Repository navigation
Merge upstream 2026-11-16 - #39
Merged
Merged
Conversation
…ewest cards behind the deck (plxnative#1405) With several servers, Home's Continue Watching opened scrolled to the first card of the server that answered first. The newer cards of a server answering after it were held behind the window, offset by their count, so DOWN from the hero landed on the wrong card. The cause was deck::place, not the row's focus or scroll. Every server's answer lands on its own. A whole-source answer (a first answer, or a refresh whose reload failed) joins the deck through settle_deck, and place seated every joiner by the rule for a deck the reader had paged along: its cards newer than the window's first card went behind the window, and the overflow was dropped from the head. place now judges whether the deck is at its head, with fit's own test on the lanes that held it before this call. At the head, a joiner's newer cards lead the window, and the overflow goes from the oldest end. A full window lets go of its oldest shown cards, which stay ahead as lookahead. A deck the reader moved keeps the old rule. A joiner with no rows (a server whose Continue Watching is empty) is seated at its head rather than past its tail, so it no longer reads as a lane the deck moved past and hides the next joiner's cards. Deck tests (red first) cover: a joiner at the head leads with its newer cards; the arrival order makes no difference; the newest cards fill a full window; a reader-moved deck keeps its window; an empty server between two others; an empty server alone. Home tests drive the real store (a new test hook, land_deck_lane_for_test) and the engine's DOWN from the hero. They cover a cold deck in either arrival order, with an empty server in between, and after a failed reload. They also check a single-server deck, and a reader who stays on the same card when cards land in front of it. Closes plxnative#1395 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…tive#1401) Bumps the python-tools group in /tools with 1 update: [markdown-it-py](https://github.com/executablebooks/markdown-it-py). Updates `markdown-it-py` from 4.2.0 to 4.3.0 - [Release notes](https://github.com/executablebooks/markdown-it-py/releases) - [Changelog](https://github.com/executablebooks/markdown-it-py/blob/master/CHANGELOG.md) - [Commits](executablebooks/markdown-it-py@v4.2.0...v4.3.0) --- updated-dependencies: - dependency-name: markdown-it-py dependency-version: 4.3.0 dependency-type: direct:production update-type: version-update:semver-minor dependency-group: python-tools ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
…e#1392) ## Draft scope Continue the new UI migration on current main, separately from already-merged plxnative#1366. No automatic merge or ready-for-review transition is requested. - Execute all eleven app-effect families through one owned dispatcher FIFO and taken-head runner, retaining sender lifetime, frozen ReturnState, budgets, cancellation/refusal and exactly-once completion. - Capture screen runtime policy, all thirteen effective profile preferences, install/running language and complete immutable player presentation at the app boundary. Controlled hosts use explicit inputs and explicit native-port capability, never ambient diagnostic/preference fallbacks. - Bind strict CURRENT V11 bootstrap/writer/reader contracts; reject unknown/noncanonical fields and underspecified unpublished profiles. Refund only dropped scalar-row JSON reservations, without changing peak/retained-data limits. - Add targeted source and negative-control gates, captured-policy/runtime tests, strict replay refusal accounting and the exact-byte current Filmography fixture. Existing recording directories and ALPHABET remain unchanged. CURRENT requires SAME; only the approved historical dynamic-Content scheduler profile expects exact early refusal, never a fixture-name/hash exception. ## Ownership and integration Based on main `8859e4bad1838f154a90bbb5885fd48d928c7d72`, independently confirmed through origin and GitHub git/ref. The patch was extracted semantically against the recovered source's original base, excluding EOL-only recovery churn and unrelated work, then applied three-way in a separate clean branch/worktree. The source worktree's HEAD, logical index, raw index and tracked bytes are independently checked unchanged. 129 owned paths: 127 from reviewed migration changes plus four newly added paths, minus two excluded paths. Exclusions retain incoming main's outgoing-stop timing fix and the entire original fixture inventory, including its README. No source-tree replacement, old-PR restack, closed-branch push, toolchain/dependency downgrade, private report or local runtime is included. Incoming Rust 1.100.0, dependency identities, Canon optimization and CI timing fixes remain intact. Generated-unit-enum proof adds only the reviewed exact merged machine identity; generator grammar, namespace/writer contracts and negative controls remain strict. ## Verification Exact proposed Git tree `f26920af41cd524121f5b725fa909b8fcb25eaab` matches the native staged tree and independent Linux mirror. On incoming pinned Rust 1.100.0 / Python 3.12.16, all inherited GIT_* variables removed and bounded two-job Rust execution: - Final full `make check`: PASS, 568.98 seconds. Cargo lint/default/hostsim and Python phases all pass; 