It is no longer possible to subscribe to bugs as
"Please confirm subscription" e-mail messages are no longer sent.

On 2024-05-10, this was working, but since 2024-05-20 at least,
this is no longer working. I had sent:

From: Vincent Lefevre <vincent@vinc17.net>
To: 1071372-subscribe@bugs.debian.org
Subject: subscribe
Date: Mon, 20 May 2024 00:56:07 +0200
Message-ID: <20240519225607.GE2665@qaa.vinc17.org>

The issue remains with attempts today (2024-05-21) at 14:26:56 +0200
and 14:47:15 +0200 for another bug.

Actually, after a look at the logs on my mail server, this was still
working at the following date:

2024-05-11T11:57:13.575225+02:00 joooj postfix/smtp[74387]: AB7C1812: to=<1070891-subscribe@bugs.debian.org>, relay=buxtehude.debian.org[2607:f8f0:614:1::1274:39]:25, delay=3.9, delays=0.04/0.11/2.8/0.89, dsn=2.0.0, status=sent (250 OK id=1s5jTe-00BPwE-Lt)

with a reply ("Please confirm subscription" e-mail):

2024-05-11T12:00:14.976146+02:00 joooj postfix/qmgr[70857]: E359C13E: from=<1070891-ignore@bugs.debian.org>, size=4289, nrcpt=1 (queue active)

But this was no longer working at the following date:

2024-05-17T15:51:22.974407+02:00 joooj postfix/smtp[164605]: 8556DACB: to=<1071274-subscribe@bugs.debian.org>, relay=buxtehude.debian.org[209.87.16.39]:25, delay=4.4, delays=0.03/0.04/3.4/0.94, dsn=2.0.0, status=sent (250 OK id=1s7xzY-00CaPs-2o)

(I did not get the reply).

I'm having the same problem. Could it be related to #1071469 which also
involves mail issues around the same time?

At https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1071469
Louis-Philippe Véronneau says

"It seems that a debian.org mail infrastructure breakage happened
sometime between 2024-05-17 and 2024-05-18"

and

"I dug a little bit more and it seems the outage happened between
2024-05-17 ~2200 and and 2024-05-18 0110UTC."

But since the outage ended on 2024-05-18 (and the issue with bug
subscription seemed to have started before the outage), this seems
to be a different problem... unless someone did a series of changes
in the mail config on 2024-05-17 and early 2024-05-18.

Thanks for the reports; I note that around this time the systems were
upgraded. It's possible that this finally broke bug mail, which was
using a totally unsupported out of date version of EOC.

Not sure what the right path forward is to fix that without totally
re-implementing bug mailing lists (which probably needs to happen
anyway).

Looking at it, this is definitely a casualty of python2->python3. I'm
working with Liw to get the most-recent version of EoC (2010!) and I'll
do the tweaks to make it work on python3.

Look for me to have it working again this weekend.

Hi Don,
the replies for subscription requests to bug be send out for those
falling inbetween the broken phase?

(For context, I have some bugs for which I have issued a subscription
request and would like to monitor the updates there, and wonder if
once fixed need to re-issue those so I can really subscribe to the
bugs).

Regards,
Salvatore

I suspect you'll have to re-issue them, but I'm not certain.

Just an update, I've gotten the last version from Lars (from 2010!), and
I'm working on finishing up testing it.

Once I've done that, I'll need to port in the patch that makes the
buglists work (automated list creation and some other pieces).

I hope to having it working again sometime this week (but my Debian time
is limited to small 30 minute bits each day, so no real promises).

This should now be working again; enemies of carlotta has been updated
to python 3 and deployed on bendel.debian.org, and I've briefly tested
it.

Please feel free to reopen this bug (or file a new one) if you find any
more issues.

Unfortunately, any messages sent between then and now are lost due to
the way buglists are implemented (and that hasn't been changed by
porting to python 3... though eventually we should fix that.)

I've tried, and I still don't receive the "Please confirm subscription"
e-mail. But perhaps I should wait longer...

Control: reopen -1

Reopening since after more than 2 hours, I still haven't received
the "Please confirm subscription" e-mail.

Looks like you had some non-ascii characters which triggered a bug; can
you try again now? [Python 3 is much more strict than python 2, and I've
tried to be as minimal with the changes as possible, so we're probably
handling utf-8 incorrectly. Hopefully it isn't an issue.]

