We reviewed the failure and found no connection to the PR's changes, which are limited to Nx configuration files (nx.json, package.json, start-plugin-core/package.json) with no modifications to @tanstack/react-router source or tests. The assertion mismatch (rootLoader called 3 times instead of 2) is consistent with inter-test state pollution, and the PR itself notes encountering flaky unit test behavior during validation. We recommend re-running the task.
Failed after NBNicolas BeaussartView latest
chore(nx): enable dependency-aware project analysis
80038d873e7Open on GitHubSelf-Healing CI
Classification: Flaky task
An empty commit was automatically applied to the branch to trigger a new CI pipeline execution to resolve the flaky task.
The PR modifies only three files: nx.json, the root package.json (adding @nx/js), and packages/start-plugin-core/package.json. None of these touch @tanstack/react-router source code or its test suite.
The failing test (tests/errorComponent.test.tsx) asserts that rootLoader is called exactly 2 times, but it was called 3 times. There is no logical path from adding an Nx source-analysis plugin or adjusting graph configuration to changing how many times a React Router loader is invoked during a test.
The touched_projects field contains "*" only because a global package.json was modified — not because any router logic changed. This is a broad signal that does not imply the failure is related to the PR's intent.
The PR body explicitly acknowledges non-deterministic test behavior during the author's own validation: "Router Generator was detected as flaky: the first unit run encountered stale empty local fixture directories; after removing those generated directories, the focused target and full unit suite passed." This confirms the test suite is sensitive to environment state and prior-run artifacts.
The call-count mismatch (3 vs 2) is a pattern strongly associated with inter-test state pollution — a prior test can leave navigation, loader, or render state that causes an extra cycle in the subsequent test. This is unrelated to any change introduced by the PR.
No similar failures were found on other branches, and the flakiness rate is historically 0%, but the PR's own testing observations confirm non-deterministic behavior is present in this run.