136 registered rest steps pass. Harness: 745 tests, OK, four existing skips. Default app 1418/0, screens 2168/0, appkit 955/0, media 2134/0; hostsim app 1621/0 and media 2414/0. Media retains its 100 existing ignored tests. - All six final configurations PASS: shipping, bare, play-uri, play-feed, devtriggers and hostsim. - Focused current/legacy bootstrap and actual fixture preflight, replay parser 24, fixture/privacy class 13, CI registry 16, generated-proof mutations, UI-boundary controls, FIFO routing and poisoned nested-Git isolation PASS. Full DepGates selector regression PASS after integrating the new rules with main's selector contract. - Full-patch outbound/private-path scan and guard controls PASS; native diff whitespace check PASS. All original fixtures and ALPHABET are untouched; added manifest/segment match their officially imported bytes exactly. Two integration-oracle failures were fixed before the final run: patch export framing at the manifest's absent final newline, and new source rules not yet registered in main's selective gate protocol. Neither fix changes fixture bytes, weakens production policy or relaxes a test. Prior recovered-source software-renderer results (six SAME plus two exact historical refusals) remain separate evidence, not a fresh renderer claim for this integrated tree. Remote exact-head CI, simulator replay and ARM artifacts are pending publication and must be reviewed independently. ## Deliberately incomplete acceptance This is a DRAFT, not whole-project UI acceptance. ARM/native build artifacts require remote CI verification; local SDK is absent. Native TV/video-plane/glyph/performance, accepted accelerated visuals, WSLg transport, the full 70-route scenario matrix, architecture census and cached-modal acceptance remain unverified. Earlier private software-renderer evidence is logical replay evidence only, not hardware or visual approval. No account/provider/TV/system changes or SDK downloads were made.
…tive#1408) (plxnative#1409) * fix(posters): store a slot's key whole, so long art paths load (plxnative#1408) The poster store copied each built request path into a fixed 256-byte array per slot and refused any key over 255 bytes, because a truncated key would never match its probe again and would thrash the store. That gate was correct, but the fixed size capped reachable content: a Live TV channel logo wrapped in a wsrv.nl resizer link built into a 326-byte /photo/:/transcode path and stayed a skeleton while PMS served it fine. Pslot::key is now a String (set_key reuses its capacity, so a warm store claims without allocating), is_fetchable rejects only an empty key, and built_key no longer refuses long paths. KEY_MAX, PT_KEYLEN and the one-shot "REFUSED art key" log are gone. The disk tier (imgcache) already names files by a SHA-256 of the request identity, so it has no length assumption. Regression test: an_art_path_longer_than_255_bytes_claims_and_keeps_a_slot builds the issue's shape (>255 bytes), claims a slot on the first draw, keeps the key intact, and finds the same slot on the next draw. It failed before the fix (the key came back empty) and passes after. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01P6ZnFfD3NYt5YCak7vvCJf * fix(posters): drop the stale KEY_MAX cap classification and lookup comment Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01P6ZnFfD3NYt5YCak7vvCJf --------- Co-authored-by: Claude <noreply@anthropic.com>
…tive#1410) Version updates (no advisory) move to a monthly schedule with a cooldown (cargo: 7 days patch, 14 minor, 30 major; actions, pip, npm: 7 days), so a release has been public for a while before it reaches a PR. The cargo PR limit drops to 3. The rust-toolchain entry stays weekly. Held crates are now held with `update-types` (all three levels) instead of a bare `dependency-name`. Per GitHub's docs, update-types only affects version updates, while a bare name rule also suppresses security updates, so the ICU4X and qrcodegen holds were silently blocking advisory fixes too. zip and ed25519-compact, both exact `=` pins, join the holds (zip 9.2.2 -> 9.5.1 in plxnative#1402 added four crates and changed nothing this binary calls). Groups carry `applies-to: version-updates`, so security updates come as their own PRs. Claude-Session: https://claude.ai/code/session_01P6ZnFfYtDRxLqhuvaX1ReC Co-authored-by: Claude <noreply@anthropic.com>
The repositories moved from GLinnik21 to the plxnative organization. GitHub redirects the old paths, but two workflow guards compared github.repository with 'GLinnik21/plx-native' and have silently skipped since the move (build-bench.yml history jobs, fuzz.yml's main run), and tools/hud_registry_dispatch.py posted to repos/GLinnik21/plxnative-huds/dispatches, which answers HTTP 307 that the dispatch does not follow. Swap the owner to plxnative in workflows, issue/discussion templates, the site, docs, README, PRIVACY, SECURITY, every locale's settings.json, ipkroot/ctl/control, the webOS Homebrew manifest, tools/ and their tests. Every changed line is that one-word swap (156 in, 156 out). Left alone: the crates' Cargo.toml repository lines, the i18n localization URL and its qr.rs test mirror (inputs to ci/proofs/platform-generated-enums.json), historical release audits, notes and measurements, FUNDING.yml and author identities, and the plex-native-poc user agents.
