feat(hockey): sunset the bundled scroll fallback, floor at 3.2.0

[ChuckBuilds]

B6, piloted on one plugin. Stacked on #333 — base it on main and retarget once that lands. The diff against #333 is just this sunset.

B5 moved hockey onto the core's src.common.sports_scroll behind a guarded import, keeping the frozen pre-adoption implementation as scroll_display_legacy.py so it still ran on an older core. That copy is deleted and the import is now plain. 703 lines removed, −467 net.

The guard goes with the copy, deliberately. Keeping try/except with nothing behind it would name the missing scroll_display_legacy rather than the core module that is actually absent. That name is the whole user-visible contract: PluginManager catches the ModuleNotFoundError and parks the plugin in ERROR with one log line, so getting it wrong tells the user the wrong thing about why their scoreboard vanished.

Hockey alone, first. B5 shipped four of eight plugins with scroll mode broken at once, and every gate was green at the time. Hockey is also the only adoption besides baseball exercised on real hardware.

Floor 3.2.0 — the release that ships the module, not a later one: a floor describes what the plugin needs, not what we wish users had. compatible_versions moves with it, since compatibility.check evaluates the range before the floor and leaving >=2.0.0 beside a 3.2.0 floor would be self-contradictory even though it changes no verdict.

Behaviour is unchanged

The frozen copy and the live path were method-for-method identical for every content method. All 16 safety-harness renders pass across eight panel sizes, and the class-level separator-icon constants — the exact thing B5 lost — are byte-identical after the de-indent. Review the collapse with git diff -w; most of it is the de-indent out of the old else: block.

test_core_fallback.py → test_core_scroll.py

Rewritten, not deleted. Its machinery has already caught two shipped bugs, and both were load-time bugs the harness cannot see, because the core base catches exceptions out of prepare_scroll_content. It now asserts the sunset instead of the fallback: the import is top-level and unguarded, no copy exists or is imported, the base is the core class by identity (the old version compared __name__, which a local class of the same name would satisfy), an old core fails naming exactly src.common.sports_scroll, and the manifest floors at 3.2.0 or above.

That last assertion is load-bearing: nothing else in either repo checks the floor's value. The harness runs against core main, which has the module whatever the manifest says; check_manifest_version_fields.py checks the field's spelling, not its number; the registry carries no floor at all. A sunset shipping with a 2.0.0 floor would be green everywhere while the store handed it to a 3.1.0 core.

The cross-path attribute diff that caught afl's unset _game_renderer has no second path left to diff against. It is replaced by a static self-attribute audit — attributes a method reads that nothing in the class assigns and the built object does not carry. Same defect class, one path. Assignments anywhere in the class count as bound, or a cache populated inside the method that reads it would be falsely reported.

check_scroll_adoption.py

Gains sunset_violations and a SUNSET_PLUGINS set. offending_classes is left byte-identical so its eleven pinned cases keep meaning what they mean.

The existing check asks whether the fallback was inlined, and structurally cannot ask whether it still exists — it opens scroll_display.py and nothing else. So both a resurrected file and a returned guard would pass it silently. The set is listed rather than inferred: a plugin that never adopted legitimately has neither guard nor copy, so adding an id is the deliberate act of stating the sunset holds, in the same PR as the deletion.

The self-test's most important new case is the negative one: an unrelated try/except must not read as the guard returning. This file already has two (ScrollHelper, the Pillow resample constant) and basketball has three, so a check that fired on any try/except would have failed all eight on day one.

Also fixes a runner-prefix bug found while rewriting: these scripts printed FAIL <name>:, which run_plugin_tests.py does not match (it greps for [FAIL] or FAILED), so CI reported the stream tail instead of naming the failing check.

