[uche001-dev]

Summary

Adds (and in one case expands) unit test coverage for four previously-untested or thinly-tested modules: the Stellar sequence cache, the account-lock utility, the on-chain progress-tracker read, and the auth attempt tracker's lockout-expiration path.

closes #588
closes #589
closes #590
closes #598

Changes

#588 - Sequence cache (src/stellar/sequence-cache.ts)

New tests/unit/stellar/sequence-cache.test.ts. Covers:

• getNextSequence() cache hit (Redis EXISTS true, atomic increment)
• Cache miss represented both as false and as null from the Lua script, falling back to Horizon via stellarClient.getAccount and seeding with the account's current sequence
• The SEED_AND_INCREMENT_SCRIPT NX-collision case where another process seeds first
• Concurrent calls for two different accounts resolving independently
• invalidate() deleting the key, writing an audit log entry, and not throwing when the audit log call rejects
• resetTo() setting the key with the configured TTL

#589 - Account lock utility (src/utils/account-lock.ts)

New tests/unit/utils/account-lock.test.ts. Covers:

• Lock acquisition/release around a successful call
• Lock release after the wrapped function throws (so the lock doesn't leak on error)
• Serialization of two concurrent calls for the same account (asserted via execution order)
• Two concurrent calls for different accounts running without serializing
• The configured 10s acquisition timeout, using fake timers to avoid a slow real-time test

#590 - Progress tracker (src/stellar/progress-tracker.ts)

New tests/unit/stellar/progress-tracker.test.ts. Covers:

• Successful hash retrieval and decode
• Returning null when the contract isn't configured (no Soroban call made)
• Returning null when the decoded value isn't a string
• Returning null and logging a warning when the simulation reports an error, or has no retval
• Returning null and logging a warning when the RPC call throws

Note: new StellarSdk.Contract(id) validates the strkey checksum, so the mocked contract ID in these tests had to be a validly-formatted (though undeployed) contract address, not an arbitrary placeholder string. An earlier draft used a bogus ID and every test silently fell into the catch block instead of exercising the intended code path.

#598 - Auth attempt tracker (src/utils/auth-attempt-tracker.ts)

Expanded the existing tests/unit/utils/auth-attempt-tracker.test.ts (already had good coverage for attempt counting, lockout triggering/escalation, and clearing). Added:

• Explicit lockout expiration cases: ttl of 0 and the missing-key -2 sentinel both report not-locked-out
• clearAuthFailures to checkAuthLockout round-trip, asserting a cleared address is reported as not locked out

Test plan

• npx vitest run on the 4 touched/added test files: 32/32 passing
• Full suite regression check: ran npx vitest run on this branch and, in a separate git worktree off upstream/main, on a clean npm ci baseline. Diffed the sorted list of FAIL lines from both runs: identical, 89 pre-existing failing test files in both, confirming these changes introduce zero regressions
• npx eslint on the 4 touched test files: no errors or warnings

Caveats

• The pre-existing failures across the rest of the suite (SEP-10 auth logger.debug issues, a waitlist-service undefined property access, a lock TTL mismatch, etc.) are not touched by this PR and were failing identically on a clean upstream/main checkout before any of these changes. They're out of scope here.
• None of the four issues' deliverables were already present upstream; all four required genuinely new or expanded test files.

[drips-wave]

@uche001-dev Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits