Subject: | Spam case |
Spam case.
https://gitorious.org/mail-deliverystatus-bounceparser-fork/mail-
deliverystatus-bounceparser-
fork/commit/062812da56b5ddc7f5dc56fd2eff36f05b0b7690
Subject: | 0001-Spam-detection-case.patch |
From 062812da56b5ddc7f5dc56fd2eff36f05b0b7690 Mon Sep 17 00:00:00 2001
From: Michael Stevens <mstevens@etla.org>
Date: Mon, 1 Aug 2011 17:18:25 +0100
Subject: [PATCH] Spam detection case
---
MANIFEST | 2 +
lib/Mail/DeliveryStatus/BounceParser.pm | 3 +-
t/corpus/spam-rejection19.msg | 98 +++++++++++++++++++++++++++++++
t/spam-rejection19.t | 29 +++++++++
4 files changed, 131 insertions(+), 1 deletions(-)
create mode 100644 t/corpus/spam-rejection19.msg
create mode 100644 t/spam-rejection19.t
diff --git a/MANIFEST b/MANIFEST
index 5672b9c..394cafe 100644
--- a/MANIFEST
+++ b/MANIFEST
@@ -87,6 +87,7 @@ t/corpus/spam-rejection15.msg
t/corpus/spam-rejection16.msg
t/corpus/spam-rejection17.msg
t/corpus/spam-rejection18.msg
+t/corpus/spam-rejection19.msg
t/corpus/spam-rejection-uribl.msg
t/corpus/spam-with-badly-parsed-email.msg
t/corpus/spam-with-image.msg
@@ -159,6 +160,7 @@ t/spam-rejection15.t
t/spam-rejection16.t
t/spam-rejection17.t
t/spam-rejection18.t
+t/spam-rejection19.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 fa0dfa4..df79094 100644
--- a/lib/Mail/DeliveryStatus/BounceParser.pm
+++ b/lib/Mail/DeliveryStatus/BounceParser.pm
@@ -1036,7 +1036,8 @@ sub _std_reason {
/Probable\s+Spam/i or
/ESETS_SMTP\s+\(spam\)/i or
/this\s+message\s+appears\s+to\s+be\s+spam/i or
- /Spam\s+score\s+\(\S+\)\s+too\s+high/i
+ /Spam\s+score\s+\(\S+\)\s+too\s+high/i or
+ /matches\s+a\s+profile\s+the\s+Internet\s+community\s+may\s+consider\s+spam/i
) {
return "spam";
}
diff --git a/t/corpus/spam-rejection19.msg b/t/corpus/spam-rejection19.msg
new file mode 100644
index 0000000..b34dc19
--- /dev/null
+++ b/t/corpus/spam-rejection19.msg
@@ -0,0 +1,98 @@
+Delivered-To: automated-bounces+5f3c8d9e-b852-11e0-a0dd-d22393a91d63@email.example.com
+Received: by 10.142.214.7 with SMTP id m7cs197072wfg;
+ Wed, 27 Jul 2011 06:14:40 -0700 (PDT)
+Received: by 10.216.229.86 with SMTP id g64mr2792952weq.50.1311772478844;
+ Wed, 27 Jul 2011 06:14:38 -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 i48si208064wed.60.2011.07.27.06.14.37;
+ Wed, 27 Jul 2011 06:14:37 -0700 (PDT)
+Received-SPF: pass (google.com: 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: domain of admin1.example.co.uk designates 94.236.45.212 as permitted sender) smtp.mail=
+Received: by admin1.example.co.uk (Postfix)
+ id CFE8A651069; Wed, 27 Jul 2011 14:14:36 +0100 (BST)
+Date: Wed, 27 Jul 2011 14:14:36 +0100 (BST)
+From: MAILER-DAEMON@admin1.example.co.uk (Mail Delivery System)
+Subject: Undelivered Mail Returned to Sender
+To: automated-bounces+5F3C8D9E-B852-11E0-A0DD-D22393A91D63@email.example.com
+Auto-Submitted: auto-replied
+MIME-Version: 1.0
+Content-Type: multipart/report; report-type=delivery-status;
+ boundary="9317465105F.1311772476/admin1.example.co.uk"
+Message-Id: <20110727131436.CFE8A651069@admin1.example.co.uk>
+
+This is a MIME-encapsulated message.
+
+--9317465105F.1311772476/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
+
+<ws@sample.net>: host inbound.registeredsite.com[66.97.46.122] said:
+ 554 5.7.1 The message from
+ (<automated-bounces+5F3C8D9E-B852-11E0-A0DD-D22393A91D63@email.example.com>)
+ with the subject of (foo) matches a profile
+ the Internet community may consider spam. Please revise your message before
+ resending. (in reply to end of DATA command)
+
+--9317465105F.1311772476/admin1.example.co.uk
+Content-Description: Delivery report
+Content-Type: message/delivery-status
+
+Reporting-MTA: dns; admin1.example.co.uk
+X-Postfix-Queue-ID: 9317465105F
+X-Postfix-Sender: rfc822; automated-bounces+5F3C8D9E-B852-11E0-A0DD-D22393A91D63@email.example.com
+Arrival-Date: Wed, 27 Jul 2011 14:14:34 +0100 (BST)
+
+Final-Recipient: rfc822; ws@sample.net
+Action: failed
+Status: 5.7.1
+Remote-MTA: dns; inbound.registeredsite.com
+Diagnostic-Code: smtp; 554 5.7.1 The message from
+ (<automated-bounces+5F3C8D9E-B852-11E0-A0DD-D22393A91D63@email.example.com>)
+ with the subject of (foo) matches a profile
+ the Internet community may consider spam. Please revise your message before
+ resending.
+
+--9317465105F.1311772476/admin1.example.co.uk
+Content-Description: Undelivered Message
+Content-Type: message/rfc822
+
+Received: by admin1.example.co.uk (Postfix, from userid 507)
+ id 9317465105F; Wed, 27 Jul 2011 14:14:34 +0100 (BST)
+MIME-Version: 1.0
+Content-Transfer-Encoding: 7bit
+Content-Type: multipart/related; boundary="_----------=_131177247421185112"
+X-Mailer: MIME::Lite 3.027 (F2.78; T1.31; A2.07; B3.13; Q3.13)
+Date: Wed, 27 Jul 2011 14:14:34 +0100
+From: Julia Stevens <julia.stevens@email.example.com>
+Subject: foo
+X-Campaignid: example20100819.2
+List-Unsubscribe: <http://www.exampleoffers.co.uk/unsubscribe.epl?uuid=5F3C8D9E-B852-11E0-A0DD-D22393A91D63>
+To: Fred Bloggs <ws@sample.net>
+Message-Id: <20110727131434.9317465105F@admin1.example.co.uk>
+
+This is a multi-part message in MIME format.
+
+--_----------=_131177247421185112
+Content-Disposition: inline
+Content-Transfer-Encoding: quoted-printable
+Content-Type: text/html
+
+HTML
+
+--_----------=_131177247421185112--
+
+
+--9317465105F.1311772476/admin1.example.co.uk--
diff --git a/t/spam-rejection19.t b/t/spam-rejection19.t
new file mode 100644
index 0000000..e890143
--- /dev/null
+++ b/t/spam-rejection19.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-rejection19.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