1 # Copyright (C) 2015-2020 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
3 # based on notmuch, but with no concept of folders, files or flags
5 # Indexes mail with Xapian and our (SQLite-based) ::Msgmap for use
6 # with the web and NNTP interfaces. This index maintains thread
7 # relationships for use by PublicInbox::SearchThread.
8 # This writes to the search index.
9 package PublicInbox::SearchIdx;
12 use base qw(PublicInbox::Search PublicInbox::Lock);
13 use PublicInbox::MIME;
14 use PublicInbox::InboxWritable;
15 use PublicInbox::MID qw/mid_clean mid_mime mids_for_index/;
16 use PublicInbox::MsgIter;
18 use POSIX qw(strftime);
19 use PublicInbox::OverIdx;
20 use PublicInbox::Spawn qw(spawn);
21 use PublicInbox::Git qw(git_unquote);
22 use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp);
23 my $X = \%PublicInbox::Search::X;
24 my ($DB_CREATE_OR_OPEN, $DB_OPEN);
26 BATCH_BYTES => defined($ENV{XAPIAN_FLUSH_THRESHOLD}) ?
27 0x7fffffff : 1_000_000,
28 DEBUG => !!$ENV{DEBUG},
31 my $xapianlevels = qr/\A(?:full|medium)\z/;
34 my ($class, $ibx, $creat, $shard) = @_;
35 ref $ibx or die "BUG: expected PublicInbox::Inbox object: $ibx";
36 my $levels = qr/\A(?:full|medium|basic)\z/;
37 my $inboxdir = $ibx->{inboxdir};
38 my $version = $ibx->version;
39 my $indexlevel = 'full';
40 my $altid = $ibx->{altid};
42 require PublicInbox::AltId;
43 $altid = [ map { PublicInbox::AltId->new($ibx, $_); } @$altid ];
45 if ($ibx->{indexlevel}) {
46 if ($ibx->{indexlevel} =~ $levels) {
47 $indexlevel = $ibx->{indexlevel};
49 die("Invalid indexlevel $ibx->{indexlevel}\n");
52 $ibx = PublicInbox::InboxWritable->new($ibx);
54 inboxdir => $inboxdir,
59 indexlevel => $indexlevel,
61 $self->{-set_indexlevel_once} = 1 if $indexlevel eq 'medium';
64 $self->{lock_path} = "$inboxdir/ssoma.lock";
65 my $dir = $self->xdir;
66 $self->{over} = PublicInbox::OverIdx->new("$dir/over.sqlite3");
67 } elsif ($version == 2) {
68 defined $shard or die "shard is required for v2\n";
70 $self->{shard} = $shard;
71 $self->{lock_path} = undef;
73 die "unsupported inbox version=$version\n";
75 $self->{creat} = ($creat || 0) == 1;
79 sub need_xapian ($) { $_[0]->{indexlevel} =~ $xapianlevels }
83 if (need_xapian($self)) {
84 my $xdb = delete $self->{xdb} or croak 'not acquired';
87 $self->lock_release if $self->{creat};
91 sub load_xapian_writable () {
92 return 1 if $X->{WritableDatabase};
93 PublicInbox::Search::load_xapian() or return;
94 my $xap = $PublicInbox::Search::Xap;
95 for (qw(Document TermGenerator WritableDatabase)) {
96 $X->{$_} = $xap.'::'.$_;
98 eval 'require '.$X->{WritableDatabase} or die;
99 *sortable_serialise = $xap.'::sortable_serialise';
100 $DB_CREATE_OR_OPEN = eval($xap.'::DB_CREATE_OR_OPEN()');
101 $DB_OPEN = eval($xap.'::DB_OPEN()');
108 my $dir = $self->xdir;
109 if (need_xapian($self)) {
110 croak 'already acquired' if $self->{xdb};
