Re: Time to revisit the message id generation algorithm?

["Johann 'Myrkraverk' Oskarsson" <myrkraverk@gmx.com>]

The following message is a courtesy copy of an article
that has been posted to gnu.emacs.gnus as well.

Lars Ingebrigtsen <larsi@gnus.org> writes:

> "Johann 'Myrkraverk' Oskarsson" <myrkraverk@yahoo.com> writes:
>
>> Depending on the seriousness of said unethical practices, that person
>> may have just lost her life to (message-unique-id).
>
> True, but it's pretty unlikely.  If you want anonymous email, you
> should go through an email anonymising service.

You're right.  However, I do have a real life story of the .fsf exposing
an author years later - ironically on the very day I posted the articlue
on gnu.emacs.gnus.

As it is, I can't recommend people to use Gnus with TOR on usenet.

> Having the uid in the Message-ID is documented, and is handy when
> scoring followups.

I'm not disputing the current id generation.  It has worked very well so
far.

>> This is unlikely to be a problem but for people setting the
>> mail-host-address to their email provider.  Say Google or Yahoo.

> Nobody has ever reported a problem like that in Gnus' 25 years, so I
> think it's pretty unlikely.

It *is* unlikely.  I am for example setting the mail-host-address to a
unique identifier for my laptop, even though there is no publicly
queriable IP address for it, asuka.myrkraverk.com.

An the other hand, if I were to set mail-host-address to mail.gmx.com
(to reflect my From field) I am running the risk of generating the same
message id as someone else (using Gnus) doing just that.


There are ways to address these issues without panicking.  For example,
the .fsf can be made configurable, and the user id as well.

Personally I'd like to see more random data in the id, but doing that
right means also not doing it in a hurry.

I'm likely to come up with some proposals in the near future.  For
example, I might bring a patch for the following configuration
variables: message-unique-id-use-uid, message-unique-id-prefix,
message-unique-id-suffix.  Where the defaults are t, "" and ".fsf"
respectively.


Johann