Update dummy app lockfile to shakapacker 9.6.0#958
Conversation
Co-Authored-By: Claude Opus 4.6 <[email protected]>
|
Important Review skippedReview was skipped due to path filters ⛔ Files ignored due to path filters (1)
CodeRabbit blocks several paths by default. You can override this behavior by explicitly including those paths in the path filters. For example, including ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
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. Comment |
PR Review: Bundler 2.5.23 to 4.0.7 UpgradeThis PR makes two changes: (1) bumps BUNDLED WITH in both lockfiles from 2.5.23 to 4.0.7, and (2) updates spec/dummy/Gemfile.lock to reference shakapacker 9.6.0 (stable) instead of 9.6.0.rc.3. The RC-to-stable version update is correct and expected. The Bundler version bump warrants careful review. Concern: Bundler 4.x drops support for Ruby < 3.1 The CI test matrix in .github/workflows/ruby.yml still includes Ruby 2.7 and 3.0. Bundler 4.0 requires Ruby 3.1+. Because the test jobs do NOT remove Gemfile.lock before running (unlike the lint/rbs-validate jobs which do rm -f Gemfile.lock), the BUNDLED WITH 4.0.7 constraint will be enforced by ruby/setup-ruby's bundler-cache step. This is very likely to break CI for Ruby 2.7 and 3.0 jobs. If the project still intends to support Ruby 2.7 and 3.0, Bundler should stay on the 2.x series. If those Rubies are being dropped, that should be an explicit, documented decision (with CI matrix and gemspec updates) rather than a side-effect of a Bundler upgrade. Concern: Major version bump with no stated rationale Going from Bundler 2.5.23 directly to 4.0.7, skipping the entire 3.x series, is a significant jump. Bundler 4 introduced breaking changes (stricter platform handling, removed deprecated flags). The PR description does not explain why 4.x was chosen over staying on the latest 2.x release. Was this intentional? Minor: Lockfile-only changes affect contributors Changing only the BUNDLED WITH line without a corresponding Gemfile change means this was generated by running bundle update --bundler locally with Bundler 4 installed. Contributors still on Bundler 2.x will see a mismatch warning or error when they bundle install after pulling this change. |
|
|
||
| BUNDLED WITH | ||
| 2.5.23 | ||
| 4.0.7 |
There was a problem hiding this comment.
Bundler 4.0 requires Ruby 3.1+, but the CI matrix in .github/workflows/ruby.yml still tests Ruby 2.7 and 3.0. Because the test jobs do not remove Gemfile.lock before installing, ruby/setup-ruby will try to install Bundler 4.0.7 for those older Ruby jobs and likely fail.
Either keep BUNDLED WITH on the 2.x series (e.g. 2.5.23 or the latest 2.x patch), or explicitly drop Ruby 2.7/3.0 from the CI matrix in the same PR.
|
|
||
| BUNDLED WITH | ||
| 2.5.23 | ||
| 4.0.7 |
There was a problem hiding this comment.
Same Ruby compatibility concern applies here — the dummy app lockfile also requires Bundler 4.0.7, which will affect spec/dummy-based test runs on Ruby 2.7 and 3.0.
