Fixes the latent thread-park found by the "same-class crash-state" audit (follow-up to the SQLCipher cold-start crash, #210).
Problem
PruneWorker and BackfillWorker were the only two pre-auth background DB entry points not gating on EncryptedCacheGuard.isCacheLocked() before opening the database (unlike SyncWorker / SendWorker / IdleService). With encryptCache + appLock on and a headless cold start where the user hasn't authenticated (WorkManager after reboot, or the periodic backfill/prune window while locked), the first DAO call runs prepareCache() → resolvePassphrase() → session.await(), parking the WorkManager thread until unlock and serializing other DB openers behind the held prepareCache mutex. Self-heals on unlock, but wastes wakelock/battery and makes zero prune/backfill progress while locked.
Fix
Switch MailPruner / MailBackfiller injection to dagger.Lazy and add the if (cacheGuard.isCacheLocked()) return Result.retry() guard before resolving it — mirroring SyncWorker / SendWorker.
Tests
New PruneWorkerTest / BackfillWorkerTest (JVM): locked → retry() with the Lazy dep never resolved (verify(exactly = 0) { lazyDep.get() }); unlocked → runs; failure → retry(). Preflight green (build, unit tests, androidTest compile, lint, ktlint, detekt).
Broader coverage (all pre-auth entry points + the suspending-prepareCache provisioner case) is tracked in #225; instrumented coverage in #226.
Fixes the latent thread-park found by the "same-class crash-state" audit (follow-up to the SQLCipher cold-start crash, #210).
## Problem
`PruneWorker` and `BackfillWorker` were the only two pre-auth background DB entry points **not** gating on `EncryptedCacheGuard.isCacheLocked()` before opening the database (unlike `SyncWorker` / `SendWorker` / `IdleService`). With `encryptCache + appLock` on and a headless cold start where the user hasn't authenticated (WorkManager after reboot, or the periodic backfill/prune window while locked), the first DAO call runs `prepareCache()` → `resolvePassphrase()` → `session.await()`, **parking the WorkManager thread** until unlock and serializing other DB openers behind the held `prepareCache` mutex. Self-heals on unlock, but wastes wakelock/battery and makes zero prune/backfill progress while locked.
## Fix
Switch `MailPruner` / `MailBackfiller` injection to `dagger.Lazy` and add the `if (cacheGuard.isCacheLocked()) return Result.retry()` guard before resolving it — mirroring `SyncWorker` / `SendWorker`.
## Tests
New `PruneWorkerTest` / `BackfillWorkerTest` (JVM): locked → `retry()` with the `Lazy` dep **never resolved** (`verify(exactly = 0) { lazyDep.get() }`); unlocked → runs; failure → `retry()`. Preflight green (build, unit tests, androidTest compile, lint, ktlint, detekt).
Broader coverage (all pre-auth entry points + the suspending-`prepareCache` provisioner case) is tracked in #225; instrumented coverage in #226.
Closes #224
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Fixes the latent thread-park found by the "same-class crash-state" audit (follow-up to the SQLCipher cold-start crash, #210).
Problem
PruneWorkerandBackfillWorkerwere the only two pre-auth background DB entry points not gating onEncryptedCacheGuard.isCacheLocked()before opening the database (unlikeSyncWorker/SendWorker/IdleService). WithencryptCache + appLockon and a headless cold start where the user hasn't authenticated (WorkManager after reboot, or the periodic backfill/prune window while locked), the first DAO call runsprepareCache()→resolvePassphrase()→session.await(), parking the WorkManager thread until unlock and serializing other DB openers behind the heldprepareCachemutex. Self-heals on unlock, but wastes wakelock/battery and makes zero prune/backfill progress while locked.Fix
Switch
MailPruner/MailBackfillerinjection todagger.Lazyand add theif (cacheGuard.isCacheLocked()) return Result.retry()guard before resolving it — mirroringSyncWorker/SendWorker.Tests
New
PruneWorkerTest/BackfillWorkerTest(JVM): locked →retry()with theLazydep never resolved (verify(exactly = 0) { lazyDep.get() }); unlocked → runs; failure →retry(). Preflight green (build, unit tests, androidTest compile, lint, ktlint, detekt).Broader coverage (all pre-auth entry points + the suspending-
prepareCacheprovisioner case) is tracked in #225; instrumented coverage in #226.Closes #224
🤖 Generated with Claude Code