Date: Fri, 21 Jun 2024 09:56:17 +0200
From: Vincent Lefevre <vincent@vinc17.net>
To: 1073053-subscribe@bugs.debian.org
Subject: subscribe
Message-ID: <20240621075617.GA6161@qaa.vinc17.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1

and a non-ASCII character as usual, i.e. not UTF-8.
Still the same problem.

In any case, I think that the body of the message should entirely
be ignored.

On 21.06.2024 10:40, Vincent Lefevre wrote:

Hi,
For me bug subscription works again fine. No clue what is the difference
to your setup.

Hilmar
--
sigfault

Hrm. Looing at this more closely, the version of EOC that we're using
doesn't handle encodings at all, and just happened to work correctly,
but modern python with bytes->string coercion is causing all sorts of
issues.

I'll dig back into this and get a real solution implemented.

Hi Don,

For me it neither work yet with subscribing to bugs. But I hope this
counts not simply as "me too" reply. I noticed that in my case even if
I just trigger a subscription mail with 'bts subscribe' command I do
not get any subscrbtion request challenge reply.

| From: carnil@debian.org
| To: 1071576-subscribe@bugs.debian.org
| Subject: subscription request for bug #1071576
| Date: Tue, 25 Jun 2024 21:59:25 +0200
| User-Agent: devscripts bts/2.23.7
| Message-ID: <1719345565-2799-bts-carnil@debian.org>
|
| thanks

Does this gives any further insights in what goes wrong?

Regards,
Salvatore

Hrm. I can see the subscription requests coming in, and in theory, the
logs show confirmation messages going out, but I'm not sure that they
actually are going out because I don't see them in the mail logs.

Turning on more debugging here to see if I can get some useful ideas.
[It's possible that it's not writing the e-mail correctly to sendmail or
something stupid like that...]

I've tried again, and it's still working for me (I'm getting the ack
message).

Would you mind trying again and I'll see if I can see something useful
from the debug logs? [I should have the times now, so hopefully I'll see
the outgoing messages too.]

Hi Don,

Sure, just done now again.

Thanks a lot for your time on this matter, it is *very* much
appreciated!

Regards,
Salvatore

For bugs I'm reporting, I've a procmail rule to automatically
subscribe, and on some new bugs, I could see that this worked (there's
no accented character in the body in these cases), but in the copy of
the mail in the "Please confirm subscription" message, I can see:

[...]
X-Spam-Flag: NO
X-Spam-Score: 2.062
X-Spam-Level: **
X-Spam-Status: No, score=2.062 tagged_above=-10000 required=6.31
        tests=[BAYES_00=-1.9, EMPTY_MESSAGE=2.32,
        HEADER_FROM_DIFFERENT_DOMAINS=0.249, PYZOR_CHECK=1.392,
        TVD_SPACE_RATIO=0.001] autolearn=no autolearn_force=no
[...]
From: Vincent Lefevre <vincent@vinc17.net>
To: 1074263-subscribe@bugs.debian.org
Subject: subscribe
Message-Id: <20240625144507.035ADEAE@joooj.vinc17.net>
Date: Tue, 25 Jun 2024 16:45:07 +0200 (CEST)
X-CrossAssassin-Score: 49740
X-Debian-PR-Message: list 1074263
X-Debian-PR-Package: otpclient
X-Debian-PR-Title: otpclient: with FreeOTP+ export, spaces in key URIs are not encoded

I'm wondering when the "X-Debian-PR-Title:" header is added (these
headers are not provided by me), and whether it can be an issue when
the bug title contains particular characters, such as non-ASCII ones.

Moreover, is there antispam filtering? (See the X-Spam-* headers,
and EMPTY_MESSAGE=2.32 might yield false positives while an empty
message is normal in this case.)

I've tried again with an acceted character, and I've received a
"Please confirm subscription" message 4 minutes later, but the
copy of my mail is incorrect:

