* mail-host-address and i-did-not-set--mail-host-address--so-tickle-me
@ 2020-12-12 22:29 yarnton
2020-12-12 22:53 ` Adam Sjøgren
0 siblings, 1 reply; 6+ messages in thread
From: yarnton @ 2020-12-12 22:29 UTC (permalink / raw)
To: ding
Hi,
As far as I understand, the function message-make-fqdn in gnus/message.el concats the string "i-did-not-set--mail-host-address--so-tickle-me" to system-name if that is not a fully qualified domain name. This is used to construct a unique Message-ID.
To be a bit more precise, it also checks the values of user-mail-address and mail-host-address before defaulting to string with the "tickle-me" substring.
I used to use Gnus, but my MUA these days is Notmuch. Notmuch and other packages use message.el, and I am trying to understand why I need to set up mail-host-address to something arbitrary (I use "gnu.org") in order to avoid the "tickle-me" string.
Is this a bug from Notmuch when calling message.el, some ancient setting from Gnus that doesn't play well with other packages, or something I'm doing wrong? In principle, I don't see why I need to set my email parameters in my .emacs. Those are a concern of Notmuch and MSTMP only.
I think Message-ID should reveal as little information as possible these days, as it can be used for identifying or at least fingerprinting messages. Whenever I use Emacs from a device where the hostname returns a fully qualified name like an IP address, that will leak into my Message-ID. Is there any technical reason for not constructing a random Message-ID like [UUID]@gnu.org?
Thanks.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: mail-host-address and i-did-not-set--mail-host-address--so-tickle-me
2020-12-12 22:29 mail-host-address and i-did-not-set--mail-host-address--so-tickle-me yarnton
@ 2020-12-12 22:53 ` Adam Sjøgren
2020-12-13 3:48 ` 황병희
2020-12-13 18:10 ` yarnton
0 siblings, 2 replies; 6+ messages in thread
From: Adam Sjøgren @ 2020-12-12 22:53 UTC (permalink / raw)
To: ding
yarnton@tutanota.com writes:
> In principle, I don't see why I need to set my email parameters in my
> .emacs. Those are a concern of Notmuch and MSTMP only.
If Emacs is generating the Message-ID, I'd say Emacs needs enough
information to do so - the recommended way is to put the host after the @.
You could argue that the recommendation is outdated.
> I think Message-ID should reveal as little information as possible
> these days, as it can be used for identifying or at least
> fingerprinting messages.
(The whole purpose of the Message-ID is to uniquely identify the
message! I guess you mean identifying... the sender? But that's also
done in From:, right? If you are thinking of identifying the machine
sending, then that is recorded in the Received headers as well, right?)
> Whenever I use Emacs from a device where the hostname returns a fully
> qualified name like an IP address, that will leak into my Message-ID.
(It is usually also recorded in the headers the MTAs add.)
> Is there any technical reason for not constructing a random Message-ID
> like [UUID]@gnu.org?
Not except for what RFC 2822 recommends:
· https://tools.ietf.org/html/rfc2822#section-3.6.4
At least according to the logic in the RFC, you should put something you
control on the right side of the @, and not somebody else's host/domain
- but as long as the whole thing is unique, you should be good.
Best regards,
Adam
--
"I could't find my perfect attendance pencil, which Adam Sjøgren
is the only pencil I turned in [for sharpening]" asjo@koldfront.dk
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: mail-host-address and i-did-not-set--mail-host-address--so-tickle-me
2020-12-12 22:53 ` Adam Sjøgren
@ 2020-12-13 3:48 ` 황병희
2020-12-13 18:10 ` yarnton
1 sibling, 0 replies; 6+ messages in thread
From: 황병희 @ 2020-12-13 3:48 UTC (permalink / raw)
To: The Gnus
> (The whole purpose of the Message-ID is to uniquely identify the
> message! I guess you mean identifying... the sender? But that's also
> done in From:, right? If you are thinking of identifying the machine
> sending, then that is recorded in the Received headers as well, right?)
I agree with Adam's comment.
