fix(deadcode): false positive on early return + multi-line dict return (closes #105)

[Wolfvin]

Closes #105

Root cause

When _detect_unreachable_code encountered a multi-line return { (open brace > close brace on the start line), it skipped the line via continue without resetting found_terminal. If a previous if x: return None had set the terminal flag, the multi-line return body (e.g. dict literals) was falsely flagged as unreachable code.

Example false positive (the exact pattern from scripts/commands/analyze.py before this fix):

def _detect_vulns(workspace, max_items):
    ...
    if total == 0:
        return None
    return {                       # ← multiline return start, was skipped
        "category": "vulnerabilities",  # ← falsely flagged as unreachable
        "total": total,
    }

Fix

In the multi-line return skip path (Python branch), check the current line's indent vs the previous terminal's indent. If the current return is in an outer scope (lower indent), the previous terminal is no longer relevant — reset found_terminal before continuing.

Also reverts the antipattern if x: return None / else: return {...} introduced in PR #96 across 8 functions in scripts/commands/analyze.py back to the PEP 8-friendly if x: return None / return {...} form. The antipattern was only there to satisfy the buggy scanner; with the fix the PEP 8 form scans cleanly.

Tests

5 new regression tests in tests/test_deadcode_engine.py:

• test_issue_105_early_return_then_final_return — simple if x: return None / return y
• test_issue_105_multiline_return_after_early_return — multi-line dict return after early return (exact repro from issue)
• test_issue_105_chained_early_returns — multiple if guards + final return
• test_issue_105_nested_if_early_return — nested if/return + outer returns
• test_issue_105_genuinely_unreachable_still_detected — sanity check that genuinely unreachable code is still reported

All 14 deadcode tests pass (9 existing + 5 new). No regressions in the broader test suite — verified zero false positives when scanning 8 major CodeLens Python files (analyze.py, deadcode_engine.py, codelens.py, scan.py, query.py, secrets_engine.py, ast_taint_engine.py, vulnscan_engine.py).

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud