On Fri, Mar 14 2025, Robert Pluim <rpl...@gmail.com> wrote:

>>>>>> On Thu, 13 Mar 2025 22:22:52 +0000, Leo Butler <leo.but...@umanitoba.ca> 
>>>>>> said:
>
>     Leo> On Thu, Mar 13 2025, gfp <g...@posteo.at> wrote:
>     >> Hi Gnus,
>     >> 
>     >> I wanted to reply to a mail and I got an error message:
>     >> 
>     >> Debugger entered--Lisp error: (error "Sending failed: 554 5.1.8
>     >> <g...@tuxedo.mail-host-ad...")
>     >> error("Sending failed: %s" "554 5.1.8
>     >> <g...@tuxedo.mail-host-address-is-not-set...")

This bogus email address, g...@tuxedo.mail-host-address-is-not-set, is
inserted by `message-make-fqdn' in message.el. It indicates that Emacs
cannot determine a non-bogus FQDN from amongst the variables:

message-user-fqdn
mail-host-address
user-mail-address

and the return value of

(system-name)

>
>     Leo> Hello,
>
>     Leo> The error message says `mail-host-address' is not set.
>     Leo> You can customize the variable thus:
>
>     Leo> M-x customize-variable mail-host-address RET
>
>     Leo> I don't think the value matters too much. You could set it to a 
> string
>     Leo> that contains a non-routable FQDN, like "localhost.lan".

Partial correction. The variable `message-bogus-system-names' is a regex
to filter bogus domain names. On my system, anything with localhost or
local is filtered.

> The value very much matters if the MTA does sender address validation,
> which this one appears to be doing. based on the "5.1.8". You should
> set it to be the same as your email addressʼ host, or set
> `user-mail-address' correctly.

Ok, fair enough.

message.el has several ways to infer the correct email address of the
sender. It is best to have its fallback contain valid information.

>
>     Leo> In Gnus, you can inspect the mail headers by typing t in the message
>     Leo> buffer. You can see what other Gnus users do by looking at the
>     Leo> Message-ID header: the text to the right of the @-sign is the
>     Leo> `mail-host-address' of that Gnus user.
>
> Yes. But those people would have set `user-mail-address' to something
> valid.

I guess you must mean syntactically valid, because my
`user-mail-address' is a syntactically valid but bogus email address.

Leo

Reply via email to