X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FSearchIdx.pm;h=8810fe76450dde64728d715e4e63f14b3c6bb741;hb=d73d783ab2cf14ba28ca63723223d8c85a68cdd5;hp=725bbd838075156f37b982384dd9bb5e077dab0f;hpb=defb6d12f5b7b63a8e95a2db2d1dff2bd8f9d83c;p=public-inbox.git
diff --git a/lib/PublicInbox/SearchIdx.pm b/lib/PublicInbox/SearchIdx.pm
index 725bbd83..eb620f44 100644
--- a/lib/PublicInbox/SearchIdx.pm
+++ b/lib/PublicInbox/SearchIdx.pm
@@ -1,4 +1,4 @@
-# Copyright (C) 2015-2018 all contributors
+# Copyright (C) 2015-2020 all contributors
# License: AGPL-3.0+
# based on notmuch, but with no concept of folders, files or flags
#
@@ -8,83 +8,74 @@
# This writes to the search index.
package PublicInbox::SearchIdx;
use strict;
-use warnings;
-use Fcntl qw(:flock :DEFAULT);
-use PublicInbox::MIME;
-use base qw(PublicInbox::Search);
-use PublicInbox::MID qw/mid_clean id_compress mid_mime mids references/;
+use v5.10.1;
+use parent qw(PublicInbox::Search PublicInbox::Lock Exporter);
+use PublicInbox::Eml;
+use PublicInbox::InboxWritable;
+use PublicInbox::MID qw(mids_for_index mids);
use PublicInbox::MsgIter;
+use PublicInbox::IdxStack;
use Carp qw(croak);
use POSIX qw(strftime);
-require PublicInbox::Git;
-
-use constant {
- PERM_UMASK => 0,
- OLD_PERM_GROUP => 1,
- OLD_PERM_EVERYBODY => 2,
- PERM_GROUP => 0660,
- PERM_EVERYBODY => 0664,
- BATCH_BYTES => 1_000_000,
- DEBUG => !!$ENV{DEBUG},
-};
-
-my %GIT_ESC = (
- a => "\a",
- b => "\b",
- f => "\f",
- n => "\n",
- r => "\r",
- t => "\t",
- v => "\013",
-);
-
-sub git_unquote ($) {
- my ($s) = @_;
- return $s unless ($s =~ /\A"(.*)"\z/);
- $s = $1;
- $s =~ s/\\([abfnrtv])/$GIT_ESC{$1}/g;
- $s =~ s/\\([0-7]{1,3})/chr(oct($1))/ge;
- $s;
-}
+use PublicInbox::OverIdx;
+use PublicInbox::Spawn qw(spawn nodatacow_dir);
+use PublicInbox::Git qw(git_unquote);
+use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp);
+our @EXPORT_OK = qw(crlf_adjust log2stack is_ancestor check_size);
+my $X = \%PublicInbox::Search::X;
+my ($DB_CREATE_OR_OPEN, $DB_OPEN);
+our $DB_NO_SYNC = 0;
+our $BATCH_BYTES = $ENV{XAPIAN_FLUSH_THRESHOLD} ? 0x7fffffff : 1_000_000;
+use constant DEBUG => !!$ENV{DEBUG};
+
+my $xapianlevels = qr/\A(?:full|medium)\z/;
+my $hex = '[a-f0-9]';
+my $OID = $hex .'{40,}';
sub new {
- my ($class, $ibx, $creat, $part) = @_;
- my $mainrepo = $ibx; # for "public-inbox-index" w/o entry in config
- my $git_dir = $mainrepo;
- my ($altid, $git);
- my $version = 1;
- if (ref $ibx) {
- $mainrepo = $ibx->{mainrepo};
- $altid = $ibx->{altid};
- $version = $ibx->{version} || 1;
- if ($altid) {
- require PublicInbox::AltId;
- $altid = [ map {
- PublicInbox::AltId->new($ibx, $_);
- } @$altid ];
+ my ($class, $ibx, $creat, $shard) = @_;
+ ref $ibx or die "BUG: expected PublicInbox::Inbox object: $ibx";
+ my $levels = qr/\A(?:full|medium|basic)\z/;
+ my $inboxdir = $ibx->{inboxdir};
+ my $version = $ibx->version;
+ my $indexlevel = 'full';
+ my $altid = $ibx->{altid};
+ if ($altid) {
+ require PublicInbox::AltId;
+ $altid = [ map { PublicInbox::AltId->new($ibx, $_); } @$altid ];
+ }
+ if ($ibx->{indexlevel}) {
+ if ($ibx->{indexlevel} =~ $levels) {
+ $indexlevel = $ibx->{indexlevel};
+ } else {
+ die("Invalid indexlevel $ibx->{indexlevel}\n");
}
- $git = $ibx->git;
- } else {
- $git = PublicInbox::Git->new($git_dir); # v1 only
}
- require Search::Xapian::WritableDatabase;
+ $ibx = PublicInbox::InboxWritable->new($ibx);
my $self = bless {
- mainrepo => $mainrepo,
- git => $git,
+ ibx => $ibx,
+ xpfx => $inboxdir, # for xpfx_init
-altid => $altid,
- version => $version,
+ ibx_ver => $version,
+ indexlevel => $indexlevel,
}, $class;
- my $perm = $self->_git_config_perm;
- my $umask = _umask_for($perm);
- $self->{umask} = $umask;
+ $self->xpfx_init;
+ $self->{-set_indexlevel_once} = 1 if $indexlevel eq 'medium';
+ if ($ibx->{-skip_docdata}) {
+ $self->{-set_skip_docdata_once} = 1;
+ $self->{-skip_docdata} = 1;
+ }
+ $ibx->umask_prepare;
if ($version == 1) {
- $self->{lock_path} = "$mainrepo/ssoma.lock";
+ $self->{lock_path} = "$inboxdir/ssoma.lock";
+ my $dir = $self->xdir;
+ $self->{oidx} = PublicInbox::OverIdx->new("$dir/over.sqlite3");
+ $self->{oidx}->{-no_fsync} = 1 if $ibx->{-no_fsync};
} elsif ($version == 2) {
- defined $part or die "partition is required for v2\n";
- # partition is a number or "all"
- $self->{partition} = $part;
+ defined $shard or die "shard is required for v2\n";
+ # shard is a number
+ $self->{shard} = $shard;
$self->{lock_path} = undef;
- $self->{msgmap_path} = "$mainrepo/msgmap.sqlite3";
} else {
die "unsupported inbox version=$version\n";
}
@@ -92,101 +83,113 @@ sub new {
$self;
}
-sub _xdb_release {
- my ($self) = @_;
- my $xdb = delete $self->{xdb} or croak 'not acquired';
- $xdb->close;
- _lock_release($self) if $self->{creat};
+sub need_xapian ($) { $_[0]->{indexlevel} =~ $xapianlevels }
+
+sub idx_release {
+ my ($self, $wake) = @_;
+ if (need_xapian($self)) {
+ my $xdb = delete $self->{xdb} or croak 'not acquired';
+ $xdb->close;
+ }
+ $self->lock_release($wake) if $self->{creat};
undef;
}
-sub _xdb_acquire {
+sub load_xapian_writable () {
+ return 1 if $X->{WritableDatabase};
+ PublicInbox::Search::load_xapian() or return;
+ my $xap = $PublicInbox::Search::Xap;
+ for (qw(Document TermGenerator WritableDatabase)) {
+ $X->{$_} = $xap.'::'.$_;
+ }
+ eval 'require '.$X->{WritableDatabase} or die;
+ *sortable_serialise = $xap.'::sortable_serialise';
+ $DB_CREATE_OR_OPEN = eval($xap.'::DB_CREATE_OR_OPEN()');
+ $DB_OPEN = eval($xap.'::DB_OPEN()');
+ my $ver = (eval($xap.'::major_version()') << 16) |
+ (eval($xap.'::minor_version()') << 8);
+ $DB_NO_SYNC = 0x4 if $ver >= 0x10400;
+ 1;
+}
+
+sub idx_acquire {
my ($self) = @_;
- croak 'already acquired' if $self->{xdb};
+ my $flag;
my $dir = $self->xdir;
- my $flag = Search::Xapian::DB_OPEN;
+ if (need_xapian($self)) {
+ croak 'already acquired' if $self->{xdb};
+ load_xapian_writable();
+ $flag = $self->{creat} ? $DB_CREATE_OR_OPEN : $DB_OPEN;
+ }
if ($self->{creat}) {
require File::Path;
- _lock_acquire($self);
- File::Path::mkpath($dir);
- $flag = Search::Xapian::DB_CREATE_OR_OPEN;
+ $self->lock_acquire;
+
+ # don't create empty Xapian directories if we don't need Xapian
+ my $is_shard = defined($self->{shard});
+ if (!-d $dir && (!$is_shard ||
+ ($is_shard && need_xapian($self)))) {
+ File::Path::mkpath($dir);
+ nodatacow_dir($dir);
+ $self->{-set_has_threadid_once} = 1;
+ }
}
- $self->{xdb} = Search::Xapian::WritableDatabase->new($dir, $flag);
-}
-
-# we only acquire the flock if creating or reindexing;
-# PublicInbox::Import already has the lock on its own.
