Encrypting outgoing message but not Gcc?

Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

Is there a way to configure Gnus to encrypt an outgoing message, but not
its Gcc?  I Gcc messages to an archive group, and unless I have the
recipient's private key, I can't view the archived message if it's been
encrypted.

I've looked around for an answer for this, but I'm sure that I'm
missing something obvious.

Thanks in advance.

-- 
 Lloyd Zusman
 ljz@asfast.com

Re: Encrypting outgoing message but not Gcc?

[Adam Sjøgren]

On Sat, 06 Dec 2003 11:17:17 -0500, Lloyd wrote:

> Is there a way to configure Gnus to encrypt an outgoing message, but
> not its Gcc?  I Gcc messages to an archive group, and unless I have
> the recipient's private key, I can't view the archived message if
> it's been encrypted.

My archived copies are encrypted with the recipients as well as my own
public-key, so I can read them myself.

I don't think I've done anything special to make Gnus do that.

Can't search through them, though, I guess.


  Best regards,

-- 
 "Do not feed the oysters under the clouds"                   Adam Sjøgren
                                                         asjo@koldfront.dk

Re: Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

spamtrap@koldfront.dk (Adam Sjøgren) writes:

> On Sat, 06 Dec 2003 11:17:17 -0500, Lloyd wrote:
>
>> Is there a way to configure Gnus to encrypt an outgoing message, but
>> not its Gcc?  I Gcc messages to an archive group, and unless I have
>> the recipient's private key, I can't view the archived message if
>> it's been encrypted.
>
> My archived copies are encrypted with the recipients as well as my own
> public-key, so I can read them myself.
>
> I don't think I've done anything special to make Gnus do that.

Hmm ... as I mentioned in my other message a few minutes ago, I seem to
need the _recipient's_ passphrase in order to decrypt my archived
copies, even with the latest CVS of Gnus.

Also, I think there's a gap somewhere in my understanding of all this.
How can it be that something encrypted twice, once with the recipient's
public key and once with yours, would only need your private key to be
decrypted?  Wouldn't there have to be two decryptions, one for each of
the previous encryptions?  Or in this case, is the encryption done with
some sort of mathematical combination of both public keys, such that
either corresponding private key can decrypt it?


> Can't search through them, though, I guess.

That's not such a problem for me.  I just want to be able to _read_
these archived copies.


-- 
 Lloyd Zusman
 ljz@asfast.com

Re: Encrypting outgoing message but not Gcc?

[Adam Sjøgren]

On Sat, 06 Dec 2003 13:35:21 -0500, Lloyd wrote:

>> My archived copies are encrypted with the recipients as well as my
>> own public-key, so I can read them myself.

>> I don't think I've done anything special to make Gnus do that.

> Hmm ... as I mentioned in my other message a few minutes ago, I seem
> to need the _recipient's_ passphrase in order to decrypt my archived
> copies, even with the latest CVS of Gnus.

Ah, now I remember - sorry, I've been misinforming about the "done
anything special part"; actually I've put:

 encrypt-to asjo@koldfront.dk

in my ~/.gnupg/options.

> Also, I think there's a gap somewhere in my understanding of all
> this.  How can it be that something encrypted twice, once with the
> recipient's public key and once with yours, would only need your
> private key to be decrypted?  Wouldn't there have to be two
> decryptions, one for each of the previous encryptions?  Or in this
> case, is the encryption done with some sort of mathematical
> combination of both public keys, such that either corresponding
> private key can decrypt it?

I'm not able to adequately and comprehensibly explain the theory, but
I think it's a proporty of either assymetric key cryptography in
general or the way PGP/GnuPG is constructed (the latter being less
likely, probably, perhaps).

,----[ From man gpg ]
| --encrypt-to name
|           Same as --recipient but this one is intended for use in the
|           options file and may be used with your own user-id as an
|           "encrypt-to-self".  These keys are only used when there are
|           other recipients given either by use of --recipient or by
|           the asked user id.  No trust checking is performed for these
|           user ids and even disabled keys can be used.
`----


  Best regards,

-- 
 "Do not feed the oysters under the clouds"                   Adam Sjøgren
                                                         asjo@koldfront.dk

Re: Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

spamtrap@koldfront.dk (Adam Sjøgren) writes:

> [ ... ]
>
> Ah, now I remember - sorry, I've been misinforming about the "done
> anything special part"; actually I've put:
>
>  encrypt-to asjo@koldfront.dk
>
> in my ~/.gnupg/options.

Aha! ... but then, how do I tell Gnus to NOT use that parameter to
encrypt the actual message that is being sent to the recipient (since I
want to use _her_ public key for that), but to indeed use it when
encrypting the message for my local Gcc?


