Description
I noticed that some CI jobs currently use continue-on-error, which allows the workflow to succeed even when build or test steps fail.
This affects the reliability of CI as a validation gate because regressions can go unnoticed while the workflow still reports success.
From my investigation, this appears to affect at least:
Why this is a problem
If a build or unit test command fails but the job is marked successful, PR validation becomes unreliable.
That can lead to:
- broken changes being merged accidentally
- false-positive CI results
- harder debugging later because failures are hidden during review
Suggested fix
Remove continue-on-error from the build and unit test jobs/steps that are expected to act as required validation checks.
The workflows should fail normally when:
- build steps fail
- unit tests fail
Additional context
I verified that failing commands in these workflows do not currently fail the overall CI result because of the continue-on-error configuration.
I can work on a PR for this if maintainers agree with the approach.
Description
I noticed that some CI jobs currently use continue-on-error, which allows the workflow to succeed even when build or test steps fail.
This affects the reliability of CI as a validation gate because regressions can go unnoticed while the workflow still reports success.
From my investigation, this appears to affect at least:
Why this is a problem
If a build or unit test command fails but the job is marked successful, PR validation becomes unreliable.
That can lead to:
Suggested fix
Remove continue-on-error from the build and unit test jobs/steps that are expected to act as required validation checks.
The workflows should fail normally when:
Additional context
I verified that failing commands in these workflows do not currently fail the overall CI result because of the continue-on-error configuration.
I can work on a PR for this if maintainers agree with the approach.