mirror of
https://github.com/ChuckBuilds/LEDMatrix.git
synced 2026-10-15 19:39:14 +00:00
6df942ec02694b486844e29084bc551d8e12cfeb
2201
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 ChuckandClaude Sonnet 5.5
|
6df942ec02 |
test: the case-spelling on-demand test patches _send_on_demand now the mailbox is gone
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Sonnet 5.5
|
e9abfba81a |
Merge origin/main into chore/remove-ipc-mailboxes
Conflicts were additive only: CHANGELOG.md (this PR's Unreleased section now sits above main's 3.8.x release notes) and test/js/run_all.js (both new unit suites kept). display.py and api_v3/__init__.py merged cleanly with the stack-trace-exposure fix (#778). Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
c5ce497c37 |
chore: prepare the 3.8.6 release (#794)
Bumps src.__version__ to 3.8.6 and turns Unreleased into ## 3.8.6: the Start Display on Boot toggle (#792), the rankings edge-case fixes (#791), and src.common.sports_live_fetch with the get_espn_scoreboard null-events fix (#793, sports family 9), which the scoreboards adopt by flooring on 3.8.6. src/common/README.md and docs/SPORTS_UNIFICATION.md say 3.8.6 for the module, and family 9's heading no longer says it is awaiting a release. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>v3.8.6 |
||
|
ChuckandClaude Opus 5.5
|
f9b0b237a3 |
feat(common): sports_live_fetch -- the reconciled live fetch and odds (sports family 9) (#793)
New hardware-free module src/common/sports_live_fetch.py, copied from
ledmatrix-plugins claude/family9-reconcile once the nine scoreboards made
_fetch_todays_games (seven bodies when measured) and _fetch_odds (three) one
body each. Both live on the plugins' SportsCore, so one mixin,
SportsLiveFetchMixin, carries exactly those two plus the odds worker
_drain_odds_queue; no manager gains a method it lacks today.
Owner decisions (2026-10-12):
- Live scores are always fresh: every poll asks ESPN through
get_espn_scoreboard and leaves the answer under the shared
espn_scoreboard_cache_key for its dates (max_age=1: only a copy another
reader fetched within the last second is reused). The plugins' 30 s
<sport_key>_scoreboard_current cache is gone.
- College basketball is asked by date like every other league (ESPN
answered single dates for both college leagues on 2026-10-12).
- Odds never block update(): one daemon worker per manager answers a queue
and attaches each line in place under _games_lock; one batch in flight per
manager, games queued meanwhile join it, every dict queued for an id gets
the one answer. show_odds, the two odds intervals and a missing odds
manager (debug, not ERROR) are respected.
- An answer whose events is not a list is no games.
Fix: get_espn_scoreboard() no longer caches an answer whose events is not a
list ("events": null), which every reader of the shared key inherited; it is
still returned to the caller.
- test/test_sports_live_fetch.py: the plugins' pinned cases (requests,
dates and keys at 10:00 and 02:00 Eastern, freshness, malformed and failed
answers, odds that never block and land later, the intervals, one batch per
manager, a worker that cannot start, update and display on two threads),
plus the host contract. Mutation-checked (11 mutants, all caught).
- test/test_sports_live_fetch_parity.py: with LEDMATRIX_PLUGINS, compares
each body with every plugin copy. Passes against claude/family9-reconcile;
fails against the pre-reconcile tree.
- test/test_espn_scoreboard_cache.py: the null-events regression.
- mypy ratchet, src/common/README.md, CHANGELOG (Unreleased: New modules,
Fixes), docs/SPORTS_UNIFICATION.md family 9 status and decisions.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
ChuckandClaude Opus 5.5
|
69d89a5d69 |
feat(web): Start Display on Boot toggle on the General tab (#792)
Turns ledmatrix.service's autostart on or off at once through the existing enable_autostart / disable_autostart system actions (sudo grants already in lib_sudoers.sh). The checkbox shows systemd's own state, read with `systemctl is-enabled`, so nothing is stored in config.json and the default stays on. Hidden when the state cannot be read; a refused change puts the box back and shows the route's message. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
886e199cae |
fix(sports_rankings): skip a malformed poll entry; rank each side by its own id (#791)
* fix(sports_rankings): skip a malformed poll entry; rank each side by its own id The two review points deferred on #789 until sports_rankings was the only copy (the nine scoreboards adopted it in ledmatrix-plugins #648): - A poll entry whose "current" is None or a string raised TypeError at `current_rank <= 0`; the broad except returned {} without setting the cache timestamp, so every team lost its rank and the standings were re-fetched on every update. The entry is now skipped. - _ranked_positions() fell back to abbreviations for BOTH sides when only one side's id was unparseable, so a colliding abbreviation could rank the wrong school. It now decides id-or-abbreviation per side. Well-formed ESPN data (ids on both sides, int ranks) ranks as before. SPORTS_UNIFICATION.md: family 8 marked done and adopted. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): a numeric-string rank is still accepted; say which values are skipped Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
ef97bd158d |
chore: prepare the 3.8.5 release (#790)
* chore: prepare the 3.8.5 release Bumps src.__version__ to 3.8.5 and turns Unreleased into ## 3.8.5: src.common.sports_rankings (#789, sports family 8), which the scoreboards adopt by flooring on 3.8.5, and #754's settled-day chunk memory, which had no CHANGELOG entry and gets one from its commit message. src/common/README.md and docs/SPORTS_UNIFICATION.md say 3.8.5 for the module. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(sports): family 8's heading says core 3.8.5, not awaiting release Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>v3.8.5 |
||
|
ChuckandClaude Opus 5.5
|
6bb6e877b4 |
feat(common): sports_rankings -- the reconciled rankings (sports family 8) (#789)
* feat(common): sports_rankings -- the reconciled rankings family (sports family 8) New hardware-free module src/common/sports_rankings.py, copied from ledmatrix-plugins claude/family8-reconcile once the nine scoreboards made _fetch_team_rankings, _choose_poll (three bodies each), _load_division_team_ids, _passes_other_filters, _best_rank and _is_ranked_game (two each) one body each. All of them live on the plugins' SportsCore, so one mixin, SportsRankingsMixin, carries exactly those six plus _ranked_positions and _poll_season_ended and the constants they read; no manager gains a method it lacks today. - Ranks are kept by ESPN team id as well as abbreviation, and a game with ids is ranked by id (ESPN lists Ohio State and Ohio State Newark both as OSU). The badge still reads the abbreviation table. - A poll whose season's endDate has passed loads nothing, so no badge shows and the "ranked" filter fails open between seasons (ESPN served April's college basketball poll on 2026-10-12); read from the payload, no calendar, no sport names. - College football's division rosters are asked for the season year (football's July rule; only that league has rosters, so no seam). - choose_top_division_poll(): the one poll rule, for the dynamic team resolvers that expand AP_TOP_n. - test/test_sports_rankings.py: the plugins' pinned cases (poll choice, which leagues are asked, payload shapes, hourly cache, the finished-season rows, the rosters' season year and caching, ranked/best rank/filter over colliding abbreviations), plus the host contract. Mutation-checked. - test/test_sports_rankings_parity.py: with LEDMATRIX_PLUGINS, compares each body and constant with every plugin copy. Passes against claude/family8-reconcile; fails against the pre-reconcile tree. - mypy ratchet, src/common/README.md, CHANGELOG (Unreleased, New modules). - sports_shared and sports_rotation docstrings: the three methods leave the "stay per-plugin" list. - docs/SPORTS_UNIFICATION.md: family 8 status and decisions; item 8 restated to match what measuring found (the standings branch never ran; the season year only matters to college football). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: cite ledmatrix-plugins #645 for the family 8 reconcile Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * ci: re-run code scanning (the Python analysis job never started) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