111 load_xapian_writable();
112 $flag = $self->{creat} ? $DB_CREATE_OR_OPEN : $DB_OPEN;
114 if ($self->{creat}) {
118 # don't create empty Xapian directories if we don't need Xapian
119 my $is_shard = defined($self->{shard});
120 if (!$is_shard || ($is_shard && need_xapian($self))) {
121 File::Path::mkpath($dir);
124 return unless defined $flag;
125 my $xdb = eval { ($X->{WritableDatabase})->new($dir, $flag) };
127 die "Failed opening $dir: ", $@;
133 my ($doc, $col, $num) = @_;
134 $num = sortable_serialise($num);
135 $doc->add_value($col, $num);
138 sub term_generator ($) { # write-only
141 $self->{term_generator} //= do {
142 my $tg = $X->{TermGenerator}->new;
143 $tg->set_stemmer($self->stemmer);
148 sub index_text ($$$$) {
149 my ($self, $text, $wdf_inc, $prefix) = @_;
150 my $tg = term_generator($self); # man Search::Xapian::TermGenerator
152 if ($self->{indexlevel} eq 'full') {
153 $tg->index_text($text, $wdf_inc, $prefix);
154 $tg->increase_termpos;
156 $tg->index_text_without_positions($text, $wdf_inc, $prefix);
160 sub index_users ($$) {
161 my ($self, $smsg) = @_;
163 my $from = $smsg->from;
167 index_text($self, $from, 1, 'A'); # A - author
168 index_text($self, $to, 1, 'XTO') if $to ne '';
169 index_text($self, $cc, 1, 'XCC') if $cc ne '';
172 sub index_diff_inc ($$$$) {
173 my ($self, $text, $pfx, $xnq) = @_;
175 index_text($self, join("\n", @$xnq), 1, 'XNQ');
178 index_text($self, $text, 1, $pfx);
181 sub index_old_diff_fn {
182 my ($self, $seen, $fa, $fb, $xnq) = @_;
184 # no renames or space support for traditional diffs,
185 # find the number of leading common paths to strip:
186 my @fa = split('/', $fa);
187 my @fb = split('/', $fb);
188 while (scalar(@fa) && scalar(@fb)) {
189 $fa = join('/', @fa);
190 $fb = join('/', @fb);
192 unless ($seen->{$fa}++) {
193 index_diff_inc($self, $fa, 'XDFN', $xnq);
203 sub index_diff ($$$) {
204 my ($self, $txt, $doc) = @_;
209 foreach (split(/\n/, $txt)) {
210 if ($in_diff && s/^ //) { # diff context
211 index_diff_inc($self, $_, 'XDFCTX', $xnq);
212 } elsif (/^-- $/) { # email signature begins
214 } elsif (m!^diff --git "?[^/]+/.+ "?[^/]+/.+\z!) {
215 # wait until "---" and "+++" to capture filenames
218 } elsif (m/^diff -(.+) (\S+) (\S+)$/) {
219 my ($opt, $fa, $fb) = ($1, $2, $3);
221 # only support unified:
222 next unless $opt =~ /[uU]/;
223 $in_diff = index_old_diff_fn($self, \%seen, $fa, $fb,
225 } elsif (m!^--- ("?[^/]+/.+)!) {
227 $fn = (split('/', git_unquote($fn), 2))[1];
228 $seen{$fn}++ or index_diff_inc($self, $fn, 'XDFN', $xnq);
230 } elsif (m!^\+\+\+ ("?[^/]+/.+)!) {
232 $fn = (split('/', git_unquote($fn), 2))[1];
233 $seen{$fn}++ or index_diff_inc($self, $fn, 'XDFN', $xnq);
235 } elsif (/^--- (\S+)/) {
238 } elsif (defined $in_diff && /^\+\+\+ (\S+)/) {
239 $in_diff = index_old_diff_fn($self, \%seen, $in_diff,
241 } elsif ($in_diff && s/^\+//) { # diff added
242 index_diff_inc($self, $_, 'XDFB', $xnq);
243 } elsif ($in_diff && s/^-//) { # diff removed
244 index_diff_inc($self, $_, 'XDFA', $xnq);
245 } elsif (m!^index ([a-f0-9]+)\.\.([a-f0-9]+)!) {
246 my ($ba, $bb) = ($1, $2);
247 index_git_blob_id($doc, 'XDFPRE', $ba);
248 index_git_blob_id($doc, 'XDFPOST', $bb);
250 } elsif (/^@@ (?:\S+) (?:\S+) @@\s*$/) {
251 # traditional diff w/o -p
252 } elsif (/^@@ (?:\S+) (?:\S+) @@\s*(\S+.*)$/) {
253 # hunk header context
254 index_diff_inc($self, $1, 'XDFHH', $xnq);
255 # ignore the following lines:
256 } elsif (/^(?:dis)similarity index/ ||
257 /^(?:old|new) mode/ ||
258 /^(?:deleted|new) file mode/ ||
259 /^(?:copy|rename) (?:from|to) / ||
260 /^(?:dis)?similarity index / ||
261 /^\\ No newline at end of file/ ||
262 /^Binary files .* differ/) {
265 # possible to be in diff context, some mail may be
266 # stripped by MUA or even GNU diff(1). "git apply"
267 # treats a bare "\n" as diff context, too
270 warn "non-diff line: $_\n" if DEBUG && $_ ne '';
275 index_text($self, join("\n", @xnq), 1, 'XNQ');
278 sub index_body ($$$) {
279 my ($self, $txt, $doc) = @_;
281 # does it look like a diff?
282 if ($txt =~ /^(?:diff|---|\+\+\+) /ms) {
283 index_diff($self, $txt, $doc);
285 index_text($self, $txt, 1, 'XNQ');
288 index_text($self, $txt, 0, 'XQUOT');
292 sub index_xapian { # msg_iter callback
293 my ($part, $depth, @idx) = @{$_[0]};
294 my ($self, $doc) = @{$_[1]};
295 my $ct = $part->content_type || 'text/plain';
296 my $fn = $part->filename;
297 if (defined $fn && $fn ne '') {
298 index_text($self, $fn, 1, 'XFN');
301 my ($s, undef) = msg_part_text($part, $ct);
302 defined $s or return;
304 # split off quoted and unquoted blocks:
305 my @sections = split(/((?:^>[^\n]*\n)+)/sm, $s);
307 index_body($self, $_, /\A>/ ? 0 : $doc) for @sections;
310 sub add_xapian ($$$$) {
311 my ($self, $mime, $smsg, $mids) = @_;
312 $smsg->{mime} = $mime; # XXX dangerous
313 my $hdr = $mime->header_obj;
314 my $doc = $X->{Document}->new;
315 my $subj = $smsg->subject;
316 add_val($doc, PublicInbox::Search::TS(), $smsg->{ts});
317 my @ds = gmtime($smsg->{ds});
318 my $yyyymmdd = strftime('%Y%m%d', @ds);
319 add_val($doc, PublicInbox::Search::YYYYMMDD(), $yyyymmdd);
320 my $dt = strftime('%Y%m%d%H%M%S', @ds);
321 add_val($doc, PublicInbox::Search::DT(), $dt);
323 my $tg = term_generator($self);
325 $tg->set_document($doc);
326 index_text($self, $subj, 1, 'S') if $subj;
327 index_users($self, $smsg);
329 msg_iter($mime, \&index_xapian, [ $self, $doc ]);
330 foreach my $mid (@$mids) {
331 index_text($self, $mid, 1, 'XM');
333 # because too many Message-IDs are prefixed with
335 if ($mid =~ /\w{12,}/) {
336 my @long = ($mid =~ /(\w{3,}+)/g);
337 index_text($self, join(' ', @long), 1, 'XM');
340 $smsg->{to} = $smsg->{cc} = ''; # WWW doesn't need these, only NNTP
341 PublicInbox::OverIdx::parse_references($smsg, $hdr, $mids);