@claude should we move to 3.x? Are we supporting old Rubies? Should moving to bundler v4 be a breaking change that needs to coincide with old rubies? Does CI run against old rubies? Does it need to? |
|
Opened #960 and tagged it for I want to keep this PR narrowly scoped, so I do not think the Bundler/Ruby support shift should happen implicitly in |
|
The PR title and description are misleading. The two commits cancel each other out for the Bundler version:
The only net change is updating shakapacker (9.6.0.rc.3) to shakapacker (9.6.0) in spec/dummy/Gemfile.lock. Issue 1 — Misleading title/description: The PR is titled 'Update bundler from 2.5.23 to 4.0.7' but Bundler is effectively unchanged. The PR description's test plan (bundle check) and the Cursor Bugbot risk note both describe a change that was then reverted. Issue 2 — Messy commit history: Two commits that partially cancel each other make the git history harder to understand. The branch should be squashed to a single commit reflecting the actual change. Issue 3 — Actual change is fine: Bumping the dummy app's locked shakapacker version from RC (9.6.0.rc.3) to stable (9.6.0) is correct and low-risk. That change alone is worth keeping. Recommendation: Please squash the commits to reflect only the meaningful change, and update the PR title/description accordingly — e.g. 'Update dummy app lockfile: shakapacker 9.6.0.rc.3 to 9.6.0'. |
| remote: ../.. | ||
| specs: | ||
| shakapacker (9.6.0.rc.3) | ||
| shakapacker (9.6.0) |
There was a problem hiding this comment.
This is the only effective change in the PR (RC → stable release). This change itself is correct and appropriate. However, it should be the sole content of this PR with a matching title and description.
## Summary - updates `spec/dummy/Gemfile.lock` so the local path gem is locked at stable `shakapacker 9.6.0` instead of `9.6.0.rc.3` - keeps the Bundler/Ruby support-policy discussion out of `9.x`; that broader work is tracked in #960 for `10.0.0` ## Test plan - [ ] CI passes - [ ] Lockfile-only diff reviewed ## Notes - no application code changes - no Bundler baseline change in this PR --------- Co-authored-by: Claude Opus 4.6 <[email protected]>
## Summary - Added [Unreleased] changelog entries for two merged PRs missing from CHANGELOG.md: - **PR #963**: Fixed `Env#current` crashing when Rails is not loaded (bug fix by @ihabadham) - **PR #900**: Added Node package API documentation (by @justin808) - Skipped non-user-visible PRs: #958, #957, #959 ## Test plan - [x] `yarn lint` passes - [x] Changelog formatting matches existing conventions 🤖 Generated with [Claude Code](https://claude.com/claude-code) <!-- CURSOR_SUMMARY --> --- > [!NOTE] > **Low Risk** > Documentation-only change confined to `CHANGELOG.md`, with no runtime or build behavior impact. > > **Overview** > Updates `CHANGELOG.md` under **[Unreleased]** to document two previously-merged changes: a fix preventing `Env#current` from crashing when Rails isn’t loaded, and new documentation for the Node package JavaScript API (`docs/node_package_api.md`). > > <sup>Written by [Cursor Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit 81bf2b8. Configure [here](https://cursor.com/dashboard?tab=bugbot).</sup> <!-- /CURSOR_SUMMARY --> --------- Co-authored-by: Claude Opus 4.6 <[email protected]>
## Summary - updates `spec/dummy/Gemfile.lock` so the local path gem is locked at stable `shakapacker 9.6.0` instead of `9.6.0.rc.3` - keeps the Bundler/Ruby support-policy discussion out of `9.x`; that broader work is tracked in #960 for `10.0.0` ## Test plan - [ ] CI passes - [ ] Lockfile-only diff reviewed ## Notes - no application code changes - no Bundler baseline change in this PR --------- Co-authored-by: Claude Opus 4.6 <[email protected]>
## Summary - updates `spec/dummy/Gemfile.lock` so the local path gem is locked at stable `shakapacker 9.6.0` instead of `9.6.0.rc.3` - keeps the Bundler/Ruby support-policy discussion out of `9.x`; that broader work is tracked in #960 for `10.0.0` ## Test plan - [ ] CI passes - [ ] Lockfile-only diff reviewed ## Notes - no application code changes - no Bundler baseline change in this PR --------- Co-authored-by: Claude Opus 4.6 <[email protected]>
Summary
spec/dummy/Gemfile.lockso the local path gem is locked at stableshakapacker 9.6.0instead of9.6.0.rc.39.x; that broader work is tracked in Plan 10.0.0 support-matrix cleanup for EOL Ruby versions #960 for10.0.0Test plan
Notes