filter: begin work on a new filter API

This filter API should be independent of Email::Filter and
hopefully less intrusive to long running processes.

6 files changed, 312 insertions, 0 deletions

diff --git a/lib/PublicInbox/Filter/Base.pm b/lib/PublicInbox/Filter/Base.pm
new file mode 100644
index 00000000..0991e874
--- /dev/null
+++ b/lib/PublicInbox/Filter/Base.pm
@@ -0,0 +1,100 @@
+# Copyright (C) 2016 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
+#
+# base class for creating per-list or per-project filters
+package PublicInbox::Filter::Base;
+use strict;
+use warnings;
+use PublicInbox::MsgIter;
+use constant MAX_MID_SIZE => 244; # max term size - 1 in Xapian
+
+my $NO_HTML = '*** We only accept plain-text mail, no HTML ***';
+our %DEFAULTS = (
+        reject_suffix => [ qw(exe bat cmd com pif scr vbs cpl zip) ],
+        reject_type => [ "text/html:$NO_HTML", "text/xhtml:$NO_HTML",
+                'application/vnd.ms-*:No proprietary data formats' ],
+);
+our $INVALID_FN = qr/\0/;
+
+sub REJECT () { 100 }
+sub ACCEPT { scalar @_ > 1 ? $_[1] : 1 }
+sub IGNORE () { 0 }
+
+my %patmap = ('*' => '.*', '?' => '.', '[' => '[', ']' => ']');
+sub glob2pat {
+        my ($glob) = @_;
+        $glob =~ s!(.)!$patmap{$1} || "\Q$1"!ge;
+        $glob;
+}
+
+sub new {
+        my ($class, %opts) = @_;
+        my $self = bless { err => '', %opts }, $class;
+        foreach my $f (qw(reject_suffix reject_type)) {
+                # allow undef:
+                $self->{$f} = $DEFAULTS{$f} unless exists $self->{$f};
+        }
+        if (defined $self->{reject_suffix}) {
+                my $tmp = $self->{reject_suffix};
+                $tmp = join('|', map { glob2pat($_) } @$tmp);
+                $self->{reject_suffix} = qr/\.($tmp)\s*\z/i;
+        }
+        my $rt = [];
+        if (defined $self->{reject_type}) {
+                my $tmp = $self->{reject_type};
+                @$rt = map {
+                        my ($type, $msg) = split(':', $_, 2);
+                        $type = lc $type;
+                        $msg ||= "Unacceptable Content-Type: $type";
+                        my $re = glob2pat($type);
+                        [ qr/\b$re\b/i, $msg ];
+                } @$tmp;
+        }
+        $self->{reject_type} = $rt;
+        $self;
+}
+
+sub reject ($$) {
+        my ($self, $reason) = @_;
+        $self->{err} = $reason;
+        REJECT;
+}
+
+sub err ($) { $_[0]->{err} }
+
+# for MDA
+sub delivery {
+        my ($self, $mime) = @_;
+
+        my $rt = $self->{reject_type};
+        my $reject_suffix = $self->{reject_suffix} || $INVALID_FN;
+        my (%sfx, %type);
+
+        msg_iter($mime, sub {
+                my ($part, $depth, @idx) = @{$_[0]};
+
+                my $ct = $part->content_type || 'text/plain';
+                foreach my $p (@$rt) {
+                        if ($ct =~ $p->[0]) {
+                                $type{$p->[1]} = 1;
+                        }
+                }
+
+                my $fn = $part->filename;
+                if (defined($fn) && $fn =~ $reject_suffix) {
+                        $sfx{$1} = 1;
+                }
+        });
+
+        my @r;
+        if (keys %type) {
+                push @r, sort keys %type;
+        }
+        if (keys %sfx) {
+                push @r, 'Rejected suffixes(s): '.join(', ', sort keys %sfx);
+        }
+
+        @r ? $self->reject(join("\n", @r)) : $self->ACCEPT;
+}
+
+1;
diff --git a/lib/PublicInbox/Filter/Mirror.pm b/lib/PublicInbox/Filter/Mirror.pm
new file mode 100644
index 00000000..d9940889
--- /dev/null
+++ b/lib/PublicInbox/Filter/Mirror.pm
@@ -0,0 +1,12 @@
