: allow YouTube embeds (frame-src was 'none'; trailers never worked)
Fixes and refactors along the way:
- fix(stalker): Advanced Search sent bare get_ordered_list requests and
skipped the auth handshake when isFullStalkerPortal was missing on the
active-playlist meta — full portals answered "Authorization failed."
and search looked empty; now mirrors the catalog request shape and
routes through makeAuthenticatedRequest with URL-based detection
- fix(stalker): TMDB fields survive info re-normalization; detail views
prefer the store copy patched by async enrichment over stale snapshots
- refactor(xtream): split oversized vod/serial detail components into
component-scoped playback services; detail routes re-initialize on
route param changes (router reuses them for detail-to-detail nav)
- i18n: all new keys translated across the 18 locales
Docs: docs/architecture/tmdb-metadata-enrichment.md + CLAUDE.md updates.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(tmdb): provide route params observable to inline collection details, linearize regexes
The global-collection inline detail host builds a fake ActivatedRoute for
VodDetailsRouteComponent/SerialDetailsComponent with only snapshot.params.
Since the detail components now read route.params via toSignal() (detail->
detail re-init), the missing observable crashed component construction and
the content hero never rendered — broke dashboard-activation, favorites and
recent Electron E2E on all platforms. Provide the params observable
alongside the snapshot and assert it in the component spec.
Also resolves both CodeQL js/polynomial-redos alerts: bracket-stripping in
normalizeTitle now excludes opening delimiters inside the classes, and
youtubeEmbedUrl extracts watch?v= ids with a linear two-pass match instead
of "watch\?.*v=". Combining-diacritics range rewritten as explicit \u
escapes (greptile note).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(tmdb): surface TMDB-only VOD score in the rating badge, drop youtube.com from CSP
Review follow-ups on PR #1123: the Xtream VOD detail badge renders
rating_imdb, but the merge wrote the TMDB score only into `rating`, so a
TMDB-only score was never displayed (Codex P2) — fill rating_imdb when the
provider left it empty, mirroring the Stalker merge. All trailer iframes
are normalized to youtube-nocookie.com, so the extra youtube.com frame-src
allowance was dead surface (greptile) — removed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(tmdb): resolve confirmed review findings — matching correctness, race guards, cache schema
Fixes the confirmed findings from the PR #1123 code review:
- Stalker search: setSelectedContentType now runs BEFORE setSelectedItem,
so the TMDB enrichment gate in the selection hook no longer sees the
content type of the previously open tab (wrong/no enrichment after
ITV -> search -> movie).
- Title normalization is now two-tier (normalizeTitleKeys): the exact
normalized form keeps a trailing year, the base form strips it and
remembers the tag. Year stripping is anchored to the end of the title
("2001: A Space Odyssey" keeps its year) and language-prefix stripping
is UPPERCASE-only ("It: Chapter Two" is no longer amputated).
- All catalog matching (similar rail, actor pages, DB worker
DB_MATCH_TITLES) compares exact forms first and only accepts
year-stripped matches when the stripped tag is year-compatible (+-1)
with the TMDB year — "Blade Runner" (1982) can no longer claim a
catalog "Blade Runner 2049". CatalogTitleMatch carries the stripped
trailingYear so the renderer can apply the guard to worker matches.
- mergedBackdrops tolerates a plain-string backdrop_path; enrichment
merge+patch blocks are wrapped in try/catch so a malformed provider
payload can no longer become an unhandled rejection.
- loadGlobalMatches (both actor routes) guards against actor->actor
navigation races — a slow match for the previous person no longer
overwrites the current one's results.
- tmdb_metadata media_type CHECK widened to ('movie','tv','person') and
person rows now use the honest 'person' type (TmdbCacheMediaType).
Pre-release dev DBs with the narrow CHECK are rebuilt in place — the
table is a pure cache, so the migration is a self-healing
drop-and-recreate keyed off sqlite_master.
Docs updated (tmdb-metadata-enrichment.md, CLAUDE.md). New regression
coverage: title-normalization.util.spec.ts, two-tier cases in
tmdb-similar.util.spec.ts and title-match.operations.spec.ts.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* feat(epg): add vertical list view for the live EPG panel
Add an EPG list view — a vertical, single-day programme list — as an
alternative rendering of the live EPG panel, selectable via a new
Settings → EPG → "Guide view" toggle (epgViewMode: 'timeline' | 'list',
default 'timeline' so existing users see no change).
