Overview - Nx

Nx

7 days30 days

Latest CI Pipeline Executions

Fuzzy

Failed
34503
Flaky task→
6d55fb8d fix(core): graceful cache write failures for project graph and lockfile caches Change cache write errors from hard throws to logged warnings with cleanup. Affected locations: - nx-deps-cache: writeCache() now warns and removes stale files instead of throwing after 5 failed attempts - js plugin: lockfile external-nodes and dependencies caches use safeWriteFileCache() with pre-serialization validation - gradle plugins: report cache writes are wrapped in try/catch with warning on failure This prevents Nx from crashing when disk writes fail (e.g. permissions, disk full) while ensuring corrupted partial files are cleaned up.
23 days ago by AgentEnder
Canceled
34503
c26a6796 fix(core): graceful cache write failures for project graph and lockfile caches Change cache write errors from hard throws to logged warnings with cleanup. Affected locations: - nx-deps-cache: writeCache() now warns and removes stale files instead of throwing after 5 failed attempts - js plugin: lockfile external-nodes and dependencies caches use safeWriteFileCache() with pre-serialization validation - gradle plugins: report cache writes are wrapped in try/catch with warning on failure This prevents Nx from crashing when disk writes fail (e.g. permissions, disk full) while ensuring corrupted partial files are cleaned up.
23 days ago by AgentEnder
Succeeded
master
eb498861 fix(repo): reset package.json files after local release (#34648) ## Current Behavior When running `pnpm nx-release --local false`, the `nx release version` step modifies `package.json` files for `angular-rspack`, `angular-rspack-compiler`, `dotnet`, and `maven` (bumping versions and resolving `workspace:*` protocols). The local release path (`--local false`, non-CI) exits early before reaching the reset logic that the CI path uses, leaving unstaged changes behind. ## Expected Behavior After the release script completes (or is interrupted), the source `package.json` files should be restored to their original state. The version bumps are only needed in `dist/` for publishing — the source files should stay at `0.0.1` with `workspace:*` protocols. This PR: - Extracts a shared `resetPackageJsons()` function used by both the local and CI code paths - Wraps the local release steps in `try/finally` so files are restored even on errors - Adds a `SIGINT` handler so files are restored on ctrl+C
23 days ago by AgentEnder
Succeeded
34582
300a826c fix(core): skip writing deps cache if already up-to-date
23 days ago by AgentEnder
Succeeded
master
bdc61b5a chore(repo): improve e2e test timeout handling and bump cache bust (#34383) ## Current Behavior E2E tests may timeout without clear error messages, making it difficult to diagnose test failures. ## Expected Behavior E2E test utilities should provide better timeout handling and logging to help diagnose test failures. ## Changes - Add timeout handling to e2e test utilities (`runCLI` and `runLernaCLI`) - Add command logging to track execution time - Improve timeout error messages with process output - Bump cache bust value ## Related Issue(s) CI stability and debugging improvements
24 days ago by AgentEnder
Succeeded
master
fcf46603 fix(core): preserve nxCloud=skip in non-interactive CNW mode (#34616) ## Current Behavior After #34580, `determineNxCloudV2()` returns `'skip'` in non-interactive mode, but the caller remaps it to `nxCloud = 'yes'` with `skipCloudConnect = true`. This causes `setupCI()` to run and generate `.github/workflows/ci.yml` in new workspaces — which didn't happen before. This breaks the `extras.test.ts` e2e snapshot test because `.github/workflows/ci.yml` now appears in the expanded default task inputs. ## Expected Behavior Keep `nxCloud = 'skip'` when the cloud choice is `'skip'`, which prevents CI file generation in non-interactive mode. This restores the behavior prior to #34580. ## Related Issue(s) Fixes the `extras.test.ts` e2e snapshot failure introduced by #34580.
24 days ago by AgentEnder
Failed
34332
Fix applied→
14ca3159 fix(core): handle string dependsOn entries
25 days ago by AgentEnder
Succeeded
master
dc81b8bb fix(gradle): ensure that atomized task targets have dependsOn (#34611)      ## Current Behavior  ## Expected Behavior  The dependsOn of atomized tasks should match the base non-atomized task. ## Related Issue(s)  Fixes Q-174
25 days ago by AgentEnder
Succeeded
master
1ecf0fb6 fix(core): reject pending promises directly when plugin worker exits unexpectedly (#34588) When a plugin worker process exits unexpectedly, the exit handler previously sent synthetic `loadResult` messages to all pending response handlers. If any handler was waiting for a different result type (e.g. `createNodesResult`), the type validation would reject with a confusing "Expected createNodesResult, got loadResult" error instead of surfacing the actual cause. Split response handlers into `onMessage` / `onError` callbacks so the exit handler can reject each pending promise directly with a clear "Plugin worker exited unexpectedly" error. Also use unique transaction IDs for `load` messages (via `generateTxId`) to avoid potential handler overwrites during worker restarts. Fixes #34564
25 days ago by AgentEnder
Succeeded
master
1ecf0fb6 fix(core): reject pending promises directly when plugin worker exits unexpectedly (#34588) When a plugin worker process exits unexpectedly, the exit handler previously sent synthetic `loadResult` messages to all pending response handlers. If any handler was waiting for a different result type (e.g. `createNodesResult`), the type validation would reject with a confusing "Expected createNodesResult, got loadResult" error instead of surfacing the actual cause. Split response handlers into `onMessage` / `onError` callbacks so the exit handler can reject each pending promise directly with a clear "Plugin worker exited unexpectedly" error. Also use unique transaction IDs for `load` messages (via `generateTxId`) to avoid potential handler overwrites during worker restarts. Fixes #34564
