chore: remove dead type re-exports, prune stale suppressions, gate unused-types

[thymikee]

What

Two related cleanups that finish the arc started in #1363, and turn the unused-types rule from warn into error. Every fallow detector is now at zero, with a suppression list that's 4x smaller.

1. Dead type re-exports (72 findings)

71 were re-export lines (export type { X } from './y.ts' with no importer of X through that path); one was a direct declaration (HelpTopicName, zero references anywhere).

Removing them chained, which is the argument for gating it:

• 8 imports that existed only to feed a re-export became unused — caught by noUnusedLocals.
• Clearing those exposed a further chain: command-projection.ts → batch/index.ts → batch/projection.ts.
• Two export type {} husks were left where every name in a block was dead — caught by oxlint's no-useless-empty-export, not by fallow.
• A comment in batch-policy.ts documented the command-surface.ts re-export path this removes; updated so it doesn't rot.

One finding wasn't what it looked like. fallow flagged DebugSymbolsOptions/DebugSymbolsResult in apple/core/debug-symbols/types.ts; removing them broke apple/core/debug-symbols.ts, which re-exported them from there. The instinct is to restore the leaf — but both ends were dead. Every real consumer imports from contracts/debug-symbols.ts, so the fix was removing both together.

2. Stale suppressions (20 blocks → 5)

The premise of #1363 was that fallow looked clean largely because of its own ignore list — so the list deserved the same audit. Emptying ignoreExports and re-running shows most entries no longer match any finding. They went stale for three different reasons:

• Code moved — assertSafeDerivedCleanup is exported from runner-cache.ts, not the runner-xctestrun.ts the suppression named.
• The exports became genuinely used — apps.ts, runner-contract.ts, runner-session.ts, cloud-webdriver.ts, the test-utils fixtures.
• Duplication — installAndroidInstallablePath was suppressed in two places at once.

app-lifecycle.ts needed a code fix rather than a suppression: it re-exported three parseAndroid* helpers from app-parsers.ts that nothing imports through it. The re-export is dropped and the block goes away; the two type re-exports on that statement are consumed and stay.

Important

The seven src/daemon/handlers/* blocks are kept (collapsed into one documented glob). They are not stale — those handlers are reached only through the dynamic import() table in request-handler-chain.ts, which the --production analysis behind check:production-exports cannot follow to a consumer. My first pass removed them because the staleness probe only exercised the default config; the repo runs fallow twice, and check:production-exports caught the mistake.

Safety: the published API is unchanged

Built main and this branch, then compared the exported names of all 11 entry points in package.json#exports:

index: identical (6)           io: identical (13)          artifacts: identical (1)
metro: identical (6)           batch: identical (1)        remote-config: identical (1)
install-source: identical (4)  android-adb: identical (13) contracts: identical (15)
selectors: identical (10)      finders: identical (3)
→ PUBLIC API IDENTICAL across all 11 published entries

Note

A raw byte-compare of dist/src/*.d.ts shows 2 of 19 files differing. That is not an API change: AndroidSnapshotBackendMetadata relocated between internal code-splitting chunks (sdk-android-adb.d.ts → index.d.ts). Chunk files aren't published entry points, so that boundary isn't consumer-visible — which is why the entry-point comparison above is the authoritative check, not byte identity.

Verification

Rebased onto main after #1363, #1362 and #1360 landed, and re-verified against the new tip:

• pnpm typecheck, pnpm lint, pnpm check:tooling, pnpm check:fallow, pnpm check:production-exports, pnpm format:check — all clean
• fallow dead-code — all detectors zero, in both the default and --production configs
• Tests: 489/489 files, 4340/4340 (serialized); smoke 14/14; provider-integration + interaction-contract + output-economy 50 files / 231 tests

[github-actions]

Size Report

Metric	Base	Current	Diff
JS raw	1.8 MB	1.8 MB	0 B
JS gzip	578.2 kB	578.2 kB	0 B
npm tarball	692.9 kB	692.9 kB	+16 B
npm unpacked	2.4 MB	2.4 MB	+46 B

Startup median (7 runs, lower is better):

Scenario	Base	Current	Diff
CLI --version	28.5 ms	26.6 ms	-1.8 ms
CLI --help	59.1 ms	57.7 ms	-1.4 ms

Top changed chunks: no changes in the largest emitted chunks.

[thymikee]

Reviewed f1275ea8: no findings. The cleanup is correctly based after #1363/#1362/#1360. Published package entrypoints remain the unchanged src/sdk/* modules, with affected selector, remote-config, and contract types still exported directly. The consolidated production suppression matches the seven handlers lazy-loaded by request-handler-chain.ts, while the removed Android/debug-symbol re-exports have no active consumers through those paths. Default and production Fallow audits, tooling, typecheck, coverage, integration, packaging, platform smokes, and CodeQL are all green. Ready for human review. Residual risk: the published-entry export comparison is author-provided rather than a permanent CI assertion.

[github-actions]

PR Preview Action v1.8.1
Preview removed because the pull request was closed.
2026-07-22 16:18 UTC