-sub _lock_acquire {
- my ($self) = @_;
- croak 'already locked' if $self->{lockfh};
- my $lock_path = $self->{lock_path} or return;
- sysopen(my $lockfh, $lock_path, O_WRONLY|O_CREAT) or
- die "failed to open lock $lock_path: $!\n";
- flock($lockfh, LOCK_EX) or die "lock failed: $!\n";
- $self->{lockfh} = $lockfh;
-}
-
-sub _lock_release {
- my ($self) = @_;
- return unless $self->{lock_path};
- my $lockfh = delete $self->{lockfh} or croak 'not locked';
- flock($lockfh, LOCK_UN) or die "unlock failed: $!\n";
- close $lockfh or die "close failed: $!\n";
+ return unless defined $flag;
+ $flag |= $DB_NO_SYNC if $self->{ibx}->{-no_fsync};
+ my $xdb = eval { ($X->{WritableDatabase})->new($dir, $flag) };
+ croak "Failed opening $dir: $@" if $@;
+ $self->{xdb} = $xdb;
}
sub add_val ($$$) {
my ($doc, $col, $num) = @_;
- $num = Search::Xapian::sortable_serialise($num);
+ $num = sortable_serialise($num);
$doc->add_value($col, $num);
}
-sub add_values ($$) {
- my ($doc, $values) = @_;
-
- my $ts = $values->[PublicInbox::Search::TS];
- add_val($doc, PublicInbox::Search::TS, $ts);
-
- my $num = $values->[PublicInbox::Search::NUM];
- defined($num) and add_val($doc, PublicInbox::Search::NUM, $num);
-
- my $bytes = $values->[PublicInbox::Search::BYTES];
- defined($bytes) and add_val($doc, PublicInbox::Search::BYTES, $bytes);
-
- my $lines = $values->[PublicInbox::Search::LINES];
- add_val($doc, PublicInbox::Search::LINES, $lines);
+sub term_generator ($) { # write-only
+ my ($self) = @_;
- my $yyyymmdd = strftime('%Y%m%d', gmtime($ts));
- add_val($doc, PublicInbox::Search::YYYYMMDD, $yyyymmdd);
+ $self->{term_generator} //= do {
+ my $tg = $X->{TermGenerator}->new;
+ $tg->set_stemmer($self->stemmer);
+ $tg;
+ }
}
-sub index_users ($$) {
- my ($tg, $smsg) = @_;
+sub index_text ($$$$) {
+ my ($self, $text, $wdf_inc, $prefix) = @_;
+ my $tg = term_generator($self); # man Search::Xapian::TermGenerator
- my $from = $smsg->from;
- my $to = $smsg->to;
- my $cc = $smsg->cc;
+ if ($self->{indexlevel} eq 'full') {
+ $tg->index_text($text, $wdf_inc, $prefix);
+ $tg->increase_termpos;
+ } else {
+ $tg->index_text_without_positions($text, $wdf_inc, $prefix);
+ }
+}
- $tg->index_text($from, 1, 'A'); # A - author
- $tg->increase_termpos;
- $tg->index_text($to, 1, 'XTO') if $to ne '';
- $tg->increase_termpos;
- $tg->index_text($cc, 1, 'XCC') if $cc ne '';
- $tg->increase_termpos;
+sub index_headers ($$) {
+ my ($self, $smsg) = @_;
+ my @x = (from => 'A', # Author
+ subject => 'S', to => 'XTO', cc => 'XCC');
+ while (my ($field, $pfx) = splice(@x, 0, 2)) {
+ my $val = $smsg->{$field};
+ index_text($self, $val, 1, $pfx) if $val ne '';
+ }
}
sub index_diff_inc ($$$$) {
- my ($tg, $text, $pfx, $xnq) = @_;
+ my ($self, $text, $pfx, $xnq) = @_;
if (@$xnq) {
- $tg->index_text(join("\n", @$xnq), 1, 'XNQ');
- $tg->increase_termpos;
+ index_text($self, join("\n", @$xnq), 1, 'XNQ');
@$xnq = ();
}
- $tg->index_text($text, 1, $pfx);
- $tg->increase_termpos;
+ index_text($self, $text, 1, $pfx);
}
sub index_old_diff_fn {
- my ($tg, $seen, $fa, $fb, $xnq) = @_;
+ my ($self, $seen, $fa, $fb, $xnq) = @_;
# no renames or space support for traditional diffs,
# find the number of leading common paths to strip:
@@ -197,7 +200,7 @@ sub index_old_diff_fn {
$fb = join('/', @fb);
if ($fa eq $fb) {
unless ($seen->{$fa}++) {
- index_diff_inc($tg, $fa, 'XDFN', $xnq);
+ index_diff_inc($self, $fa, 'XDFN', $xnq);
}
return 1;
}
@@ -208,22 +211,18 @@ sub index_old_diff_fn {
}
sub index_diff ($$$) {
- my ($tg, $lines, $doc) = @_;
+ my ($self, $txt, $doc) = @_;
my %seen;
my $in_diff;
my @xnq;
my $xnq = \@xnq;
- foreach (@$lines) {
+ foreach (split(/\n/, $txt)) {
if ($in_diff && s/^ //) { # diff context
- index_diff_inc($tg, $_, 'XDFCTX', $xnq);
+ index_diff_inc($self, $_, 'XDFCTX', $xnq);
} elsif (/^-- $/) { # email signature begins
$in_diff = undef;
- } elsif (m!^diff --git ("?a/.+) ("?b/.+)\z!) {
- my ($fa, $fb) = ($1, $2);
- my $fn = (split('/', git_unquote($fa), 2))[1];
- $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq);
- $fn = (split('/', git_unquote($fb), 2))[1];
- $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq);
+ } elsif (m!^diff --git "?[^/]+/.+ "?[^/]+/.+\z!) {
+ # wait until "---" and "+++" to capture filenames
$in_diff = 1;