e43708c878 |
docs(sports): family 7 is adopted (core 3.8.4, ledmatrix-plugins #643) (#788)
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
2e7cab7d53 |
fix(display): on_demand_request_id has a class default for controllers built without __init__
_on_demand_state now publishes it, and test_state_stream_readers builds controllers with __new__. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
3d79d7d1cd |
fix(web): a delivered on-demand start reads as starting until the display acts on it
On ledpi (three cold starts) the display acknowledged the start as its socket opened, then took ~5 s to act on it while Vegas built its first strip; the status routes meanwhile showed the display's own idle state, so a UI polling every 700 ms flashed idle. The display's on-demand state now names the request it answers (request_id). A delivered start keeps reading as status "starting" with delivered: true, in /display/on-demand/status and as on_demand_pending in /display/current-status, until the display publishes state for that request id (a display without the field: any state newer than the delivery), for at most DELIVERED_SHOWN_SECONDS (30 s). The display's startup state, which can be published after the acknowledgement, names no request and does not end it. Tests: stays starting against the startup idle state (no id, an older id); the matching active state and the matching error take over; an older display's newer state takes over; the 30 s cap; the display's state names its request. Mutation check: 11 mutants, 11 killed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
97f3432341 |
fix(web): a cancelled on-demand start never reaches the display after its replacement
CI caught a race in test_a_new_start_supersedes_the_pending_one: the dispatcher's worker could read the old start, the route cancel it and send the new one, and the worker's send of the old one land after it. cancel() now waits out a send already in flight (the worker holds a send lock while it reads the pending start and sends it), so whatever the caller sends next lands after it. Pinned by test_a_cancel_waits_for_a_send_in_flight, which fails without the wait. The Linux-only TestRealSocket test for a display that went away now expects the 202 and the start-timeout that follows, as the route answers since the background dispatcher. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
a591b20402 |
feat(web): a start that waits for the display answers 202 and is delivered in the background
The start route held a request open for up to 45 s while a cold-started display loaded its plugins; the MQTT bridge (15 s timeout) and browsers reported a failure for a request that was then delivered. Now, when no display is listening, the route starts the service if asked and answers 202 with status "starting" at once. A single worker in the web process (web_interface/on_demand_dispatch.py) sends the request until the display acknowledges it or the wait runs out (45 s cold start, 10 s for a running service without a socket yet). A newer start supersedes the pending one; a stop cancels it (and succeeds, with cancelled_request_id, even with no display listening). The outcome is reported by /display/on-demand/status (source "web": starting, or error with start-timeout or the socket's reason, until the display publishes something newer) and by /display/current-status as on_demand_pending. Callers: the web UI's on-demand modal and "Preview on display" treat "starting" as taken (an info toast); the MQTT bridge already treats any non-error 2xx as success (now pinned by a test). Tests: the dispatcher (ack, retry then ack, start-timeout, other failures, superseded, an in-flight ack for a superseded start, stop while pending, a per-start wait, outcome lifetime); the routes (202, status routes while pending and after a timeout, a later display state replacing the failure, stop while pending, a new start superseding); a JS suite for app.js. Mutation check: 20 mutants on the worker, the routes and app.js, 20 killed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
110a15d8f6 |
test: the on-demand routes' other callers ack over the socket, not the mailbox
test_api_v3_bool_coercion, test_api_v3_lazy_plugin_discovery and test_web_api's stop test checked the mailbox write; they now get an ack from a mocked control socket. The stop route touches no manager any more, so it leaves the removed-catch-all sample. A request that was sent is never 'not listening' (pins the sent check the mutation run let through). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
62ff14b972 |
feat(ipc)!: remove the cache-key mailboxes (control socket stage 5)
The control socket is now the only way the web interface sends the display a command. The display stops reading display_on_demand_request and plugin_error_clear_request, and the web interface stops writing them. - Display: no mailbox poll (MailboxWatch, the 1 s / 0.25 s cadence, _consume_on_demand_request, the deprecation log) and no persisted display_on_demand_processed_id guard; the error publisher reads no clear request. CacheManager.file_signature and MailboxWatch are removed. - A write to either retired key is dropped by CacheManager.save_cache and logged once per writer, naming the plugin from the call stack (or the request's plugin_id), with the API to move to. - Web: on-demand start with no display listening starts the service (when start_service) and sends the request again once the socket answers (45 s, 10 s for a running service without a socket yet); every other failure is a 503 (400 for invalid_args). Stop answers 503 when no display listens, unless stop_service. errors/clear answers 503 with a reason-specific message instead of writing a request; clear_pending is always false. src.ipc.client.should_fall_back is replaced by display_not_listening. - Kept: display_current_state, display_on_demand_state, plugin_runtime_snapshot and the heartbeat (read whenever the socket cannot answer), and display_on_demand_config (the display's resume record). Tests: mailbox-only tests removed (test_on_demand_mailbox.py, the mailbox cadence, file_signature and MailboxWatch tests); tests that injected requests through the mailbox now use the socket queue or a plugin's in-process request. The run-loop harness sends on-demand requests over its fake control socket, so four golden traces change: on-demand starts and stops land at the request instant instead of the next 0.25 s mailbox look (one frame fewer on the screen they end), and in vegas.json within one frame instead of 263 ms, which shifts the later 1 s-throttled WiFi-notice check by under a second. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
1c1a6822c2 |
perf(espn): remember settled day chunks between window refreshes (#754)
Since ESPN started rejecting date ranges, every scoreboard's hourly Recent/Upcoming refresh re-asks its 22-day window (14 back, 7 ahead) one day at a time. Measured on hdpi 2026-10-02 (NFL, college football, MLB, college baseball, NHL): the hourly refresh was ~270 of 321 ESPN requests and ~21 of 24.6MB in the hour. Days that ended three or more days ago cannot change, and they were 68% of the window's bytes (6.9 of 10.2MB). _fetch_one_chunk now keeps a settled chunk (last day <= UTC today - 3) in memory for 24h as zlib-compressed JSON, keyed by URL, the other params and the chunk, and answers it from there. Both range paths go through it: fetch_espn_scoreboard and BackgroundDataService._fetch_in_date_chunks. Each hit is parsed afresh, failed and capped chunks are not stored, and the memory is bounded (512 entries / 8MB compressed). Against live ESPN, a second refresh of the five hdpi windows went from 111 requests and 10.18MB to 50 requests and 3.23MB with identical events; the memory held 60 entries in 585KB. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
aa6d0a213d |
chore: prepare the 3.8.4 release (#787)
