Repair Jira fixture caching, validate the shipped FixVersion action, and gate releases

[Jamie-BitFlight]

Why this change is needed

The old Docker fixture failed before it could verify the shipped action, and the suite mostly tested helper API calls. Production code also treated Jira's project-version array as a paginated object, so existing versions were missed and concurrent issue updates could create duplicates.

What changes

• Use the shared official Jira Software 10.3.26 / PostgreSQL 16.15 fixture, pinned to commit 4e95d80b6115ef60f14bc7c351a23c6ae5f06da5. Cold preparation or cache restoration verifies a checksummed database/home pair and real Jira writes; a cold fixture is saved only after a successful restore onto fresh volumes.
• Execute the packaged dist/index.js and independently read back Jira state. Cover existing-version reuse, partially overlapping requests, preservation, one new version across multiple issues, duplicate issue inputs, reruns, and an actual Jira 404 failure.
• Run a separate native uses: ./ invocation and assert exact Jira version IDs afterward.
• Fix project-version array handling and coordinate in-flight version creation within the action process.
• Align Node 24 and the frozen lockfile, regenerate the shipped bundle, and require lint and bundle parity in CI. Remove the duplicated MySQL/setup/snapshot path.
• Gate release publication on a successful same-repository main-push Jira E2E run. Validate the exact tested revision, select the next unused stable version, and atomically push only the release commit and exact tag. Remove the masked version-expression error and lifecycle-side tag pushes.

Verified results

All five PR workflows passed at 00f945c907fa556f69d61340c8b62911836d3410:

• CI: typecheck, lint, build parity, 178 unit tests
• Standalone unit tests
• Lint
• Automated review
• Real Jira E2E and native action: all three packaged scenarios passed; independent readback confirmed the native invocation preserved version 10103, added existing version 10104, and created no duplicate.

The final E2E job explicitly reported fixture-source=cache, restoring the fixture produced by the successful cold run. It verified authentication, issue creation, fixVersions, transition, and deletion on Jira 10.3.26 before running the action. Observed whole-job times were about 8m 50s cold and 3m 39s cached.

Frozen installs, strict workflow validation, source/E2E typechecks, and the real Yarn version/documentation lifecycle also passed locally. Default-branch and release verification follow the merge.

The fixture uses synthetic data and Atlassian's public short-lived test license. It verifies single-node Jira Data Center REST behavior; Cloud-specific APIs and cluster behavior are outside this lane.

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

• Configuration used: defaults
• Review profile: CHILL
• Plan: Advanced
• Run ID: 41c9efdd-14a1-4e3a-88a1-b5963b99427d

---

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

---

_{Comment @coderabbitai help to get the list of available commands.}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-10T01:41:54.511757Z	00f945c	New commits

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.