src/scanoss/inspection/copyleft.py (1)

210-213: SUCCESS / FAIL logic reversed

run() returns FAIL when no copyleft components are found:

if len(copyleft_components) <= 0:
    return PolicyStatus.FAIL.value, results

Typically the check should succeed (SUCCESS) when no violations exist and fail only when copyleft components are present.

-        if len(copyleft_components) <= 0:
-            return PolicyStatus.FAIL.value, results
-        return PolicyStatus.SUCCESS.value, results
+        if len(copyleft_components) > 0:
+            return PolicyStatus.FAIL.value, results
+        return PolicyStatus.SUCCESS.value, results

CHANGELOG.md (1)

12-17: Changelog entry OK – minor nit

Consider adding a short “### Security” (even if empty) for consistency with older releases.

src/scanoss/inspection/copyleft.py (1)

64-68: Avoid shadowing the built-in format function

self.format = format overwrites Python’s built-in format.
Rename the attribute (e.g. format_type) to prevent confusion and future bugs.

src/scanoss/inspection/undeclared_component.py (1)

64-66: Same built-in shadowing as in copyleft module

self.format = format masks the built-in format. Rename to format_type for clarity.

src/scanoss/inspection/policy_check.py (1)

441-480: License-priority helper – minor efficiency / style notes

1. Converting licenses_by_source to dict[str, dict] then back to list duplicates memory; iterate once and return when first priority hit to save work.
2. Missing space after comma in the signature: def _get_licenses_order_by_source_priority(self, licenses_data):

Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro

tests/test_policy_inspect.py (1)

257-262: Duplicate check – same fragility comment as above

See the previous note: relying on exact whitespace makes the test noisy for benign formatting tweaks. A small helper that normalises markdown rows would keep maintenance cost low.

tests/test_policy_inspect.py (4)

117-118: Hard-coding the new expected size may hide future regressions

The update from 3 → 2 components appears correct given the refactor, but the assertion is still a magic number.
Consider asserting against the length of results['summary'].split() or explicitly checking the purls you expect to survive the new prioritisation – this future-proofs the test if another priority tweak changes the count again.

---

150-155: Brittle comparison due to literal newline character

expected_summary_output ends with a trailing \n. On Windows or when refactoring the formatter this may flip to \r\n, causing a false-negative.
Prefer stripping line endings before the equality check:

-expected_summary_output = '2 component(s) with copyleft licenses were found.\n'
+expected_summary_output = '2 component(s) with copyleft licenses were found.'
 ...
-self.assertEqual(results['summary'], expected_summary_output)
+self.assertEqual(results['summary'].strip(), expected_summary_output)

---

214-219: String-equality on Markdown table rows is fragile

The test will fail if column order or spacing changes even though the semantic information is identical.
Instead of literal equality, parse the markdown (or simply split the line on |) and assert on the tuple (purl, version, license) for each row.

---

333-337: Jira-MD expectations suffer from the same brittleness

Literal pipe-separated strings are prone to break when the renderer changes alignment. Consider tokenising the line and asserting on the fields, or at least normalising whitespace.

Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro