www: favor reading more from SQLite, and less from Xapian

[public-inbox.git] / lib / PublicInbox / SearchIdx.pm

diff --git a/lib/PublicInbox/SearchIdx.pm b/lib/PublicInbox/SearchIdx.pm
index 3412a615b66c3bdc2fd2897ad3c5ac6b2010287f..3596972faa962ebec1160a51f1cbdbf9cd799b98 100644 (file)
--- a/lib/PublicInbox/SearchIdx.pm
+++ b/lib/PublicInbox/SearchIdx.pm
@@ -76,7 +76,8 @@ sub new {
        if ($version == 1) {
                $self->{lock_path} = "$mainrepo/ssoma.lock";
                my $dir = $self->xdir;
-               $self->{over} = PublicInbox::OverIdx->new("$dir/over.sqlite3");
+               $self->{over_ro} = $self->{over} =
+                               PublicInbox::OverIdx->new("$dir/over.sqlite3");
        } elsif ($version == 2) {
                defined $part or die "partition is required for v2\n";
                # partition is a number
@@ -282,9 +283,14 @@ sub add_message {
                $smsg->{lines} = $mime->body_raw =~ tr!\n!\n!;
                defined $bytes or $bytes = length($mime->as_string);
                $smsg->{bytes} = $bytes;
+
                add_val($doc, PublicInbox::Search::TS(), $smsg->ts);
-               my $yyyymmdd = strftime('%Y%m%d', gmtime($smsg->ds));
-               add_val($doc, PublicInbox::Search::YYYYMMDD, $yyyymmdd);
+               my @ds = gmtime($smsg->ds);
+               my $yyyymmdd = strftime('%Y%m%d', @ds);
+               add_val($doc, PublicInbox::Search::YYYYMMDD(), $yyyymmdd);
+               my $dt = strftime('%Y%m%d%H%M%S', @ds);
+               add_val($doc, PublicInbox::Search::DT(), $dt);
+               my @vals = ($smsg->{ts}, $smsg->{ds});
 
                my $tg = $self->term_generator;
 
@@ -334,7 +340,7 @@ sub add_message {
                });
 
                # populates smsg->references for smsg->to_doc_data
-               my $refs = parse_references($smsg);
+               my $refs = parse_references($smsg, $mid0, $mids);
                my $data = $smsg->to_doc_data($oid, $mid0);
                foreach my $mid (@$mids) {
                        $tg->index_text($mid, 1, 'XM');
@@ -355,7 +361,7 @@ sub add_message {
 
                utf8::encode($data);
                $data = compress($data);
-               my @vals = ($smsg->ts, $num, $mids, $refs, $xpath, $data);
+               push @vals, $num, $mids, $refs, $xpath, $data;
                $self->{over}->add_over(\@vals);
                $doc->add_boolean_term('Q' . $_) foreach @$mids;
                $doc->add_boolean_term('XNUM' . $num) if defined $num;
@@ -463,22 +469,23 @@ sub term_generator { # write-only
        $self->{term_generator} = $tg;
 }
 
-sub parse_references ($) {
-       my ($smsg) = @_;
+sub parse_references ($$$) {
+       my ($smsg, $mid0, $mids) = @_;
        my $mime = $smsg->{mime};
        my $hdr = $mime->header_obj;
        my $refs = references($hdr);
+       push(@$refs, @$mids) if scalar(@$mids) > 1;
        return $refs if scalar(@$refs) == 0;
 
-       # prevent circular references via References here:
-       my %mids = map { $_ => 1 } @{mids($hdr)};
+       # prevent circular references here:
+       my %seen = ( $mid0 => 1 );
        my @keep;
        foreach my $ref (@$refs) {
                if (length($ref) > PublicInbox::MID::MAX_MID_SIZE) {
                        warn "References: <$ref> too long, ignoring\n";
                        next;
                }
-               next if $mids{$ref};
+               next if $seen{$ref}++;
                push @keep, $ref;
        }
        $smsg->{references} = '<'.join('> <', @keep).'>' if @keep;