test: coverage analysis + improvements across all identified gaps#8
Merged
Conversation
Captures jest coverage numbers, identifies the sandbox/instrumentation incompatibility, and lists prioritized gaps (leader forwarding, route error paths, raftNode failure branches, edge resilience). Co-Authored-By: Claude Opus 4.8 <[email protected]> Claude-Session: https://claude.ai/code/session_0159Rma834V7wPgft7fzoFtg
The vm determinism sandbox compiles reducers from fn.toString(), so under `jest --coverage` Istanbul's injected `cov_<hash>()` counters reached the sandbox as undefined free globals and threw ReferenceError -> apply 500. Bind a no-op sink for those counter identifiers before SANDBOX_SETUP removes Proxy, so instrumented reducers still run deterministically while a genuinely banned global (Date, etc.) still throws. Add coverageThreshold + collectCoverageFrom to jest.config.js, a `test:coverage` script, and a CI coverage step; extend the CI Node matrix to 20.x and 22.x. Adds sandbox regression tests for the instrumentation case. Co-Authored-By: Claude Opus 4.8 <[email protected]> Claude-Session: https://claude.ai/code/session_0159Rma834V7wPgft7fzoFtg
forwardToLeader / isForwarded had ~18% coverage and no dedicated tests, despite being the write path every non-leader node relies on. Add a suite driving a stub leader server: successful relay, verbatim content-type/body passthrough, request-context (X-Request-Id/X-Actor) propagation + the X-Forwarded-By anti-loop mark, and the connection-error and socket-timeout fallbacks. forward.ts now at 100% stmts / 89% branch. Co-Authored-By: Claude Opus 4.8 <[email protected]> Claude-Session: https://claude.ai/code/session_0159Rma834V7wPgft7fzoFtg
Add tests for the controller/route failure surface that the happy-path single-node suites miss: - book & module controllers on a 3-node cluster: follower writes -> 421 (forward-then-fallback), the X-Forwarded-By anti-loop branch, strong reads on the leader, and 404 on the strong-read path - raft membership admin routes: GET /members, 400 on missing id/url, 409 on removing an unknown member, 421 on a follower - app-level /ready (503 pre-leader) and /metrics (404 when disabled) - RaftNode.collectMetrics: leader per-peer replication-lag series and its reset/rebuild after a member is removed controllers 72->85% stmt / 56->79% branch; routes 76->88% / 69->83%. Co-Authored-By: Claude Opus 4.8 <[email protected]> Claude-Session: https://claude.ai/code/session_0159Rma834V7wPgft7fzoFtg
The HTTP-backed edge suites don't reach the failure lifecycle. Drive EdgeReplica through a controllable fake LogStreamSource: reconnect-after-error resumes from the applied cursor (and closes the old conn), replayed entries are not re-applied, onCaughtUp advances past filtered scoped entries, snapshot bootstrap restores state + cursor, and waitForIndex resolves / rejects on stop and timeout. Co-Authored-By: Claude Opus 4.8 <[email protected]> Claude-Session: https://claude.ai/code/session_0159Rma834V7wPgft7fzoFtg
Add tests/logger.test.ts (level filtering, JSON/pretty output, base/child + request-context enrichment, env-driven createLogger) — logger.ts 0% -> ~100%, lifting src/platform to 99% stmt / 92% branch. Ratchet the global coverageThreshold up to the new floor (86/71/86/88). Update the analysis doc with a resolution summary. Co-Authored-By: Claude Opus 4.8 <[email protected]> Claude-Session: https://claude.ai/code/session_0159Rma834V7wPgft7fzoFtg
naveed949
changed the title
The PR-open CI run executed on the initial docs commit only; the follow-up push that added the coverage gate and the Node 20.x/22.x matrix did not fire a synchronize run. Empty commit to run CI against the latest tree. Co-Authored-By: Claude Opus 4.8 <[email protected]> Claude-Session: https://claude.ai/code/session_0159Rma834V7wPgft7fzoFtg
naveed949
marked this pull request as ready for review
June 22, 2026 18:59
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Started as a test-coverage analysis (
docs/TEST_COVERAGE.md) and then implemented all six prioritized improvement areas it identified. The suite grew 236 → 273 tests, coverage is now enforced in CI, and a latent sandbox/coverage incompatibility is fixed.Coverage before → after
src/platformsrc/controllerssrc/routessrc/edgeWhat changed
ee99d16). The vm determinism sandbox compiles reducers fromfn.toString(), so underjest --coverageIstanbul's injectedcov_<hash>()counters reached the sandbox as undefined globals →ReferenceError→ apply 500 (3 tests failed only under coverage).compileReducernow binds a no-op sink for those counter identifiers (beforeProxyis stripped), so instrumented reducers run while a genuinely banned global still throws. AddedcollectCoverageFrom+coverageThreshold, atest:coveragescript, and a CI coverage step on a Node 20.x + 22.x matrix.650d5c0) —tests/forward.test.ts: relay, content-type/body passthrough, X-Request-Id/X-Actor propagation, X-Forwarded-By anti-loop, connection-error and socket-timeout fallbacks.forward.ts18% → 100%.e1429ec) —tests/httpRouting.test.ts: follower 421 (book + module), strong reads on the leader, 404 on the strong-read path, raft membership admin routes (400/409/421), app/ready503 and/metrics404.e1429ec) —tests/raftMetrics.test.ts: leader per-peer replication-lag series and its reset/rebuild after a member is removed.f6f819b) —tests/edgeResilience.test.ts: reconnect-after-error resumes from the cursor, replays aren't re-applied, onCaughtUp advances past filtered scoped entries, snapshot bootstrap, waitForIndex resolve/reject — via a controllable fake stream source.a65a03d) —tests/logger.test.ts: level filtering, JSON/pretty output, base/child + request-context enrichment, env-drivencreateLogger.logger.ts0% → ~100%.Two branches were deliberately left (they need fault injection that risks flakiness): the read-barrier 2 s timeout, and the raw HTTP/EventSource socket internals (covered end-to-end by the edge integration suites).
Test plan
npx tsc --noEmitclean;yarn testandyarn test:coverageboth green (273/273) with thresholds met (stmt 88.5%, branch 73.4%, func 89.7%, line 90.5%).🤖 Generated with Claude Code