Re: interesting creatures // ? for the open-pgp bestiary

[David Shaw <dshaw@xxxxxxxxxxxxxxx>]

On Mon, Jul 31, 2006 at 04:53:35PM -0400, vedaal@xxxxxxxx wrote:
> 
> in rfc 2440 bis-18, section 6.2 
> ASCII Armor is described, as well as the different types of header 
> lines
> 
> in commandline versions of pgp and gnupg,
> it is possible to just Armor a file,
> not armor sign, or armor encrypt, 
> just armor
> 
> this is a useful way for including any file type as part of an 
> inline message, and then signing and encrypting the entire message,
> thereby avoiding any attachments, and questions as to the safety 
> and authorship of the attachment,
> but also not leaving one's signature on the file itself
> 
> this type of pgp 'output' is not described anywhere in rfc 2440

It is a "Literal Message" and consists of a single literal data packet
(whether armored or not).  It is described in section 10.3.

> [1] from pgp classic,
> using the command: pgp -a filename

> [3] from gnupg,
> using the command: gpg -a --store filename

Both of these create literal messages.

> [2] from gnupg,
> using the command: gpg --enarmor filename

This does not create a literal message, and is not an OpenPGP feature
(and is thus not part of 2440 bis or otherwise).  It is mainly used as
a OpenPGP packet hacking tool and for the GPG selftest.

> one unusual feature in all three examples,
> is that the line of armored message block is less than 64 
> characters.
> 
> is there a minimum length to a line of pgp armor?

There is no minimum.  The maximum is 76 (section 6.3):

    The encoded output stream must be represented in lines of no more
    than 76 characters each.

David