# traditional diff:
} elsif (m/^diff -(.+) (\S+) (\S+)$/) {
@@ -231,26 +230,28 @@ sub index_diff ($$$) {
push @xnq, $_;
# only support unified:
next unless $opt =~ /[uU]/;
- $in_diff = index_old_diff_fn($tg, \%seen, $fa, $fb,
+ $in_diff = index_old_diff_fn($self, \%seen, $fa, $fb,
$xnq);
- } elsif (m!^--- ("?a/.+)!) {
- my $fn = (split('/', git_unquote($1), 2))[1];
- $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq);
+ } elsif (m!^--- ("?[^/]+/.+)!) {
+ my $fn = $1;
+ $fn = (split('/', git_unquote($fn), 2))[1];
+ $seen{$fn}++ or index_diff_inc($self, $fn, 'XDFN', $xnq);
$in_diff = 1;
- } elsif (m!^\+\+\+ ("?b/.+)!) {
- my $fn = (split('/', git_unquote($1), 2))[1];
- $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq);
+ } elsif (m!^\+\+\+ ("?[^/]+/.+)!) {
+ my $fn = $1;
+ $fn = (split('/', git_unquote($fn), 2))[1];
+ $seen{$fn}++ or index_diff_inc($self, $fn, 'XDFN', $xnq);
$in_diff = 1;
} elsif (/^--- (\S+)/) {
$in_diff = $1;
push @xnq, $_;
} elsif (defined $in_diff && /^\+\+\+ (\S+)/) {
- $in_diff = index_old_diff_fn($tg, \%seen, $in_diff, $1,
- $xnq);
+ $in_diff = index_old_diff_fn($self, \%seen, $in_diff,
+ $1, $xnq);
} elsif ($in_diff && s/^\+//) { # diff added
- index_diff_inc($tg, $_, 'XDFB', $xnq);
+ index_diff_inc($self, $_, 'XDFB', $xnq);
} elsif ($in_diff && s/^-//) { # diff removed
- index_diff_inc($tg, $_, 'XDFA', $xnq);
+ index_diff_inc($self, $_, 'XDFA', $xnq);
} elsif (m!^index ([a-f0-9]+)\.\.([a-f0-9]+)!) {
my ($ba, $bb) = ($1, $2);
index_git_blob_id($doc, 'XDFPRE', $ba);
@@ -260,7 +261,7 @@ sub index_diff ($$$) {
# traditional diff w/o -p
} elsif (/^@@ (?:\S+) (?:\S+) @@\s*(\S+.*)$/) {
# hunk header context
- index_diff_inc($tg, $1, 'XDFHH', $xnq);
+ index_diff_inc($self, $1, 'XDFHH', $xnq);
# ignore the following lines:
} elsif (/^(?:dis)similarity index/ ||
/^(?:old|new) mode/ ||
@@ -271,7 +272,9 @@ sub index_diff ($$$) {
/^Binary files .* differ/) {
push @xnq, $_;
} elsif ($_ eq '') {
- $in_diff = undef;
+ # possible to be in diff context, some mail may be
+ # stripped by MUA or even GNU diff(1). "git apply"
+ # treats a bare "\n" as diff context, too
} else {
push @xnq, $_;
warn "non-diff line: $_\n" if DEBUG && $_ ne '';
@@ -279,264 +282,184 @@ sub index_diff ($$$) {
}
}
- $tg->index_text(join("\n", @xnq), 1, 'XNQ');
- $tg->increase_termpos;
+ index_text($self, join("\n", @xnq), 1, 'XNQ');
}
-sub index_body ($$$) {
- my ($tg, $lines, $doc) = @_;
- my $txt = join("\n", @$lines);
- if ($doc) {
- # does it look like a diff?
- if ($txt =~ /^(?:diff|---|\+\+\+) /ms) {
- $txt = undef;
- index_diff($tg, $lines, $doc);
+sub index_xapian { # msg_iter callback
+ my $part = $_[0]->[0]; # ignore $depth and $idx
+ my ($self, $doc) = @{$_[1]};
+ my $ct = $part->content_type || 'text/plain';
+ my $fn = $part->filename;
+ if (defined $fn && $fn ne '') {
+ index_text($self, $fn, 1, 'XFN');
+ }
+ if ($part->{is_submsg}) {
+ my $mids = mids_for_index($part);
+ index_ids($self, $doc, $part, $mids);
+ my $smsg = bless {}, 'PublicInbox::Smsg';
+ $smsg->populate($part);
+ index_headers($self, $smsg);
+ }
+
+ my ($s, undef) = msg_part_text($part, $ct);
+ defined $s or return;
+ $_[0]->[0] = $part = undef; # free memory
+
+ # split off quoted and unquoted blocks:
+ my @sections = PublicInbox::MsgIter::split_quotes($s);
+ undef $s; # free memory
+ for my $txt (@sections) {
+ if ($txt =~ /\A>/) {
+ index_text($self, $txt, 0, 'XQUOT');
} else {
- $tg->index_text($txt, 1, 'XNQ');
+ # does it look like a diff?
+ if ($txt =~ /^(?:diff|---|\+\+\+) /ms) {
+ index_diff($self, $txt, $doc);
+ } else {
+ index_text($self, $txt, 1, 'XNQ');
+ }
}
- } else {
- $tg->index_text($txt, 0, 'XQUOT');
+ undef $txt; # free memory
}
- $tg->increase_termpos;
- @$lines = ();
}
-sub add_message {
- # mime = Email::MIME object
- my ($self, $mime, $bytes, $num, $oid, $mid0) = @_;
- my $doc_id;
- my $mids = mids($mime->header_obj);
- my $skel = $self->{skeleton};
+sub index_ids ($$$$) {
+ my ($self, $doc, $hdr, $mids) = @_;
+ for my $mid (@$mids) {
+ index_text($self, $mid, 1, 'XM');
- eval {
- my $smsg = PublicInbox::SearchMsg->new($mime);
- my $doc = $smsg->{doc};
- my $subj = $smsg->subject;
- my $xpath;
- if ($subj ne '') {
- $xpath = $self->subject_path($subj);
- $xpath = id_compress($xpath);
+ # because too many Message-IDs are prefixed with
+ # "Pine.LNX."...