Bumps src.__version__ to 3.8.4 and turns Unreleased into ## 3.8.4: the refresh-cap report (#759), a failed on-demand request ending its own session (#779), reason codes instead of exception messages in API errors (#778), and src.common.sports_rotation (#786, sports family 7), which the scoreboards adopt by flooring on 3.8.4. src/common/README.md says 3.8.4 for it, and the SPORTS_UNIFICATION module table gains its row. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>v3.8.4 |
||
|
ChuckandClaude Opus 5.5
|
e186ae8b24 |
feat(common): sports_rotation -- the reconciled other-games rotation (sports family 7) (#786)
* feat(common): sports_rotation -- the reconciled other-games rotation (sports family 7) New hardware-free module src/common/sports_rotation.py, copied from ledmatrix-plugins claude/family7-reconcile once the nine scoreboards made _by_importance, _other_games_window, _advance_other_games_if_due (two bodies each), _rotate_other_games_on_display (two) and _attach_odds_to_rotated_games (three; ufc had none) one body each. All of them live on the plugins' SportsCore, so one mixin, SportsRotationMixin, carries exactly those five plus the default _rankings_loaded, and no manager gains a method it lacks today. - The window advances under _games_lock: update() and display() both advance it, and interleaved, each added a width and skipped a window. - The display path's due-check reads the pool _compose_selection will cut, unfiltered fallback included. - Rotated-in games get odds when show_odds is on (decided 2026-10-09: ufc too), skipped without an odds manager. - _rankings_loaded() is the seam _by_importance asks: the abbreviation table by default; football overrides it to count its by-id rankings. - test/test_sports_rotation.py: the plugins' pinned cases (importance order per rankings table and the seam, the window over time, the display-path rotation incl. the card on screen, redraws and recompose counts, update() and display() in sequence and interleaved on two threads -- with and without the lock -- and rotated-in odds), plus the host contract. - test/test_sports_rotation_parity.py: with LEDMATRIX_PLUGINS, compares each body with every plugin copy (drift-report normalisation plus decorators), checks no other plugin class carries a copy, and that only football overrides _rankings_loaded. Passes against claude/family7-reconcile; fails against the pre-reconcile tree, as it should. - mypy ratchet, src/common/README.md, CHANGELOG (Unreleased, New modules). - sports_shared docstring: _by_importance and _other_games_window (and the long-promoted _is_favorite_game) leave the "stay per-plugin" list. - docs/SPORTS_UNIFICATION.md: family 7 status, decisions, the _rankings_loaded seam; families 5 and 6 marked done and adopted (3.8.1 / ledmatrix-plugins #631, 3.8.2 / #637) instead of "adoption waits". Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(sports): stage 4 is adopted (core 3.8.0, ledmatrix-plugins #594) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: cite ledmatrix-plugins #641 for the family 7 reconcile Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
65ede54cba |
fix(web): keep exception messages out of API responses (py/stack-trace-exposure) (#778)
CodeQL had ~40 open py/stack-trace-exposure alerts on main. Almost all
flowed through describe_exception(), which returned "TypeName: message"
(redacted, capped); the rest through _run_systemctl_command's str(err),
WiFiManager's `return False, str(e)`, unit_refresh's f-strings and two
str(e)/f"{err}" messages in api_v3/__init__.py.
describe_exception() now returns a reason code -- the type, plus the
errno symbol for an OSError ("OSError:EIO", "PermissionError:EACCES") --
and logs the redacted message itself. That keeps what #538 wanted (a
failing disk still says EIO in the response) without quoting paths,
URLs or library internals, and fixes every call site at once; the
test_no_api_v3_handler_discards_its_exception policy still holds.
Service results: _get_display_service_status returns active/returncode
only, and the on-demand start/stop `service` result keeps
returncode/active/started/status but drops systemctl stdout/stderr
(logged on failure). Nothing in web_interface/static, the templates or
the MQTT bridge reads those fields. The Starlark SIGKILL-restart error
no longer returns systemctl stderr as `details`.
WiFi, unit-refresh, config-save and plugin-removal failures now say
what failed with the reason code and point at the log. display.py is
untouched (draft #773 edits it).
Tests: test_api_v3_no_exception_text.py drives one route per affected
file with a marker in the exception message and asserts it never
reaches the body; all 13 fail on origin/main, and targeted mutations
(drop the service filter, put stderr back, str(e) in WiFiManager,
{e} in unit_refresh, {install_err} in system.py, message back in
describe_exception) each fail at least one. Tests that asserted the old
message-in-details contract now assert the reason code.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
ChuckandClaude Opus 5.5
|
20bae8b609 |
fix(display): a failed on-demand request drops the session it ended (#779)
_set_on_demand_error ends any running session (_reset_on_demand_fields) but left its saved copy, display_on_demand_config, in the cache. A failed request that replaced a running session therefore made the next restart resume the session that had already ended. Clear the saved copy where every error path goes through, and drop the two restore-failed callers' own clears, which this now covers. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Sonnet 5.5
|
1bcb524fc3 |
feat(scroll): report a panel that cannot reach its refresh cap, and suggest one it can hold (#759)
* feat(scroll): report a panel that cannot reach its refresh cap, and suggest one it can hold
Scroll speeds are solved against display.hardware.limit_refresh_rate_hz,
which is only a ceiling. A panel that cannot reach it still moves whole
pixels per frame, but every scroll runs slow by the shortfall and the
"smooth" ladder is the cap's, not the panel's. A user rig (Pi 4, 2x128x64,
adafruit-hat-pwm, pwm_bits 9, gpio_slowdown 5) measured 107.6-113.1 Hz under
a 120 Hz cap: 60 px/s ran at 55, and nothing said why.
- scroll_config: refresh_shortfall() (more than 3% under the planned rate),
holdable_cap() (a multiple of 10, 5% under the measurement, since the
measurement is the fast end of an uncapped panel's drift), and
describe_refresh_shortfall().
- FrameTimingRecorder.plan_refresh(): once the measured period has held for
three trusted windows, a shortfall is logged once as a warning naming the
cap to use. DisplayManager calls it only for a real panel, not the
emulator or the fallback canvas. The stats file records
planned_refresh_hz (additive).
- GET /api/v3/config/refresh-rate, plus a hint under the Display tab's
Limit Refresh Rate field (js/pages/display.js) with a button that fills in
the suggested cap.
- _panel_refresh_hz (behind the Vegas slider's advice) ignores a measurement
written under a different cap, so a changed cap stops being advised from
the old rate before the display restarts.
Verified on ledpi with a temporary 200 Hz cap: the warning logged about a
minute after the restart ("about 132 Hz ... Set Limit Refresh Rate to
120 Hz"), the endpoint returned the same shortfall, and the Display tab
showed the hint; its button filled in 120. ledpi was restored afterwards.
Rebased onto main after the Display tab became an ES-module page (#771); the
hint moved from inline script into display.js, with a jsdom test.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
* fix(scroll): count only agreeing windows toward the refresh shortfall; ignore a null planned rate
Review fixes on #759.
- A window the period estimate rejects (more than MAX_REFRESH_DROP faster
than the adopted period) no longer counts toward the shortfall check, and
it restarts the run. After a loaded start fixed a slow period, later
windows at the real, faster rate were rejected yet still counted, so the
warning could name the slow rate against a cap the panel was meeting. It
now needs REFRESH_CHECK_WINDOWS consecutive windows that agree with the
period.
- _panel_refresh_hz treats a stats file whose planned_refresh_hz key is
present but null as no measurement. The emulator and the fallback canvas
write it that way (DisplayManager never plans a refresh for them), and
their frame rate says nothing about the cap. A file with no such key (an
older display) keeps the old behaviour.