342 my $data = $smsg->to_doc_data;
343 $doc->set_data($data);
344 if (my $altid = $self->{-altid}) {
345 foreach my $alt (@$altid) {
346 my $pfx = $alt->{xprefix};
347 foreach my $mid (@$mids) {
348 my $id = $alt->mid2alt($mid);
349 next unless defined $id;
350 $doc->add_boolean_term($pfx . $id);
354 $doc->add_boolean_term('Q' . $_) foreach @$mids;
355 $self->{xdb}->replace_document($smsg->{num}, $doc);
358 sub _msgmap_init ($) {
360 die "BUG: _msgmap_init is only for v1\n" if $self->{ibx_ver} != 1;
361 $self->{mm} //= eval {
362 require PublicInbox::Msgmap;
363 PublicInbox::Msgmap->new($self->{inboxdir}, 1);
368 # mime = Email::MIME object
369 my ($self, $mime, $smsg) = @_;
370 my $hdr = $mime->header_obj;
371 my $mids = mids_for_index($hdr);
372 $smsg //= bless { blob => '' }, 'PublicInbox::Smsg'; # test-only compat
373 $smsg->{mid} //= $mids->[0]; # v1 compatibility
374 $smsg->{num} //= do { # v1
376 index_mm($self, $mime);
380 $smsg->{ds} //= msg_datestamp($hdr, $self->{autime});
381 $smsg->{ts} //= msg_timestamp($hdr, $self->{cotime});
384 # order matters, overview stores every possible piece of
385 # data in doc_data (deflated). Xapian only stores a subset
386 # of the fields which exist in over.sqlite3. We may stop
387 # storing doc_data in Xapian sometime after we get multi-inbox
389 if (my $over = $self->{over}) { # v1 only
390 $over->add_overview($mime, $smsg);
392 if (need_xapian($self)) {
393 add_xapian($self, $mime, $smsg, $mids);
398 warn "failed to index message <".join('> <',@$mids).">: $@\n";
404 # returns begin and end PostingIterator
406 my ($self, $termval) = @_;
407 my $db = $self->{xdb};
409 ($db->postlist_begin($termval), $db->postlist_end($termval));
414 my ($self, $termval, $cb) = @_;
415 my $batch_size = 1000; # don't let @ids grow too large to avoid OOM
417 my ($head, $tail) = $self->find_doc_ids($termval);
418 return if $head == $tail;
420 for (; $head != $tail && @ids < $batch_size; $head++) {
421 push @ids, $head->get_docid;
427 # v1 only, where $mid is unique
429 my ($self, $mid) = @_;
430 $mid = mid_clean($mid);
432 if (my $over = $self->{over}) {
433 my $nr = eval { $over->remove_oid(undef, $mid) };
435 warn "failed to remove <$mid> from overview: $@\n";
437 warn "<$mid> missing for removal from overview\n";
440 return unless need_xapian($self);
441 my $db = $self->{xdb};
444 batch_do($self, 'Q' . $mid, sub {
446 $db->delete_document($_) for @$ids;
451 warn "failed to remove <$mid> from Xapian: $@\n";
453 warn "<$mid> missing for removal from Xapian\n";
457 # MID is a hint in V2
459 my ($self, $oid, $mid) = @_;
461 $self->{over}->remove_oid($oid, $mid) if $self->{over};
463 return unless need_xapian($self);
464 my $db = $self->{xdb};
466 # XXX careful, we cannot use batch_do here since we conditionally
467 # delete documents based on other factors, so we cannot call
468 # find_doc_ids twice.
