Why is the v5 key prefixed with non-armored data?

Hey! I noticed that the v5 "emma" test key is not a pure ASCII armored artifact, but instead has some (GnuPG-specific?) textual metadata prepended. As a result, v5-compatibility is overshadowed by potential armor parser errors.

Imho, the prefixed text data should be removed from the test key, as the capability to process armored data with prefixed text is already tested elsewhere (Concatenated ASCII Armor Keyring).