Author: Jeremy Harris Date: To: exim-users Subject: Re: [exim] 4.94: _gnutls_sort_clist: Assertion `k == clist_size'
failed
On 17/02/2021 09:11, Evgeniy Berdnikov via Exim-users wrote: > # exim4 -d-all+transport+tls+route+timestamp -M 1lC14S-00Fr49-2d
> 11:27:57 Exim version 4.94 uid=0 gid=0 pid=3989747 D=3c000000 > 11:27:58.920 about to gnutls_handshake
> 11:27:59.112 (TLS1.2)-(DHE-CUSTOM2048)-(RSA-SHA512)-(AES-256-GCM)
> 11:27:59.112 CLIENT_RANDOM 963c22ab8f8da8a84304b62ffdb9188b616ff6522507ce73ce5fdbcfaa709ff1 fd51aff7d59df024b9f41114be80b554bf5889deff4d616892e0543230c7940a28af5d134e751a72e37a414104ed2544
> 11:27:59.112 TLS: checking peer certificate
> exim4: ../../../lib/x509/common.c:1794: _gnutls_sort_clist: Assertion `k == clist_size' failed.
> 11:27:59.115 LOG: MAIN PANIC
> 11:27:59.115 Delivery status for Aliaksandr.Xxxxxxxxxxx@???: got 0 of 7 bytes (pipeheader) from transport process 3989748 for transport smtp
> 11:27:59.115
> 11:27:59.115 LOG: MAIN
> 11:27:59.115 == aliaksandr.xxxxxxxxxxx@??? <Aliaksandr.Xxxxxxxxxxx@???> R=dnslookup T=remote_smtp defer (-1) DT=0.000s: smtp transport process returned non-zero status 0x0006: terminated by signal 6
>
> For me it looks as some bug in gnutls...
Yup. Although it's always possible we're using gnutls in some way
we shouldn't, a library shouldn't trap like that.
I'd suggest asking on the gnutls-help.lists.gnutls.org mailinglist.
--
Cheers,
Jeremy