+# Copyright (C) 2016 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
+
+# Dumb filter for blindly accepting everything
+package PublicInbox::Filter::Mirror;
+use base qw(PublicInbox::Filter::Base);
+use strict;
+use warnings;
+
+sub delivery { $_[0]->ACCEPT };
+
+1;
diff --git a/lib/PublicInbox/Filter/Vger.pm b/lib/PublicInbox/Filter/Vger.pm
new file mode 100644
index 00000000..9498081a
--- /dev/null
+++ b/lib/PublicInbox/Filter/Vger.pm
@@ -0,0 +1,33 @@
+# Copyright (C) 2016 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
+
+# Filter for vger.kernel.org list trailer
+package PublicInbox::Filter::Vger;
+use base qw(PublicInbox::Filter::Base);
+use strict;
+use warnings;
+
+my $l0 = qr/-+/; # older messages only had one '-'
+my $l1 =
+ qr/To unsubscribe from this list: send the line "unsubscribe [\w-]+" in/;
+my $l2 = qr/the body of a message to majordomo\@vger\.kernel\.org/;
+my $l3 =
+  qr!More majordomo info at +http://vger\.kernel\.org/majordomo-info\.html!;
+
+# only LKML had this, and LKML nowadays has no list trailer since Jan 2016
+my $l4 = qr!Please read the FAQ at +http://www\.tux\.org/lkml/!;
+
+sub delivery {
+        my ($self, $mime) = @_;
+        my $s = $mime->as_string;
+
+        # the vger appender seems to only work on the raw string,
+        # so in multipart (e.g. GPG-signed) messages, the list trailer
+        # becomes invisible to MIME-aware email clients.
+        if ($s =~ s/$l0\n$l1\n$l2\n$l3\n($l4\n)?\z//os) {
+                $mime = Email::MIME->new(\$s);
+        }
+        $self->ACCEPT($mime);
+}
+
+1;
diff --git a/t/filter_base.t b/t/filter_base.t
new file mode 100644
index 00000000..ee5c7307
--- /dev/null
+++ b/t/filter_base.t
@@ -0,0 +1,81 @@
+# Copyright (C) 2016 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
+use strict;
+use warnings;
+use Test::More;
+use Email::MIME;
+use_ok 'PublicInbox::Filter::Base';
+
+{
+        my $f = PublicInbox::Filter::Base->new;
+        ok($f, 'created stock object');
+        ok(defined $f->{reject_suffix}, 'rejected suffix redefined');
+        is(ref($f->{reject_suffix}), 'Regexp', 'reject_suffix should be a RE');
+}
+
+{
+        my $f = PublicInbox::Filter::Base->new(reject_suffix => undef);
+        ok($f, 'created base object q/o reject_suffix');
+        ok(!defined $f->{reject_suffix}, 'reject_suffix not defined');
+}
+
+{
+        my $f = PublicInbox::Filter::Base->new;
+        my $html_body = "<html><body>hi</body></html>";
+        my $parts = [
+                Email::MIME->create(
+                        attributes => {
+                                content_type => 'text/xhtml; charset=UTF-8',
+                                encoding => 'base64',
+                        },
+                        body => $html_body,
+                ),
+                Email::MIME->create(
+                        attributes => {
+                                content_type => 'text/plain',
+                                encoding => 'quoted-printable',
+                        },
+                        body => 'hi = "bye"',
+                )
+        ];
+        my $email = Email::MIME->create(
+                header_str => [
+                  From => 'a@example.com',
+                  Subject => 'blah',
+                  'Content-Type' => 'multipart/alternative'
+                ],
+                parts => $parts,
+        );
+        is($f->delivery($email), 100, "xhtml rejected");
+}
+
+{
+        my $f = PublicInbox::Filter::Base->new;
+        my $parts = [
+                Email::MIME->create(
+                        attributes => {
+                                content_type => 'application/vnd.ms-excel',
+                                encoding => 'base64',