Three new tests fail on the previous code and pass now.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
|
||
|
ChuckandClaude Opus 5.5
|
370c8fe273 |
chore: remove dead code, deprecate unused plugin APIs (over-engineering audit) (#783)
* chore: remove dead code, deprecate unused plugin APIs (over-engineering audit) Whole-tree audit. Every symbol was checked against core, the plugin monorepo and all eight third-party plugins in plugins.json first. - Deprecate (removal 3.10.0) plugin-facing methods nothing calls: LogoDownloader bulk download, ConfigManager backup/secret wrappers, APIHelper extras, BackgroundDataService poll API, PluginManager / PluginStateManager info readers, and a few CacheManager, FontManager, BaseOddsManager, DynamicTeamResolver methods and PluginTestCase. plugin_api_usage.py learns their receiver names; DEPRECATIONS doc regenerated. - Remove core-internal dead code: CacheMetrics, Vegas status/stats plumbing, sync "new cycle" message (followers ignore unknown types), unused operation types, test-only PluginCatalog readers, IPC to_dict and ping, _parse_form_value, CacheStrategyProtocol, ErrorAggregator callbacks, duplicate web response helpers. - Web UI: drop never-mounted json-file-manager.js, the example widget, utils/error_handler.js, four uncalled PluginAPI methods, and 29 escapeHtml shims (call window.LEDEscape directly). Public globals, BaseWidget and widget names unchanged. - Remove six one-off scripts (owner decision) and the unused markupsafe and pytest-mock pins. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(web): calendar picker error text goes in a text node, not innerHTML Same output as the escaped innerHTML it replaces; clears Codacy's XSS-pattern alerts on the line. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>v3.8.3 |
||
|
ChuckandClaude
|
358911cb20 |
chore: prepare the 3.8.3 release (#784)
Bumps src.__version__ to 3.8.3 and turns Unreleased into ## 3.8.3, adding entries for what merged since 3.8.2: the installer's desktop check (#780, #781) -- 3.8.2 stops fresh Raspberry Pi OS Lite installs with "Desktop environment detected", and the one-shot installer installs the newest release, so the fix only reaches users once it is tagged -- the ESPN scoreboard payload slimming (#749), the Vegas static-pause and store restart fixes (#753), and plugin-computed display modes in the web UI (#769). scripts/check_release_version.py v3.8.3 passes. Claude-Session: https://claude.ai/code/session_01VZNWFWprcFYGfuf1JAyrBJ Co-authored-by: Claude <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
d98b479727 |
fix(install): only a running desktop stops the install; detect desktops by metapackage (#781)
* fix(install): detect desktops by their metapackages, not name prefixes
The Lite check matched any installed package starting with gnome/kde/
xfce/lxde, so standalone parts (gnome-keyring, xfce4-terminal,
lxde-icon-theme) rejected a Lite system. Match whole names of desktop
metapackages and session managers instead.
Also catch desktops the prefixes missed: Raspberry Pi OS Trixie replaced
raspberrypi-ui-mods with rpd-wayland-core / rpd-x-core, Debian tasksel
desktops (task-*-desktop), and multi-arch names (plasma-workspace:arm64).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VZNWFWprcFYGfuf1JAyrBJ
* docs(common): list espn_payload in the common README
#749 added src/common/espn_payload.py without a summary row or section,
so test_common_readme_lists_every_module fails on main.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VZNWFWprcFYGfuf1JAyrBJ
(cherry picked from commit
|
||
|
ChuckandClaude
|
87f255b2fa |
docs(common): list espn_payload in the common README (#782)
#749 added src/common/espn_payload.py without a summary row or section, so test_common_readme_lists_every_module fails on main. Claude-Session: https://claude.ai/code/session_01VZNWFWprcFYGfuf1JAyrBJ Co-authored-by: Claude <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
6c533d62af |
fix(plugins): web mode lookups use the modes the display registered (#668) (#769)
* fix(plugins): web mode lookups use the modes the display registered (#668) A plugin may compute its display modes from its config: soccer-scoreboard registers soccer_<league>_live/recent/upcoming for every custom_leagues entry, which no manifest can list ahead of time. The display always rotated them (_register_loaded_plugin prefers plugin.modes), but the web process reads plugins as files, so /display/modes, the on-demand dialog and on-demand/start with a mode and no plugin_id saw only manifests -- a custom league's mode was missing from every list and 404'd on lookup. - PluginStateManager.record_modes(): the controller records what it registered, on the loaded record (an unload or reload forgets it) - the runtime snapshot carries it per plugin as "modes" (bounded), and PluginRuntimeView.display_modes() reports it only while live - PluginCatalog takes a runtime_source; get_plugin_display_modes and find_plugin_for_mode prefer the live modes, falling back to the manifest when the display is stopped or has not loaded the plugin. The view is read at most once a second, so a listing is one read, not one per plugin. No manifest or plugin change needed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(plugins): call the runtime view's display_modes directly Codacy flagged the getattr/callable indirection as 'lookup is not callable'. The view is a PluginRuntimeView or None; anything else raises inside the existing try and falls back to the manifest. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(plugins): address review -- no manifest fallback for live plugins, keep mode names whole, send registered spelling Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
cb06124b42 |
fix: a Vegas static pause survives a non-numeric display duration; aliased store installs ask for a restart (#753)
* fix(vegas): a display duration that is not a number no longer cancels a static pause The Vegas static pause compared plugin.get_display_duration() with the clock. clock-simple, calendar and countdown return their display_duration setting straight from config.json, so a value saved as "20" or null reached that comparison as a string or None. The TypeError went to the pause's broad except, which ended the pause: the plugin flashed up and the scroll went straight on, at every one of its turns. inf held the pause until something interrupted it, and NaN, False, 0 or a negative number ended it at once. The pause now reads the duration the way the rotation has since #739, with the same helper, then the rotation's fallbacks: 30 s for anything that is not a number or a get_display_duration() that raises, 15 s for a number at or below zero. Logged once per plugin. test_vegas_static_mode.py's pauses used 0 to mean "no wait"; they now use 0.01. The helper moves from display_controller (_finite_seconds) to base_plugin (finite_seconds), unchanged: the coordinator cannot import from display_controller, which imports src.vegas_mode at module level, and a new src module would turn ledmatrix-plugins' min-core table check red until it was listed. base_plugin is already loaded whenever either one is. Tests: test/test_vegas_static_pause_duration.py, on a fake clock, including TestSameAsTheRotation, which runs every value through both the pause and the rotation's _get_display_duration/_resolve_durations. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(web): a store install asks for a restart by the id it installed as POST /plugins/install decides restart_required from whether config.json already enables the plugin: the display loads a plugin when its enabled flag changes, so one already enabled (a reinstall, or a config carried over) keeps running the copy it loaded until a restart. The route read that flag under the registry id. Weather, Music, Stocks and Leaderboard install under the id their manifests declare (weather -> ledmatrix-weather), which is the config section's id, so reinstalling an enabled one never reported that a restart was needed. Both the queued and the direct path now look up the installed id once (#746's _installed_plugin_id) and use it for the plugin_id they answer with and for the enabled check. Tests: test/test_api_v3_install_restart_installed_id.py, through the Flask test client, both paths. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(vegas): type the static pause fallback on its own (mypy ratchet) _static_pause_duration assigned the fallback to `seconds`, which the except branch typed as float before finite_seconds() reassigned it as float | None. A separate `fallback` keeps both types exact; behaviour is unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
7c5fa9cfdb |
perf(fetch): cache ESPN scoreboard windows without the parts nothing reads (#749)
