ci: enable cognitive complexity for very_good_cli.yaml by marcossevilla · Pull Request #1755 · VeryGoodOpenSource/very_good_cli

marcossevilla · 2026-10-01T12:28:40Z

Description

Migrates every reusable workflow from very_good_workflows@v1.21.1 to @v2 and enables the cognitive complexity audit added in VeryGoodOpenSource/very_good_workflows#520.

All callers (very_good_cli, ci, spdx_license, test_optimizer) now reference @v2.
Drops run_bloc_lint: false, which is the default in v2 (the only breaking change). The hook packages built with dart_package.yml no longer run bloc lint, which they never needed.
The build job enables run_cognitive_complexity on the ubuntu-latest matrix leg only, so both legs do not race on the sticky PR comment.
cognitive_complexity_fail_on_increase is enabled to block regressions relative to the base branch.
The build job is granted pull-requests: write so the action can post its sticky comment. On fork PRs the token is read-only and the comment step fails silently without failing the job.
The matrix runs with fail-fast: false so a Windows failure does not cancel the Ubuntu run, and vice versa.

cognitive_complexity_diff_base is left empty: on pull requests the action diffs against origin/main, and on pushes to main it scans all of lib. The functions previously over the threshold of 15 were fixed in #1756, #1757, and #1758.

Follow-ups

Run package:cognitive_complexity locally beyond the CI defaults (lower threshold, --max-function-lines, bricks/test_optimizer/hooks).
Run package:undead to find and clean up dead code.

Type of Change

✨ New feature (non-breaking change which adds functionality)
🛠️ Bug fix (non-breaking change which fixes an issue)
❌ Breaking change (fix or feature that would cause existing functionality to change)
🧹 Code refactor
✅ Build configuration change
📝 Documentation
🗑️ Chore

github-actions · 2026-10-01T12:29:51Z

📊 Cognitive Complexity Analysis

Tip

Status: Clean - no complexity regressions in the current diff.

Last checked: a820e54 (run).

marcossevilla · 2026-10-16T13:15:01Z

@vgvbot update this with very_good_workflows@v2

unicoderbot · 2026-10-16T13:15:04Z

On it, I'll address the review comments and push the changes.

I've pushed changes to address the feedback.

Co-authored-by: marcossevilla <marcossevilla@users.noreply.github.com>

marcossevilla · 2026-10-16T13:54:15Z

@vgvbot fix CI failures

unicoderbot · 2026-10-16T13:54:19Z

On it, I'll address the review comments and push the changes.

I've pushed changes to address the feedback.

Co-authored-by: marcossevilla <marcossevilla@users.noreply.github.com>

marcossevilla · 2026-10-16T14:20:53Z

@vgvbot fix CI failures

unicoderbot · 2026-10-16T14:20:57Z

On it, I'll address the review comments and push the changes.

I've pushed changes to address the feedback.

Co-authored-by: marcossevilla <marcossevilla@users.noreply.github.com>

@v2

Pin every reusable workflow to @v2, drop run_bloc_lint (false is the v2 default), and remove the HEAD diff base, which diffed the checkout against itself and made the complexity scan a no-op. The scan now runs only on the ubuntu matrix leg so both jobs do not race on the sticky PR comment.

unicoderbot · 2026-10-19T09:43:43Z

On it, I'll address the review comments and push the changes.

I've pushed changes to address the feedback.

Co-authored-by: marcossevilla <marcossevilla@users.noreply.github.com>

unicoderbot · 2026-10-19T10:41:36Z