1 # Copyright (C) 2020-2021 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # Detached/external index cross inbox search indexing support
5 # read-write counterpart to PublicInbox::ExtSearch
7 # It's based on the same ideas as public-inbox-v2-format(5) using
8 # over.sqlite3 for dedupe and sharded Xapian. msgmap.sqlite3 is
9 # missing, so there is no Message-ID conflict resolution, meaning
10 # no NNTP support for now.
12 # v2 has a 1:1 mapping of index:inbox or msgmap for NNTP support.
13 # This is intended to be an M:N index:inbox mapping, but it'll likely
14 # be 1:N in common practice (M==1)
16 package PublicInbox::ExtSearchIdx;
19 use parent qw(PublicInbox::ExtSearch PublicInbox::Lock);
20 use Carp qw(croak carp);
21 use Sys::Hostname qw(hostname);
22 use POSIX qw(strftime);
23 use File::Glob qw(bsd_glob GLOB_NOSORT);
24 use PublicInbox::Search;
25 use PublicInbox::SearchIdx qw(prepare_stack is_ancestor is_bad_blob);
26 use PublicInbox::OverIdx;
27 use PublicInbox::MiscIdx;
28 use PublicInbox::MID qw(mids);
29 use PublicInbox::V2Writable;
30 use PublicInbox::InboxWritable;
31 use PublicInbox::ContentHash qw(content_hash);
33 use PublicInbox::DS qw(now add_timer);
34 use DBI qw(:sql_types); # SQL_BLOB
37 my (undef, $dir, $opt) = @_;
38 my $l = $opt->{indexlevel} // 'full';
39 $l !~ $PublicInbox::SearchIdx::INDEXLEVELS and
40 die "invalid indexlevel=$l\n";
41 $l eq 'basic' and die "E: indexlevel=basic not yet supported\n";
43 xpfx => "$dir/ei".PublicInbox::Search::SCHEMA_VERSION,
45 creat => $opt->{creat},
46 ibx_map => {}, # (newsgroup//inboxdir) => $ibx
47 ibx_cfg => [], # by config section order
53 lock_path => "$dir/ei.lock",
55 $self->{shards} = $self->count_shards ||
56 nproc_shards({ nproc => $opt->{jobs} });
57 my $oidx = PublicInbox::OverIdx->new("$self->{xpfx}/over.sqlite3");
58 $self->{-no_fsync} = $oidx->{-no_fsync} = 1 if !$opt->{fsync};
59 $self->{oidx} = $oidx;
64 my ($self, $ibx) = @_;
65 $self->{ibx_map}->{$ibx->eidx_key} //= do {
66 delete $self->{-ibx_ary}; # invalidate cache
67 push @{$self->{ibx_cfg}}, $ibx;
72 sub _ibx_attach { # each_inbox callback
73 my ($ibx, $self) = @_;
74 attach_inbox($self, $ibx);
78 my ($self, $cfg) = @_;
80 $cfg->each_inbox(\&_ibx_attach, $self);
83 sub check_batch_limit ($) {
85 my $self = $req->{self};
86 my $new_smsg = $req->{new_smsg};
87 my $n = $self->{transact_bytes} += $new_smsg->{bytes};
89 # set flag for PublicInbox::V2Writable::index_todo:
90 ${$req->{need_checkpoint}} = 1 if $n >= $self->{batch_bytes};
94 my ($req, $smsg) = @_;
95 my $self = $req->{self};
96 my $docid = $smsg->{num};
97 my $idx = $self->idx_shard($docid);
98 my $oid = $req->{oid};
99 my $xibx = $req->{ibx};
100 my $eml = $req->{eml};
101 my $eidx_key = $xibx->eidx_key;
102 if (my $new_smsg = $req->{new_smsg}) { # 'm' on cross-posted message
103 my $xnum = $req->{xnum};
104 $self->{oidx}->add_xref3($docid, $xnum, $oid, $eidx_key);
105 $idx->ipc_do('add_eidx_info', $docid, $eidx_key, $eml);
106 check_batch_limit($req);
109 my $nr = $self->{oidx}->remove_xref3($docid, $oid, $eidx_key,
112 $self->{oidx}->eidxq_del($docid);
113 $idx->ipc_do('xdb_remove', $docid);
114 } elsif ($rm_eidx_info) {
115 $idx->ipc_do('remove_eidx_info',
116 $docid, $eidx_key, $eml);
117 $self->{oidx}->eidxq_add($docid); # yes, add
122 # called by V2Writable::sync_prepare
123 sub artnum_max { $_[0]->{oidx}->eidx_max }
125 sub index_unseen ($) {
127 my $new_smsg = $req->{new_smsg} or die 'BUG: {new_smsg} unset';
128 my $eml = delete $req->{eml};
129 $new_smsg->populate($eml, $req);
130 my $self = $req->{self};
131 my $docid = $self->{oidx}->adj_counter('eidx_docid', '+');
132 $new_smsg->{num} = $docid;
133 my $idx = $self->idx_shard($docid);
134 $self->{oidx}->add_overview($eml, $new_smsg);
135 my $oid = $new_smsg->{blob};
136 my $ibx = delete $req->{ibx} or die 'BUG: {ibx} unset';
137 $self->{oidx}->add_xref3($docid, $req->{xnum}, $oid, $ibx->eidx_key);
138 $idx->index_eml($eml, $new_smsg, $ibx->eidx_key);
139 check_batch_limit($req);
142 sub do_finalize ($) {
144 if (my $indexed = $req->{indexed}) { # duplicated messages
145 do_xpost($req, $_) for @$indexed;
146 } elsif (exists $req->{new_smsg}) { # totally unseen messsage
149 # `d' message was already unindexed in the v1/v2 inboxes,
150 # so it's too noisy to warn, here.