The sports scoreboards cache their Recent/Upcoming window as the raw ESPN response, and it stays parsed in the memory tier while fresh. Measured on hdpi, most of it is never drawn: per-team stat leaders, athlete cards (featuredAthletes, probables), team and event links, headlines, video highlights and geo broadcasts. None of those keys is read by core or by any plugin in ledmatrix-plugins. BackgroundDataService now drops them from an ESPN /scoreboard response before caching and delivering it (src/common/espn_payload.py), keeping everything else. On the five hdpi windows that is 10.6MB -> 3.0MB of JSON and ~40MB -> ~12MB of parsed objects, and parsing an expired window gets 3-4x cheaper. submit_fetch_request(slim_payload=False) caches a response whole. Claude-Session: https://claude.ai/code/session_01BkfgXMqqwn2w4NN7LRzhxy Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
ec117a35a1 |
fix(install): stop libblockdev matching the desktop-environment check (#780)
* fix(install): stop libblockdev matching the desktop-environment check dpkg -l | grep -E '^ii.*kde' searched the description column too, so unrelated packages (libblockdev-*) set DESKTOP_DETECTED on Raspberry Pi OS Lite. Match on the installed package name only, anchored. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VZNWFWprcFYGfuf1JAyrBJ * docs(install): correct why the desktop check matched libblockdev The unanchored .*kde matched the package name mid-word ("bloc-kde-v"), not only description text. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VZNWFWprcFYGfuf1JAyrBJ --------- Co-authored-by: Claude <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
e4f5e49ff7 |
test(sports): treat an adopted sports_helpers copy as parity, not missing (#777)
* test(sports): treat an adopted sports_helpers copy as parity, not missing The scoreboards deleted their copies of the sports_helpers bodies and constants when they adopted SportsHelpersMixin (ledmatrix-plugins #563/#564), so the 19 parity tests in test/test_sports_helpers.py failed whenever LEDMATRIX_PLUGINS pointed at a plugins checkout. A copy that is gone now counts as adopted when the plugin imports src.common.sports_helpers, as the stage 3/4 and game-over parity tests already do; a copy that remains must still match. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(sports): _adopted checks for a real import via the AST, not a text match Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
e40bc47d28 |
chore: prepare the 3.8.2 release (#776)
Bumps src.__version__ to 3.8.2 and turns Unreleased into ## 3.8.2: the display's malloc arena cap and between-screen malloc_trim (#774), and src.common.sports_favorites (#775, sports family 6), which the scoreboards adopt by flooring on 3.8.2. src/common/README.md and docs/SPORTS_UNIFICATION.md say 3.8.2 for it; the SPORTS_UNIFICATION module table also said "next release" for the four stage 4 modules, which shipped in 3.8.0. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>v3.8.2 |
||
|
ChuckandClaude Opus 5.5
|
3bdb5bff3b |
feat(common): sports_favorites -- the reconciled favourite matching (sports family 6) (#775)
* feat(common): sports_favorites -- the reconciled favourite matching (sports family 6) New hardware-free module src/common/sports_favorites.py, copied from ledmatrix-plugins claude/family6-reconcile once the nine scoreboards made _is_favorite_game (seven bodies), _select_games_for_display (two) and _select_recent_games_for_display (three) one body each. One mixin per class that carries the methods, so adopting one gives no manager a method it did not have: - SportsFavoritesMixin (SportsCore): _is_favorite_game and _favorite_code. - SportsUpcomingFavoritesMixin: _select_games_for_display. - SportsRecentFavoritesMixin: _select_recent_games_for_display. Each side of a game is named by the 3.5.0 _favorite_key seam (SportsHelpersMixin; the abbreviation by default, nrl overrides it with the ESPN team id and None for a missing id) and compared with favorite_teams stripped and upper-cased. The selection methods give each favourite up to the per-team limit, count a game between two favourites for both, treat only games with an id as possible duplicates and log their summary at INFO. - test/test_sports_favorites.py: the plugins' pinned cases for an abbreviation host and an id-keyed (nrl-style) host -- case, spaces, ids, the NEW collision, the "None" favourite, missing keys; selection order, limits, duplicates and the id-less fix, the INFO summary; host contract, one carrier per method, and SportsGameRulesMixin reaching the shared body. - test/test_sports_favorites_parity.py: with LEDMATRIX_PLUGINS, compares each body with every plugin copy (drift-report normalisation plus decorators), checks no other plugin class carries a copy, and that only nrl overrides _favorite_key. - mypy ratchet, src/common/README.md, CHANGELOG (Unreleased, New modules). - sports_helpers docstrings: _favorite_key now has a caller and an override. - docs/SPORTS_UNIFICATION.md: family 6 status and decisions, and the seam table. SportsCoreSharedMixin._round_robin_favorites still groups by raw abbreviation or _team_in: it is not one of the plugin bodies, so it waits for a later family. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(sports): family 6 also routes the Upcoming favourites-only filter and three live boosts ledmatrix-plugins claude/family6-reconcile now sends the Upcoming update()'s favourites-only pre-filter and the basketball, hockey and lacrosse live favourite boost through _is_favorite_game, so a lower-case favourite works on a favourites-only Upcoming board. The module is unchanged (update() is not promoted); the parity test still passes against the branch. Updates the pinned row and cell counts and what is left for later families. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: cite ledmatrix-plugins #635 for the family 6 reconcile Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
e745ae8060 |
feat(display): cap malloc arenas in-process and malloc_trim between screens (#774)
* feat(display): cap malloc arenas in-process and malloc_trim between screens Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test: malloc_tuning with ctypes mocked; add to the mypy ratchet Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): malloc arena cap and malloc_trim between screens Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(changelog): spacing Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
a669d781f5 |
chore: prepare the 3.8.1 release (#756)
* chore: prepare the 3.8.1 release Bumps src.__version__ to 3.8.1 and moves the Unreleased CHANGELOG entries under a 3.8.1 heading, with an empty Unreleased above it. The reason for the release is smooth scrolling at held-frame speeds. On 3.8.0 the default 50 px/s snapped to a stepped 48 px/s on a 120 Hz panel, and any scroll slower than one pixel per refresh showed a half-pixel step across the middle of the panel. Both are fixed on main (#710, #711) but were in no release, so every stable-channel device still had them. - #710's CHANGELOG entry had been filed under 3.8.0 although it merged after the v3.8.0 tag; it moves to 3.8.1. - #711 had no CHANGELOG entry; it gets one. scripts/check_release_version.py v3.8.1 passes. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: sports_game_over and draw_text_outlined ship in 3.8.1 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>v3.8.1 |
||
|
ChuckandClaude Opus 5.5
|
c20c0beac2 |
feat(web): the Display tab is an ES-module page, with a page-visibility service (stage 4) (#771)
* feat(web): the Display tab is an ES-module page, with a page-visibility service (stage 4) - core/visibility.js: each page gets ctx.visibility (whileVisible, every, isVisible). Work registered there runs only while the page's tab is active and the browser tab visible, and ends when the page is swapped out. It reads the active tab from window.LEDVisibility, so it agrees with the classic partials. The registry gained a mountContext option for per-mount services. - pages/display.js replaces display.html's two inline scripts. The 5 s sync status poll runs through ctx.visibility.every; the status and scroll-speed hint requests go through ctx.api with ctx.signal, as does the Vegas order widget's plugin-list request. The Advanced toggle is a delegated data-action; window.updateSyncUI is a deprecated alias. - New DOM suites test_visibility_service.js and test_display_page.js; test_display_partial_ids.js imports the module. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(web): no computed keys in the Display page's readout and destroy Codacy's object-injection rule flagged v[id] and ctx.state[name]. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
6fb2dc3595 |
fix(sports): honour every pending kickoff in the idle back-off, not just the first (#772)