+                        },
+                        body => 'junk',
+                ),
+                Email::MIME->create(
+                        attributes => {
+                                content_type => 'text/plain',
+                                encoding => 'quoted-printable',
+                        },
+                        body => 'junk',
+                )
+        ];
+        my $email = Email::MIME->create(
+                header_str => [
+                  From => 'a@example.com',
+                  Subject => 'blah',
+                  'Content-Type' => 'multipart/mixed'
+                ],
+                parts => $parts,
+        );
+        is($f->delivery($email), 100, 'proprietary format rejected on glob');
+}
+
+done_testing();
diff --git a/t/filter_mirror.t b/t/filter_mirror.t
new file mode 100644
index 00000000..01be282e
--- /dev/null
+++ b/t/filter_mirror.t
@@ -0,0 +1,40 @@
+# Copyright (C) 2016 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
+use strict;
+use warnings;
+use Test::More;
+use Email::MIME;
+use_ok 'PublicInbox::Filter::Mirror';
+
+my $f = PublicInbox::Filter::Mirror->new;
+ok($f, 'created PublicInbox::Filter::Mirror object');
+{
+        my $html_body = "<html><body>hi</body></html>";
+        my $parts = [
+                Email::MIME->create(
+                        attributes => {
+                                content_type => 'text/html; charset=UTF-8',
+                                encoding => 'base64',
+                        },
+                        body => $html_body,
+                ),
+                Email::MIME->create(
+                        attributes => {
+                                content_type => 'text/plain',
+                                encoding => 'quoted-printable',
+                        },
+                        body => 'hi = "bye"',
+                )
+        ];
+        my $email = Email::MIME->create(
+                header_str => [
+                  From => 'a@example.com',
+                  Subject => 'blah',
+                  'Content-Type' => 'multipart/alternative'
+                ],
+                parts => $parts,
+        );
+        is($f->ACCEPT, $f->delivery($email), 'accept any trash that comes');
+}
+
+done_testing();
diff --git a/t/filter_vger.t b/t/filter_vger.t
new file mode 100644
index 00000000..83a4c9ee
--- /dev/null
+++ b/t/filter_vger.t
@@ -0,0 +1,46 @@
+# Copyright (C) 2016 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
+use strict;
+use warnings;
+use Test::More;
+use Email::MIME;
+use_ok 'PublicInbox::Filter::Vger';
+
+my $f = PublicInbox::Filter::Vger->new;
+ok($f, 'created PublicInbox::Filter::Vger object');
+{
+        my $lkml = <<'EOF';
+From: foo@example.com
+Subject: test
+
+keep this
+--
+To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
+the body of a message to majordomo@vger.kernel.org
+More majordomo info at  http://vger.kernel.org/majordomo-info.html
+Please read the FAQ at  http://www.tux.org/lkml/
+EOF
+
+        my $mime = Email::MIME->new($lkml);
+        $mime = $f->delivery($mime);
+        is("keep this\n", $mime->body, 'normal message filtered OK');
+}
+
+{
+        my $no_nl = <<'EOF';
+From: foo@example.com
+Subject: test
+
+OSX users :P--
+To unsubscribe from this list: send the line "unsubscribe git" in
+the body of a message to majordomo@vger.kernel.org
+More majordomo info at  http://vger.kernel.org/majordomo-info.html
+EOF
+
+        my $mime = Email::MIME->new($no_nl);
+        $mime = $f->delivery($mime);
+        is('OSX users :P', $mime->body, 'missing trailing LF in original OK');
+}
+
+
+done_testing();