469 my ($head, $tail) = $self->find_doc_ids('Q' . $mid);
470 return if $head == $tail;
472 # there is only ONE element in @delete unless we
473 # have bugs in our v2writable deduplication check
475 for (; $head != $tail; $head++) {
476 my $docid = $head->get_docid;
477 my $doc = $db->get_document($docid);
478 my $smsg = PublicInbox::Smsg->wrap($mid);
479 $smsg->load_expand($doc);
480 if ($smsg->{blob} eq $oid) {
481 push(@delete, $docid);
484 $db->delete_document($_) foreach @delete;
488 sub index_git_blob_id {
489 my ($doc, $pfx, $objid) = @_;
491 my $len = length($objid);
492 for (my $len = length($objid); $len >= 7; ) {
493 $doc->add_term($pfx.$objid);
494 $objid = substr($objid, 0, --$len);
499 my ($self, $mime) = @_;
500 my $mid = eval { mid_clean(mid_mime($mime)) };
501 $self->remove_message($mid) if defined $mid;
505 my ($self, $mime) = @_;
506 my $mid = mid_clean(mid_mime($mime));
507 my $mm = $self->{mm};
510 if (defined $self->{regen_down}) {
511 $num = $mm->num_for($mid) and return $num;
513 while (($num = $self->{regen_down}--) > 0) {
514 if ($mm->mid_set($num, $mid) != 0) {
518 } elsif (defined $self->{regen_up}) {
519 $num = $mm->num_for($mid) and return $num;
521 # this is to fixup old bugs due to add-remove-add
522 while (($num = ++$self->{regen_up})) {
523 if ($mm->mid_set($num, $mid) != 0) {
529 $num = $mm->mid_insert($mid) and return $num;
531 # fallback to num_for since filters like RubyLang set the number
536 my ($self, $mime) = @_;
537 $self->{mm}->mid_delete(mid_clean(mid_mime($mime)));
541 my ($self, $mime, $smsg) = @_;
542 my $num = index_mm($self, $mime);
544 add_message($self, $mime, $smsg);
548 my ($self, $mime) = @_;
549 unindex_blob($self, $mime);
550 unindex_mm($self, $mime);
554 my ($git, $blob, $sizeref) = @_;
556 my $str = $git->cat_file($blob, $sizeref);
557 # fixup bugs from import:
558 $$str =~ s/\A[\r\n]*From [^\r\n]*\r?\n//s;
559 PublicInbox::MIME->new($str);
564 # called by public-inbox-index
566 my ($self, $opts) = @_;
567 delete $self->{lock_path} if $opts->{-skip_lock};
568 $self->{-inbox}->with_umask(sub { $self->_index_sync($opts) })
571 sub batch_adjust ($$$$$) {
572 my ($max, $bytes, $batch_cb, $latest, $nr) = @_;
576 $batch_cb->($nr, $latest);
582 my ($self, $log, $add_cb, $del_cb, $batch_cb) = @_;
583 my $hex = '[a-f0-9]';
584 my $h40 = $hex .'{40}';
585 my $addmsg = qr!^:000000 100644 \S+ ($h40) A\t${hex}{2}/${hex}{38}$!;
586 my $delmsg = qr!^:100644 000000 ($h40) \S+ D\t${hex}{2}/${hex}{38}$!;
587 my $git = $self->{git};
590 my $max = BATCH_BYTES;
596 while (defined($line = <$log>)) {
597 if ($line =~ /$addmsg/o) {
599 if (delete $D{$blob}) {
600 if (defined $self->{regen_down}) {
601 my $num = $self->{regen_down}--;
602 $self->{mm}->num_highwater($num);
606 my $mime = do_cat_mail($git, $blob, \$bytes) or next;
