On Sun, Feb 14, 2021 at 10:49:52AM -0500, John Levine wrote:

> # virtual addresses
> 回声@xn--zbs01c.xn--5nqx41au4nqohsp3axcg.xn--fiqs8s     echo
> 回声@声.电子邮件测试.中国        echo
> 
> Feb 14 10:26:16 eaicheck postfix/smtpd[48778]: DED2E89340: 
> client=localhost[127.0.0.1]
> Feb 14 10:26:16 eaicheck postfix/smtp[48776]: CBA778933E: 
> to=<[email protected]>, 
> orig_to=<�M-^[M-^^声@声.�M-^T��M-^P�M-^B�件�M-^K�M-^U.中�M-^[�>, 
> relay=127.0.0.1[127.0.0.1]:10027, delay=0.18, delays=0.03/0.01/0.05/0.09, 
> dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as DED2E89340)

Looks good so far, as you reported.
> Feb 14 10:31:25 eaicheck postfix/postfix-script[48806]: refreshing the 
> Postfix mail system
> Feb 14 15:31:25 eaicheck postfix/master[35605]: reload -- version 3.5.8, 
> configuration /usr/local/etc/postfix

Not sure why the reload here, did the configuration change?

> Feb 14 10:31:51 eaicheck postfix/smtpd[48813]: connect from 
> gal.iecc.com[64.57.183.53]
> Feb 14 10:31:52 eaicheck postfix/smtpd[48813]: NOQUEUE: reject: RCPT from 
> gal.iecc.com[64.57.183.53]: 550 5.1.1 
> <�M-^[M-^^声@�M-^[M-^^声.�M-^T��M-^P�M-^B�件�M-^K�M-^U.中�M-^[�>: Recipient 
> address rejected: User unknown in virtual alias table; from=<[email protected]> 
> to=<�M-^[M-^^声@�M-^[M-^^声.�M-^T��M-^P�M-^B�件�M-^K�M-^U.中�M-^[�> proto=ESMTP 
> helo=<gal.iecc.com>

Comparing:

> CBA778933E: orig_to=<�M-^[M-^^声@声.�M-^T��M-^P�M-^B�件�M-^K�M-^U.中�M-^[�>, 
> relay=127.0.0.1[127.0.0.1]:10027, delay=0.18, delays=0.03/0.01/0.05/0.09, 
> dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as DED2E89340)
> reject:   550 5.1.1 
> <�M-^[M-^^声@�M-^[M-^^声.�M-^T��M-^P�M-^B�件�M-^K�M-^U.中�M-^[�>:

We observe that the reject address and the accepted addresses are not
the same in their domain parts:

             声.�M-^T��M-^P�M-^B�件�M-^K�M-^U.中�M-^[�
    �M-^[M-^^声.�M-^T��M-^P�M-^B�件�M-^K�M-^U.中�M-^[�

The second has some extra bytes at the front.  Did some friendly
application prepend a BOM?

Whatever you used to scrape the logs and insert the output into your
post seems to not have been UTF-8 transparent.  It would help if you
could avoid the transformation from:

    声.电子邮件测试.中国

to:

   声.�M-^T��M-^P�M-^B�件�M-^K�M-^U.中�M-^[�

So that at least we could tell what UTF-8 sequences were actually sent.

-- 
    Viktor.

Reply via email to