[arun279]

Part of the retroactive cleanup of #24: "PR 2: App test pruning" and "PR 3: CI test pruning in test/ci" from the 2026-10-21 cleanup audit, done as one PR because both are test pruning. Every item was re-checked against current feat/expo-native (dfad4ee) first. Each deletion below is tautological, outmoded or duplicated, and names the test that still covers its behaviour. Where only a weak test covered real behaviour, I replaced it with a better one.

App tests

Change	Why	Still covered by
no-flake-masks: drop allowlist, an empty array asserted to equal []	Tautology	The scan assertion in the same test
Four in-memory preference fakes (prefs/_storage.ts, support/stores.ts, support/up-next.tsx, and a fourth the audit missed in episode-reminders.test.tsx) become one memoryPreferenceStorage(seed?)	Duplicated	Not a behaviour change; every test that used a fake now uses it
kv-contract.ts moves from core/test/support into native/__tests__/support; support/vitest.ts and its moduleNameMapper go	Its only runner is the native suite; the shim existed so the web store could share it	The same suite still runs against both native backends
perf-up-next.ts builds entry from support/up-next.tsx's entry() with the same values	Duplicated builder	The perf tests, unchanged
theme.test.ts folds into prefs-store.test.ts	One store, one file	The same two tests
sync-strip.test.tsx "counts the backlog" goes	It repeats syncBanner's backlog copy through the strip's generic message rendering	sync-contract.test.ts "surfaces a backlog only past the threshold…", and the strip's other render tests
query-keys.test.ts's literal tuples go	They mirror the factory and change with every rename	The prefix test stays. query-cache.test.ts now also asserts that the profile, genres and first-play queries persist. A renamed key head breaks persistence, which the tuples only caught by accident
maestro-ids.test.ts: drop the "added to Watchlist" tombstone	Outmoded	The positive "In Watchlist" assertion beside it
Drop export from twelve support symbols that only their own files use	Dead exports	None needed

CI tests

Change	Why	Still covered by
Drop not.toContain checks for MAESTRO_DRIVER_STARTUP_TIMEOUT, REQUIRED:, POLL_SECONDS, --dimensions=ALL, PLAY_SIZE_LIMIT_BYTES, AAB_SIZE_LIMIT_BYTES, density 480, summarize-atlas, restore-base-metrics/cue-footprint-$, measure-play-size, cue-native-ios-sizes, the response-timing variables, "Measure the merge-base simulator app", IPA_SIZE_LIMIT_BYTES and .size-limit.json	Outmoded: each token appears nowhere in the repository except its own tombstone	The positive assertions in the same tests (runner start line, require-green-ci.sh and the required run-id inputs, density 640, footprint's no-build checks including xcodebuild, one xcodebuild in native-ios, the Firebase limit)
Drop three continue-on-error guards in fast-validation	Duplicated	no-flake-masks.test.ts scans every workflow for continue-on-error
ios-daemons: drop the disable-daemons-then-boot check	Duplicated	ios-accessibility.test.ts, whose expected boot block already has disable-ios-daemons.sh right before simctl boot. The host-resources half stays
fast-validation and publish-update read jobs through test/support/workflow-jobs.ts	One parser instead of three	publish-update used to slice the release workflow from the leg to the end of the file, so the android check could search the ios leg's steps; it now reads one job
core/test/ci/render-counts.test.ts moves into test/ci/render-performance.test.ts, and the "keeps zero render-count deviation" case goes	It tests a root script, and that case duplicated "rejects an introduced scenario with a different count"	That test and the rest of the gate's cases

Kept, against the audit:

• The force-avd-creation and --flatten-debug-output negatives encode live choices. The second is the only check behind "leaving screenshots in place".
• ios-maestro-readiness pins the boot order relative to xcode-select and the runner, which no other test does.

Mutation check

Each break was applied alone, and the named test failed:

Break	Failing test
syncBanner never returns the backlog banner	sync-contract.test.ts "surfaces a backlog only past the threshold and the grace window"
SyncStrip stops rendering the banner message	sync-strip.test.tsx "says offline in a sentence…" and others
Rename the users key head	query-cache.test.ts "keeps the screens a user lands on"
Rename the genres key head	query-cache.test.ts "keeps the screens a user lands on"
continue-on-error: true on native-e2e-ios-light	no-flake-masks.test.ts
Boot without disable-ios-daemons.sh	ios-accessibility.test.ts "turns on simulator accessibility before every iOS lane boots its simulator"
Render gate allows a deviation of 1	render-performance.test.ts "rejects an introduced scenario with a different count"

Footprint

No product code changes, so no bundle changes. Test code: 219 lines added, 358 deleted, net -139, including the moved render-count test.

pnpm check passes on Node 22.23.4.

[github-actions]

Pull request footprint

measurement	base to head
Product code lines in core/src, native/src, native/app, and native/modules	+0
Test lines in test, packages/*/{test,__tests__,e2e}, and .maestro	-128
CI and tooling lines in .github and scripts	+0
Product comment lines identified by a comment prefix	+0

User-delivered artifact sizes

measurement	base	head	delta
Expo iOS JavaScript bundle, raw file	5.44 MB	5.44 MB	0 B
Expo Android JavaScript bundle, raw file	5.48 MB	5.48 MB	0 B
Firebase tester APK, arm64-v8a and all densities	41.16 MB	41.16 MB	0 B
Play download estimate, XXXHDPI arm64-v8a English Android 15	19.83 MB	19.83 MB	0 B

Most complex product function

measurement	base	head
Cyclomatic complexity, oxlint complexity	10	10
Cognitive complexity, Biome noExcessiveCognitiveComplexity	15	15