From 7715c70e5a8667fbd9eade0cffb6ab05a714dd5d Mon Sep 17 00:00:00 2001
From: Eric Wong <e@80x24.org>
Date: Fri, 28 Dec 2018 19:17:36 +0000
Subject: add filter for gmane archives

Extracted from import_slrnspool, since some spools get converted
to mbox or what not.
---
 lib/PublicInbox/Filter/Gmane.pm | 41 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 41 insertions(+)
 create mode 100644 lib/PublicInbox/Filter/Gmane.pm

(limited to 'lib/PublicInbox')

diff --git a/lib/PublicInbox/Filter/Gmane.pm b/lib/PublicInbox/Filter/Gmane.pm
new file mode 100644
index 00000000..f9042dfc
--- /dev/null
+++ b/lib/PublicInbox/Filter/Gmane.pm
@@ -0,0 +1,41 @@
+# Copyright (C) 2018 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
+
+# Filter for importing some archives from gmane
+package PublicInbox::Filter::Gmane;
+use base qw(PublicInbox::Filter::Base);
+use strict;
+use warnings;
+
+sub scrub {
+	my ($self, $mime) = @_;
+	my $hdr = $mime->header_obj;
+
+	# gmane rewrites Received headers, which increases spamminess
+	# Some older archives set Original-To
+	foreach my $x (qw(Received To)) {
+		my @h = $hdr->header_raw("Original-$x");
+		if (@h) {
+			$hdr->header_set($x, @h);
+			$hdr->header_set("Original-$x");
+		}
+	}
+
+	# Approved triggers for the SA HEADER_SPAM rule,
+	# X-From is gmane specific
+	foreach my $drop (qw(Approved X-From)) {
+		$hdr->header_set($drop);
+	}
+
+	# appears to be an old gmane bug:
+	$hdr->header_set('connect()');
+
+	$self->ACCEPT($mime);
+}
+
+sub delivery {
+	my ($self, $mime) = @_;
+	$self->scrub($mime);
+}
+
+1;
-- 
cgit v1.2.3-24-ge0c7