 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

b'Return-Path: <1@buxtehude.debian.org>\nX-Original-To: lists-bugs=1073053-subscribe@bendel.debian.org\nDelivered-To: lists-bugs=1073053-subscribe@bendel.debian.org\nReceived: from localhost (localhost [127.0.0.1])\n\tby bendel.debian.org (Postfix) with ESMTP id 914D920631\n\tfor <lists-bugs=1073053-subscribe@bendel.debian.org>; Wed, 26 Jun 2024 08:06:09 +0000 (UTC)\nX-Virus-Scanned: Debian amavisd-new at bendel.debian.org\nX-Spam-Flag: NO\nX-Spam-Score: -1.661\nX-Spam-Level:\nX-Spam-Status: No, score=-1.661 tagged_above=-10000 required=6.31\n\ttests=[BAYES_00=-1.9, HEADER_FROM_DIFFERENT_DOMAINS=0.249,\n\tT_SCC_BODY_TEXT_LINE=-0.01] autolearn=no autolearn_force=no\nReceived: from bendel.debian.org ([127.0.0.1])\n\tby localhost (lists.debian.org [127.0.0.1]) (amavisd-new, port 2525)\n\twith ESMTP id OqkWwY62Gpce\n\tfor <lists-bugs=1073053-subscribe@bendel.debian.org>;\n\tWed, 26 Jun 2024 08:06:06 +0000 (UTC)\nReceived: from buxtehude.debian.org (buxtehude.debian.org [IPv6:2607:f8f0:61
 4:1::1274:39])\n\t(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)\n\t key-exchange ECDHE (P-256) server-signature RSA-PSS (2048 bits) server-digest SHA256\n\t client-signature RSA-PSS (2048 bits) client-digest SHA256)\n\t(Client CN "buxtehude.debian.org", Issuer "Debian SMTP CA" (not verified))\n\tby bendel.debian.org (Postfix) with ESMTPS id 4882920628\n\tfor <bugs=1073053-subscribe@lists.debian.org>; Wed, 26 Jun 2024 08:06:06 +0000 (UTC)\nReceived: from debbugs by buxtehude.debian.org with local (Exim 4.94.2)\n\t(envelope-from <1@buxtehude.debian.org>)\n\tid 1sMNfG-00DUKf-Em\n\tfor bugs=1073053-subscribe@lists.debian.org; Wed, 26 Jun 2024 08:06:02 +0000\nReceived: (at 1073053-subscribe) by bugs.debian.org; 26 Jun 2024 08:02:35 +0000\nX-Spam-Bayes: score:0.0000 Tokens: new, 8; hammy, 124; neutral, 25; spammy, 0.\n\tspammytokens: hammytokens:0.000-+--H*F:D*vinc17.net,\n\t0.000-+--H*rp:D*vinc17.net, 0.000-+--HX-Mailer-Info:mutt,\n\t0.000-+--HX-Mailer-Info:www.vinc17.
 net, 0.000-+--H*MI:vinc17\nReceived: from joooj.vinc17.net ([2001:4b99:1:3:216:3eff:fe20:ac98]:34110)\n\tby buxtehude.debian.org with esmtps (TLS1.3:ECDHE_X25519__RSA_PSS_RSAE_SHA256__AES_256_GCM:256)\n\t(Exim 4.94.2)\n\t(envelope-from <vincent@vinc17.net>)\n\tid 1sMNbu-00DToS-Dn\n\tfor 1073053-subscribe@bugs.debian.org; Wed, 26 Jun 2024 08:02:35 +0000\nReceived: from smtp-qaa.vinc17.net (2a02-8428-1b1d-4d01-96a9-491d-7b48-ba31.rev.sfr.net [IPv6:2a02:8428:1b1d:4d01:96a9:491d:7b48:ba31])\n\tby joooj.vinc17.net (Postfix) with ESMTPSA id 4F23D49E;\n\tWed, 26 Jun 2024 10:02:30 +0200 (CEST)\nReceived: by qaa.vinc17.org (Postfix, from userid 1000)\n\tid C677ECA00F4; Wed, 26 Jun 2024 10:02:29 +0200 (CEST)\nDate: Wed, 26 Jun 2024 10:02:29 +0200\nFrom: Vincent Lefevre <vincent@vinc17.net>\nTo: 1073053-subscribe@bugs.debian.org\nSubject: subscribe\nMessage-ID: <20240626080229.GB3052@qaa.vinc17.org>\nMIME-Version: 1.0\nContent-Type: text/plain; charset=iso-8859-1\nContent-Disposition: inline\n
 Content-Transfer-Encoding: 8bit\nX-Mailer-Info: https://www.vinc17.net/mutt/\nUser-Agent: Mutt/2.2.13+76 (1f3da810) vl-167818 (2024-04-20)\nX-Debian-PR-Message: list 1073053\nX-Debian-PR-Package: unbound\nX-Debian-PR-Title: unbound returns SERVFAIL for every request\n\n-- \nVincent Lef\xe8vre <vincent@vinc17.net> - Web: <https://www.vinc17.net/>\n100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>\nWork: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)\n\n\n'

[...]

BTW, this incorrect copy breaks my procmail rule to confirm the
subscription, because this rule does some check on this copy to
ensure that the subscription request comes from me.