CC: | michael.stevens [...] dianomi.com, michael.smith [...] dianomi.com |
Subject: | another spam detection case |
Another spam detection.
https://gitorious.org/mail-deliverystatus-bounceparser-fork/mail-
deliverystatus-bounceparser-
fork/commit/ecc190a34f60d817409595e0de27a3d4e91676d9
or attached.
Subject: | 0001-Add-a-new-spam-detection-case.patch |
From ecc190a34f60d817409595e0de27a3d4e91676d9 Mon Sep 17 00:00:00 2001
From: Michael Stevens <mstevens@etla.org>
Date: Tue, 10 May 2011 14:38:52 +0100
Subject: [PATCH] Add a new spam detection case.
---
MANIFEST | 2 +
lib/Mail/DeliveryStatus/BounceParser.pm | 3 +-
t/corpus/spam-rejection13.msg | 94 +++++++++++++++++++++++++++++++
t/spam-rejection13.t | 29 ++++++++++
4 files changed, 127 insertions(+), 1 deletions(-)
create mode 100644 t/corpus/spam-rejection13.msg
create mode 100644 t/spam-rejection13.t
diff --git a/MANIFEST b/MANIFEST
index 0e29369..2960490 100644
--- a/MANIFEST
+++ b/MANIFEST
@@ -81,6 +81,7 @@ t/corpus/spam-rejection9.msg
t/corpus/spam-rejection10.msg
t/corpus/spam-rejection11.msg
t/corpus/spam-rejection12.msg
+t/corpus/spam-rejection13.msg
t/corpus/spam-rejection-uribl.msg
t/corpus/spam-with-badly-parsed-email.msg
t/corpus/spam-with-image.msg
@@ -146,6 +147,7 @@ t/spam-rejection9.t
t/spam-rejection10.t
t/spam-rejection11.t
t/spam-rejection12.t
+t/spam-rejection13.t
t/spam-rejection-uribl.t
t/spam-with-image.t
t/spamassassin.t
diff --git a/lib/Mail/DeliveryStatus/BounceParser.pm b/lib/Mail/DeliveryStatus/BounceParser.pm
index f71facf..d61d7c6 100644
--- a/lib/Mail/DeliveryStatus/BounceParser.pm
+++ b/lib/Mail/DeliveryStatus/BounceParser.pm
@@ -1029,7 +1029,8 @@ sub _std_reason {
/breaches\s+local\s+URIBL\s+policy/i or
/Your\s+email\s+had\s+spam-like\s+header\s+contents/i or
/detected\s+as\s+spam/i or
- /Denied\s+due\s+to\s+spam\s+list/i
+ /Denied\s+due\s+to\s+spam\s+list/i or
+ /appears\s+to\s+be\s+unsolicited/i
) {
return "spam";
}
diff --git a/t/corpus/spam-rejection13.msg b/t/corpus/spam-rejection13.msg
new file mode 100644
index 0000000..922135e
--- /dev/null
+++ b/t/corpus/spam-rejection13.msg
@@ -0,0 +1,94 @@
+Delivered-To: automated-bounces+ca616b18-7a99-11e0-9137-b6e182a371be@email.example.com
+Received: by 10.204.122.204 with SMTP id m12cs106554bkr;
+ Mon, 9 May 2011 17:09:40 -0700 (PDT)
+Received: by 10.227.7.146 with SMTP id d18mr4785673wbd.96.1304986180166;
+ Mon, 09 May 2011 17:09:40 -0700 (PDT)
+Return-Path: <>
+Received: from admin1.example.co.uk (admin1.example.co.uk [94.236.45.212])
+ by mx.google.com with ESMTP id y20si17683363wbh.0.2011.05.09.17.09.38;
+ Mon, 09 May 2011 17:09:39 -0700 (PDT)
+Received-SPF: pass (google.com: best guess record for domain of admin1.example.co.uk designates 94.236.45.212 as permitted sender) client-ip=94.236.45.212;
+Authentication-Results: mx.google.com; spf=pass (google.com: best guess record for domain of admin1.example.co.uk designates 94.236.45.212 as permitted sender) smtp.mail=
+Received: by admin1.example.co.uk (Postfix)
+ id A92D665106E; Tue, 10 May 2011 01:09:38 +0100 (BST)
+Date: Tue, 10 May 2011 01:09:38 +0100 (BST)
+From: MAILER-DAEMON@admin1.example.co.uk (Mail Delivery System)
+Subject: Undelivered Mail Returned to Sender
+To: automated-bounces+CA616B18-7A99-11E0-9137-B6E182A371BE@email.example.com
+Auto-Submitted: auto-replied
+MIME-Version: 1.0
+Content-Type: multipart/report; report-type=delivery-status;
+ boundary="82FE9651062.1304986178/admin1.example.co.uk"
+Message-Id: <20110510000938.A92D665106E@admin1.example.co.uk>
+
+This is a MIME-encapsulated message.
