Skip to content

Reuse the gateway setup in PostgreSQL credential-withdrawal tests - #2790

Merged
freeqaz-openai merged 1 commit into
mainfrom
test/reuse-revision-gateway-start
Oct 17, 2026
Merged

freeqaz-openai merged 1 commit into
mainfrom
test/reuse-revision-gateway-start

Conversation

@freeqaz-openai

Copy link
Copy Markdown
Collaborator

Ten worker starts repeated the gateway driver and 50 ms convergence setting already provided by startWithGateway. They now pass just the Compute behavior each test needs. The case that also records worker events keeps its explicit setup.

All 36 tests in the file passed before and after with PostgreSQL 18.6. In a separate copy, making the first withdrawal report revoked instead of pending failed the existing retry assertion: one attempt instead of two. These are worker tests with a simulated gateway, not a live credential operation.

@freeqaz-openai
freeqaz-openai requested a review from a team as a code owner October 17, 2026 21:57
@freeqaz-openai freeqaz-openai added the agent-created PR opened by an AI agent; does not identify the reviewer or approver. label Oct 17, 2026
@clawsweeper

clawsweeper Bot commented Oct 17, 2026 •

Copy link
Copy Markdown
Contributor

🦞👀
ClawSweeper picked this up.

Pull request received. I will update this pull request when review starts.

ClawSweeper review complete

ClawSweeper finished reviewing this revision. The review result is being finalized.

View the workflow run.

@clawsweeper clawsweeper Bot added P3 Low-risk cleanup, docs, polish, ergonomics, or speculative feature. rating: 🦐 gold shrimp Decent PR readiness signal, but merge confidence is limited. status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR. labels Oct 17, 2026
@clawsweeper

clawsweeper Bot commented Oct 17, 2026

Copy link
Copy Markdown
Contributor

Codex review: needs maintainer review before merge.

What this changes

This PR replaces ten repeated worker configurations with the existing gateway setup helper.

Example: A credential-withdrawal test starts its worker with a 50 ms convergence timeout.

  • Before: The test repeats the timeout and gateway Driver configuration.
  • After: startWithGateway supplies the identical configuration.

Review scores

Measure Result What it means
Overall readiness 🦐 gold shrimp (3/6) Focused, correct test maintenance supported by PostgreSQL harness validation; evidence does not exercise a live gateway and makes no such claim.
Proof confidence 🦐 gold shrimp (3/6) This collaborator-authored test-only refactor changes no shipped behavior; PostgreSQL CI supports fixture equivalence, and simulated gateway results are explicitly distinguished from live credential operations.
Patch quality 🐚 platinum hermit (4/6) No actionable review findings were identified.

Product

Kind: Test only · Worth it: Yes · Fix scope: Complete
Reason: The collaborator-authored direction reduces duplicated fixture configuration using its established owner without changing product contracts or test expectations.

Merge readiness

✅ Ready for maintainer review

Keep this PR open: it is a focused, behavior-preserving test cleanup with no actionable findings, and repository policy reserves closure for maintainer triage.

Priority: P3
Reviewed head: 632d78b53aab13f850685c66fa1988b92ac784ca

Before merge

None.

Findings

None.

Agent review details

How this fits together

PostgreSQL integration fixtures start the real controller worker with simulated Compute and gateway Drivers, then inspect durable work, authorization outcomes, and withdrawal state.

flowchart LR
  A[Withdrawal test] --> B[Existing gateway setup helper]
  B --> C[Worker fixture]
  C --> D[Controller worker]
  D --> E[PostgreSQL state]
  D --> F[Simulated Compute Driver]
Loading

Technical review

Best possible solution:

Reuse the existing helper while retaining each scenario’s Compute behavior and the event-recording case’s explicit options.

Do we have a high-confidence way to reproduce the issue?

No product bug is claimed; source inspection verifies that the helper supplies the exact options removed from each call site.

Is this the best way to solve the issue?

The existing helper is the smallest complete owner for this repeated setup, and the diff preserves scenarios, assertions, and special event capture.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning medium; reviewed against ee9391867036.

Provenance checked

Testing

Proof path: in-process harness.

Security

None.

Evidence

What I checked:

Likely related people:

  • Free Wortley: Raw commit f890b9d adds tests/integration/postgres-worker-agent-revision-credential-sources.test.mjs:260 relative to its recorded parents. This identifies author metadata, not feature responsibility or a PR merger. (role: source-line author; confidence: high; commits: f890b9d48ddb; files: tests/integration/postgres-worker-agent-revision-credential-sources.test.mjs)

Review metrics

Metric Value Why it matters
Scope Production +0/-0; tests +104/-134 Reuses an existing helper at ten call sites, removing 30 net lines without adding machinery or assertions.

Labels

Label changes:

  • add P3: Low-risk test maintenance with no user-visible behavior change.
  • add rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🦐 gold shrimp and patch quality is 🐚 platinum hermit.
  • add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR.

Label justifications:

  • P3: Low-risk test maintenance with no user-visible behavior change.
  • rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🦐 gold shrimp and patch quality is 🐚 platinum hermit.
  • status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR.

Rating scale

6/6 🦀 challenger crab · 5/6 🦞 diamond lobster · 4/6 🐚 platinum hermit · 3/6 🦐 gold shrimp · 2/6 🦪 silver shellfish · 1/6 🧂 unranked krab. Overall follows the weaker of proof and patch quality; ✨ marks media proof (a screenshot, video, or linked artifact) that directly shows the changed behavior.

Workflow

ClawSweeper edits this one comment on every review. Comment @clawsweeper re-review for a fresh review only. Repair and merge execution lanes are retired.

Reviewed October 17, 2026, 6:09 PM ET / 22:09 UTC.

@freeqaz-openai
freeqaz-openai merged commit 2ce4c66 into main Oct 17, 2026
30 of 31 checks passed
@freeqaz-openai
freeqaz-openai deleted the test/reuse-revision-gateway-start branch October 17, 2026 22:12

This branch was previously deployed

1 inactive deployment
integration-qa-pr — 632d78b5 Deployed Oct 17, 2026 by freeqaz-openai via QA advisory (kubernetes) #2044
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

agent-created PR opened by an AI agent; does not identify the reviewer or approver. P3 Low-risk cleanup, docs, polish, ergonomics, or speculative feature. rating: 🦐 gold shrimp Decent PR readiness signal, but merge confidence is limited. status: 👀 ready for maintainer look ClawSweeper has no concrete contributor-facing blocker left for this PR.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant