X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FContentId.pm;h=8d77934f20a51665f974da408d7195f26509ec15;hb=95bdac7f09c69036efed537a4d03d5bdd2ae4eb6;hp=8347de2d2846654ef2042c4d34aa6d4595e92dc1;hpb=9e9863aa5fb74358cd2b5960e4d8d16f1ee9fece;p=public-inbox.git

diff --git a/lib/PublicInbox/ContentId.pm b/lib/PublicInbox/ContentId.pm
index 8347de2d..8d77934f 100644
--- a/lib/PublicInbox/ContentId.pm
+++ b/lib/PublicInbox/ContentId.pm
@@ -1,18 +1,56 @@
-# Copyright (C) 2018 all contributors <meta@public-inbox.org>
+# Copyright (C) 2018-2020 all contributors <meta@public-inbox.org>
 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 
+# Unstable internal API.
+# Used for on-the-fly duplicate detection in V2 inboxes.
+# This is not stored in any database anywhere and may change
+# as changes in duplicate detection are needed.
+# See L<public-inbox-v2-format(5)> manpage for more details.
 package PublicInbox::ContentId;
 use strict;
 use warnings;
 use base qw/Exporter/;
 our @EXPORT_OK = qw/content_id content_digest/;
 use PublicInbox::MID qw(mids references);
+use PublicInbox::MsgIter;
 
 # not sure if less-widely supported hash families are worth bothering with
 use Digest::SHA;
 
-# Content-* headers are often no-ops, so maybe we don't need them
-my @ID_HEADERS = qw(Subject From Date To Cc);
+sub digest_addr ($$$) {
+	my ($dig, $h, $v) = @_;
+	$v =~ tr/"//d;
+	$v =~ s/@([a-z0-9\_\.\-\(\)]*([A-Z])\S*)/'@'.lc($1)/ge;
+	utf8::encode($v);
+	$dig->add("$h\0$v\0");
+}
+
+sub content_dig_i {
+	my ($dig) = $_[1];
+	my ($part, $depth, @idx) = @{$_[0]};
+	$dig->add("\0$depth:".join('.', @idx)."\0");
+	my $fn = $part->filename;
+	if (defined $fn) {
+		utf8::encode($fn);
+		$dig->add("fn\0$fn\0");
+	}
+	my @d = $part->header('Content-Description');
+	foreach my $d (@d) {
+		utf8::encode($d);
+		$dig->add("d\0$d\0");
+	}
+	$dig->add("b\0");
+	my $ct = $part->content_type || 'text/plain';
+	my ($s, undef) = msg_part_text($part, $ct);
+	if (defined $s) {
+		$s =~ s/\r\n/\n/gs;
+		$s =~ s/\s*\z//s;
+		utf8::encode($s);
+	} else {
+		$s = $part->body;
+	}
+	$dig->add($s);
+}
 
 sub content_digest ($) {
 	my ($mime) = @_;
@@ -22,20 +60,35 @@ sub content_digest ($) {
 	# References: and In-Reply-To: get used interchangeably
 	# in some "duplicates" in LKML.  We treat them the same
 	# in SearchIdx, so treat them the same for this:
-	my %seen;
-	foreach my $mid (@{mids($hdr)}) {
-		$dig->add('mid: '.$mid);
-		$seen{$mid} = 1;
-	}
+	# do NOT consider the Message-ID as part of the content_id
+	# if we got here, we've already got Message-ID reuse
+	my %seen = map { $_ => 1 } @{mids($hdr)};
 	foreach my $mid (@{references($hdr)}) {
-		next if $seen{$mid};
-		$dig->add('ref: '.$mid);
+		$dig->add("ref\0$mid\0") unless $seen{$mid}++;
+	}
+
+	# Only use Sender: if From is not present
+	foreach my $h (qw(From Sender)) {
+		my @v = $hdr->header($h);
+		if (@v) {
+			digest_addr($dig, $h, $_) foreach @v;
+		}
+	}
+	foreach my $h (qw(Subject Date)) {
+		my @v = $hdr->header($h);
+		foreach my $v (@v) {
+			utf8::encode($v);
+			$dig->add("$h\0$v\0");
+		}
 	}
-	foreach my $h (@ID_HEADERS) {
-		my @v = $hdr->header_raw($h);
-		$dig->add("$h: $_") foreach @v;
+	# Some mail processors will add " to unquoted names that were
+	# not in the original message.  For the purposes of deduplication,
+	# do not take it into account:
+	foreach my $h (qw(To Cc)) {
+		my @v = $hdr->header($h);
+		digest_addr($dig, $h, $_) foreach @v;
 	}
-	$dig->add($mime->body_raw);
+	msg_iter($mime, \&content_dig_i, $dig);
 	$dig;
 }