Make the output of each linter available to users for further processing

[ferrarimarco]

Is there an existing issue for this?

• I have searched the existing issues

Current Behavior

Super-linter generates a text log file with its output and the output (stdout and stderr) of each linter and code formatter it runs. This log file might be too fragile to parse for further processing.

After we switched to GNU parallel to run tasks, we are already producing and parsing the JSON files it generates to get stdout, stderr, and the exit code of each linter. These JSON files also include some more data points, such as the runtime of each process. These JSON files aren't currently available for users to process because super-linter stores them in a temporary location inside the container.

Expected Behavior

Make output files available for users to parse. By doing this, users can also benefit from whatever output format they can configure for each linter, without super-linter getting in their way.

Anything else?

The lowest hanging fruit would be to just expose the JSON files that GNU parallel produces, although we might want to implement some basic validation to ensure that updates to GNU parallel don't introduce breaking changes to the format of these JSON files.

/cc @ChaosEternal

[jhonkan]

Would be great if a user can configure and output the reports files, so they could be uploaded to Sonar:

• https://docs.sonarsource.com/sonarqube/latest/analyzing-source-code/importing-external-issues/external-analyzer-reports/
• https://docs.sonarsource.com/sonarqube/latest/analyzing-source-code/importing-external-issues/importing-issues-from-sarif-reports/

[ferrarimarco]

Hi @jhonkan !

If the linter supports configuring the output format using its configuration file, that would be trivial to do, once we expose output files.