152 # cur_cmt may be undef for unindex_oid, set by V2Writable::index_todo
153 if (defined(my $cur_cmt = $req->{cur_cmt})) {
154 ${$req->{latest_cmt}} = $cur_cmt;
158 sub do_step ($) { # main iterator for adding messages to the index
160 my $self = $req->{self} // die 'BUG: {self} missing';
162 if (my $next_arg = $req->{next_arg}) {
163 if (my $smsg = $self->{oidx}->next_by_mid(@$next_arg)) {
164 $req->{cur_smsg} = $smsg;
165 $self->git->cat_async($smsg->{blob},
166 \&ck_existing, $req);
167 return; # ck_existing calls do_step
169 delete $req->{next_arg};
171 die "BUG: {cur_smsg} still set" if $req->{cur_smsg};
172 my $mid = shift(@{$req->{mids}}) // last;
174 $req->{next_arg} = [ $mid, \$id, \$prev ];
180 sub _blob_missing ($$) { # called when $smsg->{blob} is bad
181 my ($req, $smsg) = @_;
182 my $self = $req->{self};
183 my $xref3 = $self->{oidx}->get_xref3($smsg->{num});
184 my @keep = grep(!/:$smsg->{blob}\z/, @$xref3);
186 $keep[0] =~ /:([a-f0-9]{40,}+)\z/ or
187 die "BUG: xref $keep[0] has no OID";
189 $self->{oidx}->remove_xref3($smsg->{num}, $smsg->{blob});
190 my $upd = $self->{oidx}->update_blob($smsg, $oidhex);
191 my $saved = $self->{oidx}->get_art($smsg->{num});
193 $self->{oidx}->delete_by_num($smsg->{num});
197 sub ck_existing { # git->cat_async callback
198 my ($bref, $oid, $type, $size, $req) = @_;
199 my $smsg = delete $req->{cur_smsg} or die 'BUG: {cur_smsg} missing';
200 if ($type eq 'missing') {
201 _blob_missing($req, $smsg);
202 } elsif (!is_bad_blob($oid, $type, $size, $smsg->{blob})) {
203 my $self = $req->{self} // die 'BUG: {self} missing';
204 local $self->{current_info} = "$self->{current_info} $oid";
205 my $cur = PublicInbox::Eml->new($bref);
206 if (content_hash($cur) eq $req->{chash}) {
207 push @{$req->{indexed}}, $smsg; # for do_xpost
208 } # else { index_unseen later }
213 # is the messages visible in the inbox currently being indexed?
214 # return the number if so
215 sub cur_ibx_xnum ($$) {
216 my ($req, $bref) = @_;
217 my $ibx = $req->{ibx} or die 'BUG: current {ibx} missing';
219 $req->{eml} = PublicInbox::Eml->new($bref);
220 $req->{chash} = content_hash($req->{eml});
221 $req->{mids} = mids($req->{eml});
222 for my $mid (@{$req->{mids}}) {
224 while (my $x = $ibx->over->next_by_mid($mid, \$id, \$prev)) {
225 return $x->{num} if $x->{blob} eq $req->{oid};
231 sub index_oid { # git->cat_async callback for 'm'
232 my ($bref, $oid, $type, $size, $req) = @_;
233 my $self = $req->{self};
234 local $self->{current_info} = "$self->{current_info} $oid";
235 return if is_bad_blob($oid, $type, $size, $req->{oid});
236 my $new_smsg = $req->{new_smsg} = bless {
238 }, 'PublicInbox::Smsg';
239 $new_smsg->set_bytes($$bref, $size);
240 defined($req->{xnum} = cur_ibx_xnum($req, $bref)) or return;
245 sub unindex_oid { # git->cat_async callback for 'd'
246 my ($bref, $oid, $type, $size, $req) = @_;
247 my $self = $req->{self};
248 local $self->{current_info} = "$self->{current_info} $oid";
249 return if is_bad_blob($oid, $type, $size, $req->{oid});
250 return if defined(cur_ibx_xnum($req, $bref)); # was re-added
254 # overrides V2Writable::last_commits, called by sync_ranges via sync_prepare
256 my ($self, $sync) = @_;
258 my $ekey = $sync->{ibx}->eidx_key;
259 my $uv = $sync->{ibx}->uidvalidity;
260 for my $i (0..$sync->{epoch_max}) {
261 $heads->[$i] = $self->{oidx}->eidx_meta("lc-v2:$ekey//$uv;$i");
266 sub _ibx_index_reject ($) {
268 $ibx->mm // return 'unindexed, no msgmap.sqlite3';
269 $ibx->uidvalidity // return 'no UIDVALIDITY';
270 $ibx->over // return 'unindexed, no over.sqlite3';
274 sub _sync_inbox ($$$) {
275 my ($self, $sync, $ibx) = @_;
276 my $ekey = $ibx->eidx_key;
277 if (defined(my $err = _ibx_index_reject($ibx))) {
278 return "W: skipping $ekey ($err)";
281 $sync->{nr} = \(my $nr = 0);
282 my $v = $ibx->version;
284 $sync->{epoch_max} = $ibx->max_git_epoch // return;
285 sync_prepare($self, $sync); # or return # TODO: once MiscIdx is stable
287 my $uv = $ibx->uidvalidity;
288 my $lc = $self->{oidx}->eidx_meta("lc-v1:$ekey//$uv");
289 my $head = $ibx->mm->last_commit //
290 return "E: $ibx->{inboxdir} is not indexed";
291 my $stk = prepare_stack($sync, $lc ? "$lc..$head" : $head);
292 my $unit = { stack => $stk, git => $ibx->git };
293 push @{$sync->{todo}}, $unit;
295 return "E: $ekey unsupported inbox version (v$v)";
297 for my $unit (@{delete($sync->{todo}) // []}) {
298 last if $sync->{quit};
299 index_todo($self, $sync, $unit);
301 $self->{midx}->index_ibx($ibx) unless $sync->{quit};
302 $ibx->git->cleanup; # done with this inbox, now
306 sub gc_unref_doc ($$$$) {
307 my ($self, $ibx_id, $eidx_key, $docid) = @_;
308 my $dbh = $self->{oidx}->dbh;
310 # for debug/info purposes, oids may no longer be accessible
311 my $sth = $dbh->prepare_cached(<<'', undef, 1);
