Fix Expat license notice detection

[nikitaijari]

Fixes #4971

Summary

The Expat license notice in the pep8-naming LICENSE file is detected as both mit and unknown-license-reference.
The problematic notice is:
Licensed under the terms of the Expat License

The issue was originally reported with ScanCode 32.5.0 installed through pip on Linux. I reproduced the reported behavior using ScanCode 32.5.0 in Windows.

Investigation

The root cause is that license-intro_20.RULE matches the generic introduction Licensed under the terms of but does not identify the Expat license itself. I also looked at how these matches are handled during license detection. License-introduction matches can be filtered when calculating the final detected license expression but the original match can still be reported in reference_matches. Therefore, changing only the final license-expression calculation would not remove the reported unknown-license-reference.

Changes

• Added mit_expat.RULE to specifically recognize:
  Licensed under the terms of the Expat License
• Added a regression test using the Expat license notice and the corresponding expected output
• Updated CHANGELOG.rst
• Added myself to AUTHORS.rst

With the new rule, the Expat notice is recognized as mit instead of being reported through the generic unknown-license-reference rule.

Testing

Targeted regression test:

pytest -q tests/licensedcode/test_plugin_license_detection.py -k "unknown_license_intro_expat"

1 passed, 20 deselected

Licensedcode test suite:

pytest -q tests/licensedcode

474 passed, 2 failed

The two failures are unrelated to this change:

• test_match_has_correct_line_positions_in_automake_perl_file
• test_dump_license - failed to load w3c-docs-19990405 from the local Windows AppData license data.

Full test suite:

pytest

3838 passed, 91 failed, 42 skipped, 4 xfailed

The full-suite failures occurred in unrelated areas of the test suite. The targeted regression test passes.

Note: I'm new to open-source contribution so please give me feedback if there is a better way to approach this, structure the regression test or handle any part of the contribution process.

Tasks

• Reviewed contribution guidelines
• PR is descriptively titled and links the original issue
• Tests pass - full local test suite has unrelated failures; CI checks are pending
• Commit is in a uniquely-named feature branch with no merge conflicts
• Updated documentation pages (not applicable)
• Updated CHANGELOG.rst