unsatisfying feedback for wrong hello-world solution

[glennj]

I tested a wrong hello-world string and got "an error occurred" instead of "test failure".

Then I downloaded the exercises locally, and submitted a wrong solution. The online editor changed to:

Investigate why.

[glennj]

Via the undocumented Exercism API, we can get the metadata for the last test run.

{
  "test_run": {
    "uuid": "d946a94a-7b01-46a3-a99b-486db272693e",
    "submission_uuid": "05fb2bc5f9dd4384a08d5b5c6cd7d97b",
    "version": 3,
    "status": "fail",
    "message": null,
    "message_html": null,
    "output": null,
    "output_html": null,
    "tests": [
      {
        "name": "Say Hi!",
        "status": "fail",
        "test_code": "expected := \"Hello, World!\"\ntesting.expect_value(t, hello_world(), expected)",
        "message": "unknown",
        "message_html": "unknown",
        "expected": null,
        "output": null,
        "output_html": null,
        "task_id": null
      }
    ],
    "tasks": [],
    "highlightjs_language": "odin",
    "links": {
      "self": "https://exercism.org/api/v2/solutions/e6e598169c26471fa84588f1a2ff8a18/submissions/05fb2bc5f9dd4384a08d5b5c6cd7d97b/test_run"
    }
  },
  "test_runner": {
    "average_test_duration": 3,
    "status": {
      "exercise": true,
      "track": true
    }
  }
}

I'm surprised that the message for the first test is "unknown"

[glennj]

Found it: tracing the run.sh script, I see this

+ [[ Finished 1 test in 204µs. The test failed. == Finished * tests\. * test*failed\. ]]

Due to only one test, the regex fails.

I'll take this issue.