312 SELECT oidbin FROM xref3 WHERE docid = ? AND ibx_id = ?
314 $sth->execute($docid, $ibx_id);
315 my @oid = map { unpack('H*', $_->[0]) } @{$sth->fetchall_arrayref};
317 $dbh->prepare_cached(<<'')->execute($docid, $ibx_id);
318 DELETE FROM xref3 WHERE docid = ? AND ibx_id = ?
320 my $remain = $self->{oidx}->get_xref3($docid);
321 if (scalar(@$remain)) {
322 $self->{oidx}->eidxq_add($docid); # enqueue for reindex
324 warn "I: unref #$docid $eidx_key $oid\n";
327 warn "I: remove #$docid $eidx_key @oid\n";
328 $self->idx_shard($docid)->ipc_do('xdb_remove', $docid);
333 my ($self, $opt) = @_;
334 $self->{cfg} or die "E: GC requires ->attach_config\n";
336 $self->idx_init($opt); # acquire lock via V2Writable::_idx_init
338 my $dbh = $self->{oidx}->dbh;
339 $dbh->do('PRAGMA case_sensitive_like = ON'); # only place we use LIKE
340 my $x3_doc = $dbh->prepare('SELECT docid FROM xref3 WHERE ibx_id = ?');
341 my $ibx_ck = $dbh->prepare('SELECT ibx_id,eidx_key FROM inboxes');
342 my $lc_i = $dbh->prepare(<<'');
343 SELECT key FROM eidx_meta WHERE key LIKE ? ESCAPE ?
346 while (my ($ibx_id, $eidx_key) = $ibx_ck->fetchrow_array) {
347 next if $self->{ibx_map}->{$eidx_key};
348 $self->{midx}->remove_eidx_key($eidx_key);
349 warn "I: deleting messages for $eidx_key...\n";
350 $x3_doc->execute($ibx_id);
351 while (defined(my $docid = $x3_doc->fetchrow_array)) {
352 gc_unref_doc($self, $ibx_id, $eidx_key, $docid);
354 $dbh->prepare_cached(<<'')->execute($ibx_id);
355 DELETE FROM inboxes WHERE ibx_id = ?
357 # drop last_commit info
359 $pat =~ s/([_%\\])/\\$1/g;
360 $lc_i->execute("lc-%:$pat//%", '\\');
361 while (my ($key) = $lc_i->fetchrow_array) {
362 next if $key !~ m!\Alc-v[1-9]+:\Q$eidx_key\E//!;
363 warn "I: removing $key\n";
364 $dbh->prepare_cached(<<'')->execute($key);
365 DELETE FROM eidx_meta WHERE key = ?
369 warn "I: $eidx_key removed\n";
372 # it's not real unless it's in `over', we use parallelism here,
373 # shards will be reading directly from over, so commit
374 $self->{oidx}->commit_lazy;
375 $self->{oidx}->begin_lazy;
377 for my $idx (@{$self->{idx_shards}}) {
378 warn "I: cleaning up shard #$idx->{shard}\n";
379 $idx->shard_over_check($self->{oidx});
381 my $nr = $dbh->do(<<'');
382 DELETE FROM xref3 WHERE docid NOT IN (SELECT num FROM over)
384 warn "I: eliminated $nr stale xref3 entries\n" if $nr != 0;
390 my ($self, $sync, $smsg) = @_;
391 my $ibx_id = delete($smsg->{ibx_id}) // die '{ibx_id} unset';
392 my $pos = $sync->{id2pos}->{$ibx_id} // die "$ibx_id no pos";
393 $self->{-ibx_ary}->[$pos] // die "BUG: ibx for $smsg->{blob} not mapped"
396 sub _fd_constrained ($) {
398 $self->{-fd_constrained} //= do {
400 if (eval { require BSD::Resource; 1 }) {
401 my $NOFILE = BSD::Resource::RLIMIT_NOFILE();
402 ($soft, undef) = BSD::Resource::getrlimit($NOFILE);
404 chomp($soft = `sh -c 'ulimit -n'`);
406 if (defined($soft)) {
407 my $want = scalar(@{$self->{-ibx_ary}}) + 64; # estimate
408 my $ret = $want > $soft;
411 RLIMIT_NOFILE=$soft insufficient (want: $want), will close DB handles early
416 warn "Unable to determine RLIMIT_NOFILE: $@\n";
422 sub _reindex_finalize ($$$) {
423 my ($req, $smsg, $eml) = @_;
424 my $sync = $req->{sync};
425 my $self = $sync->{self};
426 my $by_chash = delete $req->{by_chash} or die 'BUG: no {by_chash}';
427 my $nr = scalar(keys(%$by_chash)) or die 'BUG: no content hashes';
428 my $orig_smsg = $req->{orig_smsg} // die 'BUG: no {orig_smsg}';
429 my $docid = $smsg->{num} = $orig_smsg->{num};
430 $self->{oidx}->add_overview($eml, $smsg); # may rethread
431 check_batch_limit({ %$sync, new_smsg => $smsg });
432 my $chash0 = $smsg->{chash} // die "BUG: $smsg->{blob} no {chash}";
433 my $stable = delete($by_chash->{$chash0}) //
434 die "BUG: $smsg->{blob} chash missing";
435 my $idx = $self->idx_shard($docid);
436 my $top_smsg = pop @$stable;
437 $top_smsg == $smsg or die 'BUG: top_smsg != smsg';
438 my $ibx = _ibx_for($self, $sync, $smsg);
439 $idx->index_eml($eml, $smsg, $ibx->eidx_key);
440 for my $x (reverse @$stable) {