607 my $smsg = bless {}, 'PublicInbox::Smsg';
608 batch_adjust(\$max, $bytes, $batch_cb, $latest, ++$nr);
609 $smsg->{blob} = $blob;
610 $smsg->{bytes} = $bytes;
611 $add_cb->($self, $mime, $smsg);
612 } elsif ($line =~ /$delmsg/o) {
615 } elsif ($line =~ /^commit ($h40)/o) {
618 } elsif ($line =~ /^author .*? ([0-9]+) [\-\+][0-9]+$/) {
619 $self->{autime} = $1;
620 } elsif ($line =~ /^committer .*? ([0-9]+) [\-\+][0-9]+$/) {
621 $self->{cotime} = $1;
624 close($log) or die "git log failed: \$?=$?";
626 foreach my $blob (keys %D) {
627 my $mime = do_cat_mail($git, $blob, \$bytes) or next;
628 $del_cb->($self, $mime);
630 $batch_cb->($nr, $latest, $newest);
634 my ($self, $opts, $range) = @_;
635 my $git = $self->{git};
637 if (index($range, '..') < 0) {
638 # don't show annoying git errrors to users who run -index
640 $git->qx(qw(rev-parse -q --verify), "$range^0");
642 open my $fh, '<', '/dev/null' or
643 die "failed to open /dev/null: $!\n";
648 # Count the new files so they can be added newest to oldest
649 # and still have numbers increasing from oldest to newest
651 my $pr = $opts->{-progress};
652 $pr->("counting changes\n\t$range ... ") if $pr;
653 # can't use 'rev-list --count' if we use --diff-filter
654 my $fh = $git->popen(qw(log --pretty=tformat:%h
655 --no-notes --no-color --no-renames
656 --diff-filter=AM), $range);
657 ++$fcount while <$fh>;
658 close $fh or die "git log failed: \$?=$?";
659 my $high = $self->{mm}->num_highwater;
660 $pr->("$fcount\n") if $pr; # continue previous line
661 $self->{ntodo} = $fcount;
663 if (index($range, '..') < 0) {
664 if ($high && $high == $fcount) {
665 # fix up old bugs in full indexes which caused messages to
666 # not appear in Msgmap
667 $self->{regen_up} = $high;
669 # normal regen is for for fresh data
670 $self->{regen_down} = $fcount;
673 # Give oldest messages the smallest numbers
674 $self->{regen_down} = $high + $fcount;
677 $git->popen(qw/log --pretty=raw --no-notes --no-color --no-renames
678 --raw -r --no-abbrev/, $range);
681 # --is-ancestor requires git 1.8.0+
682 sub is_ancestor ($$$) {
683 my ($git, $cur, $tip) = @_;
684 return 0 unless $git->check($cur);
685 my $cmd = [ 'git', "--git-dir=$git->{git_dir}",
686 qw(merge-base --is-ancestor), $cur, $tip ];
687 my $pid = spawn($cmd);
688 waitpid($pid, 0) == $pid or die join(' ', @$cmd) .' did not finish';
692 sub need_update ($$$) {
693 my ($self, $cur, $new) = @_;
694 my $git = $self->{git};
695 return 1 if $cur && !is_ancestor($git, $cur, $new);
696 my $range = $cur eq '' ? $new : "$cur..$new";
697 chomp(my $n = $git->qx(qw(rev-list --count), $range));
698 ($n eq '' || $n > 0);
701 # The last git commit we indexed with Xapian or SQLite (msgmap)
702 # This needs to account for cases where Xapian or SQLite is
703 # out-of-date with respect to the other.
