fix: add lightweight recovery framework with retry‑backoff

[saidai-bhuvanesh]

Description:

This PR introduces a small, evidence‑driven recovery layer to improve the app’s resilience against transient network failures and crashes.

Key Additions

• RecoveryCoordinator.kt
  • Provides runWithRecovery for any suspend operation.
  • Implements exponential back‑off (default 3 attempts, starting at 500 ms) and returns the project’s Result<T> type.
• WorkerRecoveryHelper.kt
  • Offers safeExecute for CoroutineWorker implementations.
  • Retries only on network‑related AppExceptions, preserving existing worker semantics.

Repository Updates

• IptvRepository.kt
  • Wrapped fetchChannelsForPlaylistWithRetries with RecoveryCoordinator to centralize retry logic and convert exceptions to Result.Error.
• (Pending) StreamRepository.kt and TraktSyncWorker.kt will be refactored similarly to use the new helpers (already staged in this branch).

Tests

• New unit tests (RecoveryCoordinatorTest.kt & WorkerRecoveryHelperTest.kt) verify retry counts, back‑off behavior, and correct error handling.

Benefits

• Prevents crashes caused by uncaught throw statements in network code.
• Provides consistent, configurable retry handling across repositories and workers.
• Keeps the implementation lightweight and modular, aligning with the evidence‑driven approach.

Please review the changes, approve, and merge.