441 $ibx = _ibx_for($self, $sync, $x);
442 my $hdr = delete $x->{hdr} // die 'BUG: no {hdr}';
443 $idx->ipc_do('add_eidx_info', $docid, $ibx->eidx_key, $hdr);
445 return if $nr == 1; # likely, all good
447 warn "W: #$docid split into $nr due to deduplication change\n";
449 for my $ary (values %$by_chash) {
450 for my $x (reverse @$ary) {
451 warn "removing #$docid xref3 $x->{blob}\n";
452 my $n = $self->{oidx}->remove_xref3($docid, $x->{blob});
453 die "BUG: $x->{blob} invalidated #$docid" if $n == 0;
455 my $x = pop(@$ary) // die "BUG: #$docid {by_chash} empty";
456 $x->{num} = delete($x->{xnum}) // die '{xnum} unset';
457 $ibx = _ibx_for($self, $sync, $x);
458 if (my $over = $ibx->over) {
459 my $e = $over->get_art($x->{num});
460 $e->{blob} eq $x->{blob} or die <<EOF;
461 $x->{blob} != $e->{blob} (${\$ibx->eidx_key}:$e->{num});
463 push @todo, $ibx, $e;
464 $over->dbh_close if _fd_constrained($self);
466 die "$ibx->{inboxdir}: over.sqlite3 unusable: $!\n";
470 while (my ($ibx, $e) = splice(@todo, 0, 2)) {
471 reindex_unseen($self, $sync, $ibx, $e);
475 sub _reindex_oid { # git->cat_async callback
476 my ($bref, $oid, $type, $size, $req) = @_;
477 my $sync = $req->{sync};
478 my $self = $sync->{self};
479 my $orig_smsg = $req->{orig_smsg} // die 'BUG: no {orig_smsg}';
480 my $expect_oid = $req->{xr3r}->[$req->{ix}]->[2];
481 my $docid = $orig_smsg->{num};
482 if (is_bad_blob($oid, $type, $size, $expect_oid)) {
483 my $remain = $self->{oidx}->remove_xref3($docid, $expect_oid);
485 warn "W: #$docid gone or corrupted\n";
486 $self->idx_shard($docid)->ipc_do('xdb_remove', $docid);
487 } elsif (my $next_oid = $req->{xr3r}->[++$req->{ix}]->[2]) {
488 $self->git->cat_async($next_oid, \&_reindex_oid, $req);
490 warn "BUG: #$docid gone (UNEXPECTED)\n";
491 $self->idx_shard($docid)->ipc_do('xdb_remove', $docid);
495 my $ci = $self->{current_info};
496 local $self->{current_info} = "$ci #$docid $oid";
497 my $re_smsg = bless { blob => $oid }, 'PublicInbox::Smsg';
498 $re_smsg->set_bytes($$bref, $size);
499 my $eml = PublicInbox::Eml->new($bref);
500 $re_smsg->populate($eml, { autime => $orig_smsg->{ds},
501 cotime => $orig_smsg->{ts} });
502 my $chash = content_hash($eml);
503 $re_smsg->{chash} = $chash;
504 $re_smsg->{xnum} = $req->{xr3r}->[$req->{ix}]->[1];
505 $re_smsg->{ibx_id} = $req->{xr3r}->[$req->{ix}]->[0];
506 $re_smsg->{hdr} = $eml->header_obj;
507 push @{$req->{by_chash}->{$chash}}, $re_smsg;
508 if (my $next_oid = $req->{xr3r}->[++$req->{ix}]->[2]) {
509 $self->git->cat_async($next_oid, \&_reindex_oid, $req);
510 } else { # last $re_smsg is the highest priority xref3
511 local $self->{current_info} = "$ci #$docid";
512 _reindex_finalize($req, $re_smsg, $eml);
516 sub _reindex_smsg ($$$) {
517 my ($self, $sync, $smsg) = @_;
518 my $docid = $smsg->{num};
519 my $xr3 = $self->{oidx}->get_xref3($docid, 1);
520 if (scalar(@$xr3) == 0) { # _reindex_check_stale should've covered this
522 BUG? #$docid $smsg->{blob} is not referenced by inboxes during reindex
524 $self->{oidx}->delete_by_num($docid);
525 $self->idx_shard($docid)->ipc_do('xdb_remove', $docid);
529 # we sort {xr3r} in the reverse order of ibx_sorted so we can
530 # hit the common case in _reindex_finalize without rereading
531 # from git (or holding multiple messages in memory).
532 my $id2pos = $sync->{id2pos}; # index in ibx_sorted
534 $id2pos->{$b->[0]} <=> $id2pos->{$a->[0]}
536 $b->[1] <=> $a->[1] # break ties with {xnum}
538 @$xr3 = map { [ $_->[0], $_->[1], unpack('H*', $_->[2]) ] } @$xr3;
539 my $req = { orig_smsg => $smsg, sync => $sync, xr3r => $xr3, ix => 0 };
540 $self->git->cat_async($xr3->[$req->{ix}]->[2], \&_reindex_oid, $req);
543 sub checkpoint_due ($) {
545 ${$sync->{need_checkpoint}} || (now() > $sync->{next_check});
549 # I've copied FS images and only changed the hostname before,
550 # so prepend hostname. Use `state' since these a BOFH can change
551 # these while this process is running and we always want to be
552 # able to release locks taken by this process.
