Re: nmh query

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: nmh query

Bill Wohler
[Trying to get through some backlog before upcoming travel builds it
back up again.]

Hey, you started using MH a couple of years before me. Awesome.

I'm still using 1.6, so I'm unfamiliar with the keepbcc and .msmtprc
file. Please use [hidden email] for your question.

[hidden email] wrote:

> i read the faq - honest injun!
>
> i started using mh in 1982 and it still works - thanks in part to folks like
> you!
>
> if there is a small nit in nmh-1.7.1 (even with keepbcc on in .msmtprc
> BCC: line does not show up to bcc recipients) who do i ask?
>
> thanks much!
> cheers
> bala
>

--
Bill Wohler <[hidden email]> aka <[hidden email]>
http://www.newt.com/wohler/, GnuPG ID:610BD9AD

Reply | Threaded
Open this post in threaded view
|

Re: nmh query

David Levine-3
Hi Bala,

> [hidden email] wrote:
>
> > if there is a small nit in nmh-1.7.1 (even with keepbcc on in .msmtprc
> > BCC: line does not show up to bcc recipients) who do i ask?

.msmtprc looks like something used by msmtp.  nmh does not use it,
not does it have any kind of keepbcc setting.

And it looks like BCC: has never listed the bcc recipients.  This
is from the MH 6.8.5 post.c:

  fprintf (out, "BCC:\n");

If you want us to look into this further, which mail transport are
you using (smtp, sendmail/smtp, or sendmail/pipe)?  That would be
set either in nmh's mts.conf, your ~/.mh_profile, or in your
invocation of send with the -mts switch.

David

Reply | Threaded
Open this post in threaded view
|

Re: nmh query

David Levine-3
In reply to this post by Bill Wohler
Bala writes:

> hi david
>
> thanks for any help! here is my mh_profile entry
>
> send: -mts sendmail/pipe -sendmail /home/bala/bin/msmtp -noverbose -alias /home/bala/.mhrc

OK, with sendmail/pipe, all addresses are provided in the message
itself to the sendmail program.  BSD sendmail deletes Bcc: lines:

   -t     Read message for recipients.  To:, Cc:, and
          Bcc: lines will be scanned for recipient
          addresses.  The Bcc: line will be deleted
          before transmission.

I don't know what msmtp does with Bcc: lines.  Perhaps the msmtp
documention offers information on the keepbcc that you mentioned.

While it doesn't have anything to do with this, note that nmh's Dcc:
is not supported with sendmail/pipe.  I would use sendmail/smtp if
you want support for Dcc:.

David

Reply | Threaded
Open this post in threaded view
|

Re: nmh query

Valdis Klētnieks
In reply to this post by David Levine-3
On Sun, 03 Nov 2019 15:08:12 -0500, David Levine said:

> And it looks like BCC: has never listed the bcc recipients.  This
> is from the MH 6.8.5 post.c:
>
>   fprintf (out, "BCC:\n");

That's proper behavior.  BCC is *blind* carbon copy, specifically intended
to *not* show who else got copies.  I'd argue that the only time it's acceptable
to list recipients there is if you are feeding to an MSA that's like 'sendmail -t' that
needs it to get additional recipients because you can't speak SMTP and hand
off RCPT TO:<...> for the Bcc people.

And of course, you trust your MTA to then promptly eat that header and not
promulgate it any further.

I'll go further and assert that if there is still a Bcc: header in the RFC822 headers
once the MSA has accepted the mail for further processing, that somebody has
dropped the ball because sometimes, even a Bcc: that shows that there *were*
other recipients may be an unacceptable information leak....

attachment0 (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: nmh query

Ken Hornstein-2
>That's proper behavior.  BCC is *blind* carbon copy, specifically
>intended to *not* show who else got copies.  I'd argue that the only
>time it's acceptable to list recipients there is if you are feeding
>to an MSA that's like 'sendmail -t' that needs it to get additional
>recipients because you can't speak SMTP and hand off RCPT TO:<...> for
>the Bcc people.

To get back to the larger point ... with sendmail/pipe, you are accepting
reduced functionality.  MH-style BCC will continue to work.  But Dcc
(which behaves like everyone other MUA's Bcc) will not.

--Ken

Reply | Threaded
Open this post in threaded view
|

Re: nmh query

Robert Elz
In reply to this post by Valdis Klētnieks
    Date:        Sun, 03 Nov 2019 20:56:45 -0500
    From:        "Valdis Kl=?utf-8?Q?=c4=93?=tnieks" <[hidden email]>
    Message-ID:  <103064.1572832605@turing-police>


  | I'll go further and assert that if there is still a Bcc: header in the
  | RFC822 headers once the MSA has accepted the mail for further processing,
  | that somebody has dropped the ball

Not always, sometimes there is only the Bcc header containing recipients.

At least one recipient field used to be required, when Bcc is the only one,
it had to be be retained (it didn't need to, and shouldn't, contain any
addresses, but the field had to remain).   This requirement seems to have
been deleted, and now a message with no recipient fields is OK, but for
compat with older MUAs (potentially even MTAs) it is still a good idea to
include an empty Bcc: field when there are no To or Cc fields.

  | because sometimes, even a Bcc: that shows that there *were*
  | other recipients may be an unacceptable information leak....

When there is another recipient field, and this copy of the message
is to an address that is in that other field, that's reasonable.  But when
the message is being delivered to someone on the Bcc list (which is all
the time when that was the only recipient field) there needs to be
something to explain to the recipient why the message was delivered to
them - a Bcc field ("one or more others received this message - you were
one of them") achieves that purpose, whether it contains the recipient's
address, or not.

kre


Reply | Threaded
Open this post in threaded view
|

Re: nmh query

David Levine-3
Robert writes:

> At least one recipient field used to be required, when Bcc is the only one,
> it had to be be retained (it didn't need to, and shouldn't, contain any
> addresses, but the field had to remain).   This requirement seems to have
> been deleted, and now a message with no recipient fields is OK, but for
> compat with older MUAs (potentially even MTAs) it is still a good idea to
> include an empty Bcc: field when there are no To or Cc fields.

Thanks for that explanation.  nmh does retain the (always blank) BCC: in the blind copies.

Valdis, the non-blind message does not retain any Bcc fields, so no leak there.

I included the following in this message draft:

Bcc: [hidden email],
     Robert Elz <[hidden email]>,
     "Valdis Kl=?utf-8?Q?ē?=tnieks" <[hidden email]>

I'd be interested to see how much of that makes it through to each of you.

David

Reply | Threaded
Open this post in threaded view
|

Re: nmh query

Robert Elz
    Date:        Mon, 04 Nov 2019 21:30:26 -0500
    From:        David Levine <[hidden email]>
    Message-ID:  <[hidden email]>

  | I'd be interested to see how much of that makes it through to each of you.

This (the list message) was just as would be expected.

The bcc with Received, X-Received, X-Google-DKIM-Signature, X-Gm-Message-State,
and X-Google-Smtp-Source fields omitted here; they were all between the
Return-Path (added when the message arrived, of course) and the From fields:

        Return-Path: <[hidden email]>
        From: David Levine <[hidden email]>
        Date: Mon, 04 Nov 2019 21:30:26 -0500
        Message-ID: <[hidden email]>
        Subject: Re: nmh query
        BCC:

        ------- Blind-Carbon-Copy

        To: [hidden email]
        From: David Levine <[hidden email]>
        Subject: Re: nmh query
        In-Reply-To: Your message of Mon, 04 Nov 2019 12:41:21 +0700
        References: <103064.1572832605@turing-police>
          <[hidden email]>
          <[hidden email]>
          <[hidden email]>
          <[hidden email]>
        MIME-Version: 1.0
        Content-Type: text/plain; charset="UTF-8"
        Content-Transfer-Encoding: quoted-printable
        Date: Mon, 04 Nov 2019 21:30:26 -0500
        Message-ID: <[hidden email]>

        [message body just the same as the list message]

        ------- End of Blind-Carbon-Copy

Looks fine to me, I see no reason to change anything.

kre


Reply | Threaded
Open this post in threaded view
|

Re: nmh query

Ralph Corderoy
In reply to this post by David Levine-3
Hi David,

> I don't know what msmtp does with Bcc: lines.

It removes them by default.
https://marlam.de/msmtp/msmtp.html#remove_005fbcc_005fheaders

> Perhaps the msmtp documention offers information on the keepbcc that
> you mentioned.

keepbcc was renamed to remove_bcc_headers in msmtp 1.6.0.
https://marlam.de/msmtp/news-archive.html

--
Cheers, Ralph.