Bug #69938 for Mail-DeliveryStatus-BounceParser: Another spam case

Another spam case. Patch attached, or https://gitorious.org/mail-deliverystatus- bounceparser-fork/mail-deliverystatus-bounceparser- fork/commit/56dd1b7e0f68efcf30860848aa6f998ed3848b2e

From 56dd1b7e0f68efcf30860848aa6f998ed3848b2e Mon Sep 17 00:00:00 2001 From: Michael Stevens <mstevens@etla.org> Date: Mon, 1 Aug 2011 14:30:10 +0100 Subject: [PATCH] Another spam detection case. --- MANIFEST | 2 + lib/Mail/DeliveryStatus/BounceParser.pm | 3 +- t/corpus/spam-rejection15.msg | 100 +++++++++++++++++++++++++++++++ t/spam-rejection15.t | 29 +++++++++ 4 files changed, 133 insertions(+), 1 deletions(-) create mode 100644 t/corpus/spam-rejection15.msg create mode 100644 t/spam-rejection15.t diff --git a/MANIFEST b/MANIFEST index d5b0bd0..5d446d1 100644 --- a/MANIFEST +++ b/MANIFEST @@ -83,6 +83,7 @@ t/corpus/spam-rejection11.msg t/corpus/spam-rejection12.msg t/corpus/spam-rejection13.msg t/corpus/spam-rejection14.msg +t/corpus/spam-rejection15.msg t/corpus/spam-rejection-uribl.msg t/corpus/spam-with-badly-parsed-email.msg t/corpus/spam-with-image.msg @@ -151,6 +152,7 @@ t/spam-rejection11.t t/spam-rejection12.t t/spam-rejection13.t t/spam-rejection14.t +t/spam-rejection15.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 afc21de..c2d7887 100644 --- a/lib/Mail/DeliveryStatus/BounceParser.pm +++ b/lib/Mail/DeliveryStatus/BounceParser.pm @@ -1032,7 +1032,8 @@ sub _std_reason { /detected\s+as\s+spam/i or /Denied\s+due\s+to\s+spam\s+list/i or /appears\s+to\s+be\s+unsolicited/i or - /antispam\s+checks/i + /antispam\s+checks/i or + /Probable\s+Spam/i ) { return "spam"; } diff --git a/t/corpus/spam-rejection15.msg b/t/corpus/spam-rejection15.msg new file mode 100644 index 0000000..d3c9814 --- /dev/null +++ b/t/corpus/spam-rejection15.msg @@ -0,0 +1,100 @@ +Delivered-To: automated-bounces+13a16098-bab5-11e0-b15e-be5c69b90385@email.example.com +Received: by 10.142.230.13 with SMTP id c13cs10840wfh; + Sat, 30 Jul 2011 07:06:14 -0700 (PDT) +Received: from mr.google.com ([10.227.58.130]) + by 10.227.58.130 with SMTP id g2mr4677769wbh.44.1312034774006 (num_hops = 1); + Sat, 30 Jul 2011 07:06:14 -0700 (PDT) +Received: by 10.227.58.130 with SMTP id g2mr3601178wbh.44.1312034773040; + Sat, 30 Jul 2011 07:06:13 -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 fm12si6029405wbb.50.2011.07.30.07.06.11; + Sat, 30 Jul 2011 07:06:11 -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 4774E6510F7; Sat, 30 Jul 2011 15:06:11 +0100 (BST) +Date: Sat, 30 Jul 2011 15:06:11 +0100 (BST) +From: MAILER-DAEMON@admin1.example.co.uk (Mail Delivery System) +Subject: Undelivered Mail Returned to Sender +To: automated-bounces+13A16098-BAB5-11E0-B15E-BE5C69B90385@email.example.com +Auto-Submitted: auto-replied +MIME-Version: 1.0 +Content-Type: multipart/report; report-type=delivery-status; + boundary="40F186510F5.1312034771/admin1.example.co.uk" Message-Id: <20110730140611.4774E6510F7@admin1.example.co.uk> + +This is a MIME-encapsulated message. + +--40F186510F5.1312034771/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 + +<gf@sample.co.uk>: host mailhost.example2.co.uk[88.208.238.130] said: 550 + Probable Spam (in reply to end of DATA command) + +--40F186510F5.1312034771/admin1.example.co.uk +Content-Description: Delivery report +Content-Type: message/delivery-status + +Reporting-MTA: dns; admin1.example.co.uk +X-Postfix-Queue-ID: 40F186510F5 +X-Postfix-Sender: rfc822; automated-bounces+13A16098-BAB5-11E0-B15E-BE5C69B90385@email.example.com +Arrival-Date: Sat, 30 Jul 2011 15:06:10 +0100 (BST) + +Final-Recipient: rfc822; gf@sample.co.uk +Action: failed +Status: 5.0.0 +Remote-MTA: dns; mailhost.example2.co.uk +Diagnostic-Code: smtp; 550 Probable Spam + +--40F186510F5.1312034771/admin1.example.co.uk +Content-Description: Undelivered Message +Content-Type: message/rfc822 + +Received: by admin1.example.co.uk (Postfix, from userid 507) + id 40F186510F5; Sat, 30 Jul 2011 15:06:10 +0100 (BST) +DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=email.example.com; + s=default10; t=1312034770; + bh=YhIgAd9h+zPrVAdaRozHBFzTKR7pEZ9DCmeceTMYjnM=; + h=MIME-Version:Content-Transfer-Encoding:Content-Type:Date:From: + Subject:List-Unsubscribe:To:Message-Id; + b=Llt2w50Y531qpYGT4W2k8ByIukLB1KyvRYRRqvAUQlGC7nF4v10OQ3JJ1xyTrelDI + nUrcAg6zmbj4rIfHALg4UUwwVHMmKT+ewzKRJNCmUqc+LYvyuKgtLOEdbexkMFSe5A + QeDxaqro0RQkMflhX+PA3BQm2HxGX3whjTe51DfY= +MIME-Version: 1.0 +Content-Transfer-Encoding: 7bit +Content-Type: multipart/related; boundary="_----------=_131203477022648100" +X-Mailer: MIME::Lite 3.027 (F2.78; T1.31; A2.07; B3.13; Q3.13) +Date: Sat, 30 Jul 2011 15:06:10 +0100 +From: Fred Bloggs <fred.bloggs@email.example.com> +Subject: subject +X-Campaignid: example20100819.2 +List-Unsubscribe: <http://www.exampleoffers.co.uk/unsubscribe.epl?uuid=13A16098-BAB5-11E0-B15E-BE5C69B90385> +To: John Sample <gf@sample.co.uk> +Message-Id: <20110730140610.40F186510F5@admin1.example.co.uk> + +This is a multi-part message in MIME format. + +--_----------=_131203477022648100 +Content-Disposition: inline +Content-Transfer-Encoding: quoted-printable +Content-Type: text/html + +HTML + +--_----------=_131203477022648100-- + + +--40F186510F5.1312034771/admin1.example.co.uk-- diff --git a/t/spam-rejection15.t b/t/spam-rejection15.t new file mode 100644 index 0000000..25efdc3 --- /dev/null +++ b/t/spam-rejection15.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-rejection15.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