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,
63 $self->{lock_path} = "$inboxdir/ssoma.lock";
64 my $dir = $self->xdir;
65 $self->{over} = PublicInbox::OverIdx->new("$dir/over.sqlite3");
66 } elsif ($version == 2) {
67 defined $shard or die "shard is required for v2\n";
69 $self->{shard} = $shard;
70 $self->{lock_path} = undef;
72 die "unsupported inbox version=$version\n";
74 $self->{creat} = ($creat || 0) == 1;
78 sub need_xapian ($) { $_[0]->{indexlevel} =~ $xapianlevels }
82 if (need_xapian($self)) {
83 my $xdb = delete $self->{xdb} or croak 'not acquired';
86 $self->lock_release if $self->{creat};
90 sub load_xapian_writable () {
91 return 1 if $X->{WritableDatabase};
92 PublicInbox::Search::load_xapian() or return;
93 my $xap = $PublicInbox::Search::Xap;
94 for (qw(Document TermGenerator WritableDatabase)) {
95 $X->{$_} = $xap.'::'.$_;
97 eval 'require '.$X->{WritableDatabase} or die;
98 *sortable_serialise = $xap.'::sortable_serialise';
99 $DB_CREATE_OR_OPEN = eval($xap.'::DB_CREATE_OR_OPEN()');
100 $DB_OPEN = eval($xap.'::DB_OPEN()');
107 my $dir = $self->xdir;
108 if (need_xapian($self)) {
109 croak 'already acquired' if $self->{xdb};
110 load_xapian_writable();
111 $flag = $self->{creat} ? $DB_CREATE_OR_OPEN : $DB_OPEN;
113 if ($self->{creat}) {
117 # don't create empty Xapian directories if we don't need Xapian
118 my $is_shard = defined($self->{shard});
119 if (!$is_shard || ($is_shard && need_xapian($self))) {
120 File::Path::mkpath($dir);
123 return unless defined $flag;
124 my $xdb = eval { ($X->{WritableDatabase})->new($dir, $flag) };
126 die "Failed opening $dir: ", $@;
132 my ($doc, $col, $num) = @_;
133 $num = sortable_serialise($num);
134 $doc->add_value($col, $num);
137 sub term_generator ($) { # write-only
140 $self->{term_generator} //= do {
141 my $tg = $X->{TermGenerator}->new;
142 $tg->set_stemmer($self->stemmer);
147 sub index_text ($$$$) {
148 my ($self, $text, $wdf_inc, $prefix) = @_;
149 my $tg = term_generator($self); # man Search::Xapian::TermGenerator
151 if ($self->{indexlevel} eq 'full') {
152 $tg->index_text($text, $wdf_inc, $prefix);
153 $tg->increase_termpos;
155 $tg->index_text_without_positions($text, $wdf_inc, $prefix);
159 sub index_users ($$) {
160 my ($self, $smsg) = @_;
162 my $from = $smsg->from;
166 index_text($self, $from, 1, 'A'); # A - author
167 index_text($self, $to, 1, 'XTO') if $to ne '';
168 index_text($self, $cc, 1, 'XCC') if $cc ne '';
171 sub index_diff_inc ($$$$) {
172 my ($self, $text, $pfx, $xnq) = @_;
174 index_text($self, join("\n", @$xnq), 1, 'XNQ');
177 index_text($self, $text, 1, $pfx);
180 sub index_old_diff_fn {
181 my ($self, $seen, $fa, $fb, $xnq) = @_;
183 # no renames or space support for traditional diffs,
184 # find the number of leading common paths to strip:
185 my @fa = split('/', $fa);
186 my @fb = split('/', $fb);
187 while (scalar(@fa) && scalar(@fb)) {
188 $fa = join('/', @fa);
189 $fb = join('/', @fb);
191 unless ($seen->{$fa}++) {
192 index_diff_inc($self, $fa, 'XDFN', $xnq);
202 sub index_diff ($$$) {
203 my ($self, $txt, $doc) = @_;
208 foreach (split(/\n/, $txt)) {
209 if ($in_diff && s/^ //) { # diff context
210 index_diff_inc($self, $_, 'XDFCTX', $xnq);
211 } elsif (/^-- $/) { # email signature begins