+ if ($mid =~ /\w{12,}/) {
+ my @long = ($mid =~ /(\w{3,}+)/g);
+ index_text($self, join(' ', @long), 1, 'XM');
}
+ }
+ $doc->add_boolean_term('Q' . $_) for @$mids;
+ for my $l ($hdr->header_raw('List-Id')) {
+ $l =~ /<([^>]+)>/ or next;
+ my $lid = $1;
+ $doc->add_boolean_term('G' . $lid);
+ index_text($self, $lid, 1, 'XL'); # probabilistic
+ }
+}
- my $lines = $mime->body_raw =~ tr!\n!\n!;
- my @values = ($smsg->ts, $num, $bytes, $lines);
- add_values($doc, \@values);
-
- my $tg = $self->term_generator;
-
- $tg->set_document($doc);
- $tg->index_text($subj, 1, 'S') if $subj;
- $tg->increase_termpos;
-
- index_users($tg, $smsg);
-
- msg_iter($mime, sub {
- my ($part, $depth, @idx) = @{$_[0]};
- my $ct = $part->content_type || 'text/plain';
- my $fn = $part->filename;
- if (defined $fn && $fn ne '') {
- $tg->index_text($fn, 1, 'XFN');
- }
-
- return if $ct =~ m!\btext/x?html\b!i;
-
- my $s = eval { $part->body_str };
- if ($@) {
- if ($ct =~ m!\btext/plain\b!i) {
- # Try to assume UTF-8 because Alpine
- # seems to do wacky things and set
- # charset=X-UNKNOWN
- $part->charset_set('UTF-8');
- $s = eval { $part->body_str };
- $s = $part->body if $@;
- }
- }
- defined $s or return;
-
- my (@orig, @quot);
- my $body = $part->body;
- my @lines = split(/\n/, $body);
- while (defined(my $l = shift @lines)) {
- if ($l =~ /^>/) {
- index_body($tg, \@orig, $doc) if @orig;
- push @quot, $l;
- } else {
- index_body($tg, \@quot, 0) if @quot;
- push @orig, $l;
- }
- }
- index_body($tg, \@quot, 0) if @quot;
- index_body($tg, \@orig, $doc) if @orig;
- });
-
- # populates smsg->references for smsg->to_doc_data
- my $refs = parse_references($smsg);
- $mid0 = $mids->[0] unless defined $mid0;
- my $data = $smsg->to_doc_data($oid, $mid0);
- foreach my $mid (@$mids) {
- $tg->index_text($mid, 1, 'XM');
- }
+sub add_xapian ($$$$) {
+ my ($self, $eml, $smsg, $mids) = @_;
+ my $doc = $X->{Document}->new;
+ add_val($doc, PublicInbox::Search::TS(), $smsg->{ts});
+ 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);
+ add_val($doc, PublicInbox::Search::BYTES(), $smsg->{bytes});
+ add_val($doc, PublicInbox::Search::UID(), $smsg->{num});
+ add_val($doc, PublicInbox::Search::THREADID, $smsg->{tid});
+
+ my $tg = term_generator($self);
+ $tg->set_document($doc);
+ index_headers($self, $smsg);
+
+ msg_iter($eml, \&index_xapian, [ $self, $doc ]);
+ index_ids($self, $doc, $eml, $mids);
+
+ # by default, we maintain compatibility with v1.5.0 and earlier
+ # by writing to docdata.glass, users who never exect to downgrade can
+ # use --skip-docdata
+ if (!$self->{-skip_docdata}) {
+ # WWW doesn't need {to} or {cc}, only NNTP
+ $smsg->{to} = $smsg->{cc} = '';
+ PublicInbox::OverIdx::parse_references($smsg, $eml, $mids);
+ my $data = $smsg->to_doc_data;
$doc->set_data($data);
- if (my $altid = $self->{-altid}) {
- foreach my $alt (@$altid) {
- my $pfx = $alt->{xprefix};
- foreach my $mid (@$mids) {
- my $id = $alt->mid2alt($mid);
- next unless defined $id;
- $doc->add_boolean_term($pfx . $id);
- }
+ }
+
+ if (my $altid = $self->{-altid}) {
+ foreach my $alt (@$altid) {
+ my $pfx = $alt->{xprefix};
+ foreach my $mid (@$mids) {
+ my $id = $alt->mid2alt($mid);
+ next unless defined $id;
+ $doc->add_boolean_term($pfx . $id);
}
}
+ }
+ $self->{xdb}->replace_document($smsg->{num}, $doc);
+}
- 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,
- $num, $xpath);
- }
+sub _msgmap_init ($) {
+ my ($self) = @_;
+ die "BUG: _msgmap_init is only for v1\n" if $self->{ibx_ver} != 1;
+ $self->{mm} //= eval {
+ require PublicInbox::Msgmap;
+ my $rw = $self->{ibx}->{-no_fsync} ? 2 : 1;
+ PublicInbox::Msgmap->new($self->{ibx}->{inboxdir}, $rw);
};
-
- if ($@) {
- warn "failed to index message <".join('> <',@$mids).">: $@\n";
- return undef;
- }
- $doc_id;
}
-# returns deleted doc_id on success, undef on missing
-sub remove_message {
- my ($self, $mid) = @_;
- my $db = $self->{xdb};
- my $doc_id;
- $mid = mid_clean($mid);
+sub add_message {
+ # mime = PublicInbox::Eml or Email::MIME object
+ my ($self, $mime, $smsg, $sync) = @_;
+ my $mids = mids_for_index($mime);
+ $smsg //= bless { blob => '' }, 'PublicInbox::Smsg'; # test-only compat
+ $smsg->{mid} //= $mids->[0]; # v1 compatibility
+ $smsg->{num} //= do { # v1
+ _msgmap_init($self);
+ index_mm($self, $mime, $smsg->{blob}, $sync);
+ };
+
+ # v1 and tests only:
+ $smsg->populate($mime, $sync);
+ $smsg->{bytes} //= length($mime->as_string);
eval {
- my ($head, $tail) = $self->find_doc_ids('Q' . $mid);
- if ($head->equal($tail)) {
- warn "cannot remove non-existent <$mid>\n";
+ # order matters, overview stores every possible piece of
+ # data in doc_data (deflated). Xapian only stores a subset
+ # of the fields which exist in over.sqlite3. We may stop
+ # storing doc_data in Xapian sometime after we get multi-inbox
+ # search working.