Sincerely, Gnus fan Byung-Hee
--
^고맙습니다 _布德天下_ 감사합니다_^))//
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: mail-host-address and i-did-not-set--mail-host-address--so-tickle-me
2020-12-12 22:53 ` Adam Sjøgren
2020-12-13 3:48 ` 황병희
@ 2020-12-13 18:10 ` yarnton
2020-12-13 21:28 ` Adam Sjøgren
1 sibling, 1 reply; 6+ messages in thread
From: yarnton @ 2020-12-13 18:10 UTC (permalink / raw)
To: Adam Sjøgren; +Cc: ding
> (The whole purpose of the Message-ID is to uniquely identify the
> message! I guess you mean identifying... the sender? But that's also
> done in From:, right? If you are thinking of identifying the machine
> sending, then that is recorded in the Received headers as well, right?)
>
Sure, but identifying a message in a unique way does not require leaking unnecessary information about the host. That was my point. The vast majority of the email clients put something pretty non-unique on the right hand side of the Message-ID like @emailservice.com or @localhost.localdomain (literally).
This still respects RFC 2822, as the left hand side of Message-ID is what makes the whole Message-ID likely to be unique.
From: and my IP address can be fairly anonymous, but that is not useful if Message-ID leaks my exact domain. For example, say my localhost gets set to some local domain name that is quite representative of where I am by dhcpd e.g., lab21.bio.dtu.dk. I am inadvertendly leaking more or less who I am.
An additional problem is that message-make-fqdn will concat the string ".i-did-not-set-mail-host-address--so-tickle-me" into Message-ID if it cannot construct a valid Message-ID. Gnus checks the Message-ID before sending, but other Emacs email clients do not. A different way of flagging an error might be more appropriate here.
In other words, I think the logic inside the message-mode function message-make-fqdn is a bit outdated for today's Internet. Furthermore, returning a special value to flag an error is not playing very will with other Emacs email clients aside from Gnus, as these do not check for the presence of "[...]-tickle-me" to flag an error before sending an email, nor show the Message-ID header when composing a message.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: mail-host-address and i-did-not-set--mail-host-address--so-tickle-me
2020-12-13 18:10 ` yarnton
@ 2020-12-13 21:28 ` Adam Sjøgren
2020-12-14 11:14 ` 황병희
0 siblings, 1 reply; 6+ messages in thread
From: Adam Sjøgren @ 2020-12-13 21:28 UTC (permalink / raw)
To: ding
yarnton@tutanota.com writes:
> I think the logic inside the message-mode function message-make-fqdn
> is a bit outdated for today's Internet.
What do you think it should do instead?
Best regards,
Adam
--
"As a statue wearing complicated clothes Adam Sjøgren
A monument to power, but on reconstrucdead toes asjo@koldfront.dk
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: mail-host-address and i-did-not-set--mail-host-address--so-tickle-me
2020-12-13 21:28 ` Adam Sjøgren
@ 2020-12-14 11:14 ` 황병희
0 siblings, 0 replies; 6+ messages in thread
From: 황병희 @ 2020-12-14 11:14 UTC (permalink / raw)
To: The Gnus
Adam Sjøgren <asjo@koldfront.dk> writes:
> yarnton@tutanota.com writes:
>
>> I think the logic inside the message-mode function message-make-fqdn
>> is a bit outdated for today's Internet.
>
> What do you think it should do instead?
Also i'm waiting for yarnton's comments. Because Message-ID is my favorite thing
in Gnus.
Sincerely, Gnus fan Byung-Hee
--
^고맙습니다 _白衣從軍_ 감사합니다_^))//
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2020-12-14 11:14 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-12-12 22:29 mail-host-address and i-did-not-set--mail-host-address--so-tickle-me yarnton
2020-12-12 22:53 ` Adam Sjøgren
2020-12-13 3:48 ` 황병희
2020-12-13 18:10 ` yarnton
2020-12-13 21:28 ` Adam Sjøgren
2020-12-14 11:14 ` 황병희
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).