553 state $retval = hostname . '-' . do {
554 my $m; # machine-id(5) is systemd
555 if (open(my $fh, '<', '/etc/machine-id')) { $m = <$fh> }
556 # (g)hostid(1) is in GNU coreutils, kern.hostid is most BSDs
557 chomp($m ||= `{ sysctl -n kern.hostid ||
558 hostid || ghostid; } 2>/dev/null`
559 || "no-machine-id-or-hostid-on-$^O");
566 my $expect = delete($self->{-eidxq_locked}) or return;
567 my ($owner_pid, undef) = split(/-/, $expect);
568 return if $owner_pid != $$; # shards may fork
569 my $oidx = $self->{oidx};
571 my $cur = $oidx->eidx_meta('eidxq_lock') // '';
572 if ($cur eq $expect) {
573 $oidx->eidx_meta('eidxq_lock', '');
575 } elsif ($cur ne '') {
576 warn "E: eidxq_lock($expect) stolen by $cur\n";
578 warn "E: eidxq_lock($expect) released by another process\n";
585 eidxq_release($self) and $self->{oidx}->commit_lazy;
588 sub _eidxq_take ($) {
590 my $val = "$$-${\time}-$>-".host_ident;
591 $self->{oidx}->eidx_meta('eidxq_lock', $val);
592 $self->{-eidxq_locked} = $val;
595 sub eidxq_lock_acquire ($) {
597 my $oidx = $self->{oidx};
599 my $cur = $oidx->eidx_meta('eidxq_lock') || return _eidxq_take($self);
600 if (my $locked = $self->{-eidxq_locked}) { # be lazy
601 return $locked if $locked eq $cur;
603 my ($pid, $time, $euid, $ident) = split(/-/, $cur, 4);
604 my $t = strftime('%Y-%m-%d %k:%M:%S', gmtime($time));
605 if ($euid == $> && $ident eq host_ident) {
608 I: PID:$pid (re)indexing Xapian since $t, it will continue our work
612 warn "I: eidxq_lock is stale ($cur), clobbering\n";
613 return _eidxq_take($self);
615 warn "E: kill(0, $pid) failed: $!\n"; # fall-through:
617 my $fn = $oidx->dbh->sqlite_db_filename;
619 W: PID:$pid, UID:$euid on $ident is indexing Xapian since $t
620 W: If this is unexpected, delete `eidxq_lock' from the `eidx_meta' table:
621 W: sqlite3 $fn 'DELETE FROM eidx_meta WHERE key = "eidxq_lock"'
628 $self->{-ibx_ary} //= do {
629 # highest boost first, stable for config-ordering tiebreaker
632 ($b->{boost} // 0) <=> ($a->{boost} // 0)
633 } @{$self->{ibx_cfg}} ];
637 sub eidxq_process ($$) { # for reindexing
638 my ($self, $sync) = @_;
640 return unless eidxq_lock_acquire($self);
641 my $dbh = $self->{oidx}->dbh;
642 my $tot = $dbh->selectrow_array('SELECT COUNT(*) FROM eidxq') or return;
644 local $sync->{-regen_fmt} = "%u/$tot\n";
645 my $pr = $sync->{-opt}->{-progress};
647 my $min = $dbh->selectrow_array('SELECT MIN(docid) FROM eidxq');
648 my $max = $dbh->selectrow_array('SELECT MAX(docid) FROM eidxq');
649 $pr->("Xapian indexing $min..$max (total=$tot)\n");
651 $sync->{id2pos} //= do {
654 $id2pos{$_->{-ibx_id}} = $pos++ for (@{ibx_sorted($self)});
659 $del = $dbh->prepare('DELETE FROM eidxq WHERE docid = ?');
660 $iter = $dbh->prepare('SELECT docid FROM eidxq ORDER BY docid ASC');
662 while (defined(my $docid = $iter->fetchrow_array)) {
663 last if $sync->{quit};
664 if (my $smsg = $self->{oidx}->get_art($docid)) {
665 _reindex_smsg($self, $sync, $smsg);
667 warn "E: #$docid does not exist in over\n";
669 $del->execute($docid);
672 if (checkpoint_due($sync)) {
673 $dbh = $del = $iter = undef;
674 reindex_checkpoint($self, $sync); # release lock
675 $dbh = $self->{oidx}->dbh;
679 $self->git->async_wait_all;
680 $pr->("reindexed ${$sync->{nr}}/$tot\n") if $pr;
683 sub _reindex_unseen { # git->cat_async callback
684 my ($bref, $oid, $type, $size, $req) = @_;
685 return if is_bad_blob($oid, $type, $size, $req->{oid});
686 my $self = $req->{self} // die 'BUG: {self} unset';
687 local $self->{current_info} = "$self->{current_info} $oid";
688 my $new_smsg = bless { blob => $oid, }, 'PublicInbox::Smsg';
689 $new_smsg->set_bytes($$bref, $size);
690 my $eml = $req->{eml} = PublicInbox::Eml->new($bref);
691 $req->{new_smsg} = $new_smsg;
692 $req->{chash} = content_hash($eml);
693 $req->{mids} = mids($eml); # do_step iterates through this
694 do_step($req); # enter the normal indexing flow
697 # --reindex may catch totally unseen messages, this handles them
698 sub reindex_unseen ($$$$) {
699 my ($self, $sync, $ibx, $xsmsg) = @_;
702 autime => $xsmsg->{ds},
703 cotime => $xsmsg->{ts},
704 oid => $xsmsg->{blob},
706 xnum => $xsmsg->{num},
707 # {mids} and {chash} will be filled in at _reindex_unseen
709 warn "I: reindex_unseen ${\$ibx->eidx_key}:$req->{xnum}:$req->{oid}\n";
710 $self->git->cat_async($xsmsg->{blob}, \&_reindex_unseen, $req);
713 sub _reindex_check_unseen ($$$) {
714 my ($self, $sync, $ibx) = @_;
715 my $ibx_id = $ibx->{-ibx_id};
717 my ($beg, $end) = (1, $slice);
719 # first, check if we missed any messages in target $ibx
721 my $pr = $sync->{-opt}->{-progress};
722 my $ekey = $ibx->eidx_key;
723 local $sync->{-regen_fmt} =
724 "$ekey checking unseen %u/".$ibx->over->max."\n";
727 while (scalar(@{$msgs = $ibx->over->query_xover($beg, $end)})) {
728 ${$sync->{nr}} = $beg;
729 $beg = $msgs->[-1]->{num} + 1;
730 $end = $beg + $slice;
731 if (checkpoint_due($sync)) {
732 reindex_checkpoint($self, $sync); # release lock
735 my $inx3 = $self->{oidx}->dbh->prepare_cached(<<'', undef, 1);
736 SELECT DISTINCT(docid) FROM xref3 WHERE
737 ibx_id = ? AND xnum = ? AND oidbin = ?