705 my ($self, $mm) = @_;
706 my $lm = $mm->last_commit || '';
708 if (need_xapian($self)) {
709 $lx = $self->{xdb}->get_metadata('last_commit') || '';
713 # Use last_commit from msgmap if it is older or unset
714 if (!$lm || ($lx && $lm && is_ancestor($self->{git}, $lm, $lx))) {
720 sub reindex_from ($$) {
721 my ($reindex, $last_commit) = @_;
722 return $last_commit unless $reindex;
723 ref($reindex) eq 'HASH' ? $reindex->{from} : '';
726 # indexes all unindexed messages (v1 only)
728 my ($self, $opts) = @_;
729 my $tip = $opts->{ref} || 'HEAD';
730 my ($last_commit, $lx, $xlog);
731 my $git = $self->{git};
733 my $pr = $opts->{-progress};
735 my $xdb = $self->begin_txn_lazy;
736 my $mm = _msgmap_init($self);
739 close($xlog) or die "git log failed: \$?=$?";
742 $last_commit = _last_x_commit($self, $mm);
743 $lx = reindex_from($opts->{reindex}, $last_commit);
745 $self->{over}->rollback_lazy;
746 $self->{over}->disconnect;
749 $xdb->cancel_transaction if $xdb;
750 $xdb = _xdb_release($self);
752 # ensure we leak no FDs to "git log" with Xapian <= 1.2
753 my $range = $lx eq '' ? $tip : "$lx..$tip";
754 $xlog = _git_log($self, $opts, $range);
756 $xdb = $self->begin_txn_lazy;
757 } while (_last_x_commit($self, $mm) ne $last_commit);
759 my $dbh = $mm->{dbh} if $mm;
761 my ($nr, $commit, $newest) = @_;
764 my $cur = $mm->last_commit || '';
765 if (need_update($self, $cur, $newest)) {
766 $mm->last_commit($newest);
771 if ($newest && need_xapian($self)) {
772 my $cur = $xdb->get_metadata('last_commit');
773 if (need_update($self, $cur, $newest)) {
774 $xdb->set_metadata('last_commit', $newest);
777 $self->commit_txn_lazy;
779 $xdb = _xdb_release($self);
780 # let another process do some work... <
781 $pr->("indexed $nr/$self->{ntodo}\n") if $pr && $nr;
783 $xdb = $self->begin_txn_lazy;
784 $dbh->begin_work if $dbh;
789 read_log($self, $xlog, *index_both, *unindex_both, $cb);
793 # order matters for unlocking
794 $_[0]->{xdb} = undef;
795 $_[0]->{lockfh} = undef;
798 # remote_* subs are only used by SearchIdxPart
801 if (my $w = $self->{w}) {
802 print $w "commit\n" or die "failed to write commit: $!";
804 $self->commit_txn_lazy;
810 if (my $w = delete $self->{w}) {
811 my $pid = delete $self->{pid} or die "no process to wait on\n";
812 print $w "close\n" or die "failed to write to pid:$pid: $!\n";
813 close $w or die "failed to close pipe for pid:$pid: $!\n";
814 waitpid($pid, 0) == $pid or die "remote process did not finish";
815 $? == 0 or die ref($self)." pid:$pid exited with: $?";
817 die "transaction in progress $self\n" if $self->{txn};
818 $self->_xdb_release if $self->{xdb};
823 my ($self, $oid, $mid) = @_;
824 if (my $w = $self->{w}) {
825 # triggers remove_by_oid in a shard
826 print $w "D $oid $mid\n" or die "failed to write remove $!";
828 $self->begin_txn_lazy;
829 $self->remove_by_oid($oid, $mid);
835 return if $self->{txn};
837 $self->{-inbox}->with_umask(sub {
838 my $xdb = $self->{xdb} || $self->_xdb_acquire;
839 $self->{over}->begin_lazy if $self->{over};
840 $xdb->begin_transaction if $xdb;
846 # store 'indexlevel=medium' in v2 shard=0 and v1 (only one shard)
847 # This metadata is read by Admin::detect_indexlevel:
851 if (!$self->{shard} && # undef or 0, not >0
852 delete($self->{-set_indexlevel_once})) {
853 my $xdb = $self->{xdb};
854 my $level = $xdb->get_metadata('indexlevel');
855 if (!$level || $level ne 'medium') {
856 $xdb->set_metadata('indexlevel', 'medium');
861 sub commit_txn_lazy {
863 delete $self->{txn} or return;
864 $self->{-inbox}->with_umask(sub {
865 if (my $xdb = $self->{xdb}) {
866 set_indexlevel($self);
867 $xdb->commit_transaction;
869 $self->{over}->commit_lazy if $self->{over};
875 if (need_xapian($self)) {
876 die "$$ $0 xdb not released\n" if $self->{xdb};
878 die "$$ $0 still in transaction\n" if $self->{txn};