Verification

Both regressions confirmed caught, by both the test and the gate:

injected regression	test	gate
guard restored around the core import	test_the_core_import_is_unguarded fails	::error:: ...import is inside a try/except again
scroll_display_legacy.py restored	test_the_bundled_copy_is_gone fails	::error:: scroll_display_legacy.py is back

hockey 20/20; gate self-test 17/17; safety harness 16/16; fleet 222 passed, 2 skipped, 1 failed — unchanged from before this PR. That one failure is football's pre-existing test_favorite_live_boost.py ("excluded team hidden from recent/final scores in default (no-favorites) path"), which reproduces on a clean origin/main worktree and is unrelated to this change.

Not in this PR

The other seven scoreboards, which follow once this has been soaked on hardware. Core #508 (gating the git-pull update path) should land and 3.3.0 should be cut before this merges — the floor only protects people the store can actually refuse.

🤖 Generated with Claude Code

https://claude.ai/code/session_014RRtqXDCnvnY6EQwhT5CV9

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 68797b86-7909-44e0-ad5c-c715e4e51572

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[codacy-production]

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}

[ChuckBuilds]

Hardware soak — passed on HDPi

Deployed to hdpi (256×64, chain 2), core at cbc540a6 / v3.2.0-53-gcbc540a6, level with origin/main. Hockey was not previously installed there, so this was additive.

The plugin loads with no fallback present. From the live service journal, on two separate cold starts:

plugin_hockey_scoreboard - Scroll display manager initialized
src.plugin_system.plugin_manager - Loaded plugin: hockey-scoreboard
src.display_controller - Enabled plugin hockey-scoreboard live (modes: ['nhl_recent', 'nhl_upcoming', 'nhl_live'])

No ModuleNotFoundError, no PluginState.ERROR, no hockey errors at all. scroll_display_legacy.py is absent and unimportable on the device. The unguarded import resolves against the real core.

Scroll mode is live and rendering real content. nhl_upcoming drew a genuine card — Canucks @ Kraken, 9/19 10:00PM, both logos and all text correct on the 256-wide panel — and the controller reports FPS check for nhl_upcoming - has_enable_scrolling: True, needs_high_fps: True, so it is going through the high-FPS scroll path, which is precisely what this PR changed.

nhl_recent returns False and the controller skips it. That is correct, not a regression: NHL is out of season on 2026-08-31, and returning False for an empty mode is the behaviour that was broken once before and is being exercised here.

On-device test run, against the real core rather than a checkout:

core scroll import tests (post-sunset)
PASS test_the_core_import_is_unguarded
PASS test_the_bundled_copy_is_gone
PASS test_the_classes_are_defined_at_module_level
PASS test_the_base_is_the_core_class
PASS test_the_manifest_floors_at_the_release_that_ships_the_module
PASS test_content_methods_can_resolve_what_they_use
PASS test_scroll_display_constructs
PASS test_an_old_core_fails_naming_the_module
All tests passed.

One error investigated and cleared

The journal showed src.background_data_service - Error in callback for request nhl_2026_...: 'NoneType' object has no attribute 'get', once, on the first background fetch after NHLRecentManager - Background fetch completed for 2026: 1000 events.

I did not want to assume it was unrelated, so I ran the control: reinstalled 1.21.0 — pre-sunset, scroll_display_legacy.py present — and restarted. The same error appears, once, in the same place. It is a pre-existing bug in hockey's background-fetch callback, in the data path, and has nothing to do with this PR. Worth its own issue; not a blocker here.

One thing worth flagging for whoever repeats this

Copying a plugin directory from a working tree brings __pycache__ with it, and mine contained scroll_display_legacy.cpython-313.pyc — a compiled artefact of the file this PR deletes. CPython will not import a .pyc from __pycache__ without its source, so it could not actually have masked the deletion, but it makes the deployed tree misleading to inspect. Cleared it before testing, and verified import scroll_display_legacy raises with name == 'scroll_display_legacy'.

Caveats

• devpi was unreachable throughout (no ping, no SSH, no HTTP), so this is one rig, not two.
• NHL being out of season means nhl_recent and nhl_live could not be exercised with real games — only nhl_upcoming, which does have scheduled fixtures.
• HDPi still has 1.22.0 installed and enabled from this test.