+
+--82FE9651062.1304986178/admin1.example.co.uk
+Content-Description: Notification
+Content-Type: text/plain; charset=us-ascii
+
+This is the mail system at host admin1.example.co.uk.
+
+I'm sorry to have to inform you that your message could not
+be delivered to one or more recipients. It's attached below.
+
+For further assistance, please send mail to <postmaster>
+
+If you do so, please include this problem report. You can
+delete your own text from the attached returned message.
+
+ The mail system
+
+<recipient@example.net>: host mailb.example.net[195.248.90.32] said: 554 5.7.1 Mail
+ (13049-38149) appears to be unsolicited - Score (59) over
+ MessageScoringUpperLimit (50) - send error reports to postmaster@example.net
+ (in reply to end of DATA command)
+
+--82FE9651062.1304986178/admin1.example.co.uk
+Content-Description: Delivery report
+Content-Type: message/delivery-status
+
+Reporting-MTA: dns; admin1.example.co.uk
+X-Postfix-Queue-ID: 82FE9651062
+X-Postfix-Sender: rfc822; automated-bounces+CA616B18-7A99-11E0-9137-B6E182A371BE@email.example.com
+Arrival-Date: Tue, 10 May 2011 01:09:36 +0100 (BST)
+
+Final-Recipient: rfc822; recipient@example.net
+Action: failed
+Status: 5.7.1
+Remote-MTA: dns; mailb.example.net
+Diagnostic-Code: smtp; 554 5.7.1 Mail (13049-38149) appears to be unsolicited -
+ Score (59) over MessageScoringUpperLimit (50) - send error reports to
+ postmaster@example.net
+
+--82FE9651062.1304986178/admin1.example.co.uk
+Content-Description: Undelivered Message
+Content-Type: message/rfc822
+
+Received: by admin1.example.co.uk (Postfix, from userid 507)
+ id 82FE9651062; Tue, 10 May 2011 01:09:36 +0100 (BST)
+MIME-Version: 1.0
+Content-Transfer-Encoding: 7bit
+Content-Type: multipart/related; boundary="_----------=_13049861762689733"
+X-Mailer: MIME::Lite 3.027 (F2.78; T1.31; A2.07; B3.13; Q3.13)
+Date: Tue, 10 May 2011 01:09:36 +0100
+From: Fred Bloggs <fred.bloggs@email.example.com>
+Subject: foo
+X-Campaignid: example20101019.1
+List-Unsubscribe: <http://www.exampleoffers.co.uk/unsubscribe.epl?uuid=CA616B18-7A99-11E0-9137-B6E182A371BE>
+To: John Smith <recipient@example.net>
+Message-Id: <20110510000936.82FE9651062@admin1.example.co.uk>
+
+This is a multi-part message in MIME format.
+
+--_----------=_13049861762689733
+Content-Disposition: inline
+Content-Transfer-Encoding: quoted-printable
+Content-Type: text/html
+
+HTML
+
+--_----------=_13049861762689733--
+
+
+--82FE9651062.1304986178/admin1.example.co.uk--
diff --git a/t/spam-rejection13.t b/t/spam-rejection13.t
new file mode 100644
index 0000000..dc11871
--- /dev/null
+++ b/t/spam-rejection13.t
@@ -0,0 +1,29 @@
+#!perl -wT
+use strict;
+
+use Test::More tests => 3;
+
+use Mail::DeliveryStatus::BounceParser;
+
+# FH because we're being backcompat to pre-lexical
+sub readfile {
+ my $fn = shift;
+ open FH, "$fn" or die $!;
+ local $/;
+ my $text = <FH>;
+ close FH;
+ return $text;
+}
+
+my $message = readfile('t/corpus/spam-rejection13.msg');
+
+my $bounce = Mail::DeliveryStatus::BounceParser->new($message);
+
+isa_ok($bounce, 'Mail::DeliveryStatus::BounceParser');
+ok($bounce->is_bounce, "This is a bounce");
+
+my ($report) = $bounce->reports;
+
+my $std_reason = $report->get("std_reason");
+
+is($std_reason, "spam", "std reason is spam");
--
1.7.4.1