Description
See https://github.com/jupyterlab/jupyter-chat/actions/runs/30491550601/job/90710371003?pr=473
In #473 I added a workflow that confirms the v0.23.0 release fails on JupyterLab <4.6. However, when the tests fail, they take >1 hour to fail.
Digging a bit deeper, apparently we are re-running tests on failure. We should be doing the opposite - the E2E tests should be reliable and we should re-run on success to confirm they are not flaky.
We'll also want to improve the test fixtures to fail fast instead of always waiting 60 seconds for any element to appear.
Description
See https://github.com/jupyterlab/jupyter-chat/actions/runs/30491550601/job/90710371003?pr=473
In #473 I added a workflow that confirms the v0.23.0 release fails on JupyterLab <4.6. However, when the tests fail, they take >1 hour to fail.
Digging a bit deeper, apparently we are re-running tests on failure. We should be doing the opposite - the E2E tests should be reliable and we should re-run on success to confirm they are not flaky.
We'll also want to improve the test fixtures to fail fast instead of always waiting 60 seconds for any element to appear.