[openpgp] dropping the sample "ascii-armored message"

Daniel Kahn Gillmor <dkg@fifthhorseman.net> Tue, 30 May 2023 04:32 UTC

From: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
To: openpgp@ietf.org
Autocrypt: addr=dkg@fifthhorseman.net; prefer-encrypt=mutual; keydata= mDMEX+i03xYJKwYBBAHaRw8BAQdACA4xvL/xI5dHedcnkfViyq84doe8zFRid9jW7CC9XBiI0QQf FgoAgwWCX+i03wWJBZ+mAAMLCQcJEOCS6zpcoQ26RxQAAAAAAB4AIHNhbHRAbm90YXRpb25zLnNl cXVvaWEtcGdwLm9yZ/tr8E9NA10HvcAVlSxnox6z62KXCInWjZaiBIlgX6O5AxUKCAKbAQIeARYh BMKfigwB81402BaqXOCS6zpcoQ26AADZHQD/Zx9nc3N2kj13AUsKMr/7zekBtgfSIGB3hRCU74Su G44A/34Yp6IAkndewLxb1WdRSokycnaCVyrk0nb4imeAYyoPtBc8ZGtnQGZpZnRoaG9yc2VtYW4u bmV0PojRBBMWCgCDBYJf6LTfBYkFn6YAAwsJBwkQ4JLrOlyhDbpHFAAAAAAAHgAgc2FsdEBub3Rh dGlvbnMuc2VxdW9pYS1wZ3Aub3JnL0Gwxvypz2tu1IPG+yu1zPjkiZwpscsitwrVvzN3bbADFQoI ApsBAh4BFiEEwp+KDAHzXjTYFqpc4JLrOlyhDboAAPkXAP0Z29z7jW+YzLzPTQML4EQLMbkHOfU4 +s+ki81Czt0WqgD/SJ8RyrqDCtEP8+E4ZSR01ysKqh+MUAsTaJlzZjehiQ24MwRf6LTfFgkrBgEE AdpHDwEBB0DkKHOW2kmqfAK461+acQ49gc2Z6VoXMChRqobGP0ubb4kBiAQYFgoBOgWCX+i03wWJ BZ+mAAkQ4JLrOlyhDbpHFAAAAAAAHgAgc2FsdEBub3RhdGlvbnMuc2VxdW9pYS1wZ3Aub3Jnfvo+ nHoxDwaLaJD8XZuXiaqBNZtIGXIypF1udBBRoc0CmwICHgG+oAQZFgoAbwWCX+i03wkQPp1xc3He VlxHFAAAAAAAHgAgc2FsdEBub3RhdGlvbnMuc2VxdW9pYS1wZ3Aub3JnaheiqE7Pfi3Atb3GGTw+ jFcBGOaobgzEJrhEuFpXREEWIQQttUkcnfDcj0MoY88+nXFzcd5WXAAAvrsBAIJ5sBg8Udocv25N stN/zWOiYpnjjvOjVMLH4fV3pWE1AP9T6hzHz7hRnAA8d01vqoxOlQ3O6cb/kFYAjqx3oMXSBhYh BMKfigwB81402BaqXOCS6zpcoQ26AADX7gD/b83VObe14xrNP8xcltRrBZF5OE1rQSPkMNy+eWpk eCwA/1hxiS8ZxL5/elNjXiWuHXEvUGnRoVj745Vl48sZPVYMuDgEX+i03xIKKwYBBAGXVQEFAQEH QIGex1WZbH6xhUBve5mblScGYU+Y8QJOomXH+rr5tMsMAwEICYjJBBgWCgB7BYJf6LTfBYkFn6YA CRDgkus6XKENukcUAAAAAAAeACBzYWx0QG5vdGF0aW9ucy5zZXF1b2lhLXBncC5vcmcEAx9vTD3b J0SXkhvcRcCr6uIDJwic3KFKxkH1m4QW0QKbDAIeARYhBMKfigwB81402BaqXOCS6zpcoQ26AAAX mwD8CWmukxwskU82RZLMk5fm1wCgMB5z8dA50KLw3rgsCykBAKg1w/Y7XpBS3SlXEegIg1K1e6dR fRxL7Z37WZXoH8AH
Date: Tue, 30 May 2023 06:07:15 +0200
Message-ID: <874jnu5uoc.fsf@fifthhorseman.net>
MIME-Version: 1.0
Content-Type: multipart/signed; boundary="=-=-="; micalg="pgp-sha256"; protocol="application/pgp-signature"
Archived-At: <https://mailarchive.ietf.org/arch/msg/openpgp/kNnXbNHDCtR6GrQRkgFtzN08bKs>
Subject: [openpgp] dropping the sample "ascii-armored message"
Precedence: list

Yesterday i reviewed all of the test vectors currently in the draft, and
i'd like to propose dropping the "Example of an ASCII Armored Message".

That's this one:

-----BEGIN PGP MESSAGE-----

yDgBO22WxBHv7O8X7O/jygAEzol56iUKiXmV+XmpCtmpqQUKiQrFqclFqUDBovzS
vBSFjNSiVHsuAA==
-----END PGP MESSAGE-----


This sample message is a ZIP-compressed binary literal data packet that
claims its filename is "_CONSOLE", containing the ASCII text "Can't
anyone keep a secret around here?\n".

This has at least three problems:

 - ZIP isn't a high priority compression algorithm (it's not even
   "SHOULD implement", it's just "SHOULD decompress")

 - The contents of the literal data packet include a unix-style newline,
   but the literal datta packet is marked as binary.

 - The _CONSOLE filename (meaning "for eyes only") is explicitly
   deprecated.  Even if it weren't deprecated, the message is not
   encrypted, and it's deeply weird to say "here is a cleartext message
   for your eyes only".

It's also unclear in the OpenPGP ecosystem where such a packet
sequence would be useful.  For example, it's not encrypted, and it's
not signed.  It's not even well-compressed, since it's got
ASCII-armoring added back onto it.  The original string is shorter!

From the perspective of SOP, SOP doesn't even have a datatype or
subcommand that would be capable of consuming or generate this packet
sequence.  No one has asked for anything like it in SOP, and i can't
imagine a good argument for adding it.

If anything, this particular ASCII-armored message should exist in a
bestiary of "things the OpenPGP spec allows you to create but you really
probably shouldn't".  I don't think such a message belongs as a
highlighted example in the draft.

By contrast, the rest of the draft contains many useful ASCII-armored
OpenPGP test vectors, so i don't think we lose anything with just
cutting this one out.

I've opened a merge request for this non-substantive change:

   https://gitlab.com/openpgp-wg/rfc4880bis/-/merge_requests/315

      --dkg

Attachment: signature.asc

[openpgp] dropping the sample "ascii-armored mess… Daniel Kahn Gillmor

[openpgp] dropping the sample "ascii-armored message"

Attachment: signature.asc