25 days ago by AgentEnder
Succeeded
master
1ecf0fb6 fix(core): reject pending promises directly when plugin worker exits unexpectedly (#34588) When a plugin worker process exits unexpectedly, the exit handler previously sent synthetic `loadResult` messages to all pending response handlers. If any handler was waiting for a different result type (e.g. `createNodesResult`), the type validation would reject with a confusing "Expected createNodesResult, got loadResult" error instead of surfacing the actual cause. Split response handlers into `onMessage` / `onError` callbacks so the exit handler can reject each pending promise directly with a clear "Plugin worker exited unexpectedly" error. Also use unique transaction IDs for `load` messages (via `generateTxId`) to avoid potential handler overwrites during worker restarts. Fixes #34564
25 days ago by AgentEnder
Canceled
master
1ecf0fb6 fix(core): reject pending promises directly when plugin worker exits unexpectedly (#34588) When a plugin worker process exits unexpectedly, the exit handler previously sent synthetic `loadResult` messages to all pending response handlers. If any handler was waiting for a different result type (e.g. `createNodesResult`), the type validation would reject with a confusing "Expected createNodesResult, got loadResult" error instead of surfacing the actual cause. Split response handlers into `onMessage` / `onError` callbacks so the exit handler can reject each pending promise directly with a clear "Plugin worker exited unexpectedly" error. Also use unique transaction IDs for `load` messages (via `generateTxId`) to avoid potential handler overwrites during worker restarts. Fixes #34564
25 days ago by AgentEnder
Succeeded
master
1ecf0fb6 fix(core): reject pending promises directly when plugin worker exits unexpectedly (#34588) When a plugin worker process exits unexpectedly, the exit handler previously sent synthetic `loadResult` messages to all pending response handlers. If any handler was waiting for a different result type (e.g. `createNodesResult`), the type validation would reject with a confusing "Expected createNodesResult, got loadResult" error instead of surfacing the actual cause. Split response handlers into `onMessage` / `onError` callbacks so the exit handler can reject each pending promise directly with a clear "Plugin worker exited unexpectedly" error. Also use unique transaction IDs for `load` messages (via `generateTxId`) to avoid potential handler overwrites during worker restarts. Fixes #34564
25 days ago by AgentEnder
Canceled
master
1ecf0fb6 fix(core): reject pending promises directly when plugin worker exits unexpectedly (#34588) When a plugin worker process exits unexpectedly, the exit handler previously sent synthetic `loadResult` messages to all pending response handlers. If any handler was waiting for a different result type (e.g. `createNodesResult`), the type validation would reject with a confusing "Expected createNodesResult, got loadResult" error instead of surfacing the actual cause. Split response handlers into `onMessage` / `onError` callbacks so the exit handler can reject each pending promise directly with a clear "Plugin worker exited unexpectedly" error. Also use unique transaction IDs for `load` messages (via `generateTxId`) to avoid potential handler overwrites during worker restarts. Fixes #34564
25 days ago by AgentEnder
Canceled
34332
d231c220 fix(core): handle string dependsOn entries
25 days ago by AgentEnder
Succeeded
34588
60e16181 fix(core): reject pending promises directly when plugin worker exits unexpectedly When a plugin worker process exits unexpectedly, the exit handler previously sent synthetic `loadResult` messages to all pending response handlers. If any handler was waiting for a different result type (e.g. `createNodesResult`), the type validation would reject with a confusing "Expected createNodesResult, got loadResult" error instead of surfacing the actual cause. Split response handlers into `onMessage` / `onError` callbacks so the exit handler can reject each pending promise directly with a clear "Plugin worker exited unexpectedly" error. Also use unique transaction IDs for `load` messages (via `generateTxId`) to avoid potential handler overwrites during worker restarts. Fixes #34564
26 days ago by AgentEnder
Failed
34582
274f22ee fix(core): skip writing deps cache if already up-to-date
26 days ago by AgentEnder
Canceled
34582
8af63d5b fix(core): skip writing deps cache if already up-to-date
26 days ago by AgentEnder
Succeeded
master
6d905722 fix(core): show the correct status for stopped continuous tasks (#34226) ## Current Behavior ### 1. Continuous tasks missing from `postTasksExecution` hook When running continuous tasks (e.g., `nx serve app`) and stopping them with Ctrl+C, the `postTasksExecution` lifecycle hook does not include them in `taskResults`. This breaks plugins that rely on post-run statistics (e.g., uploading task stats to DataDog). ### 2. Confusing TUI status when sibling continuous task exits When multiple continuous tasks run together and one exits unexpectedly, its sibling is marked as "failed" even though it was intentionally terminated/stopped by the task orchestrator. ## Expected Behavior 1. All tasks, including continuous ones, are included in `taskResults` for the `postTasksExecution` hook 2. Continuous tasks that are intentionally stopped (because dependent tasks completed or during graceful shutdown) report as `success` with `Stopped` display status 3. Continuous tasks that exit unexpectedly (crash) report as `failure` 4. TUI summary shows correct status: success when all tasks completed successfully, square icon for stopped tasks ## Related Issue(s) Fixes https://github.com/nrwl/nx/issues/33561 Supersedes: - https://github.com/nrwl/nx/pull/33562 - https://github.com/nrwl/nx/pull/34132
26 days ago by AgentEnder
Succeeded
34332
3382e74f chore(core): review feedback
26 days ago by AgentEnder