package PublicInbox::SearchMsg;
use strict;
use warnings;
-use Search::Xapian;
use PublicInbox::MID qw/mid_clean mid_mime/;
use PublicInbox::Address;
use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp);
+use Time::Local qw(timegm);
sub new {
my ($class, $mime) = @_;
my $doc = Search::Xapian::Document->new;
- $doc->add_boolean_term('T' . 'mail');
-
- bless { type => 'mail', doc => $doc, mime => $mime }, $class;
+ bless { doc => $doc, mime => $mime }, $class;
}
sub wrap {
bless { doc => $doc, mime => undef, mid => $mid }, $class;
}
+sub get {
+ my ($class, $head, $db, $mid) = @_;
+ my $doc_id = $head->get_docid;
+ my $doc = $db->get_document($doc_id);
+ load_expand(wrap($class, $doc, $mid))
+}
+
sub get_val ($$) {
my ($doc, $col) = @_;
Search::Xapian::sortable_unserialise($doc->get_value($col));
}
+sub to_doc_data {
+ my ($self, $oid, $mid0) = @_;
+ $oid = '' unless defined $oid;
+ join("\n",
+ $self->subject,
+ $self->from,
+ $self->references,
+ $self->to,
+ $self->cc,
+ $oid,
+ $mid0,
+ $self->{bytes} || '',
+ $self->{lines} || ''
+ );
+}
+
sub load_from_data ($$) {
my ($self) = $_[0]; # data = $_[1]
- my ($subj, $from, $refs, $to, $cc, $blob, $mid0) = split(/\n/, $_[1]);
- $self->{subject} = $subj;
- $self->{from} = $from;
- $self->{references} = $refs;
- $self->{to} = $to;
- $self->{cc} = $cc;
- $self->{blob} = $blob;
- $self->{mid} = $mid0;
+ (
+ $self->{subject},
+ $self->{from},
+ $self->{references},
+
+ # To: and Cc: are stored to optimize HDR/XHDR in NNTP since
+ # some NNTP clients will use that for message displays.
+ $self->{to},
+ $self->{cc},
+
+ $self->{blob},
+ $self->{mid},
+ $self->{bytes},
+ $self->{lines}
+ ) = split(/\n/, $_[1]);
}
sub load_expand {
my ($self) = @_;
my $doc = $self->{doc};
my $data = $doc->get_data or return;
- $self->{ts} = get_val($doc, &PublicInbox::Search::TS);
- $self->{ds} = get_val($doc, &PublicInbox::Search::DS);
+ $self->{ts} = get_val($doc, PublicInbox::Search::TS());
+ my $dt = get_val($doc, PublicInbox::Search::DT());
+ my ($yyyy, $mon, $dd, $hh, $mm, $ss) = unpack('A4A2A2A2A2A2', $dt);
+ $self->{ds} = timegm($ss, $mm, $hh, $dd, $mon - 1, $yyyy);
utf8::decode($data);
load_from_data($self, $data);
$self;
}
# :bytes and :lines metadata in RFC 3977
-sub bytes ($) { get_val($_[0]->{doc}, &PublicInbox::Search::BYTES) }
-sub lines ($) { get_val($_[0]->{doc}, &PublicInbox::Search::LINES) }
-sub num ($) {
- $_[0]->{num} ||= get_val($_[0]->{doc}, PublicInbox::Search::NUM)
-}
+sub bytes ($) { $_[0]->{bytes} }
+sub lines ($) { $_[0]->{lines} }
sub __hdr ($$) {
my ($self, $field) = @_;
my $mime = $self->{mime} or return;
$val = $mime->header($field);
$val = '' unless defined $val;
- $val =~ tr/\n/ /;
+ $val =~ tr/\t\n/ /;
$val =~ tr/\r//d;
$self->{$field} = $val;
}
$self->{ds} ||= eval { msg_datestamp($self->{mime}->header_obj); } || 0;
}
-sub to_doc_data {
- my ($self, $oid, $mid0) = @_;
- my @rows = ($self->subject, $self->from, $self->references,
- $self->to, $self->cc);
- $oid = '' unless defined $oid;
- join("\n", @rows, $oid, $mid0);
-}
-
sub references {
my ($self) = @_;
my $x = $self->{references};
$self->{mid} = $mid;
} elsif (my $rv = $self->{mid}) {
$rv;
+ } elsif ($self->{doc}) {
+ $self->{mid} = _get_term_val($self, 'Q', qr/\AQ/);
} else {
- $self->{mid} = _get_term_val($self, 'Q', qr/\AQ/) ||
- $self->_extract_mid;
+ $self->_extract_mid; # v1 w/o Xapian
}
}
sub _extract_mid { mid_clean(mid_mime($_[0]->{mime})) }
-sub thread_id {
- my ($self) = @_;
- my $tid = $self->{thread};
- return $tid if defined $tid;
- $self->{thread} = _get_term_val($self, 'G', qr/\AG/); # *G*roup
-}
-
-# XXX: consider removing this, we can phrase match subject
-sub path {
- my ($self) = @_;
- my $path = $self->{path};
- return $path if defined $path;
- $self->{path} = _get_term_val($self, 'XPATH', qr/\AXPATH/); # path
-}
-
-sub type {
- my ($self) = @_;
- my $type = $self->{type};
- return $type if defined $type;
- $self->{type} = _get_term_val($self, 'T', qr/\AT/);
-}
-
1;