-- * Some assumption on messages were wrong. Now the tests use more reasonable sizes. * bigString now uses the full utf8-extent, with the exception of U+0000. This code gets dropped during the encryption-decryption process.