Create new job to make cancelled jobs as failure jobs in CI runner by ThreeMonth03 · Pull Request #780 · solvcon/modmesh

ThreeMonth03 · 2026-05-14T04:57:41Z

In issue #772, we notice that when a job times out, it would be cancelled instead of fail.
Therefore, this pull request is to create new job, that could monitor cancelled jobs, and make cancelled jobs as failure jobs in CI runner.

ThreeMonth03

@yungyuc Please review this pull request when you are available. Thanks.

ThreeMonth03 · 2026-05-14T04:59:26Z

+  send_email_on_failure:
+    needs: [nouse_install, fail_on_cancelled_jobs]
+    # Run if any of the dependencies failed or timed out in master branch
    if: ${{ always() && contains(needs.*.result, 'failure') && github.ref_name == 'master' && github.event.repository.fork == false }}
    uses: ./.github/workflows/send_email_on_fail.yml
    with:
      job_results: |
        - nouse_install: ${{ needs.nouse_install.result }}
+        - fail_on_cancelled_jobs: ${{ needs.fail_on_cancelled_jobs.result }}


When a job is cancelled, a notified email would be sent now.

ThreeMonth03 · 2026-05-14T05:06:32Z

-  send_email_on_failure:
+  fail_on_cancelled_jobs:
    needs: [nouse_install]
-    # Run if any of the dependencies failed in master branch
+    if: ${{ always() }}
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Fail on cancelled jobs
+        run: |
+          result="${{ needs.nouse_install.result }}"
+          echo "nouse_install: ${result}"
+          if [ "${result}" = "cancelled" ]; then
+            exit 1
+          fi


We cannot make a time-out job as a failure job easily. There are two workaround method.
The first opinion is to create a new job to monitor the cancelled jobs, and it would raise failure when there is a cancelled job.
The second opinion is to time the process by a timer. If the process times out, the timer would return 1.

I prefer the first opinion, so I create a new job to monitor the cancelled jobs.

This would work, but is not as elegant as I hoped for.

If there is not a better way, I would take it. But I hope there is.

yungyuc

@ExplorerRay @chestercheng Could you please help brainstorm?

yungyuc · 2026-05-14T14:32:24Z

+  fail_on_cancelled_jobs:
    needs: [standalone_buffer, build, build_windows]
-    # Run if any of the dependencies failed in master branch
+    if: ${{ always() }}


Is it necessary to write the if, if it always() happens?

Yes, it is. It's an interesting technique. See this reference.

If a job fails or is skipped, all jobs that need it are skipped unless the jobs use a conditional expression that causes the job to continue. If a run contains a series of jobs that need each other, a failure or skip applies to all jobs in the dependency chain from the point of failure or skip onwards. If you would like a job to run even if a job it is dependent on did not succeed, use the always() conditional expression in jobs.<job_id>.if.

In other words, if we remove if: ${{ always() }}, this job would be skipped when the jobs in needs are cancelled or fail. However, we don't want to skip this job at any time, so we need to add if: ${{ always() }}.

This is cool. Could you please add a code comment above the line for the link to the reference? It look so innocent and future maintainers will thank for the point.

No problem, I've added the code comment to make the statement clear.

yungyuc · 2026-05-14T14:34:01Z

-  send_email_on_failure:
+  fail_on_cancelled_jobs:
    needs: [nouse_install]
-    # Run if any of the dependencies failed in master branch
+    if: ${{ always() }}
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Fail on cancelled jobs
+        run: |
+          result="${{ needs.nouse_install.result }}"
+          echo "nouse_install: ${result}"
+          if [ "${result}" = "cancelled" ]; then
+            exit 1
+          fi


This would work, but is not as elegant as I hoped for.

If there is not a better way, I would take it. But I hope there is.

ExplorerRay · 2026-05-16T06:03:36Z

I think this PR can use https://github.com/marketplace/actions/run-and-post-run-action to solve this so that we don't need to handle a new job in CI.

I am considering adding a step like the following for each job that might timeout. And we may differentiate return code for failure and timeout (I write 2 in the following example).

- name: Failure when job timeout cancelled
  uses: gacts/run-and-post-run@v1
  with:
    post: |
      if [ "${{ job.status }}" == "cancelled" ]; then
        echo "::error::CI job timed out! Forcing failure."
        exit 2
      fi

ExplorerRay

I think it's better to adding post check step into setup dependency composite action.

If there isn't a composite action to set up dependency like Windows, add an extra step like what you original did.

ExplorerRay · 2026-05-16T06:34:36Z

      job_results: |
        - standalone_buffer: ${{ needs.standalone_buffer.result }}
        - build: ${{ needs.build.result }}
+        - build_windows: ${{ needs.build_windows.result }}


Nice catch. I forget this part.

ExplorerRay · 2026-05-16T06:38:26Z


    steps:

+      - name: Fail if job was cancelled


I think these lines can be moved into the beginning of setup dependency composite action so that we can set this when running that action.

ThreeMonth03 commented May 14, 2026

View reviewed changes

yungyuc assigned ThreeMonth03 May 14, 2026

yungyuc added the build Build system and automation label May 14, 2026

yungyuc reviewed May 14, 2026

View reviewed changes

yungyuc requested review from ExplorerRay and chestercheng May 14, 2026 14:34

ThreeMonth03 force-pushed the issue772 branch 2 times, most recently from 26947d3 to d82bb1d Compare May 15, 2026 07:20

ThreeMonth03 force-pushed the issue772 branch 2 times, most recently from e2a463d to 187499c Compare May 16, 2026 06:25

ExplorerRay suggested changes May 16, 2026

View reviewed changes

ThreeMonth03 force-pushed the issue772 branch 3 times, most recently from ef8a471 to 3256453 Compare May 16, 2026 07:19

ExplorerRay mentioned this pull request May 16, 2026

Timed-out CI job should be marked as failed #772

Open

Create new job to make cancelled jobs as failure jobs in CI runner

a1f9435

ThreeMonth03 force-pushed the issue772 branch from 3256453 to a1f9435 Compare May 18, 2026 05:40

Conversation

ThreeMonth03 commented May 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ThreeMonth03 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

yungyuc left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ThreeMonth03 May 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ExplorerRay commented May 16, 2026

Uh oh!

ExplorerRay left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ThreeMonth03 commented May 14, 2026 •

edited

Loading

ThreeMonth03 May 14, 2026 •

edited

Loading