- New EpgListViewComponent (app-epg-list-view) mirrors
EpgTimelineComponent's input/output contract 1:1, so all four live
hosts (M3U player, unified live tab, Xtream, Stalker) swap the panel
with a plain @if and identical bindings.
- Reuses the shared view-agnostic EPG modules (classifyTimelineWhen,
hasProgramsForDateKey, epg-archive.util catch-up gating,
epg-summary.util collapsed-summary maths, epg-date helpers,
EpgProgrammeDialogService, app-epg-timeline-empty-state) — no
duplicated logic.
- Rows show time range, title, optional description, live progress on
the on-air row, catch-up "Watch" on past rows when archive playback
is available, and a details dialog; keyboard activation guards
nested buttons (target === currentTarget).
- Auto-focuses the on-air row on channel select, restores it across
collapse/expand remounts, and shows a sticky in-flow "On now" strip
(never overlaying rows) when the current programme is scrolled away;
all scroll maths is rect-based relative to the scroller.
- List mode raises only the inline panel height via an epg--list
modifier (--epg-inline-height clamp); timeline and collapsed heights
are unchanged.
- Setting flows end-to-end (Settings interface → DEFAULT_SETTINGS →
SettingsStore/StorageMap → segmented control in the EPG section);
Electron-only UI, PWA stays on the timeline default. i18n keys added
to all 18 locales.
- Tests: new component/row/utils/scroll-controller specs, settings
persistence spec, swap tests in all four host specs, and Electron
E2E for the settings round-trip and the rendered list view. Docs
updated (m3u-playlist-module.md).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(epg): address list-view review findings from Codex and Greptile
- Reset the list view to today when a new channel's programme set
arrives while the user is parked on another day (timeline parity):
the scroll controller now keys by the full programme-set identity
(programsFocusKey) and commits today before focusing, instead of
silently stranding the new channel on the stale day. (Codex P2)
- Centralise the 'timeline' fallback as a resolvedEpgViewMode computed
on SettingsStore; the four live hosts consume the derived signal
instead of duplicating the `?? 'timeline'` expression. (Greptile P2)
- Extract the component's reactive plumbing into
registerEpgListViewEffects(), bringing the component back under the
300-line guideline (290). (Greptile P2)
- Controller spec rewritten around programme-set fixtures with new
coverage: return-to-today on channel switch, day navigation left
alone, no-takeover when today has no data, empty-set no-op.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* fix(epg): drop malformed programmes from the list-view day filter
Reject programmes whose stop is not after their start in
buildEpgListRows — same as the timeline's buildTimelineBlocks. Bad
provider data would otherwise render impossible time ranges and could
even be offered as catch-up playable. (Codex P2 on e6fd0d08)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
* fix(security): harden Electron IPC against MITM, SSRF, path and injection risks
S1 TLS: validate certs by default on playlist/EPG fetches (opt-out via IPTVNATOR_ALLOW_INSECURE_TLS); new util/secure-https.ts.
S2: write-file IPC restricted to save-dialog-authorized paths.
S3: XTREAM_PROBE_URL guarded by assertRemoteUrlAllowed + maxRedirects:0; new events/url-safety.ts (+19 tests).
S4: EPG titles rendered via interpolation, not [innerHTML].
S5: downloads reveal/play limited to recorded download paths.
S6: Stalker cmd encoded (slash-preserving) to block query injection.
EPG-worker and Stalker fetches reject file://-style/credentialed URLs; LAN/self-hosted targets remain allowed.
* perf(player): lazy-load web video players via @defer
Wrap Video.js/HTML5/ArtPlayer in @defer (on immediate) so video.js, hls.js,
artplayer and mpegts.js split into a deferred chunk loaded on first playback
instead of eagerly on the player route. Embedded MPV (native) stays eager.
Spec uses DeferBlockBehavior.Playthrough.
* fix(player): remove leaked HTML video listeners on destroy
volumechange used a mismatched removeEventListener reference, while
loadedmetadata and timeupdate were never removed at all. Bind all three to
stable handler fields used for both add and remove, and add a teardown
regression test asserting each listener is detached on destroy.
* refactor(dashboard): extract pure navigation helpers from DashboardDataService
Move the 8 stateless link/navigation-state/type-kind helpers into a new
dashboard-navigation.util.ts so the routing logic is independently testable and
the 1260-line god-service shrinks. DashboardDataService keeps the public methods
as thin delegators (facade) so the public API and the single consumer
(workspace-dashboard-rails) are unchanged. First slice of the DashboardDataService
decomposition; verified by the existing service spec (33/33) and the app typecheck.