>> Also, I think there's a gap somewhere in my understanding of all
>> this.  How can it be that something encrypted twice, once with the
>> recipient's public key and once with yours, would only need your
>> private key to be decrypted?  [ ... ]
>
> I'm not able to adequately and comprehensibly explain the theory, but
> I think it's a proporty of either assymetric key cryptography in
> general or the way PGP/GnuPG is constructed (the latter being less
> likely, probably, perhaps).

I think, perhaps, I understand it, if I'm following you, maybe. :)

Thank you.


> ,----[ From man gpg ]
> | [ ... etc. ... ]
> `----
>
>
>   Best regards,

-- 
 Lloyd Zusman
 ljz@asfast.com

Re: Encrypting outgoing message but not Gcc?

[Adam Sjøgren]

On Sat, 06 Dec 2003 21:09:27 -0500, Lloyd wrote:

>> encrypt-to asjo@koldfront.dk

> Aha! ... but then, how do I tell Gnus to NOT use that parameter to
> encrypt the actual message that is being sent to the recipient
> (since I want to use _her_ public key for that), but to indeed use
> it when encrypting the message for my local Gcc?

With the above configuration-option for GnuPG the message will be
encrypted for both of you (i.e. so both your and her private key can
be used to read it); so the email sent and the copy archived are both
readable by you as well as and the recipient.

That also makes sense, right?


  Best regards,

-- 
 "Do not feed the oysters under the clouds"                   Adam Sjøgren
                                                         asjo@koldfront.dk

Re: Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

spamtrap@koldfront.dk (Adam Sjøgren) writes:

> On Sat, 06 Dec 2003 21:09:27 -0500, Lloyd wrote:
>
>>> encrypt-to asjo@koldfront.dk
>
>> Aha! ... but then, how do I tell Gnus to NOT use that parameter to
>> encrypt the actual message that is being sent to the recipient
>> (since I want to use _her_ public key for that), but to indeed use
>> it when encrypting the message for my local Gcc?
>
> With the above configuration-option for GnuPG the message will be
> encrypted for both of you (i.e. so both your and her private key can
> be used to read it); so the email sent and the copy archived are both
> readable by you as well as and the recipient.
>
> That also makes sense, right?

Yes, now I _finally_ get it.  This is the solution I've been looking
for.

Thank you again ... especially for your patience.


-- 
 Lloyd Zusman
 ljz@asfast.com

Re: Encrypting outgoing message but not Gcc?

[Nathan Sullivan]

spamtrap@koldfront.dk (Adam Sjøgren) writes:

> On Sat, 06 Dec 2003 13:35:21 -0500, Lloyd wrote:
>
>> Also, I think there's a gap somewhere in my understanding of all
>> this.  How can it be that something encrypted twice, once with the
>> recipient's public key and once with yours, would only need your
>> private key to be decrypted?  Wouldn't there have to be two
>> decryptions, one for each of the previous encryptions?  Or in this
>> case, is the encryption done with some sort of mathematical
>> combination of both public keys, such that either corresponding
>> private key can decrypt it?
>
> I'm not able to adequately and comprehensibly explain the theory, but
> I think it's a proporty of either assymetric key cryptography in
> general or the way PGP/GnuPG is constructed (the latter being less
> likely, probably, perhaps).

It's a property of the way almost all assymetric cryptography is
implemented.  The problem is that public key encryption is grotesquely
slow, so programs like PGP/GPG don't encrypt the entire message with
the public key.  Instead, they encrypt the message with a symmetric
cypher, like IDEA or Blowfish, then encrypt the symmetric key with the
public key for the desired recipient.  If you have multiple
recipients, then multiple copies of the encrypted symmetric key are
included, each encrypted for a specific public key pair.

-- 
	    Nathan Sullivan                     alfonso@pants.nu
-------------------------------------------------------------------------------
          Conscience is what hurts when everthing else feels so good.

Re: Encrypting outgoing message but not Gcc?

[Adam Sjøgren]

On Sun, 07 Dec 2003 03:07:03 -0800, Nathan wrote:

> It's a property of the way almost all assymetric cryptography is
> implemented.  The problem is that public key encryption is
> grotesquely slow, so programs like PGP/GPG don't encrypt the entire
> message with the public key.  Instead, they encrypt the message with
> a symmetric cypher, like IDEA or Blowfish, then encrypt the
> symmetric key with the public key for the desired recipient.  If you
> have multiple recipients, then multiple copies of the encrypted
> symmetric key are included, each encrypted for a specific public key
> pair.