_note_scheduled_start_candidate kept one kickoff. While it was inside its 15-minute grace every later kickoff was refused, and by the time the grace ended the later one had passed and was refused again as already past. So of two favourites kicking off within 15 minutes of each other, the second lost its own grace: if the first game was not live by then (a rain delay, a postponement, ESPN slow to flip it) and ESPN had not flipped the second either, the back-off went straight back to its ceiling and the second game was noticed up to that late. Later kickoffs now wait in a short queue (_later_scheduled_starts, the earliest 8). When the current kickoff's grace ends, the earliest queued one still inside its own grace takes over -- including one that has already passed. A kickoff still holds the live cadence for at most its own grace, so a postponed game costs the same quarter of an hour as before, and _next_scheduled_start_ts keeps its meaning for anything that reads or sets it. The promotion is a module function, so the mixin's method set is unchanged. Table tests replay the idle loop on a fake clock over kickoff schedules; mutation-checked (the old code fails 10 of the new tests). Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
b638b91169 |
feat(common): sports_game_over -- the reconciled game-over check (sports family 5) (#770)
* feat(common): sports_game_over -- the reconciled game-over check (sports family 5) New hardware-free module src/common/sports_game_over.py with SportsGameOverMixin._is_game_really_over, the scoreboards' SportsLive check that drops a game ESPN still lists as live, copied from ledmatrix-plugins claude/family5-reconcile once the nine copies (five bodies) became one. Over on a final period text; from period FINAL_PERIOD on, also on a 0:00 clock string unless the score is level (a tie at the end of regulation goes to overtime; a game that ends tied ends on its final status). FINAL_PERIOD is the one per-sport seam, a class attribute defaulting to None (the clock never ends a game); the scoreboards declare 3 (hockey), 4 (basketball, football, lacrosse) or None (afl, nrl, soccer, baseball, ufc). - test/test_sports_game_over.py: the plugins' pinned matrix folded to the three FINAL_PERIOD values, edge shapes, the tie guard, ufc's recorded ESPN MMA states, an override deferring through super() (baseball), the base order with SportsLiveSharedMixin._detect_stale_games, host contract. - test/test_sports_game_over_parity.py: with LEDMATRIX_PLUGINS, compares the body with every plugin copy (drift-report normalisation plus decorators) and each plugin's FINAL_PERIOD with the owner's decision. - mypy ratchet, src/common/README.md, CHANGELOG (Unreleased, New modules). - docs/SPORTS_UNIFICATION.md: family 5 status and decisions; the seam tables now match the code (FINAL_PERIOD defaults to None; the CLOCK_COUNTS_DOWN seam never existed and is gone from the doc). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: cite ledmatrix-plugins #625 for the family 5 reconcile Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
bb475a79ea |
fix(plugins): keep a plugin's tokens and local files across store updates (#755)
* fix(plugins): keep a plugin's tokens and local files across store updates A monorepo plugin update replaces the plugin directory with the fresh download and deletes the old copy, taking with it everything the plugin wrote beside itself. On 2026-10-04 updating calendar 1.2.9 -> 1.2.12 deleted token.pickle and credentials.json, and the calendar stopped until they were restored from a backup. Before the set-aside copy is discarded (store update, reinstall over an existing copy, install_from_url replace), carry over files the plugin's .gitignore excludes plus known secret/state files (*.pickle, token.json, credentials.json, config_secrets.json, .pkce_code_verifier). Files the new release ships win; byte code and .git are not carried; if a copy fails the old copy is kept. The git-pull path no longer sweeps untracked tokens into its auto-stash, which is never popped. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(plugins): find the new copy via _existing_install, as install_plugin does Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(on-demand): find the write under test by key, not by position TestARestoreWithNothingToResume took the last cache_manager.set call to be the on-demand state, but the controller's font-usage publisher thread writes font_usage_snapshot to the same mock, and on a slow runner it lands last. Failing on main since #748 (Python 3.11 job). Same fix for the named-mode restart test, which had the same race. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(starlark): fake only the editor launch, not every Popen in the request TestPixletEditorHostDefaultsButDoesNotOverride patched subprocess.Popen for the whole request. When the captive-portal before_request hook's 30s AP-mode cache had expired, its `systemctl is-active hostapd` check went through subprocess.run, got the fake process, and raised TypeError (run() uses the process as a context manager): a 500 instead of 200. Seen on the Python 3.13 job; reproduced locally by forcing the cache to expire. Other calls now reach the real Popen. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
3f920f2899 |
fix(fonts): unloading a plugin forgets its manifest fonts (#757)
* fix(fonts): unloading a plugin forgets its manifest fonts PluginManager.unload_plugin() and the failed-load cleanup only called FontManager.forget_manager_fonts(), which drops usage data. The plugin's manifest registrations stayed: plugin_fonts / plugin_font_catalogs, its plugin_id::family entries in font_catalog, and cached font objects for them -- so its fonts kept resolving after unload and a family a reinstalled manifest dropped stayed registered. The deprecated unregister_plugin_fonts did this cleanup but nothing called it; it was removed in #708. Add FontManager.forget_plugin_fonts(plugin_id) and call it from both paths alongside forget_manager_fonts (each guarded on its own, so a font manager stub with only one still works). FontManager takes no locks, so like forget_manager_fonts it uses atomic pops over snapshots. A reload (unload + load) registers the manifest fonts again and they resolve. Raised by CodeRabbit on #709. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(fonts): call the font manager's forget methods without a None-able local Pylint E1102 (Codacy) read getattr(..., None) as possibly not callable. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
0577c807eb |
feat(plugins): request_on_demand() / end_on_demand() -- plugins ask for the screen in-process (#768)
* feat(plugins): request_on_demand() / end_on_demand() -- plugins ask for the screen in-process Four plugins (birdnet-go, mqtt-notifications, on-air, pomodoro-timer) take the screen by writing the display_on_demand_request mailbox, which the display reads once a second while the control socket is up and which stage 5 removes. This is the in-process way in that stage needed. - BasePlugin.request_on_demand(mode=None, duration=None, pinned=False) and end_on_demand(), safe from any thread, go through PluginManager to DisplayController.submit_plugin_on_demand, which only queues (at most 32) and wakes the render thread through ControlServer.wake(). The render thread applies them in _drain_control_commands, after socket commands, through _handle_on_demand_request, so they land within a frame; without a socket, on the next pending-changes pass. - A plugin's stop ends only its own session; a mailbox stop still ends any. - Both answer the request id, or None with no display in the process (web interface, check_plugin.py), a full queue, or a mock manager -- a plugin's cue to write the mailbox, which the display still reads. - docs/PLUGIN_API_REFERENCE.md documents the hasattr pattern for plugins that must keep working on older cores; IPC_CONTROL_SOCKET.md and the CHANGELOG are updated. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(display): wire the on-demand handler only on a manager that has it Tests and the golden traces stand in simpler plugin managers. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
5a7893b11a |
feat(web): Schedule and General become ES-module pages (stage 3) (#767)
* feat(web): Schedule and General become ES-module pages (stage 3) Schedule and General follow stage 2 (#727): no inline scripts or inline handlers in either partial. Their code moves to static/v3/js/pages/schedule.js and pages/general.js, started per swap-in by the page registry. - Schedule: both pickers are drawn from the saved config carried as JSON in data-* attributes. The forms' hx-on save handlers become one htmx:afterRequest listener on the page; the forms are marked data-reports-result, which app.js now treats like an hx-on after-request handler, so a save still shows one notification. - General: the timezone picker reads data-timezone. The Security section's forms and buttons are delegated data-actions; requests go through core/api.js, so the login redirect is quiet, and a change made just before a swap is still reported. - handleScheduleResponse, handleDimScheduleResponse and webLogin stay as deprecated aliases through window.LEDMatrix. - New DOM suites test_schedule_page.js and test_general_page.js; the web login unit suite imports the module; test_es_modules.py pins the pages, the aliases, and the schedule config's round trip through its attribute. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(web): no unused catch bindings or computed writes in the stage 3 modules Codacy flagged two unused catch variables and dynamic-key writes in pages/schedule.js and boot.js. The schedule config is read with getAttribute, and the default days and the webLogin alias object are built with Object.fromEntries. No behaviour change. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