739 for my $xsmsg (@$msgs) {
740 my $oidbin = pack('H*', $xsmsg->{blob});
741 $inx3->bind_param(1, $ibx_id);
742 $inx3->bind_param(2, $xsmsg->{num});
743 $inx3->bind_param(3, $oidbin, SQL_BLOB);
745 my $docids = $inx3->fetchall_arrayref;
746 # index messages which were totally missed
747 # the first time around ASAP:
748 if (scalar(@$docids) == 0) {
749 reindex_unseen($self, $sync, $ibx, $xsmsg);
750 } else { # already seen, reindex later
751 for my $r (@$docids) {
752 $self->{oidx}->eidxq_add($r->[0]);
755 last if $sync->{quit};
757 last if $sync->{quit};
761 sub _reindex_check_stale ($$$) {
762 my ($self, $sync, $ibx) = @_;
764 my $pr = $sync->{-opt}->{-progress};
766 my $ekey = $ibx->eidx_key;
767 local $sync->{-regen_fmt} =
768 "$ekey check stale/missing %u/".$ibx->over->max."\n";
771 if (checkpoint_due($sync)) {
772 reindex_checkpoint($self, $sync); # release lock
774 # now, check if there's stale xrefs
775 my $iter = $self->{oidx}->dbh->prepare_cached(<<'', undef, 1);
776 SELECT docid,xnum,oidbin FROM xref3 WHERE ibx_id = ? AND docid > ?
777 ORDER BY docid,xnum ASC LIMIT 10000
779 $iter->execute($ibx->{-ibx_id}, $min);
782 while (my ($docid, $xnum, $oidbin) = $iter->fetchrow_array) {
783 return if $sync->{quit};
784 ${$sync->{nr}} = $xnum;
786 $fetching = $min = $docid;
787 my $smsg = $ibx->over->get_art($xnum);
791 } elsif (pack('H*', $smsg->{blob}) ne $oidbin) {
792 $err = "mismatch (!= $smsg->{blob})";
794 next; # likely, all good
796 # current_info already has eidx_key
797 my $oidhex = unpack('H*', $oidbin);
798 warn "$xnum:$oidhex (#$docid): $err\n";
799 my $del = $self->{oidx}->dbh->prepare_cached(<<'');
800 DELETE FROM xref3 WHERE ibx_id = ? AND xnum = ? AND oidbin = ?
802 $del->bind_param(1, $ibx->{-ibx_id});
803 $del->bind_param(2, $xnum);
804 $del->bind_param(3, $oidbin, SQL_BLOB);
807 # get_xref3 over-fetches, but this is a rare path:
808 my $xr3 = $self->{oidx}->get_xref3($docid);
809 my $idx = $self->idx_shard($docid);
810 if (scalar(@$xr3) == 0) { # all gone
811 $self->{oidx}->delete_by_num($docid);
812 $self->{oidx}->eidxq_del($docid);
813 $idx->ipc_do('xdb_remove', $docid);
814 } else { # enqueue for reindex of remaining messages
815 $idx->ipc_do('remove_eidx_info',
816 $docid, $ibx->eidx_key);
817 $self->{oidx}->eidxq_add($docid); # yes, add
820 } while (defined $fetching);
823 sub _reindex_inbox ($$$) {
824 my ($self, $sync, $ibx) = @_;
825 my $ekey = $ibx->eidx_key;
826 local $self->{current_info} = $ekey;
827 if (defined(my $err = _ibx_index_reject($ibx))) {
828 warn "W: cannot reindex $ekey ($err)\n";
830 _reindex_check_unseen($self, $sync, $ibx);
831 _reindex_check_stale($self, $sync, $ibx) unless $sync->{quit};
833 delete @$ibx{qw(over mm search git)}; # won't need these for a bit
837 my ($self, $sync) = @_;
839 # acquire eidxq_lock early because full reindex takes forever
840 # and incremental -extindex processes can run during our checkpoints
841 if (!eidxq_lock_acquire($self)) {
842 warn "E: aborting --reindex\n";
845 for my $ibx (@{ibx_sorted($self)}) {
846 _reindex_inbox($self, $sync, $ibx);
847 last if $sync->{quit};
849 $self->git->async_wait_all; # ensure eidxq gets filled completely
850 eidxq_process($self, $sync) unless $sync->{quit};
854 my ($self, $sync, $ibx) = @_;
855 my $err = _sync_inbox($self, $sync, $ibx);
856 delete @$ibx{qw(mm over)};
857 warn $err, "\n" if defined($err);
860 sub dd_smsg { # git->cat_async callback
861 my ($bref, $oid, $type, $size, $dd) = @_;
862 my $smsg = $dd->{smsg} // die 'BUG: dd->{smsg} missing';
863 my $self = $dd->{self} // die 'BUG: {self} missing';
864 my $per_mid = $dd->{per_mid} // die 'BUG: {per_mid} missing';
865 if ($type eq 'missing') {
866 _blob_missing($dd, $smsg);
867 } elsif (!is_bad_blob($oid, $type, $size, $smsg->{blob})) {
868 local $self->{current_info} = "$self->{current_info} $oid";
869 my $chash = content_hash(PublicInbox::Eml->new($bref));
870 push(@{$per_mid->{dd_chash}->{$chash}}, $smsg);
872 return if $per_mid->{last_smsg} != $smsg;
873 while (my ($chash, $ary) = each %{$per_mid->{dd_chash}}) {
874 my $keep = shift @$ary;
875 next if !scalar(@$ary);
876 $per_mid->{sync}->{dedupe_cull} += scalar(@$ary);
878 "# <$keep->{mid}> keeping #$keep->{num}, dropping ",
879 join(', ', map { "#$_->{num}" } @$ary),"\n";
880 next if $per_mid->{sync}->{-opt}->{'dry-run'};
881 my $oidx = $self->{oidx};
882 for my $smsg (@$ary) {
883 my $gone = $smsg->{num};
884 $oidx->merge_xref3($keep->{num}, $gone, $smsg->{blob});
885 $self->idx_shard($gone)->ipc_do('xdb_remove', $gone);
886 $oidx->delete_by_num($gone);
891 sub eidx_dedupe ($$$) {
892 my ($self, $sync, $msgids) = @_;
893 $sync->{dedupe_cull} = 0;
896 return unless eidxq_lock_acquire($self);
897 my ($iter, $cur_mid);