Ah, the first part I'd heard before (but blissfully forgotten), the
second was new to me. Thanks for the clarification!


  Best regards,

-- 
 "Do not feed the oysters under the clouds"                   Adam Sjøgren
                                                         asjo@koldfront.dk

Re: Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

Nathan Sullivan <alfonso@pants.nu> writes:

> spamtrap@koldfront.dk (Adam Sjøgren) writes:
>
>> On Sat, 06 Dec 2003 13:35:21 -0500, Lloyd wrote:
>>
>>> Also, I think there's a gap somewhere in my understanding of all
>>> this.  How can it be that something encrypted twice, once with the
>>> recipient's public key and once with yours, would only need your
>>> private key to be decrypted?  [ ... ]
>>
>> I'm not able to adequately and comprehensibly explain the theory, but
>> I think it's a proporty of either assymetric key cryptography in
>> general or the way PGP/GnuPG is constructed (the latter being less
>> likely, probably, perhaps).
>
> It's a property of the way almost all assymetric cryptography is
> implemented.  The problem is that public key encryption is grotesquely
> slow, so programs like PGP/GPG don't encrypt the entire message with
> the public key.  Instead, they encrypt the message with a symmetric
> cypher, like IDEA or Blowfish, then encrypt the symmetric key with the
> public key for the desired recipient.  If you have multiple
> recipients, then multiple copies of the encrypted symmetric key are
> included, each encrypted for a specific public key pair.

Thanks ... I had always wondered what symmetric cyphers like IDEA or
Blowfish had to do with public key encryption.

-- 
 Lloyd Zusman
 ljz@asfast.com

Re: Encrypting outgoing message but not Gcc?

[Reiner Steib]

On Sat, Dec 06 2003, Lloyd Zusman wrote:

> I Gcc messages to an archive group, and unless I have the
> recipient's private key, I can't view the archived message if it's
> been encrypted.

,----[ C-h v pgg-encrypt-for-me RET ]
| pgg-encrypt-for-me's value is t
| 
| Documentation:
| If t, encrypt all outgoing messages with user's public key.
`----

The default value has bee changed to `t' recently (2003-11-10).

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo--- PGP key available via WWW   http://rsteib.home.pages.de/

Re: Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

Reiner Steib <4.uce.03.r.s@nurfuerspam.de> writes:

> On Sat, Dec 06 2003, Lloyd Zusman wrote:
>
>> I Gcc messages to an archive group, and unless I have the
>> recipient's private key, I can't view the archived message if it's
>> been encrypted.
>
> ,----[ C-h v pgg-encrypt-for-me RET ]
> | pgg-encrypt-for-me's value is t
> | 
> | Documentation:
> | If t, encrypt all outgoing messages with user's public key.
> `----
>
> The default value has bee changed to `t' recently (2003-11-10).
>
> Bye, Reiner.

Thank you.  My last previous update was before 11/10.  I just now got
the latest gnus from CVS, and I see this setting.

This helps, but it isn't quite what I want, either.

Here's a typical situation of mine:

I have a hypothetical user's public key in my key ring.  Suppose this
user's address is `hypo@thetical.us'.

I want to send an encrypted message to that user, and I want a copy of
that message to be archived locally.  I use the following to do so:

  Gcc: nnml+archive:email-archive

I write the message to hypo@thetical.us and I enter `C-c C-m c o'
(mml-secure-message-encrypt-pgp) to flag it to be digitally encrypted
via GPG.

I then invoke `C-c C-c' (message-send-and-exit).

I get prompted for the passphrase, and the message gets sent and
archived (with a second passphrase prompt).

Here's what I want to occur:

- The message goes to hypo@thetical.us encrypted by her public key.

- The message goes into nnml+archive:email-archive, either not encrypted
  at all (my preferred choice), or at worst case, encrypted with _my_
  public key.

I just tested this using today's latest CVS, and the outgoing message
_and_ the Gcc are both encrypted with the the _recipient's_ public key
(I need the recipient password to view the message in my nnml archive).

But even if I correct this, my preferred outcome is that the Gcc does
not get encrypted at all.  Is there any way to do this?

Thanks again, in advance.


-- 
 Lloyd Zusman
 ljz@asfast.com

Re: Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

Reiner Steib <4.uce.03.r.s@nurfuerspam.de> writes:

> On Sat, Dec 06 2003, Lloyd Zusman wrote:
>
>> I Gcc messages to an archive group, and unless I have the
>> recipient's private key, I can't view the archived message if it's
>> been encrypted.
>
> ,----[ C-h v pgg-encrypt-for-me RET ]
> | pgg-encrypt-for-me's value is t
> | 
> | Documentation:
> | If t, encrypt all outgoing messages with user's public key.
> `----
>
> The default value has bee changed to `t' recently (2003-11-10).
>
> Bye, Reiner.