3866aa4519 |
feat(ipc): the control socket carries every web command; mailboxes are a fallback (stage 4) (#765)
- client: ControlError.sent says whether the display had the request; should_fall_back() allows a mailbox write only when it did not, or when the display is too old to know the command (upgrade case) - on-demand start/stop: a display that had the request and failed it is answered 503 (400 for invalid_args), no mailbox copy - errors.clear: new socket command, answered on the connection thread by a handler the display registers; applied and republished before the answer; plugin_error_clear_request only on fallback - display: on-demand mailbox looked at once a second while the socket is up (0.25 s without), read only when its file changed (one stat via CacheManager.file_signature / MailboxWatch); socket commands no longer touch the mailbox; a processed duplicate is consumed; writers logged once - error publisher: mailbox read only when changed; snapshot carries applied_clear_cutoff so an older mailbox request is not shown pending - docs and CHANGELOG (mailboxes kept for one release) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
26cae3e5d6 |
refactor(display): run loop stage 3 -- ScreenRunner, PREEMPTED, OnDemand/Live/Rotation Sources (#762)
* refactor(display): run each screen through a ScreenRunner (run loop stage 3) The two frame loops, the make-up dwell and the dynamic-duration exit move out of DisplayController.run() into src/screen_runner.py. ScreenRunner paces with an injected FrameClock (production: this module's time, looked up per call so the golden harness's fake clock still drives it) and returns one Outcome whose ExitReason is DURATION, CYCLE_COMPLETE, EMPTY, ERROR, DISPLAY_FALSE, RELOAD or PREEMPTED. PREEMPTED replaces the re-checks that used to follow each frame loop and the make-up dwell (current_display_mode != active_mode, the schedule, a pending WiFi notice): the runner asks its host at named service points (FRAME, AFTER_LOOP, after_dwell, FINAL), and on PREEMPTED run() goes to the next pass without advancing the rotation, as each `continue` did. RELOAD is the one early end that still advances, as a reload always did. Each service point reads the WiFi notice file exactly when the loop did (NoticeRead), because the read is throttled and deletes expired files. The host answers still use the old checks; the following commits move them to the Arbiter. _screen_preempted is gone (folded into the FRAME check); _wait_frame_interval returns the preempting plan instead of a bool. The frame pacing (8 ms deadline, 1 ms minimum yield, 1 Hz wait with socket wake) is the same code, moved. Golden traces unchanged. A capture of every harness run (all 67, with every sleep, display() call, wifi read, live scan, publish and dwell logged) is identical to origin/main except for throttled WiFi reads that returned the cached answer (no side effect) after a notice preempted. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(display): on-demand is an Arbiter Source (run loop stage 3) Arbiter.decide() now answers for an active on-demand session itself (Source.ON_DEMAND) instead of returning LEGACY: - ArbiterState gains the session: its mode list, index, expiry and pin, plus current_mode, snapshotted from the controller's fields by _arbiter_state(). The controller's attributes stay the record that the web UI, the control socket and the cache read. - The OnDemand plan is the session's current mode (an index past the end of a shortened list starts again at 0), with what is left of a timed session at `now` as max_duration and the expiry as deadline. A session with no modes left is a plan with no mode; the controller ends it and shows the rotation's mode, as _resolve_active_mode did. - on_demand_bound() is _clamp_to_on_demand made pure. It is still applied after the first frame, with the clock read there. - ArbiterState.next_on_demand() is the step _advance_on_demand takes. run() asks decide() for the screen at the point it used to call _resolve_active_mode (after any Vegas iteration, so a session that started mid-iteration still shows next), and _take_plan() applies it. Golden traces and the 67-run capture identical to origin/main. Adds TestOnDemand and the bound table to test_display_arbiter.py; the stage-2 table's on-demand rows now name ON_DEMAND instead of LEGACY. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(display): live priority is an Arbiter Source (run loop stage 3) The live-priority step of run() (step 7) and the live checks around the Vegas iteration become the Arbiter's Live Source: - ArbiterInputs gains live_modes (the scan, None where run() made none), vegas_enabled, vegas_live_in_ticker and vegas_yielded. ArbiterState gains the rotation and its index, the live resume point and the "takeover not shown yet" flag. - Live picks the next live mode round-robin (live_pick, now also what _check_live_priority returns), not advancing past a mid-screen takeover that has not shown. It outranks Vegas unless the ticker keeps live content, in which case it has no say at all, as before. With nothing live, a plan below it carries ends_live and the interrupted rotation resumes. - ArbiterState.claim_live/release_live are _apply_live_priority's bookkeeping made pure; _apply_live_priority applies them. run() reads the inputs below the WiFi notice where it always did (_arbiter_inputs_below_wifi: the Vegas check, then the scan), asks decide() once more, and _take_plan applies the claim or the resume. The Vegas iteration moves to _run_vegas_iteration, which re-decides with vegas_yielded after a yield, so a game that stopped the ticker or an on-demand session that started mid-iteration still shows next. LEGACY now means Vegas or the rotation. One redundant call is gone: a Vegas pass scanned the live plugins twice at the same instant (step 7, then step 8's "is anything live?"); it scans once. Golden traces unchanged. The 67-run capture is identical to origin/main once that duplicate scan and _apply_live_priority(None) calls that changed nothing are left out. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(display): the rotation is an Arbiter Source; LEGACY means Vegas (run loop stage 3) decide() now names the screen for every pass: the Rotation Source (Source.ROTATION) answers with the rotation's current mode, after the resume when live priority just ended. LEGACY is left meaning only Vegas, whose iteration is still run()'s own code until stage 4. Once this pass's iteration has yielded (vegas_yielded), Vegas passes and the screen it fell through to is decided like any other. The rotation's mode is state.current_mode rather than rotation[rotation_index]: they agree except where something moved the panel off the list and the rotation carries on from there (a live mode no entry names, or None after a session ended with nothing to resume to), and run() always showed current_display_mode. ArbiterState.after(outcome) is _advance_after_screen's step: an on-demand session moves to its next mode; otherwise the rotation advances unless the mode just shown is still live. The Outcome carries the two facts only the controller can see at the end of the screen (on_demand_active, the live hold from _still_live). Ending a session with no modes left stays in the controller, because it is not pure. Golden traces unchanged; the 67-run capture is identical to origin/main with the same two exclusions as the previous commit. Adds the Vegas / Rotation table and TestAfter; the stage-2 rows that said LEGACY for "live, Vegas or rotation" now say ROTATION. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(display): one decide() call at each of the runner's service points (run loop stage 3) The three mid-screen checks the frame loops made one after another -- _check_live_takeover, then _screen_preempted with _wifi_notice_pending in it -- become one call: Arbiter.decide(state, inputs, now, running=plan). It returns `running` itself while the screen holds, else the plan that ends it, from these rules in the order the loops checked them: 1. Live: a game went live while a non-live screen runs. First because it is the one preemption that changes the state (the rotation moves to the live mode and remembers where it was), and it is still claimed when a WiFi notice is pending too; the next pass shows the notice, then the game, as before. 