900 my ($max_id) = $self->{oidx}->dbh->selectrow_array(<<EOS);
901 SELECT MAX(id) FROM msgid
903 local $sync->{-regen_fmt} = "dedupe %u/$max_id\n";
905 # note: we could write this query more intelligently,
906 # but that causes lock contention with read-only processes
908 $cur_mid = $msgids->[$idx];
909 if ($cur_mid eq '') { # all Message-IDs
910 $iter = $self->{oidx}->dbh->prepare(<<EOS);
911 SELECT mid,id FROM msgid WHERE id > ? ORDER BY id ASC
913 $iter->execute($min_id);
915 $iter = $self->{oidx}->dbh->prepare(<<EOS);
916 SELECT mid,id FROM msgid WHERE mid = ? AND id > ? ORDER BY id ASC
918 $iter->execute($cur_mid, $min_id);
920 while (my ($mid, $id) = $iter->fetchrow_array) {
921 last if $sync->{quit};
922 $self->{current_info} = "dedupe $mid";
923 ${$sync->{nr}} = $min_id = $id;
925 while (my $x = $self->{oidx}->next_by_mid($mid, \$id, \$prv)) {
928 next if scalar(@smsg) < 2;
930 dd_chash => {}, # chash => [ary of smsgs]
931 last_smsg => $smsg[-1],
935 $candidates += scalar(@smsg) - 1;
936 for my $smsg (@smsg) {
942 $self->git->cat_async($smsg->{blob}, \&dd_smsg, $dd);
944 # need to wait on every single one @smsg contents can get
945 # invalidated inside dd_smsg for messages with multiple
947 $self->git->async_wait_all;
949 if (checkpoint_due($sync)) {
951 reindex_checkpoint($self, $sync);
955 goto dedupe_restart if defined($msgids->[++$idx]);
957 my $n = delete $sync->{dedupe_cull};
958 if (my $pr = $sync->{-opt}->{-progress}) {
959 $pr->("culled $n/$candidates candidates ($nr_mid msgids)\n");
964 sub eidx_sync { # main entry point
965 my ($self, $opt) = @_;
967 my $warn_cb = $SIG{__WARN__} || \&CORE::warn;
968 local $self->{current_info} = '';
969 local $SIG{__WARN__} = sub {
970 return if PublicInbox::Eml::warn_ignore(@_);
971 $warn_cb->($self->{current_info}, ': ', @_);
973 $self->idx_init($opt); # acquire lock via V2Writable::_idx_init
974 $self->{oidx}->rethread_prepare($opt);
976 need_checkpoint => \(my $need_checkpoint = 0),
978 next_check => now() + 10,
980 # DO NOT SET {reindex} here, it's incompatible with reused
981 # V2Writable code, reindex is totally different here
982 # compared to v1/v2 inboxes because we have multiple histories
984 -regen_fmt => "%u/?\n",
986 local $SIG{USR1} = sub { $need_checkpoint = 1 };
987 my $quit = PublicInbox::SearchIdx::quit_cb($sync);
988 local $SIG{QUIT} = $quit;
989 local $SIG{INT} = $quit;
990 local $SIG{TERM} = $quit;
991 for my $ibx (@{ibx_sorted($self)}) {
992 $ibx->{-ibx_id} //= $self->{oidx}->ibx_id($ibx->eidx_key);
994 if (my $msgids = delete($opt->{dedupe})) {
995 local $sync->{checkpoint_unlocks} = 1;
996 eidx_dedupe($self, $sync, $msgids);
998 if (delete($opt->{reindex})) {
999 local $sync->{checkpoint_unlocks} = 1;
1000 eidx_reindex($self, $sync);
1003 # don't use $_ here, it'll get clobbered by reindex_checkpoint
1004 if ($opt->{scan} // 1) {
1005 for my $ibx (@{ibx_sorted($self)}) {
1006 last if $sync->{quit};
1007 sync_inbox($self, $sync, $ibx);
1010 $self->{oidx}->rethread_done($opt) unless $sync->{quit};
1011 eidxq_process($self, $sync) unless $sync->{quit};
1013 eidxq_release($self);
1015 $sync; # for eidx_watch
1018 sub update_last_commit { # overrides V2Writable
1019 my ($self, $sync, $stk) = @_;
1020 my $unit = $sync->{unit} // return;
1021 my $latest_cmt = $stk ? $stk->{latest_cmt} : ${$sync->{latest_cmt}};
1022 defined($latest_cmt) or return;
1023 my $ibx = $sync->{ibx} or die 'BUG: {ibx} missing';
1024 my $ekey = $ibx->eidx_key;
1025 my $uv = $ibx->uidvalidity;
1026 my $epoch = $unit->{epoch};
1028 my $v = $ibx->version;
1030 die 'No {epoch} for v2 unit' unless defined $epoch;
1031 $meta_key = "lc-v2:$ekey//$uv;$epoch";
1033 die 'Unexpected {epoch} for v1 unit' if defined $epoch;
1034 $meta_key = "lc-v1:$ekey//$uv";
1036 die "Unsupported inbox version: $v";
1038 my $last = $self->{oidx}->eidx_meta($meta_key);
1039 if (defined $last && is_ancestor($self->git, $last, $latest_cmt)) {
1040 my @cmd = (qw(rev-list --count), "$last..$latest_cmt");
1041 chomp(my $n = $unit->{git}->qx(@cmd));
1042 return if $n ne '' && $n == 0;
1044 $self->{oidx}->eidx_meta($meta_key, $latest_cmt);
1047 sub _idx_init { # with_umask callback
1048 my ($self, $opt) = @_;
1049 PublicInbox::V2Writable::_idx_init($self, $opt); # acquires ei.lock
1050 $self->{midx} = PublicInbox::MiscIdx->new($self);
1053 sub symlink_packs ($$) {
1054 my ($ibx, $pd) = @_;
1056 my $glob = "$ibx->{inboxdir}/git/*.git/objects/pack/*.idx";
1057 for my $idx (bsd_glob($glob, GLOB_NOSORT)) {
1058 my $src = substr($idx, 0, -length('.idx'));
1059 my $dst = $pd . substr($src, rindex($src, '/'));
1060 if (-f "$src.pack" and
1061 symlink("$src.pack", "$dst.pack") and
1062 symlink($idx, "$dst.idx") and
1065 # .promisor, .bitmap, .rev and .keep are optional
1066 # XXX should we symlink .keep here?
