Redundant document data increases our database size, pull the
smsg->mid off the unique term, the smsg->ts off the value, and
only generate the formatted display date off smsg->ts.
package PublicInbox::Search;
use strict;
use warnings;
package PublicInbox::Search;
use strict;
use warnings;
-use PublicInbox::SearchMsg;
use Search::Xapian qw/:standard/;
use Search::Xapian qw/:standard/;
+use PublicInbox::SearchMsg;
use Email::MIME;
use PublicInbox::MID qw/mid_clean mid_compress/;
use Email::MIME;
use PublicInbox::MID qw/mid_clean mid_compress/;
our $LANG = 'english';
use constant {
our $LANG = 'english';
use constant {
# SCHEMA_VERSION history
# 0 - initial
# 1 - subject_path is lower-cased
# SCHEMA_VERSION history
# 0 - initial
# 1 - subject_path is lower-cased
# 5 - subject_path drops trailing '.'
# 6 - preserve References: order in document data
# 7 - remove references and inreplyto terms
# 5 - subject_path drops trailing '.'
# 6 - preserve References: order in document data
# 7 - remove references and inreplyto terms
+ # 8 - remove redundant/unneeded document data
+ SCHEMA_VERSION => 8,
QP_FLAGS => FLAG_PHRASE|FLAG_BOOLEAN|FLAG_LOVEHATE|FLAG_WILDCARD,
};
QP_FLAGS => FLAG_PHRASE|FLAG_BOOLEAN|FLAG_LOVEHATE|FLAG_WILDCARD,
};
$doc->add_term(xpfx('path') . mid_compress($path));
}
$doc->add_term(xpfx('path') . mid_compress($path));
}
- my $from = $smsg->from_name;
- my $date = $smsg->date;
my $ts = Search::Xapian::sortable_serialise($smsg->ts);
$doc->add_value(PublicInbox::Search::TS, $ts);
my $ts = Search::Xapian::sortable_serialise($smsg->ts);
$doc->add_value(PublicInbox::Search::TS, $ts);
use Encode qw/find_encoding/;
my $enc_utf8 = find_encoding('UTF-8');
our $PFX2TERM_RE = undef;
use Encode qw/find_encoding/;
my $enc_utf8 = find_encoding('UTF-8');
our $PFX2TERM_RE = undef;
+use constant EPOCH_822 => 'Thu, 01 Jan 1970 00:00:00 +0000';
sub new {
my ($class, $mime) = @_;
sub new {
my ($class, $mime) = @_;
sub load_doc {
my ($class, $doc) = @_;
my $data = $doc->get_data;
sub load_doc {
my ($class, $doc) = @_;
my $data = $doc->get_data;
+ my $ts = eval {
+ no strict 'subs';
+ $doc->get_value(PublicInbox::Search::TS);
+ };
+ $ts = Search::Xapian::sortable_unserialise($ts);
$data = $enc_utf8->decode($data);
$data = $enc_utf8->decode($data);
- my ($mid, $subj, $from, $date, $refs) = split(/\n/, $data);
+ my ($subj, $from, $refs) = split(/\n/, $data);
from_name => $from,
references_sorted => $refs,
}, $class;
from_name => $from,
references_sorted => $refs,
}, $class;
sub ts {
my ($self) = @_;
sub ts {
my ($self) = @_;
- my $ts = $self->{ts};
- return $ts if $ts;
- $self->{ts} = eval {
- str2time($self->date || $self->mime->header('Date'))
- } || 0;
-}
-
-sub date {
- my ($self) = @_;
- my $date = $self->{date};
- return $date if $date;
- my $ts = eval { str2time($self->mime->header('Date')) };
- $self->{date} = POSIX::strftime('%Y-%m-%d %H:%M', gmtime($ts));
+ $self->{ts} ||= eval { str2time($self->mime->header('Date')) } || 0;
}
sub to_doc_data {
my ($self) = @_;
}
sub to_doc_data {
my ($self) = @_;
PublicInbox::Search::subject_summary($self->subject) . "\n" .
$self->from_name . "\n".
PublicInbox::Search::subject_summary($self->subject) . "\n" .
$self->from_name . "\n".
$self->references_sorted;
}
$self->references_sorted;
}
my @h = (
Subject => $self->subject,
'X-PI-From' => $self->from_name,
my @h = (
Subject => $self->subject,
'X-PI-From' => $self->from_name,
- 'X-PI-Date' => $self->date,
'X-PI-TS' => $self->ts,
'Message-ID' => "<$self->{mid}>",
'X-PI-TS' => $self->ts,
'Message-ID' => "<$self->{mid}>",
+
+ # prevent Email::Simple::Creator from running,
+ # this header is useless for threading as we use X-PI-TS
+ # for sorting and display:
+ 'Date' => EPOCH_822,
);
my $refs = $self->{references_sorted};
my $mime = Email::MIME->create(header_str => \@h);
);
my $refs = $self->{references_sorted};
my $mime = Email::MIME->create(header_str => \@h);
- $mime->header_set('References', $refs) if (defined $refs);
+ my $h = $mime->header_obj;
+ $h->header_set('References', $refs) if (defined $refs);
+
+ # drop useless headers Email::MIME set for us
+ $h->header_set('Date');
+ $h->header_set('MIME-Version');
if (defined $mid) {
$self->{mid} = $mid;
if (defined $mid) {
$self->{mid} = $mid;
+ } elsif (my $rv = $self->{mid}) {
+ $rv;
} else {
$self->ensure_metadata; # needed for ghosts
$self->{mid} ||= $self->_extract_mid;
} else {
$self->ensure_metadata; # needed for ghosts
$self->{mid} ||= $self->_extract_mid;
$subj = "<u\nid=\"u\">$subj</u>";
}
$subj = "<u\nid=\"u\">$subj</u>";
}
- my $ts = $mime->header('X-PI-TS');
- unless (defined $ts) {
- $ts = msg_timestamp($mime);
- }
- $ts = POSIX::strftime('%Y-%m-%d %H:%M', gmtime($ts));
-
+ my $ts = _msg_date($mime);
my $rv = "<table\nsummary=l$level><tr>";
if ($level) {
$rv .= '<td><pre>' . (' ' x $level) . '</pre></td>';
my $rv = "<table\nsummary=l$level><tr>";
if ($level) {
$rv .= '<td><pre>' . (' ' x $level) . '</pre></td>';
+sub _msg_date {
+ my ($mime) = @_;
+ my $ts = $mime->header('X-PI-TS') || msg_timestamp($mime);
+ POSIX::strftime('%Y-%m-%d %H:%M', gmtime($ts));
+}
+
sub _inline_header {
my ($dst, $state, $mime, $level) = @_;
my $pfx = ' ' x $level;
sub _inline_header {
my ($dst, $state, $mime, $level) = @_;
my $pfx = ' ' x $level;
my $cur = $state->{cur};
my $mid = $mime->header('Message-ID');
my $f = $mime->header('X-PI-From');
my $cur = $state->{cur};
my $mid = $mime->header('Message-ID');
my $f = $mime->header('X-PI-From');
- my $d = $mime->header('X-PI-Date');
+ my $d = _msg_date($mime);
$f = PublicInbox::Hval->new($f);
$d = PublicInbox::Hval->new($d);
$f = $f->as_html;
$f = PublicInbox::Hval->new($f);
$d = PublicInbox::Hval->new($d);
$f = $f->as_html;