* fix(review): address PR feedback (IPv6 link-local, write-path cap, @defer placeholder)
- url-safety: broaden IPv6 link-local detection to the full fe80::/10 range
(fe80:: through febf::), not just the fe80:: prefix (+ regression tests).
- playlist.events: cap authorizedWritePaths (evict oldest past 32) so a save
dialog opened without a following write cannot accumulate entries until restart.
- web-player-view: add a @placeholder to each @defer (on immediate) player block
to avoid the one-frame blank/layout-shift before the chunk resolves.
* fix(security): close Electron network and download gaps
* test(downloads): cover cancellation and restart cleanup
* fix(downloads): address Greptile review gaps
* test(security): reproduce remaining Greptile findings
* fix(security): close remaining Greptile findings
* test(downloads): reproduce early database queue stall
* fix(downloads): release queue after setup failures
* test(downloads): reproduce completion queue stall
* fix(downloads): release queue after completion failures
* feat(ui): add custom title bar window controls for Windows and Linux
Hide the native title bar on win32/linux (titleBarStyle: 'hidden', frame
untouched so native resize borders and snapping keep working) and render
minimize / maximize-restore / close buttons in the renderer, mirroring the
existing macOS traffic-light setup.
- New WINDOW:* IPC contract (minimize, toggle-maximize, close, get-state)
handled in window.events.ts, resolved from the sender WebContents;
close goes through win.close() so window-bounds persistence still runs.
- WINDOW:STATE_CHANGED pushed on maximize/unmaximize/fullscreen so the
maximize/restore glyph stays correct for OS-triggered changes; controls
hide while fullscreen.
- WindowControlsComponent mounts once in app-root as a manual popover so
it stays in the browser top layer above CDK overlays (dialogs,
multi-EPG) - same behavior as macOS traffic lights.
- Theme-aware via CSS vars (--app-on-surface, --app-hover-overlay);
Windows-red close hover. Drag regions get right padding through a
body-level frameless-platform class.
- Gated by RuntimeCapabilitiesService.usesCustomWindowControls; PWA and
macOS never mount the controls.
Includes unit specs for the component and IPC handlers, an Electron E2E
suite (window-controls.e2e.ts), and a window-chrome section in
docs/architecture/workspace-shell.md.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(linux): upgrade Electron to 41 for frameless window decorations on Wayland
With the native title bar hidden, Linux windows lost the WM-drawn shadow
and rounded corners. Electron draws client-side decorations only on
native Wayland, and frameless-window CSD (GTK drop shadow + extended
resize boundaries) landed in Electron 41 - before that, frameless
windows render as plain rectangles.
- electron ^39.8.5 -> ^41.7.2 (Wayland auto-detected since 38.2; X11
sessions remain undecorated, matching other frameless Electron apps;
Windows keeps its DWM shadow and rounded corners).
- better-sqlite3 pinned to exactly 12.9.0: the last release shipping
prebuilt binaries for both Node 20 (ABI 115, Jest) and Electron 41
(ABI 145, runtime). 12.10.0 dropped the Node 20 prebuilds, forcing a
from-source build that fails without a C++ toolchain.
- pnpm override node-abi 3.85.0 -> 3.92.0 so electron-builder
install-app-deps can map Electron 41 to ABI 145.
Reviewed Electron 40/41 breaking changes: only the renderer clipboard
deprecation, which this app does not use.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(e2e): address review feedback and window-managerless Linux CI
- Skip the three window-manager-dependent E2E assertions (maximize
toggle, main-process state sync, minimize) on Linux CI: GitHub's
ubuntu runners drive Electron under xvfb without a window manager, so
maximize/minimize state never materializes there. Windows CI and
local Linux/macOS runs keep the coverage.
- WINDOW:TOGGLE_MAXIMIZE now returns the requested state instead of
re-reading isMaximized() right after the call, which races on Linux
window managers where maximize()/unmaximize() complete
asynchronously; the WINDOW:STATE_CHANGED push stays authoritative.
- Skip attaching window-state push listeners on macOS, where the
custom controls never mount and the IPC traffic had no subscriber.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(ui): gate custom window controls on the full bridge surface
Include getWindowState and onWindowStateChange in the
usesCustomWindowControls capability check — the controls rely on both
for initial state and for keeping the maximize/restore glyph in sync
with OS-triggered changes, so a partial bridge should not mount them.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
