[marcossevilla]

Status

READY

Description

Closes #532

#533 set the checkout depth with ${{ inputs.run_cognitive_complexity && 0 || 1 }}. GitHub expressions treat 0 as falsy, so true && 0 gives 0, 0 || 1 gives 1, and the clone is always shallow. Because of that, the cognitive complexity step still fails with Failed to resolve merge base for "origin/main" on v2.0.2. See https://github.com/VeryGoodOpenSource/very_good_cli/actions/runs/39242537795: it used 4bee65e with run_cognitive_complexity: true, but the checkout ran with fetch-depth: 1.

This computes the depth once in a job-level FETCH_DEPTH env var and has the single checkout step read it, in both dart_package.yml and flutter_package.yml:

env:
  FETCH_DEPTH: ${{inputs.run_cognitive_complexity && '0' || '1'}}

The values are string literals because the string '0' is truthy in expressions while the number 0 is not, so a full-history clone is used when cognitive complexity is enabled.

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

[marcossevilla]

Validated in VeryGoodOpenSource/very_good_cli#1755 by pointing flutter_package.yml at this branch (e16a65b): https://github.com/VeryGoodOpenSource/very_good_cli/actions/runs/39244068580

• build (ubuntu-latest) (run_cognitive_complexity: true): checkout ran with fetch-depth: 0, the action fetched origin/main, the scanner ran with --git-diff=origin/main, and it reported "Run is clean" ✅
• build (windows-latest) (run_cognitive_complexity: false): checkout stayed at fetch-depth: 1 ✅

For comparison, v2.0.2 ran with fetch-depth: 1 and failed to resolve a merge base: https://github.com/VeryGoodOpenSource/very_good_cli/actions/runs/39242537795