Re: Gnus 5.10.6 problems with PGP/MIME (test cases)

["Mark D. Baushke" <mdb@gnu.org>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Daiki Ueno <ueno@unixuser.org> writes:

> Thanks Mark for summarizing the discussion.

You are welcome.

> >>>>> In <b4mek2fa9wu.fsf@jpl.org> 
> >>>>>	Katsumi Yamaoka <yamaoka@jpl.org> wrote:
> > `pgg-gpg-encrypt-region' always inserts CR in every line ending
> > of text using `(pgg-as-lbt START END 'CRLF BODY)'.  So, if the
> > --textmode option is not specified, gpg doesn't treat sending
> > data as text and recipients will see ^Ms in a decrypted message.
> 
> > In addition, do you have an idea to solve the ^Ms problem that I
> > mentioned above?
> 
> Is it not enough that we simply omit pgg-as-lbt from
> pgg-gpg-encrypt-region?
> 
> For MIME encryption, it is MUA (not PGP libraries) that should be
> responsible for converting LF to CRLF in encrypted messages since MIME
> encryption is only applicable to messages in MIME canonical format.

Could you please provide the RFC reference for this?

I see in RFC 2015 where it talks about CRLF being canonical for
signatures validation, but it also says that the MUA needs to convert
line endings to the canonical <CR><LF> sequence before the signature can
be verified. This is necessary since the local MTA may ahe converted to
a local end of line convention.

However, I do NOT see it specify that arbitrary data which is being
encrypted should have conversions forced on it at all.

If it is text, then there is a packet type specified in RFC 2440
which specifies the <CR><LF> as the canonical format and a marker that
the packet is text (this is what the --textmode switch to gpg does).

> For non-MIME encryption, line-ending conversion is not needed at all.

I disagree. This is what started the debate in the first place.

If I am sending a text message, then changing the line endings to CRLF
as is done in pgg-gpg-encrypt-region should also tell the remote end
that a text packet is coming rather than arbitrary binary data.

This is also true if I am on a CRLF line ending system and sending an
encrypted text message to a LF line ending system.

I contend that the Gnus system should know if the user is encrypted an
attachment or normal text and tell the underlying transport to do the
right thing so that all permutations of line endings (CR, CRLF, LF,
line-oriented byte count) on each end of the system will know what to
do with the encrypted data.

	Thanks,
	-- Mark
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQFD6Gr0Cg7APGsDnFERAq3cAKCORDhuF1d3IfXA9lvQJt3GK4BjQQCguS/B
bVtmekIA4N7BS8dTh4c92TI=
=v/pF
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642