mid: truncate excessively long MIDs early

Since we support duplicate MIDs in v2, we can safely truncate
long MID terms in the database and let other normal duplicate
resolution sort it out.  It seems only spammers use excessively
long MIDs, and there'll always be abuse/misuse vectors for causing
mis-threaded messages, so it's not worth worrying about
excessively long MIDs.

3 files changed, 16 insertions, 13 deletions

diff --git a/lib/PublicInbox/MID.pm b/lib/PublicInbox/MID.pm
index 96085399..422902f5 100644
--- a/lib/PublicInbox/MID.pm
+++ b/lib/PublicInbox/MID.pm
@@ -10,7 +10,10 @@ our @EXPORT_OK = qw/mid_clean id_compress mid2path mid_mime mid_escape MID_ESC
         mids references/;
 use URI::Escape qw(uri_escape_utf8);
 use Digest::SHA qw/sha1_hex/;
-use constant MID_MAX => 40; # SHA-1 hex length
+use constant {
+        MID_MAX => 40, # SHA-1 hex length # TODO: get rid of this
+        MAX_MID_SIZE => 244, # max term size (Xapian limitation) - length('Q')
+};
 
 sub mid_clean {
         my ($mid) = @_;
@@ -61,6 +64,12 @@ sub mids ($) {
                         push(@mids, $v);
                 }
         }
+        foreach my $i (0..$#mids) {
+                next if length($mids[$i]) <= MAX_MID_SIZE;
+                warn "Message-ID: <$mids[$i]> too long, truncating\n";
+                $mids[$i] = substr($mids[$i], 0, MAX_MID_SIZE);
+        }
+
         uniq_mids(\@mids);
 }
 
diff --git a/lib/PublicInbox/SearchIdx.pm b/lib/PublicInbox/SearchIdx.pm
index 3ef444d6..a70e1ebf 100644
--- a/lib/PublicInbox/SearchIdx.pm
+++ b/lib/PublicInbox/SearchIdx.pm
@@ -19,7 +19,6 @@ use POSIX qw(strftime);
 require PublicInbox::Git;
 
 use constant {
-        MAX_MID_SIZE => 244, # max term size (Xapian limitation) - length('Q')
         PERM_UMASK => 0,
         OLD_PERM_GROUP => 1,
         OLD_PERM_EVERYBODY => 2,
@@ -311,12 +310,6 @@ sub add_message {
         eval {
                 my $smsg = PublicInbox::SearchMsg->new($mime);
                 my $doc = $smsg->{doc};
-                foreach my $mid (@$mids) {
-                        # FIXME: may be abused to prevent archival
-                        length($mid) > MAX_MID_SIZE and
-                                die 'Message-ID too long';
-                        $doc->add_term('Q' . $mid);
-                }
                 my $subj = $smsg->subject;
                 my $xpath;
                 if ($subj ne '') {
@@ -392,9 +385,11 @@ sub add_message {
                                 }
                         }
                 }
+
                 if ($skel) {
                         push @values, $mids, $xpath, $data;
                         $skel->index_skeleton(\@values);
+                        $doc->add_boolean_term('Q' . $_) foreach @$mids;
                         $doc_id = $self->{xdb}->add_document($doc);
                 } else {
                         $doc_id = link_and_save($self, $doc, $mids, $refs,
@@ -469,9 +464,9 @@ sub parse_references ($) {
         my %mids = map { $_ => 1 } @{mids($hdr)};
         my @keep;
         foreach my $ref (@$refs) {
-                # FIXME: this is an archive-prevention vector like X-No-Archive
-                if (length($ref) > MAX_MID_SIZE) {
+                if (length($ref) > PublicInbox::MID::MAX_MID_SIZE) {
                         warn "References: <$ref> too long, ignoring\n";
+                        next;
                 }
                 next if $mids{$ref};
                 push @keep, $ref;
@@ -510,6 +505,8 @@ sub link_and_save {
         my $doc_id;
         $doc->add_boolean_term('XNUM' . $num) if defined $num;
         $doc->add_boolean_term('XPATH' . $xpath) if defined $xpath;
+        $doc->add_boolean_term('Q' . $_) foreach @$mids;
+
         my $vivified = 0;
         foreach my $mid (@$mids) {
                 $self->each_smsg_by_mid($mid, sub {
diff --git a/lib/PublicInbox/SearchIdxSkeleton.pm b/lib/PublicInbox/SearchIdxSkeleton.pm
index 4066b591..40b28c51 100644
--- a/lib/PublicInbox/SearchIdxSkeleton.pm
+++ b/lib/PublicInbox/SearchIdxSkeleton.pm
@@ -98,9 +98,6 @@ sub index_skeleton_real ($$) {
         my $ts = $values->[PublicInbox::Search::TS];
         my $smsg = PublicInbox::SearchMsg->new(undef);
         my $doc = $smsg->{doc};
-        foreach my $mid (@$mids) {
-                $doc->add_term('Q' . $mid);
-        }
         PublicInbox::SearchIdx::add_values($doc, $values);
         $doc->set_data($doc_data);
         $smsg->{ts} = $ts;