+ if (my $oidx = $self->{oidx}) { # v1 only
+ $oidx->add_overview($mime, $smsg);
}
- for (; $head != $tail; $head->inc) {
- my $docid = $head->get_docid;
- $db->delete_document($docid);
+ if (need_xapian($self)) {
+ add_xapian($self, $mime, $smsg, $mids);
}
};
if ($@) {
- warn "failed to remove message <$mid>: $@\n";
+ warn "failed to index message <".join('> <',@$mids).">: $@\n";
return undef;
}
- $doc_id;
-}
-
-sub term_generator { # write-only
- my ($self) = @_;
-
- my $tg = $self->{term_generator};
- return $tg if $tg;
-
- $tg = Search::Xapian::TermGenerator->new;
- $tg->set_stemmer($self->stemmer);
-
- $self->{term_generator} = $tg;
-}
-
-# increments last_thread_id counter
-# returns a 64-bit integer represented as a decimal string
-sub next_thread_id {
- my ($self) = @_;
- my $db = $self->{xdb};
- my $last_thread_id = int($db->get_metadata('last_thread_id') || 0);
-
- $db->set_metadata('last_thread_id', ++$last_thread_id);
-
- $last_thread_id;
+ $smsg->{num};
}
-sub parse_references ($) {
- my ($smsg) = @_;
- my $mime = $smsg->{mime};
- my $hdr = $mime->header_obj;
- my $refs = references($hdr);
- return $refs if scalar(@$refs) == 0;
-
- # prevent circular references via References here:
- my %mids = map { $_ => 1 } @{mids($hdr)};
- my @keep;
- foreach my $ref (@$refs) {
- if (length($ref) > PublicInbox::MID::MAX_MID_SIZE) {
- warn "References: <$ref> too long, ignoring\n";
+sub xdb_remove {
+ my ($self, $oid, @removed) = @_;
+ my $xdb = $self->{xdb} or return;
+ for my $num (@removed) {
+ my $doc = eval { $xdb->get_document($num) };
+ unless ($doc) {
+ warn "E: $@\n" if $@;
+ warn "E: #$num $oid missing in Xapian\n";
next;
}
- next if $mids{$ref};
- push @keep, $ref;
+ my $smsg = bless {}, 'PublicInbox::Smsg';
+ $smsg->load_expand($doc);
+ my $blob = $smsg->{blob} // '(unset)';
+ if ($blob eq $oid) {
+ $xdb->delete_document($num);
+ } else {
+ warn "E: #$num $oid != $blob in Xapian\n";
+ }
}
- $smsg->{references} = '<'.join('> <', @keep).'>' if @keep;
- \@keep;
}
-sub link_doc {
- my ($self, $doc, $refs, $old_tid) = @_;
- my $tid;
-
- if (@$refs) {
- # first ref *should* be the thread root,
- # but we can never trust clients to do the right thing
- my $ref = shift @$refs;
- $tid = resolve_mid_to_tid($self, $ref);
- merge_threads($self, $tid, $old_tid) if defined $old_tid;
-
- # the rest of the refs should point to this tid:
- foreach $ref (@$refs) {
- my $ptid = resolve_mid_to_tid($self, $ref);
- merge_threads($self, $tid, $ptid);
- }
- } else {
- $tid = defined $old_tid ? $old_tid : $self->next_thread_id;
- }
- $doc->add_boolean_term('G' . $tid);
- $tid;
-}
-
-sub link_and_save {
- my ($self, $doc, $mids, $refs, $num, $xpath) = @_;
- my $db = $self->{xdb};
- my $old_tid;
- 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 {
- my ($cur) = @_;
- my $type = $cur->type;
- my $cur_tid = $cur->thread_id;
- $old_tid = $cur_tid unless defined $old_tid;
- if ($type eq 'mail') {
- # do not break existing mail messages,
- # just merge the threads
- merge_threads($self, $old_tid, $cur_tid);
- return 1;
- }
- if ($type ne 'ghost') {
- die "<$mid> has a bad type: $type\n";
- }
- my $tid = link_doc($self, $doc, $refs, $old_tid);
- $old_tid = $tid unless defined $old_tid;
- $doc_id = $cur->{doc_id};
- $self->{xdb}->replace_document($doc_id, $doc);
- ++$vivified;
- 1;
- });
- }
- # not really important, but we return any vivified ghost docid, here:
- return $doc_id if defined $doc_id;
- link_doc($self, $doc, $refs, $old_tid);
- $self->{xdb}->add_document($doc);
+sub remove_by_oid {
+ my ($self, $oid, $num) = @_;
+ die "BUG: remove_by_oid is v2-only\n" if $self->{oidx};
+ $self->begin_txn_lazy;
+ xdb_remove($self, $oid, $num) if need_xapian($self);
}
sub index_git_blob_id {
@@ -549,349 +472,389 @@ sub index_git_blob_id {
}
}
-sub unindex_blob {
- my ($self, $mime) = @_;
- my $mid = eval { mid_clean(mid_mime($mime)) };
- $self->remove_message($mid) if defined $mid;
+# v1 only
+sub unindex_eml {
+ my ($self, $oid, $eml) = @_;
+ my $mids = mids($eml);
+ my $nr = 0;
+ my %tmp;
+ for my $mid (@$mids) {
+ my @removed = $self->{oidx}->remove_oid($oid, $mid);
+ $nr += scalar @removed;
+ $tmp{$_}++ for @removed;
+ }
+ if (!$nr) {
+ $mids = join('> <', @$mids);
+ warn "W: <$mids> missing for removal from overview\n";
+ }
+ while (my ($num, $nr) = each %tmp) {
+ warn "BUG: $num appears >1 times ($nr) for $oid\n" if $nr != 1;
+ }
+ if ($nr) {
+ $self->{mm}->num_delete($_) for (keys %tmp);
+ } else { # just in case msgmap and over.sqlite3 become desynched:
+ $self->{mm}->mid_delete($mids->[0]);
+ }
+ xdb_remove($self, $oid, keys %tmp) if need_xapian($self);
}
sub index_mm {
- my ($self, $mime) = @_;
- my $mid = mid_clean(mid_mime($mime));
+ my ($self, $mime, $oid, $sync) = @_;
+ my $mids = mids($mime);
my $mm = $self->{mm};
- my $num = $mm->mid_insert($mid);
- return $num if defined $num;
-
- # fallback to num_for since filters like RubyLang set the number
- $mm->num_for($mid);
-}
-
-sub unindex_mm {
- my ($self, $mime) = @_;
- $self->{mm}->mid_delete(mid_clean(mid_mime($mime)));
-}
-
-sub index_mm2 {
- my ($self, $mime, $bytes, $blob) = @_;
- my $num = $self->{mm}->num_for(mid_clean(mid_mime($mime)));
- add_message($self, $mime, $bytes, $num, $blob);
-}
-
-sub unindex_mm2 {
- my ($self, $mime) = @_;
- $self->{mm}->mid_delete(mid_clean(mid_mime($mime)));
- unindex_blob($self, $mime);
+ if ($sync->{reindex}) {
+ my $oidx = $self->{oidx};
+ for my $mid (@$mids) {
+ my ($num, undef) = $oidx->num_mid0_for_oid($oid, $mid);
+ return $num if defined $num;
+ }
+ $mm->num_for($mids->[0]) // $mm->mid_insert($mids->[0]);
+ } else {
+ # fallback to num_for since filters like RubyLang set the number
+ $mm->mid_insert($mids->[0]) // $mm->num_for($mids->[0]);
+ }
}
-sub index_both {
- my ($self, $mime, $bytes, $blob) = @_;
- my $num = index_mm($self, $mime);
- add_message($self, $mime, $bytes, $num, $blob);
+# returns the number of bytes to add if given a non-CRLF arg
+sub crlf_adjust ($) {
+ if (index($_[0], "\r\n") < 0) {
+ # common case is LF-only, every \n needs an \r;
+ # so favor a cheap tr// over an expensive m//g
+ $_[0] =~ tr/\n/\n/;
+ } else { # count number of '\n' w/o '\r', expensive:
+ scalar(my @n = ($_[0] =~ m/(?cat_async callback
+ my ($bref, $oid, $type, $size, $sync) = @_;
+ my ($nr, $max) = @$sync{qw(nr max)};
+ ++$$nr;
+ $$max -= $size;
+ $size += crlf_adjust($$bref);
+ my $smsg = bless { bytes => $size, blob => $oid }, 'PublicInbox::Smsg';
+ my $self = $sync->{sidx};
+ my $eml = PublicInbox::Eml->new($bref);
+ $smsg->{num} = index_mm($self, $eml, $oid, $sync) or
+ die "E: could not generate NNTP article number for $oid";
+ add_message($self, $eml, $smsg, $sync);
}
-sub do_cat_mail {
- my ($git, $blob, $sizeref) = @_;
- my $mime = eval {
- my $str = $git->cat_file($blob, $sizeref);
- # fixup bugs from import:
- $$str =~ s/\A[\r\n]*From [^\r\n]*\r?\n//s;
- PublicInbox::MIME->new($str);
- };
- $@ ? undef : $mime;
+sub unindex_both { # git->cat_async callback
+ my ($bref, $oid, $type, $size, $self) = @_;
+ unindex_eml($self, $oid, PublicInbox::Eml->new($bref));
}
+# called by public-inbox-index
sub index_sync {
- my ($self, $opts) = @_;
- with_umask($self, sub { $self->_index_sync($opts) });
+ my ($self, $opt) = @_;
+ delete $self->{lock_path} if $opt->{-skip_lock};
+ $self->{ibx}->with_umask(\&_index_sync, $self, $opt);
+ if ($opt->{reindex}) {
+ my %again = %$opt;
+ delete @again{qw(rethread reindex)};
+ index_sync($self, \%again);
+ }
}
-sub batch_adjust ($$$$) {
- my ($max, $bytes, $batch_cb, $latest) = @_;
- $$max -= $bytes;
- if ($$max <= 0) {
- $$max = BATCH_BYTES;
- $batch_cb->($latest, 1);
+sub check_size { # check_async cb for -index --max-size=...