I was wondering why this option didn't work for me, and I found the code
below in yesterday's CVS of pgg-gpg.el.  Notice that it makes my user ID
known to gpg by means of the `--remote-user' option.  However, this
option is no longer used in recent versions of gpg ... it's been
replaced with `--recipient'.

I think that we need a configuration variable that allows us to specify
one or the other of these option flags, so that the code can be made to
work with both older and newer versions of gpg.

  (defun pgg-gpg-encrypt-region (start end recipients &optional sign)
    "Encrypt the current region between START and END.
  If optional argument SIGN is non-nil, do a combined sign and encrypt."
    (let* ((pgg-gpg-user-id (or pgg-gpg-user-id pgg-default-user-id))
           (passphrase
            (when sign
              (pgg-read-passphrase
               (format "GnuPG passphrase for %s: " pgg-gpg-user-id)
               pgg-gpg-user-id)))
           (args
            (append
             (list "--batch" "--armor" "--always-trust" "--encrypt")
             (if sign (list "--sign" "--local-user" pgg-gpg-user-id))
             (if recipients
                 (apply #'nconc
                        (mapcar (lambda (rcpt)
                                  (list "--remote-user" rcpt))
                                (append recipients
                                        (if pgg-encrypt-for-me
                                            (list pgg-gpg-user-id)))))))))
      (pgg-as-lbt start end 'CRLF
        (pgg-gpg-process-region start end passphrase pgg-gpg-program args))
      (when sign
        (with-current-buffer pgg-errors-buffer
          ;; Possibly cache passphrase under, e.g. "jas", for future sign.
          (pgg-gpg-possibly-cache-passphrase passphrase pgg-gpg-user-id)
          ;; Possibly cache passphrase under, e.g. B565716F, for future decrypt.
          (pgg-gpg-possibly-cache-passphrase passphrase)))
      (pgg-process-when-success)))


-- 
 Lloyd Zusman
 ljz@asfast.com

Re: Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

[ NOTE:  this is a repost which has been slightly modified
         from the original post, in an attempt to get past
         the "partial message checksum" message I got from
         the moderator-bot. ]

Reiner Steib <4.uce.03.r.s@nurfuerspam.de> writes:

> [ ... ]
>
> ,----[ C-h v pgg-encrypt-for-me RET ]
> | pgg-encrypt-for-me's value is t
> | 
> | Documentation:
> | If t, encrypt all outgoing messages with user's public key.
> `----
>
> The default value has bee changed to `t' recently (2003-11-10).

I was wondering why this option didn't work for me, and I found the code
below in yesterday's CVS.  Notice that it makes my user ID known to gpg
by means of the `--remote-user' option.  However, this option is no
longer used in recent versions of gpg ... it's been replaced with
`--recipient'.

I think that we need a configuration variable that allows us to specify
one or the other of these option flags, so that the code can be made to
work with both older and newer versions of gpg.


From pgg-gpg.el ...

  (defun pgg-gpg-encrypt-region (start end recipients &optional sign)
    "Encrypt the current region between START and END.
  ;;; ... etc. ...
             (if recipients
                 (apply #'nconc
                        (mapcar (lambda (rcpt)
                                  (list "--remote-user" rcpt))
                                (append recipients
                                        (if pgg-encrypt-for-me
                                            (list pgg-gpg-user-id)))))))))
  ;;; ... etc. ...



-- 
 Lloyd Zusman
 ljz@asfast.com

Re: Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

[-- Attachment #1: Type: text/plain, Size: 666 bytes --]

Lloyd Zusman <ljz@asfast.com> writes:

> [ ... ]
>
> I was wondering why this option didn't work for me, and I found the code
> below in yesterday's CVS.  Notice that it makes my user ID known to gpg
> by means of the `--remote-user' option.  However, this option is no
> longer used in recent versions of gpg ... it's been replaced with
> `--recipient'.
>
> I think that we need a configuration variable that allows us to specify
> one or the other of these option flags, so that the code can be made to
> work with both older and newer versions of gpg.

Here's a patch for this.  What do you folks think?  If it's suitable,
feel free to apply it to CVS.

