It seems some changes to the `unicode` standard package break some of our tests, and more importantly, some Unicode Line Break rules conformity tests. (Discovered in CI tests for #201)