+ my ($oid, $type, $size, $arg, $git) = @_;
+ (($type // '') eq 'blob') or die "E: bad $oid in $git->{git_dir}";
+ if ($size <= $arg->{max_size}) {
+ $git->cat_async($oid, $arg->{index_oid}, $arg);
+ } else {
+ warn "W: skipping $oid ($size > $arg->{max_size})\n";
}
}
-# only for v1
-sub rlog {
- my ($self, $log, $add_cb, $del_cb, $batch_cb) = @_;
- my $hex = '[a-f0-9]';
- my $h40 = $hex .'{40}';
- my $addmsg = qr!^:000000 100644 \S+ ($h40) A\t${hex}{2}/${hex}{38}$!;
- my $delmsg = qr!^:100644 000000 ($h40) \S+ D\t${hex}{2}/${hex}{38}$!;
- my $git = $self->{git};
- my $latest;
- my $bytes;
- my $max = BATCH_BYTES;
- local $/ = "\n";
- my $line;
- while (defined($line = <$log>)) {
- if ($line =~ /$addmsg/o) {
- my $blob = $1;
- my $mime = do_cat_mail($git, $blob, \$bytes) or next;
- batch_adjust(\$max, $bytes, $batch_cb, $latest);
- $add_cb->($self, $mime, $bytes, $blob);
- } elsif ($line =~ /$delmsg/o) {
- my $blob = $1;
- my $mime = do_cat_mail($git, $blob, \$bytes) or next;
- batch_adjust(\$max, $bytes, $batch_cb, $latest);
- $del_cb->($self, $mime);
- } elsif ($line =~ /^commit ($h40)/o) {
- $latest = $1;
+sub v1_checkpoint ($$;$) {
+ my ($self, $sync, $stk) = @_;
+ $self->{ibx}->git->check_async_wait;
+ $self->{ibx}->git->cat_async_wait;
+
+ # latest_cmt may be undef
+ my $newest = $stk ? $stk->{latest_cmt} : undef;
+ if ($newest) {
+ my $cur = $self->{mm}->last_commit || '';
+ if (need_update($self, $cur, $newest)) {
+ $self->{mm}->last_commit($newest);
}
+ } else {
+ ${$sync->{max}} = $self->{batch_bytes};
}
- $batch_cb->($latest, 0);
-}
-sub _msgmap_init {
- my ($self) = @_;
- $self->{mm} ||= eval {
- require PublicInbox::Msgmap;
- my $msgmap_path = $self->{msgmap_path};
- if (defined $msgmap_path) { # v2
- PublicInbox::Msgmap->new_file($msgmap_path, 1);
- } else {
- PublicInbox::Msgmap->new($self->{mainrepo}, 1);
+ $self->{mm}->{dbh}->commit;
+ if ($newest && need_xapian($self)) {
+ my $xdb = $self->{xdb};
+ my $cur = $xdb->get_metadata('last_commit');
+ if (need_update($self, $cur, $newest)) {
+ $xdb->set_metadata('last_commit', $newest);
}
- };
-}
-sub _git_log {
- my ($self, $range) = @_;
- $self->{git}->popen(qw/log --reverse --no-notes --no-color
- --raw -r --no-abbrev/, $range);
-}
-
-# indexes all unindexed messages
-sub _index_sync {
- my ($self, $opts) = @_;
- my $tip = $opts->{ref} || 'HEAD';
- my $reindex = $opts->{reindex};
- my ($mkey, $last_commit, $lx, $xlog);
- $self->{git}->batch_prepare;
- my $xdb = _xdb_acquire($self);
- $xdb->begin_transaction;
- do {
- $xlog = undef;
- $mkey = 'last_commit';
- $last_commit = $xdb->get_metadata('last_commit');
- $lx = $last_commit;
- if ($reindex) {
- $lx = '';
- $mkey = undef if $last_commit ne '';
+ # let SearchView know a full --reindex was done so it can
+ # generate ->has_threadid-dependent links
+ if ($sync->{reindex} && !ref($sync->{reindex})) {
+ my $n = $xdb->get_metadata('has_threadid');
+ $xdb->set_metadata('has_threadid', '1') if $n ne '1';
}
- $xdb->cancel_transaction;
- $xdb = _xdb_release($self);
-
- # ensure we leak no FDs to "git log"
- my $range = $lx eq '' ? $tip : "$lx..$tip";
- $xlog = _git_log($self, $range);
+ }
- $xdb = _xdb_acquire($self);
- $xdb->begin_transaction;
- } while ($xdb->get_metadata('last_commit') ne $last_commit);
+ $self->{oidx}->rethread_done($sync->{-opt}) if $newest; # all done
+ commit_txn_lazy($self);
+ $self->{ibx}->git->cleanup;
+ my $nr = ${$sync->{nr}};
+ idx_release($self, $nr);
+ # let another process do some work...
+ if (my $pr = $sync->{-opt}->{-progress}) {
+ $pr->("indexed $nr/$sync->{ntodo}\n") if $nr;
+ }
+ if (!$stk) { # more to come
+ begin_txn_lazy($self);
+ $self->{mm}->{dbh}->begin_work;
+ }
+}
- my $mm = _msgmap_init($self);
- my $dbh = $mm->{dbh} if $mm;
- my $mm_only;
- my $cb = sub {
- my ($commit, $more) = @_;
- if ($dbh) {
- $mm->last_commit($commit) if $commit;
- $dbh->commit;
- }
- if (!$mm_only) {
- $xdb->set_metadata($mkey, $commit) if $mkey && $commit;
- $xdb->commit_transaction;
- $xdb = _xdb_release($self);
+# only for v1
+sub process_stack {
+ my ($self, $sync, $stk) = @_;
+ my $git = $self->{ibx}->git;
+ my $max = $self->{batch_bytes};
+ my $nr = 0;
+ $sync->{nr} = \$nr;
+ $sync->{max} = \$max;
+ $sync->{sidx} = $self;
+
+ $self->{mm}->{dbh}->begin_work;
+ if (my @leftovers = keys %{delete($sync->{D}) // {}}) {
+ warn('W: unindexing '.scalar(@leftovers)." leftovers\n");
+ for my $oid (@leftovers) {
+ $oid = unpack('H*', $oid);
+ $git->cat_async($oid, \&unindex_both, $self);
}
- # let another process do some work... <
- if ($more) {
- if (!$mm_only) {
- $xdb = _xdb_acquire($self);
- $xdb->begin_transaction;
+ }
+ if ($sync->{max_size} = $sync->{-opt}->{max_size}) {
+ $sync->{index_oid} = \&index_both;
+ }
+ while (my ($f, $at, $ct, $oid) = $stk->pop_rec) {
+ if ($f eq 'm') {
+ my $arg = { %$sync, autime => $at, cotime => $ct };
+ if ($sync->{max_size}) {
+ $git->check_async($oid, \&check_size, $arg);
+ } else {
+ $git->cat_async($oid, \&index_both, $arg);
}
- $dbh->begin_work if $dbh;
+ v1_checkpoint($self, $sync) if $max <= 0;
+ } elsif ($f eq 'd') {
+ $git->cat_async($oid, \&unindex_both, $self);
}
- };
+ }
+ v1_checkpoint($self, $sync, $stk);
+}
- if ($mm) {
- $dbh->begin_work;
- my $lm = $mm->last_commit || '';
- if ($lm eq $lx) {
- # Common case is the indexes are synced,
- # we only need to run git-log once:
- rlog($self, $xlog, *index_both, *unindex_both, $cb);
- } else {
- # Uncommon case, msgmap and xapian are out-of-sync
- # do not care for performance (but git is fast :>)
- # This happens if we have to reindex Xapian since
- # msgmap is a frozen format and our Xapian format
- # is evolving.