2. The panel's mode moved under the screen (on-demand started, ended or changed; the rotation was rebuilt). 3. The schedule turned the panel off. 4. A WiFi notice arrived (on-demand outranks it; compared with expiry). 5. A plugin reload is waiting (between frames only). Each rule is gated by plan.preemptible_by: every screen may be preempted by the gate, OnDemand, Wifi, Live, Rotation and a reload, except that a live screen leaves Live out. A follower and Vegas never preempt mid-screen. The pure helper live_takeover() is the Live rule, shared with the dwell sleep's _check_live_takeover. The controller only gathers and applies. _screen_service applies pending changes and makes the live scan when one is due (_scan_for_takeover: the same throttle and gates as before); _screen_check reads the WiFi notice exactly where the loop did (the read is throttled and deletes an expired file, so an extra read would move both), calls decide() once, and claims a live takeover. _screen_preempted is gone; _check_live_takeover and _wifi_notice_pending remain for the dwell sleep and the Vegas yield path, built on the same rules. Golden traces unchanged. The 67-run capture is identical to origin/main (with the earlier two exclusions) except for one event: in the 125 Hz loop the live scan still runs before the frame's sleep, but the claim is now made by the service point after it, so the "live" state change is logged 8 ms later (test_live_game_cuts_a_scrolling_screen_short: 8.064 -> 8.072). The screen still ends at the same frame (8.072) and every frame, sleep and pass is unchanged. Adds the mid-screen table (24 rows), live_takeover's table, and test/test_screen_runner.py (the runner on a scripted host, plus the controller's service point: which reads it makes at which checkpoint). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs, tests: run loop stage 3 -- doc, changelog, mutation survivors docs/RUN_LOOP_REDESIGN.md describes run() as it now is (two decide() calls per pass, the runner and its service points, the state snapshot and its transitions), records what stage 3 shipped and how it was checked, and adds one open "may be wrong" behaviour the mutation run surfaced: a Vegas iteration stopped for a sync follower falls through to a full rotation screen before the follower gets the panel (pinned by test_vegas_yielding_to_a_follower_shows_a_rotation_screen_first; passes on origin/main too). docs/IPC_CONTROL_SOCKET.md no longer names _screen_preempted. CHANGELOG entry under Unreleased. A mutation run broke 46 moved or new pieces once each (OnDemand, Live, Vegas/Rotation, after(), each mid-screen rule, the runner's pacing, exits and service points, the controller's gathering and claims). Three survived and get a test here: - the after-loop service point not reading the WiFi notice: the completed-loop checkpoint gets its own name, and a run-loop test has a notice pending when a later frame comes back empty; - the Vegas yield path not marking vegas_yielded: the follower test above; - _take_plan not writing back a reset on-demand index: a controller test with an index past a shortened list. All 46 now fail at least one test. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
294bd522bd |
fix(web): the installed-plugins list never waits on the registry download (#766)
GET /plugins/installed called get_registry_info() per plugin. Despite the "no network call" comment, a cold or expired cache made that download plugins.json (10 s timeout, three attempts), and with no cached copy each plugin's lookup repeated it -- offline, every load waited out the timeouts. The route now reads the registry copy already in memory, however old, via get_cached_registry_info(). A missing or expired copy starts a single background refresh (backing off after an offline failure), so a later load gets update and verified badges. Store, install and update paths still fetch. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
80048fe78e |
test(starlark): keep the route tests off the network, out of the repo, and order-independent (#764)
Three hygiene problems in test_starlark_pixlet_routes.py: - GET /plugins/installed calls plugin_store_manager.get_registry_info() per plugin, which calls fetch_registry(); on a cold cache that downloads plugins.json from raw.githubusercontent.com (10s timeout plus retries). The three listing tests in TestInstalledAppsAppearWithTheOtherPlugins therefore went to the network and stalled without a connection. The client fixture now stubs fetch_registry on the app's store manager. - _STARLARK_APPS_DIR, _STARLARK_MANIFEST_FILE and _STARLARK_MANIFEST_LOCK_FILE are three separate module constants. Four fixtures redirected only the first two, so _starlark_manifest_lock() still took its lock at the repo's starlark-apps/manifest.json.lock (Linux only; Windows has no fcntl). The path is gitignored, so `git status` never showed it. One autouse fixture now redirects all three to tmp_path for every test in the file, and the class fixtures build on it. - The TestTheStoreUsesTheTokenTheUserConfigured unit tests patch plugin_store_manager / config_manager on the api_v3 blueprint. Those attributes only exist once web_interface/app.py has been imported, which happens as a side effect of the first test that uses the client fixture. Run before any such test, the four non-client tests raised AttributeError. They now patch with create=True. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
378478124c |
test(on-demand): find the config write by key, not by position (#763)
The font-usage publisher thread writes its own cache key at its own pace; the last set() call is not always the on-demand config. Same race #751 fixed for the restore-failed state test. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
d447cdd965 |
test(starlark): stop the pixlet-editor route tests leaking a fake Popen into the captive-portal check (#761)
TestPixletEditorHostDefaultsButDoesNotOverride patched Popen on the shared subprocess module. The app's before_request captive-portal hook runs subprocess.run (`with Popen(...) as p`) for `systemctl is-active hostapd` whenever its 30s AP-mode cache is cold, so on a Linux host with systemctl - the CI runner - the hook got the FakeProcess and the request 500'd with "'FakeProcess' object does not support the context manager protocol". Whether it happened depended on how long ago the previous request ran, hence the intermittency; Windows has no systemctl and never reached it. - Patch the starlark route module's own `subprocess` binding instead of the global Popen. - Pin is_ap_mode_active to False in this file's client fixture, so no request here shells out to systemctl/nmcli at all. Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
ee7c3389f9 |
feat(fetch): count bytes on the wire as well as decoded (#760)
fetch-stats now reports wire_bytes (urllib3's raw socket byte count) beside the decoded bytes in every counter set. ESPN gzips its scoreboards, so the decoded count overstated real traffic ~10-14x: ledpi measured football at 33.2 MB/h decoded vs 3.15 MB/h on the wire. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
caec9f5bf5 |
feat(display): report a scrolling screen held by its plugin's update() (#758)
While a plugin's update() runs it holds the plugin's lock and its frames are skipped -- on a scroller, a frozen strip -- with nothing logged. The high-FPS loop now times each run of skipped frames (report_hold=True); one of 250 ms or more logs 'Display of X held N ms by its update()' (rate-limited per plugin) and is recorded as a 'display hold' busy skip, which never touches the circuit breaker. The 1 Hz loop is left out: one skipped frame there measures the loop interval on a screen that did not visibly freeze (seen on ledpi as ~1000 ms reports on clock-simple and switch-mode football). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|
ChuckandClaude Opus 5.5
|
a74b5a2f0f |
fix(espn): fetch a window's edge months whole, cap chunk requests per process (#751)
- fetch_espn_date_chunks() asks for a window's partial edge month whole when the window covers ESPN_MONTH_COVER_MIN_DAYS (7) or more of its days, trimmed to the window by US Eastern start date. New espn_request_chunks(). - Chunk requests share one process-wide cap of ESPN_CHUNK_WORKERS (6) in flight. - A new process starts as if a range had just been rejected, so it no longer spends a doomed 400 per window at start. - Also: _eastern_zone() without try/except/pass (Codacy), and test_on_demand_live_and_restore reads the last on-demand state write rather than the last cache write (the font-usage publisher raced it; main CI had failed on it since #748). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |