On Thu, Sep 26, 2019 at 01:14:46PM -0400, Viktor Dukhovni via Exim-users wrote: > > On Sep 26, 2019, at 12:50 PM, Evgeniy Berdnikov via Exim-users <exim-users@???> wrote:
> >
> >> at least one MUA strips the
> >> BCC headers before submitting the message, but fails to do so when
> >> "resending" the message. (I'm talking about Mutt, and its "bounce"
> >> capability).
> >
> > Because "bounce" function for Mutt means "forward this mail as it is",
> > it's not a regular submission, isn't it?
>
> It is still submission, not relay. The message gets new ("Resent-")
> Date, Message-Id, From, ...
Indeed, Mutt adds Recent-From, Resent-Date, Recent-Message-Id and Resent-To.
However, mutt's documentation states:
-------------------------------------------------------------------------
7. Forwarding and Bouncing Mail
Bouncing and forwarding let you send an existing message to recipients that you
specify. Bouncing a message sends a verbatim copy of a message to alternative
addresses as if they were the message's original recipients specified in the
Bcc header. Forwarding a message, on the other hand, allows you to modify the
message before it is resent (for example, by adding your own comments).
Bouncing is done using the <bounce> function and forwarding using the <forward>
function bound to “b” and “f” respectively.
-------------------------------------------------------------------------
There is also "resend-message" function in Mutt, which keeps all the set
of headers, but allows to edit body preserving the hole MIME structure
of the message. Unlike "bounce" function, it can be used to
add/delete/modify headers. If Bcc is added, it then be stripped out and
converted to additional arguments to MTA binary (/usr/sbin/sendmail).
This function does not add Resent-* headers, however.
The "bounce" function promtps only for a list of addresses, so it seems
strange to add Bcc contents implicitly.
So, Mutt has several functions with different functionality. Function
"resend-message" allows to edit headers, it handles Bcc properly.
> FWIW, Postfix drops top-level (not
> attachedas message/rfc822) Bcc headers, so there's some precedent
> for that behaviour. --
Eugene Berdnikov