213 } elsif (m!^diff --git "?[^/]+/.+ "?[^/]+/.+\z!) {
214 # wait until "---" and "+++" to capture filenames
217 } elsif (m/^diff -(.+) (\S+) (\S+)$/) {
218 my ($opt, $fa, $fb) = ($1, $2, $3);
220 # only support unified:
221 next unless $opt =~ /[uU]/;
222 $in_diff = index_old_diff_fn($self, \%seen, $fa, $fb,
224 } elsif (m!^--- ("?[^/]+/.+)!) {
226 $fn = (split('/', git_unquote($fn), 2))[1];
227 $seen{$fn}++ or index_diff_inc($self, $fn, 'XDFN', $xnq);
229 } elsif (m!^\+\+\+ ("?[^/]+/.+)!) {
231 $fn = (split('/', git_unquote($fn), 2))[1];
232 $seen{$fn}++ or index_diff_inc($self, $fn, 'XDFN', $xnq);
234 } elsif (/^--- (\S+)/) {
237 } elsif (defined $in_diff && /^\+\+\+ (\S+)/) {
238 $in_diff = index_old_diff_fn($self, \%seen, $in_diff,
240 } elsif ($in_diff && s/^\+//) { # diff added
241 index_diff_inc($self, $_, 'XDFB', $xnq);
242 } elsif ($in_diff && s/^-//) { # diff removed
243 index_diff_inc($self, $_, 'XDFA', $xnq);
244 } elsif (m!^index ([a-f0-9]+)\.\.([a-f0-9]+)!) {
245 my ($ba, $bb) = ($1, $2);
246 index_git_blob_id($doc, 'XDFPRE', $ba);
247 index_git_blob_id($doc, 'XDFPOST', $bb);
249 } elsif (/^@@ (?:\S+) (?:\S+) @@\s*$/) {
250 # traditional diff w/o -p
251 } elsif (/^@@ (?:\S+) (?:\S+) @@\s*(\S+.*)$/) {
252 # hunk header context
253 index_diff_inc($self, $1, 'XDFHH', $xnq);
254 # ignore the following lines:
255 } elsif (/^(?:dis)similarity index/ ||
256 /^(?:old|new) mode/ ||
257 /^(?:deleted|new) file mode/ ||
258 /^(?:copy|rename) (?:from|to) / ||
259 /^(?:dis)?similarity index / ||
260 /^\\ No newline at end of file/ ||
261 /^Binary files .* differ/) {
264 # possible to be in diff context, some mail may be
265 # stripped by MUA or even GNU diff(1). "git apply"
266 # treats a bare "\n" as diff context, too
269 warn "non-diff line: $_\n" if DEBUG && $_ ne '';
274 index_text($self, join("\n", @xnq), 1, 'XNQ');
277 sub index_body ($$$) {
278 my ($self, $txt, $doc) = @_;
280 # does it look like a diff?
281 if ($txt =~ /^(?:diff|---|\+\+\+) /ms) {
282 index_diff($self, $txt, $doc);
284 index_text($self, $txt, 1, 'XNQ');
287 index_text($self, $txt, 0, 'XQUOT');
291 sub index_xapian { # msg_iter callback
292 my ($part, $depth, @idx) = @{$_[0]};
293 my ($self, $doc) = @{$_[1]};
294 my $ct = $part->content_type || 'text/plain';
295 my $fn = $part->filename;
296 if (defined $fn && $fn ne '') {
297 index_text($self, $fn, 1, 'XFN');
300 my ($s, undef) = msg_part_text($part, $ct);
301 defined $s or return;
303 # split off quoted and unquoted blocks:
304 my @sections = split(/((?:^>[^\n]*\n)+)/sm, $s);
306 index_body($self, $_, /\A>/ ? 0 : $doc) for @sections;
309 sub add_xapian ($$$$) {
310 my ($self, $mime, $smsg, $mids) = @_;
311 $smsg->{mime} = $mime; # XXX dangerous
312 my $hdr = $mime->header_obj;
313 my $doc = $X->{Document}->new;
314 my $subj = $smsg->subject;
315 add_val($doc, PublicInbox::Search::TS(), $smsg->{ts});
316 my @ds = gmtime($smsg->{ds});
317 my $yyyymmdd = strftime('%Y%m%d', @ds);
318 add_val($doc, PublicInbox::Search::YYYYMMDD(), $yyyymmdd);
319 my $dt = strftime('%Y%m%d%H%M%S', @ds);
320 add_val($doc, PublicInbox::Search::DT(), $dt);
322 my $tg = term_generator($self);
324 $tg->set_document($doc);
325 index_text($self, $subj, 1, 'S') if $subj;
326 index_users($self, $smsg);
328 msg_iter($mime, \&index_xapian, [ $self, $doc ]);
329 foreach my $mid (@$mids) {
330 index_text($self, $mid, 1, 'XM');
332 # because too many Message-IDs are prefixed with
334 if ($mid =~ /\w{12,}/) {
335 my @long = ($mid =~ /(\w{3,}+)/g);
336 index_text($self, join(' ', @long), 1, 'XM');
339 $smsg->{to} = $smsg->{cc} = ''; # WWW doesn't need these, only NNTP
340 PublicInbox::OverIdx::parse_references($smsg, $hdr, $mids);
341 my $data = $smsg->to_doc_data;
342 $doc->set_data($data);
343 if (my $altid = $self->{-altid}) {
344 foreach my $alt (@$altid) {
345 my $pfx = $alt->{xprefix};
346 foreach my $mid (@$mids) {
347 my $id = $alt->mid2alt($mid);
348 next unless defined $id;
349 $doc->add_boolean_term($pfx . $id);
353 $doc->add_boolean_term('Q' . $_) foreach @$mids;
354 $self->{xdb}->replace_document($smsg->{num}, $doc);
357 sub _msgmap_init ($) {
359 die "BUG: _msgmap_init is only for v1\n" if $self->{ibx_ver} != 1;
360 $self->{mm} //= eval {
361 require PublicInbox::Msgmap;
362 PublicInbox::Msgmap->new($self->{inboxdir}, 1);
367 # mime = Email::MIME object
368 my ($self, $mime, $smsg) = @_;
369 my $hdr = $mime->header_obj;
370 my $mids = mids_for_index($hdr);
371 $smsg //= bless { blob => '' }, 'PublicInbox::Smsg'; # test-only compat
372 $smsg->{mid} //= $mids->[0]; # v1 compatibility
373 $smsg->{num} //= do { # v1
375 index_mm($self, $mime);
379 $smsg->{ds} //= msg_datestamp($hdr, $self->{autime});
380 $smsg->{ts} //= msg_timestamp($hdr, $self->{cotime});
383 # order matters, overview stores every possible piece of
384 # data in doc_data (deflated). Xapian only stores a subset
385 # of the fields which exist in over.sqlite3. We may stop
386 # storing doc_data in Xapian sometime after we get multi-inbox
388 if (my $over = $self->{over}) { # v1 only
389 $over->add_overview($mime, $smsg);
391 if (need_xapian($self)) {
392 add_xapian($self, $mime, $smsg, $mids);
397 warn "failed to index message <".join('> <',@$mids).">: $@\n";
403 # returns begin and end PostingIterator
405 my ($self, $termval) = @_;
406 my $db = $self->{xdb};
408 ($db->postlist_begin($termval), $db->postlist_end($termval));
413 my ($self, $termval, $cb) = @_;
414 my $batch_size = 1000; # don't let @ids grow too large to avoid OOM
416 my ($head, $tail) = $self->find_doc_ids($termval);
417 return if $head == $tail;
419 for (; $head != $tail && @ids < $batch_size; $head++) {
420 push @ids, $head->get_docid;
426 # v1 only, where $mid is unique
428 my ($self, $mid) = @_;
429 $mid = mid_clean($mid);
431 if (my $over = $self->{over}) {
432 my $nr = eval { $over->remove_oid(undef, $mid) };
434 warn "failed to remove <$mid> from overview: $@\n";
436 warn "<$mid> missing for removal from overview\n";
439 return unless need_xapian($self);
440 my $db = $self->{xdb};
443 batch_do($self, 'Q' . $mid, sub {
445 $db->delete_document($_) for @$ids;
450 warn "failed to remove <$mid> from Xapian: $@\n";
452 warn "<$mid> missing for removal from Xapian\n";
456 # MID is a hint in V2
458 my ($self, $oid, $mid) = @_;
460 $self->{over}->remove_oid($oid, $mid) if $self->{over};
462 return unless need_xapian($self);
463 my $db = $self->{xdb};
465 # XXX careful, we cannot use batch_do here since we conditionally
466 # delete documents based on other factors, so we cannot call
467 # find_doc_ids twice.
468 my ($head, $tail) = $self->find_doc_ids('Q' . $mid);
469 return if $head == $tail;
471 # there is only ONE element in @delete unless we
472 # have bugs in our v2writable deduplication check
474 for (; $head != $tail; $head++) {
475 my $docid = $head->get_docid;
476 my $doc = $db->get_document($docid);
477 my $smsg = PublicInbox::Smsg->wrap($mid);
478 $smsg->load_expand($doc);
479 if ($smsg->{blob} eq $oid) {
480 push(@delete, $docid);
483 $db->delete_document($_) foreach @delete;
487 sub index_git_blob_id {
488 my ($doc, $pfx, $objid) = @_;
490 my $len = length($objid);
491 for (my $len = length($objid); $len >= 7; ) {
492 $doc->add_term($pfx.$objid);
493 $objid = substr($objid, 0, --$len);
498 my ($self, $mime) = @_;
499 my $mid = eval { mid_clean(mid_mime($mime)) };
500 $self->remove_message($mid) if defined $mid;
504 my ($self, $mime) = @_;
505 my $mid = mid_clean(mid_mime($mime));
506 my $mm = $self->{mm};
509 if (defined $self->{regen_down}) {
510 $num = $mm->num_for($mid) and return $num;
512 while (($num = $self->{regen_down}--) > 0) {
513 if ($mm->mid_set($num, $mid) != 0) {
517 } elsif (defined $self->{regen_up}) {
518 $num = $mm->num_for($mid) and return $num;
520 # this is to fixup old bugs due to add-remove-add
521 while (($num = ++$self->{regen_up})) {
522 if ($mm->mid_set($num, $mid) != 0) {
528 $num = $mm->mid_insert($mid) and return $num;
530 # fallback to num_for since filters like RubyLang set the number
535 my ($self, $mime) = @_;
536 $self->{mm}->mid_delete(mid_clean(mid_mime($mime)));
540 my ($self, $mime, $smsg) = @_;
541 my $num = index_mm($self, $mime);
543 add_message($self, $mime, $smsg);
547 my ($self, $mime) = @_;
548 unindex_blob($self, $mime);
549 unindex_mm($self, $mime);
553 my ($git, $blob, $sizeref) = @_;
555 my $str = $git->cat_file($blob, $sizeref);
556 # fixup bugs from import:
557 $$str =~ s/\A[\r\n]*From [^\r\n]*\r?\n//s;
558 PublicInbox::MIME->new($str);
563 # called by public-inbox-index
565 my ($self, $opts) = @_;
566 delete $self->{lock_path} if $opts->{-skip_lock};
567 $self->{-inbox}->with_umask(sub { $self->_index_sync($opts) })
570 sub batch_adjust ($$$$$) {
571 my ($max, $bytes, $batch_cb, $latest, $nr) = @_;
575 $batch_cb->($nr, $latest);
581 my ($self, $log, $add_cb, $del_cb, $batch_cb) = @_;
582 my $hex = '[a-f0-9]';
583 my $h40 = $hex .'{40}';
584 my $addmsg = qr!^:000000 100644 \S+ ($h40) A\t${hex}{2}/${hex}{38}$!;
585 my $delmsg = qr!^:100644 000000 ($h40) \S+ D\t${hex}{2}/${hex}{38}$!;
586 my $git = $self->{git};
589 my $max = BATCH_BYTES;
595 while (defined($line = <$log>)) {
596 if ($line =~ /$addmsg/o) {
598 if (delete $D{$blob}) {
599 if (defined $self->{regen_down}) {
600 my $num = $self->{regen_down}--;
601 $self->{mm}->num_highwater($num);
605 my $mime = do_cat_mail($git, $blob, \$bytes) or next;
606 my $smsg = bless {}, 'PublicInbox::Smsg';
607 batch_adjust(\$max, $bytes, $batch_cb, $latest, ++$nr);
608 $smsg->{blob} = $blob;
609 $smsg->{bytes} = $bytes;
610 $add_cb->($self, $mime, $smsg);
611 } elsif ($line =~ /$delmsg/o) {
614 } elsif ($line =~ /^commit ($h40)/o) {
617 } elsif ($line =~ /^author .*? ([0-9]+) [\-\+][0-9]+$/) {
618 $self->{autime} = $1;
619 } elsif ($line =~ /^committer .*? ([0-9]+) [\-\+][0-9]+$/) {
620 $self->{cotime} = $1;
623 close($log) or die "git log failed: \$?=$?";
625 foreach my $blob (keys %D) {
626 my $mime = do_cat_mail($git, $blob, \$bytes) or next;
627 $del_cb->($self, $mime);
629 $batch_cb->($nr, $latest, $newest);
633 my ($self, $opts, $range) = @_;
634 my $git = $self->{git};
636 if (index($range, '..') < 0) {
637 # don't show annoying git errrors to users who run -index
639 $git->qx(qw(rev-parse -q --verify), "$range^0");
641 open my $fh, '<', '/dev/null' or
642 die "failed to open /dev/null: $!\n";
647 # Count the new files so they can be added newest to oldest
648 # and still have numbers increasing from oldest to newest
650 my $pr = $opts->{-progress};
651 $pr->("counting changes\n\t$range ... ") if $pr;
652 # can't use 'rev-list --count' if we use --diff-filter
653 my $fh = $git->popen(qw(log --pretty=tformat:%h
654 --no-notes --no-color --no-renames
655 --diff-filter=AM), $range);
656 ++$fcount while <$fh>;
657 close $fh or die "git log failed: \$?=$?";
658 my $high = $self->{mm}->num_highwater;
659 $pr->("$fcount\n") if $pr; # continue previous line
660 $self->{ntodo} = $fcount;
662 if (index($range, '..') < 0) {
663 if ($high && $high == $fcount) {
664 # fix up old bugs in full indexes which caused messages to
665 # not appear in Msgmap
666 $self->{regen_up} = $high;
668 # normal regen is for for fresh data
669 $self->{regen_down} = $fcount;
672 # Give oldest messages the smallest numbers
673 $self->{regen_down} = $high + $fcount;
676 $git->popen(qw/log --pretty=raw --no-notes --no-color --no-renames
677 --raw -r --no-abbrev/, $range);
680 # --is-ancestor requires git 1.8.0+
681 sub is_ancestor ($$$) {
682 my ($git, $cur, $tip) = @_;
683 return 0 unless $git->check($cur);
684 my $cmd = [ 'git', "--git-dir=$git->{git_dir}",
685 qw(merge-base --is-ancestor), $cur, $tip ];
686 my $pid = spawn($cmd);
687 waitpid($pid, 0) == $pid or die join(' ', @$cmd) .' did not finish';
691 sub need_update ($$$) {
692 my ($self, $cur, $new) = @_;
693 my $git = $self->{git};
694 return 1 if $cur && !is_ancestor($git, $cur, $new);
695 my $range = $cur eq '' ? $new : "$cur..$new";
696 chomp(my $n = $git->qx(qw(rev-list --count), $range));
697 ($n eq '' || $n > 0);
700 # The last git commit we indexed with Xapian or SQLite (msgmap)
701 # This needs to account for cases where Xapian or SQLite is
702 # out-of-date with respect to the other.