- my $r = $lm eq '' ? $tip : "$lm..$tip";
-
- # first, ensure msgmap is up-to-date:
- my $mkey_prev = $mkey;
- $mkey = undef; # ignore xapian, for now
- my $mlog = _git_log($self, $r);
- $mm_only = 1;
- rlog($self, $mlog, *index_mm, *unindex_mm, $cb);
- $mm_only = $mlog = undef;
-
- # now deal with Xapian
- $mkey = $mkey_prev;
- $dbh = undef;
- rlog($self, $xlog, *index_mm2, *unindex_mm2, $cb);
- }
+sub log2stack ($$$$) {
+ my ($sync, $git, $range, $ibx) = @_;
+ my $D = $sync->{D}; # OID_BIN => NR (if reindexing, undef otherwise)
+ my ($add, $del);
+ if ($ibx->version == 1) {
+ my $path = $hex.'{2}/'.$hex.'{38}';
+ $add = qr!\A:000000 100644 \S+ ($OID) A\t$path$!;
+ $del = qr!\A:100644 000000 ($OID) \S+ D\t$path$!;
} else {
- # user didn't install DBD::SQLite and DBI
- rlog($self, $xlog, *add_message, *unindex_blob, $cb);
+ $del = qr!\A:\d{6} 100644 $OID ($OID) [AM]\td$!;
+ $add = qr!\A:\d{6} 100644 $OID ($OID) [AM]\tm$!;
}
-}
-# this will create a ghost as necessary
-sub resolve_mid_to_tid {
- my ($self, $mid) = @_;
- my $tid;
- $self->each_smsg_by_mid($mid, sub {
- my ($smsg) = @_;
- my $cur_tid = $smsg->thread_id;
- if (defined $tid) {
- merge_threads($self, $tid, $cur_tid);
- } else {
- $tid = $smsg->thread_id;
+ # Count the new files so they can be added newest to oldest
+ # and still have numbers increasing from oldest to newest
+ my $fh = $git->popen(qw(log --raw -r --pretty=tformat:%at-%ct-%H
+ --no-notes --no-color --no-renames --no-abbrev),
+ $range);
+ my ($at, $ct, $stk);
+ while (<$fh>) {
+ if (/\A([0-9]+)-([0-9]+)-($OID)$/o) {
+ ($at, $ct) = ($1 + 0, $2 + 0);
+ $stk //= PublicInbox::IdxStack->new($3);
+ } elsif (/$del/) {
+ my $oid = $1;
+ if ($D) { # reindex case
+ $D->{pack('H*', $oid)}++;
+ } else { # non-reindex case:
+ $stk->push_rec('d', $at, $ct, $oid);
+ }
+ } elsif (/$add/) {
+ my $oid = $1;
+ if ($D) {
+ my $oid_bin = pack('H*', $oid);
+ my $nr = --$D->{$oid_bin};
+ delete($D->{$oid_bin}) if $nr <= 0;
+
+ # nr < 0 (-1) means it never existed
+ $stk->push_rec('m', $at, $ct, $oid) if $nr < 0;
+ } else {
+ $stk->push_rec('m', $at, $ct, $oid);
+ }
}
- 1;
- });
- return $tid if defined $tid;
-
- $self->create_ghost($mid)->thread_id;
+ }
+ close $fh or die "git log failed: \$?=$?";
+ $stk //= PublicInbox::IdxStack->new;
+ $stk->read_prepare;
}
-sub create_ghost {
- my ($self, $mid) = @_;
-
- my $tid = $self->next_thread_id;
- my $doc = Search::Xapian::Document->new;
- $doc->add_boolean_term('Q' . $mid);
- $doc->add_boolean_term('G' . $tid);
- $doc->add_boolean_term('T' . 'ghost');
+sub prepare_stack ($$$) {
+ my ($self, $sync, $range) = @_;
+ my $git = $self->{ibx}->git;
- my $smsg = PublicInbox::SearchMsg->wrap($doc, $mid);
- $self->{xdb}->add_document($doc);
-
- $smsg;
+ if (index($range, '..') < 0) {
+ # don't show annoying git errors to users who run -index
+ # on empty inboxes
+ $git->qx(qw(rev-parse -q --verify), "$range^0");
+ return PublicInbox::IdxStack->new->read_prepare if $?;
+ }
+ $sync->{D} = $sync->{reindex} ? {} : undef; # OID_BIN => NR
+ log2stack($sync, $git, $range, $self->{ibx});
+}
+
+# --is-ancestor requires git 1.8.0+
+sub is_ancestor ($$$) {
+ my ($git, $cur, $tip) = @_;
+ return 0 unless $git->check($cur);
+ my $cmd = [ 'git', "--git-dir=$git->{git_dir}",
+ qw(merge-base --is-ancestor), $cur, $tip ];
+ my $pid = spawn($cmd);
+ waitpid($pid, 0) == $pid or die join(' ', @$cmd) .' did not finish';
+ $? == 0;
+}
+
+sub need_update ($$$) {
+ my ($self, $cur, $new) = @_;
+ my $git = $self->{ibx}->git;
+ return 1 if $cur && !is_ancestor($git, $cur, $new);
+ my $range = $cur eq '' ? $new : "$cur..$new";
+ chomp(my $n = $git->qx(qw(rev-list --count), $range));
+ ($n eq '' || $n > 0);
+}
+
+# The last git commit we indexed with Xapian or SQLite (msgmap)
+# This needs to account for cases where Xapian or SQLite is
+# out-of-date with respect to the other.
