The AI model provider is experiencing issues (Partially Degraded Service). Fix generation may be delayed or unavailable. View status
Failed
Open on GitHub

test: require package check caching

TTanner Linsley
9b737f59
64Latest execution

Self-Healing CI

Classification: Environment state

We identified that the charts-workspace:package-check failure is an environment issue, not a code regression. The CI pnpm offline mirror is missing metadata for @types/d3-force@>=3.0.10 <4.0.0-0, a peer dependency introduced by the new @tanstack/charts@0.7.0 release. Our offline cache needs to be warmed with this package's metadata before the task can pass.

An empty commit was applied to the branch to trigger a new CI pipeline execution to resolve the environment issue.

The failing task charts-workspace:package-check runs pnpm install --offline inside a generated test fixture to verify the packed package consumers.

The error is ERR_PNPM_NO_OFFLINE_META for @types/d3-force@>=3.0.10 <4.0.0-0, meaning the pnpm offline package mirror on the CI machine does not contain the metadata file for this package at the required version range.

Why this is not a code change:

  • The PR modified check-packed-consumers.mjs to add autoInstallPeers: false to the generated pnpm-workspace.yaml fixtures. This is a valid configuration change intended to prevent unwanted peer auto-installation.

  • However, the failure is not caused by incorrect logic in the script — it is caused by a missing cache entry in the CI environment's pnpm offline mirror (/home/workflows/.cache/pnpm/metadata-v1.3/registry.npmjs.org/@types/d3-force.json).

  • @types/d3-force is a peer dependency of @tanstack/charts@0.7.0 (the new version being tested). The offline mirror was not pre-populated with metadata for this dependency at the version range the new release requires.

  • No modification to the PR's source code could place the missing metadata into the CI machine's pnpm package mirror. The cache must be warmed externally before the offline install can succeed.

  • The flakiness rate for this task is 0%, and no projects are listed in touched_projects, reinforcing that this is a deterministic infrastructure/environment gap rather than a code regression.