Thanks.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: recipient/remote-user option choice --]
[-- Type: text/x-patch, Size: 1025 bytes --]

*** pgg-gpg.el.orig	Sun Dec  7 09:09:01 2003
--- pgg-gpg.el	Sun Dec  7 10:11:05 2003
***************
*** 43,48 ****
--- 43,54 ----
    :group 'pgg-gpg
    :type '(repeat (string :tag "Argument")))
  
+ (defcustom pgg-gpg-recipient-arg "--recipient"
+   "Recipient option for gpg."
+   :group 'pgg-gpg
+   :type '(choice (const :tag "New `--recipient' option" "--recipient")
+                  (const :tag "Old `--remote-user' option" "--remote-user")))
+ 
  (defvar pgg-gpg-user-id nil
    "GnuPG ID of your default identity.")
  
***************
*** 154,160 ****
  	   (if recipients
  	       (apply #'nconc
  		      (mapcar (lambda (rcpt)
! 				(list "--remote-user" rcpt))
  			      (append recipients
  				      (if pgg-encrypt-for-me
  					  (list pgg-gpg-user-id)))))))))
--- 160,166 ----
  	   (if recipients
  	       (apply #'nconc
  		      (mapcar (lambda (rcpt)
! 				(list pgg-gpg-recipient-arg rcpt))
  			      (append recipients
  				      (if pgg-encrypt-for-me
  					  (list pgg-gpg-user-id)))))))))

[-- Attachment #3: Type: text/plain, Size: 36 bytes --]



-- 
 Lloyd Zusman
 ljz@asfast.com

Re: Encrypting outgoing message but not Gcc?

[Simon Josefsson]

Lloyd Zusman <ljz@asfast.com> writes:

> Lloyd Zusman <ljz@asfast.com> writes:
>
>> [ ... ]
>>
>> I was wondering why this option didn't work for me, and I found the code
>> below in yesterday's CVS.  Notice that it makes my user ID known to gpg
>> by means of the `--remote-user' option.  However, this option is no
>> longer used in recent versions of gpg ... it's been replaced with
>> `--recipient'.
>>
>> I think that we need a configuration variable that allows us to specify
>> one or the other of these option flags, so that the code can be made to
>> work with both older and newer versions of gpg.
>
> Here's a patch for this.  What do you folks think?  If it's suitable,
> feel free to apply it to CVS.

It looks good.  --remote-user was marked as obsolete, and not
documented even for GnuPG 1.0.0, the oldest GnuPG release I could find
on ftp.gnug.org, so the patch looks safe.  I have installed it.
Perhaps making it customizable wasn't needed, but probably doesn't
hurt.

Re: Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

Simon Josefsson <jas@extundo.com> writes:

> Lloyd Zusman <ljz@asfast.com> writes:
>
>> Lloyd Zusman <ljz@asfast.com> writes:
>>
>>> [ ... ]
>>>
>>> I think that we need a configuration variable that allows us to specify
>>> one or the other of these option flags, so that the code can be made to
>>> work with both older and newer versions of gpg.
>>
>> Here's a patch for this.  What do you folks think?  If it's suitable,
>> feel free to apply it to CVS.
>
> It looks good.  --remote-user was marked as obsolete, and not
> documented even for GnuPG 1.0.0, the oldest GnuPG release I could find
> on ftp.gnug.org, so the patch looks safe.  I have installed it.
> Perhaps making it customizable wasn't needed, but probably doesn't
> hurt.

Thank you.  That was fast!

-- 
 Lloyd Zusman
 ljz@asfast.com

Re: Encrypting outgoing message but not Gcc?

[Simon Josefsson]

Lloyd Zusman <ljz@asfast.com> writes:

> Is there a way to configure Gnus to encrypt an outgoing message, but not
> its Gcc?

It should be easy to add that functionality by writing a function,
similar to mml-tweak-externalize-attachments, and it to
mml-tweak-sexp-alist.

Re: Encrypting outgoing message but not Gcc?

[Lloyd Zusman]

Simon Josefsson <jas@extundo.com> writes:

> Lloyd Zusman <ljz@asfast.com> writes:
>
>> Is there a way to configure Gnus to encrypt an outgoing message, but not
>> its Gcc?
>
> It should be easy to add that functionality by writing a function,
> similar to mml-tweak-externalize-attachments, and it to
> mml-tweak-sexp-alist.

Thanks.  I've looked through mml.el, and I get the general idea about
`mml-tweak-sexp-alist'.  However, I don't know enough about the internal
logic of this whole mml thing to see how this can be used to prevent
encryption of messages destined to Gcc.

Specifically, I have the following questions:

1. When the specifed sexp is being invoked, what variables do I look at
   which will tell me that the part is on its way to a Gcc destination,
   as opposed to somewhere else?

2. Once I determine the answer to number 1, what should this sexp do to
   tell the mml system not to do any encryption on this part?


-- 
 Lloyd Zusman
 ljz@asfast.com