Because multiple checks like linting, formatting, and unit tests are bundled together into one GitHub action job, it's not immediately clear which tests are failing on PRs. We are letting ESlint errors through some PRs as we slowly improve our code quality, but it seems like we might be inadvertently allowing some test failures to pass unnoticed because of how errors are reported. Whether one or all four fail, it just shows as an overall failure in the PR.
This issue entails making a clearer distinction between types of failures in PRs. Some ideas:
Showing ESlint errors as warnings rather than errors?
Running an individual job for each of the 4 types of tests
Auto commenting on PRs with a summary of the test outputs
Because multiple checks like linting, formatting, and unit tests are bundled together into one GitHub action job, it's not immediately clear which tests are failing on PRs. We are letting ESlint errors through some PRs as we slowly improve our code quality, but it seems like we might be inadvertently allowing some test failures to pass unnoticed because of how errors are reported. Whether one or all four fail, it just shows as an overall failure in the PR.
This issue entails making a clearer distinction between types of failures in PRs. Some ideas:
Related to: #2425