Comment 0 for bug 1491926

In bug 1433372, we determined that we should percent encode Fernet tokens, because the padding characters (=) aren't considered URL safe by some RFCs.

We also fail some tempest tests because clients sometimes decode or encode responses [0]. We should just remove the padding, that way clients don't have to worry about it. When we go to validate a token, we can determine what the padding is based on the length of the token:

missing_padding = 4 - len(token) % 4
if missing_padding:
    token += b'=' * missing_padding

[0] http://cdn.pasteraw.com/es3j52dpfgem4nom62e7vktk7g5u2j1