704 my ($self, $mm) = @_;
705 my $lm = $mm->last_commit || '';
707 if (need_xapian($self)) {
708 $lx = $self->{xdb}->get_metadata('last_commit') || '';
712 # Use last_commit from msgmap if it is older or unset
713 if (!$lm || ($lx && $lm && is_ancestor($self->{git}, $lm, $lx))) {
719 sub reindex_from ($$) {
720 my ($reindex, $last_commit) = @_;
721 return $last_commit unless $reindex;
722 ref($reindex) eq 'HASH' ? $reindex->{from} : '';
725 # indexes all unindexed messages (v1 only)
727 my ($self, $opts) = @_;
728 my $tip = $opts->{ref} || 'HEAD';
729 my ($last_commit, $lx, $xlog);
730 my $git = $self->{git};
732 my $pr = $opts->{-progress};
734 my $xdb = $self->begin_txn_lazy;
735 my $mm = _msgmap_init($self);
738 close($xlog) or die "git log failed: \$?=$?";
741 $last_commit = _last_x_commit($self, $mm);
742 $lx = reindex_from($opts->{reindex}, $last_commit);
744 $self->{over}->rollback_lazy;
745 $self->{over}->disconnect;
748 $xdb->cancel_transaction if $xdb;
749 $xdb = _xdb_release($self);
751 # ensure we leak no FDs to "git log" with Xapian <= 1.2
752 my $range = $lx eq '' ? $tip : "$lx..$tip";
753 $xlog = _git_log($self, $opts, $range);
755 $xdb = $self->begin_txn_lazy;
756 } while (_last_x_commit($self, $mm) ne $last_commit);
758 my $dbh = $mm->{dbh} if $mm;
760 my ($nr, $commit, $newest) = @_;
763 my $cur = $mm->last_commit || '';
764 if (need_update($self, $cur, $newest)) {
765 $mm->last_commit($newest);
770 if ($newest && need_xapian($self)) {
771 my $cur = $xdb->get_metadata('last_commit');
772 if (need_update($self, $cur, $newest)) {
773 $xdb->set_metadata('last_commit', $newest);
776 $self->commit_txn_lazy;
778 $xdb = _xdb_release($self);
779 # let another process do some work... <
780 $pr->("indexed $nr/$self->{ntodo}\n") if $pr && $nr;
782 $xdb = $self->begin_txn_lazy;
783 $dbh->begin_work if $dbh;
788 read_log($self, $xlog, *index_both, *unindex_both, $cb);
792 # order matters for unlocking
793 $_[0]->{xdb} = undef;
794 $_[0]->{lockfh} = undef;
797 # remote_* subs are only used by SearchIdxPart
800 if (my $w = $self->{w}) {
801 print $w "commit\n" or die "failed to write commit: $!";
803 $self->commit_txn_lazy;
809 if (my $w = delete $self->{w}) {
810 my $pid = delete $self->{pid} or die "no process to wait on\n";
811 print $w "close\n" or die "failed to write to pid:$pid: $!\n";
812 close $w or die "failed to close pipe for pid:$pid: $!\n";
813 waitpid($pid, 0) == $pid or die "remote process did not finish";
814 $? == 0 or die ref($self)." pid:$pid exited with: $?";
816 die "transaction in progress $self\n" if $self->{txn};
817 $self->_xdb_release if $self->{xdb};
822 my ($self, $oid, $mid) = @_;
823 if (my $w = $self->{w}) {
824 # triggers remove_by_oid in a shard
825 print $w "D $oid $mid\n" or die "failed to write remove $!";
827 $self->begin_txn_lazy;
828 $self->remove_by_oid($oid, $mid);
834 return if $self->{txn};
836 $self->{-inbox}->with_umask(sub {
837 my $xdb = $self->{xdb} || $self->_xdb_acquire;
838 $self->{over}->begin_lazy if $self->{over};
839 $xdb->begin_transaction if $xdb;
845 sub commit_txn_lazy {
847 delete $self->{txn} or return;
848 $self->{-inbox}->with_umask(sub {
849 if (my $xdb = $self->{xdb}) {
851 # store 'indexlevel=medium' in v2 shard=0 and
852 # v1 (only one shard)
853 # This metadata is read by Admin::detect_indexlevel:
854 if (!$self->{shard} # undef or 0, not >0
855 && $self->{indexlevel} eq 'medium') {
856 $xdb->set_metadata('indexlevel', 'medium');
859 $xdb->commit_transaction;
861 $self->{over}->commit_lazy if $self->{over};
867 if (need_xapian($self)) {
868 die "$$ $0 xdb not released\n" if $self->{xdb};
870 die "$$ $0 still in transaction\n" if $self->{txn};