+sub _last_x_commit {
+ my ($self, $mm) = @_;
+ my $lm = $mm->last_commit || '';
+ my $lx = '';
+ if (need_xapian($self)) {
+ $lx = $self->{xdb}->get_metadata('last_commit') || '';
+ } else {
+ $lx = $lm;
+ }
+ # Use last_commit from msgmap if it is older or unset
+ if (!$lm || ($lx && $lm && is_ancestor($self->{ibx}->git, $lm, $lx))) {
+ $lx = $lm;
+ }
+ $lx;
}
-sub merge_threads {
- my ($self, $winner_tid, $loser_tid) = @_;
- return if $winner_tid == $loser_tid;
- my $db = $self->{xdb};
+sub reindex_from ($$) {
+ my ($reindex, $last_commit) = @_;
+ return $last_commit unless $reindex;
+ ref($reindex) eq 'HASH' ? $reindex->{from} : '';
+}
- my $batch_size = 1000; # don't let @ids grow too large to avoid OOM
- while (1) {
- my ($head, $tail) = $self->find_doc_ids('G' . $loser_tid);
- return if $head == $tail;
- my @ids;
- for (; $head != $tail && @ids < $batch_size; $head->inc) {
- push @ids, $head->get_docid;
- }
- foreach my $docid (@ids) {
- my $doc = $db->get_document($docid);
- $doc->remove_term('G' . $loser_tid);
- $doc->add_boolean_term('G' . $winner_tid);
- $db->replace_document($docid, $doc);
+# indexes all unindexed messages (v1 only)
+sub _index_sync {
+ my ($self, $opt) = @_;
+ my $tip = $opt->{ref} || 'HEAD';
+ my $git = $self->{ibx}->git;
+ $self->{batch_bytes} = $opt->{batch_size} // $BATCH_BYTES;
+ $git->batch_prepare;
+ my $pr = $opt->{-progress};
+ my $sync = { reindex => $opt->{reindex}, -opt => $opt };
+ my $xdb = $self->begin_txn_lazy;
+ $self->{oidx}->rethread_prepare($opt);
+ my $mm = _msgmap_init($self);
+ if ($sync->{reindex}) {
+ my $last = $mm->last_commit;
+ if ($last) {
+ $tip = $last;
+ } else {
+ # somebody just blindly added --reindex when indexing
+ # for the first time, allow it:
+ undef $sync->{reindex};
}
}
+ my $last_commit = _last_x_commit($self, $mm);
+ my $lx = reindex_from($sync->{reindex}, $last_commit);
+ my $range = $lx eq '' ? $tip : "$lx..$tip";
+ $pr->("counting changes\n\t$range ... ") if $pr;
+ my $stk = prepare_stack($self, $sync, $range);
+ $sync->{ntodo} = $stk ? $stk->num_records : 0;
+ $pr->("$sync->{ntodo}\n") if $pr; # continue previous line
+ process_stack($self, $sync, $stk);
}
-sub _read_git_config_perm {
- my ($self) = @_;
- my @cmd = qw(config);
- if ($self->{version} == 2) {
- push @cmd, "--file=$self->{mainrepo}/all.git/config";
- }
- my $fh = $self->{git}->popen(@cmd, 'core.sharedRepository');
- local $/ = "\n";
- my $perm = <$fh>;
- chomp $perm if defined $perm;
- $perm;
+sub DESTROY {
+ # order matters for unlocking
+ $_[0]->{xdb} = undef;
+ $_[0]->{lockfh} = undef;
}
-sub _git_config_perm {
- my $self = shift;
- my $perm = scalar @_ ? $_[0] : _read_git_config_perm($self);
- return PERM_GROUP if (!defined($perm) || $perm eq '');
- return PERM_UMASK if ($perm eq 'umask');
- return PERM_GROUP if ($perm eq 'group');
- if ($perm =~ /\A(?:all|world|everybody)\z/) {
- return PERM_EVERYBODY;
- }
- return PERM_GROUP if ($perm =~ /\A(?:true|yes|on|1)\z/);
- return PERM_UMASK if ($perm =~ /\A(?:false|no|off|0)\z/);
-
- my $i = oct($perm);
- return PERM_UMASK if ($i == PERM_UMASK);
- return PERM_GROUP if ($i == OLD_PERM_GROUP);
- return PERM_EVERYBODY if ($i == OLD_PERM_EVERYBODY);
+sub _begin_txn {
+ my ($self) = @_;
+ my $xdb = $self->{xdb} || idx_acquire($self);
+ $self->{oidx}->begin_lazy if $self->{oidx};
+ $xdb->begin_transaction if $xdb;
+ $self->{txn} = 1;
+ $xdb;
+}
- if (($i & 0600) != 0600) {
- die "core.sharedRepository mode invalid: ".
- sprintf('%.3o', $i) . "\nOwner must have permissions\n";
- }
- ($i & 0666);
+sub begin_txn_lazy {
+ my ($self) = @_;
+ $self->{ibx}->with_umask(\&_begin_txn, $self) if !$self->{txn};
}
-sub _umask_for {
- my ($perm) = @_; # _git_config_perm return value
- my $rv = $perm;
- return umask if $rv == 0;
+# store 'indexlevel=medium' in v2 shard=0 and v1 (only one shard)
+# This metadata is read by Admin::detect_indexlevel:
+sub set_metadata_once {
+ my ($self) = @_;
- # set +x bit if +r or +w were set
- $rv |= 0100 if ($rv & 0600);
- $rv |= 0010 if ($rv & 0060);
- $rv |= 0001 if ($rv & 0006);
- (~$rv & 0777);
-}
+ return if $self->{shard}; # only continue if undef or 0, not >0
+ my $xdb = $self->{xdb};
-sub with_umask {
- my ($self, $cb) = @_;
- my $old = umask $self->{umask};
- my $rv = eval { $cb->() };
- my $err = $@;
- umask $old;
- die $err if $err;
- $rv;
+ if (delete($self->{-set_has_threadid_once})) {
+ $xdb->set_metadata('has_threadid', '1');
+ }
+ if (delete($self->{-set_indexlevel_once})) {
+ my $level = $xdb->get_metadata('indexlevel');
+ if (!$level || $level ne 'medium') {
+ $xdb->set_metadata('indexlevel', 'medium');
+ }
+ }
+ if (delete($self->{-set_skip_docdata_once})) {
+ $xdb->get_metadata('skip_docdata') or
+ $xdb->set_metadata('skip_docdata', '1');
+ }
}
-sub DESTROY {
- # order matters for unlocking
- $_[0]->{xdb} = undef;
- $_[0]->{lockfh} = undef;
+sub _commit_txn {
+ my ($self) = @_;
+ if (my $xdb = $self->{xdb}) {
+ set_metadata_once($self);
+ $xdb->commit_transaction;
+ }
+ $self->{oidx}->commit_lazy if $self->{oidx};
}
-# remote_* subs are only used by SearchIdxPart and SearchIdxSkeleton
-sub remote_commit {
+sub commit_txn_lazy {
my ($self) = @_;
- print { $self->{w} } "commit\n" or die "failed to write commit: $!";
+ delete($self->{txn}) and
+ $self->{ibx}->with_umask(\&_commit_txn, $self);
}
-sub remote_close {
+sub worker_done {
my ($self) = @_;
- my $pid = delete $self->{pid} or die "no process to wait on\n";
- my $w = delete $self->{w} or die "no pipe to write to\n";
- print $w "close\n" or die "failed to write to pid:$pid: $!\n";
- close $w or die "failed to close pipe for pid:$pid: $!\n";
- waitpid($pid, 0) == $pid or die "remote process did not finish";
- $? == 0 or die ref($self)." pid:$pid exited with: $?";
+ if (need_xapian($self)) {
+ die "$$ $0 xdb not released\n" if $self->{xdb};
+ }
+ die "$$ $0 still in transaction\n" if $self->{txn};
}
1;