Nightly releases move to [plxnative/plx-native-nightly](https://github.com/plxnative/plx-native-nightly), so watching this repository's releases announces real releases only. That repository was created for this on 2026-11-14 and so far holds only its README. ## What changes - **Publish** (`nightly.yml`): - Pushes the `nightly/v*` tag **here**, at the built commit. - Then creates the prerelease in `plxnative/plx-native-nightly` with the new `NIGHTLY_RELEASE_TOKEN` secret. - If the release fails, the tag is deleted again, so a re-run isn't refused as a same-day rebuild. - The post-publish checks read from the new repository. - The "uploader is `github-actions[bot]`" check is dropped: with a token, the uploader is the token's owner. - **Tags stay here.** `plan`'s skip decision diffs against the previous tag, and the notes' compare link spans two tags. The release repository has none of this repository's commits, so its own copy of each tag sits on its `main`. - **`ci/nightly.py`:** - `latest-json`, `repo-json` and `prune` read **both** repositories, so the switchover never empties the Homebrew repository. The site keeps serving the newest pre-move nightly until the first one lands in the new place, and `prune` retires the old ones here on the same 30-day clock. - A release's commit now comes from this repository's tag, not from `target_commitish`, because a release in the new repository targets that repository's `main`. - New `release-repo` subcommand so the workflow and the reader share one constant. - **Docs:** `README.md` and `docs/distribution.md` say where nightlies live. ## Needs before merge - [ ] A `NIGHTLY_RELEASE_TOKEN` Actions secret on plx-native: a fine-grained token owned by **plxnative**, scoped to `plx-native-nightly` only, with **Contents: read and write**. Without it the next scheduled nightly fails at the "Refuse to overwrite" step with a clear error; nothing is published half-way. - [ ] `make check` on the Mac. This PR changes no Rust, Makefile or Cargo files, so no build-bench table is needed. ## Verified - `python3 ci/nightly.py --selftest` passes, including new cases for which commit a release names. - `python3 -m unittest ci.test_ci_workflows ci.test_ci_timeouts`: 99 + 1 new pass. The new test fails against the old workflow. - Run live, read-only, against the real repositories: - `latest-json` and `repo-json` still serve `0.11.0-nightly-20261115` with commit `92e63a8`. - The tag lookup resolves that same commit. - `prune --dry-run` would delete only `nightly/v0.9.0-nightly-20261017` here. - Not verified: a real publish into the new repository. That first happens on the schedule, or a `workflow_dispatch` after merge, once the secret exists. This PR's own nightly dry run builds and stages but publishes nothing. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
…-11-16 # Conflicts: # .github/DISCUSSION_TEMPLATE/compatibility.yml # .github/ISSUE_TEMPLATE/bug_report.yml # .github/ISSUE_TEMPLATE/config.yml # .github/dependabot.yml # .github/workflows/nightly.yml # PRIVACY.md # README.md # SECURITY.md # ci/allow/forks/screens.txt # ci/rc.py # ci/readonly_generated.py # ci/verify-published.sh # docs/known-issues.md # ipkroot/ctl/control # rust-modules/screens/src/preferences.rs # rust-modules/src/app/adapters/poster.rs # rust-modules/src/app/run.rs # tests/test_harness.py
…ive's names The nightly is published to its own repository now (plxnative#1411). Keep FinNative's package names and its no-telemetry build on upstream's new steps, so ci/test_ci_workflows.py reads the shape it pins. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JLzd9SJCCZ5zFGv23XMA8h
Since plxnative#1392 the Playback page reads its settings from the captured inputs, so the picker test hands Convert audio's stored value in the capture like every other row. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JLzd9SJCCZ5zFGv23XMA8h
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Merges PlxNative
mainup to8e8330ab(plxnative#1411), 9 commits, as a merge commit. No upstream release is pending, so this is not release-aligned; PLXNATIVE_BASE stays0.11.0-dev.Resolutions:
ipkroot/ctl/control,ci/verify-published.sh,ci/rc.py). docs/known-issues.md takes the new address for PlxNative issue Playback crash - 43UM7400PLB plxnative/plx-native#74..github/workflows/nightly.yml(disabled on this repository): upstream's new release-repository flow, with FinNative's package names and no telemetry..github/dependabot.yml: FinNative's version updates stay off (limit 0); bumps arrive with upstream merges.inputs.preference::<AudioConversionPref>()like every other row, and the picker test captures it too.alt_keydisk identity kept.pageshandle.ci/proofsregenerated; FinNative's i18n source hash kept.Gates: CI green;
make check-nodefok;cargo test --lib -p plx_jellyfin162 passed; plx_screens 2173 passed.