1067 for my $s (qw(promisor bitmap rev)) {
1068 symlink("$src.$s", "$dst.$s") if -f "$src.$s";
1070 } elsif (!$!{EEXIST}) {
1071 warn "W: ln -s $src.{pack,idx} => $dst.*: $!\n";
1072 unlink "$dst.pack", "$dst.idx";
1078 sub idx_init { # similar to V2Writable
1079 my ($self, $opt) = @_;
1080 return if $self->{idx_shards};
1082 $self->git->cleanup;
1084 my $ALL = $self->git->{git_dir}; # ALL.git
1086 if ($opt->{-private}) { # LeiStore
1089 umask 077; # don't bother restoring
1090 PublicInbox::Import::init_bare($ALL);
1091 $self->git->qx(qw(config core.sharedRepository 0600));
1094 PublicInbox::Import::init_bare($ALL) unless $old;
1096 my $info_dir = "$ALL/objects/info";
1097 my $alt = "$info_dir/alternates";
1098 my (@old, @new, %seen); # seen: st_dev + st_ino
1100 open(my $fh, '<', $alt) or die "open $alt: $!";
1101 $mode = (stat($fh))[2] & 07777;
1102 while (my $line = <$fh>) {
1103 chomp(my $d = $line);
1105 # expand relative path (/local/ stuff)
1106 substr($d, 0, 3) eq '../' and
1107 $d = "$ALL/objects/$d";
1108 if (my @st = stat($d)) {
1109 next if $seen{"$st[0]\0$st[1]"}++;
1111 warn "W: stat($d) failed (from $alt): $!\n";
1112 next if $opt->{-idx_gc};
1118 # for LeiStore, and possibly some mirror-only state
1119 if (opendir(my $dh, my $local = "$self->{topdir}/local")) {
1120 # highest numbered epoch first
1121 for my $n (sort { $b <=> $a } map { substr($_, 0, -4) + 0 }
1122 grep(/\A[0-9]+\.git\z/, readdir($dh))) {
1123 my $d = "$local/$n.git/objects"; # absolute path
1124 if (my @st = stat($d)) {
1125 next if $seen{"$st[0]\0$st[1]"}++;
1126 # favor relative paths for rename-friendliness
1127 push @new, "../../local/$n.git/objects\n";
1129 warn "W: stat($d) failed: $!\n";
1133 # git-multi-pack-index(1) can speed up "git cat-file" startup slightly
1136 my $pd = "$ALL/objects/pack";
1137 if (!mkdir($pd) && $!{EEXIST} && opendir($dh, $pd)) {
1138 # drop stale symlinks
1139 while (defined(my $dn = readdir($dh))) {
1140 if ($dn =~ /\.(?:idx|pack|promisor|bitmap|rev)\z/) {
1142 unlink($f) if -l $f && !-e $f;
1147 for my $ibx (@{ibx_sorted($self)}) {
1148 # create symlinks for multi-pack-index
1149 $git_midx += symlink_packs($ibx, $pd);
1150 # add new lines to our alternates file
1151 my $line = $ibx->git->{git_dir} . "/objects\n";
1152 chomp(my $d = $line);
1153 if (my @st = stat($d)) {
1154 next if $seen{"$st[0]\0$st[1]"}++;
1156 warn "W: stat($d) failed (from $ibx->{inboxdir}): $!\n";
1157 next if $opt->{-idx_gc};
1164 PublicInbox::V2Writable::write_alternates($info_dir, $mode, $o);
1166 $git_midx and $self->with_umask(sub {
1167 my @cmd = ('multi-pack-index');
1168 push @cmd, '--no-progress' if ($opt->{quiet}//0) > 1;
1169 my $lk = $self->lock_for_scope;
1170 system('git', "--git-dir=$ALL", @cmd, 'write');
1171 # ignore errors, fairly new command, may not exist
1173 $self->parallel_init($self->{indexlevel});
1174 $self->with_umask(\&_idx_init, $self, $opt);
1175 $self->{oidx}->begin_lazy;
1176 $self->{oidx}->eidx_prep;
1177 $self->{midx}->create_xdb if @new;
1180 sub _watch_commit { # PublicInbox::DS::add_timer callback
1182 delete $self->{-commit_timer};
1183 eidxq_process($self, $self->{-watch_sync});
1184 eidxq_release($self);
1185 my $fmt = delete $self->{-watch_sync}->{-regen_fmt};
1186 reindex_checkpoint($self, $self->{-watch_sync});
1187 $self->{-watch_sync}->{-regen_fmt} = $fmt;
1189 # call event_step => done unless commit_timer is armed
1190 PublicInbox::DS::requeue($self);
1193 sub on_inbox_unlock { # called by PublicInbox::InboxIdle
1194 my ($self, $ibx) = @_;
1195 my $opt = $self->{-watch_sync}->{-opt};
1196 my $pr = $opt->{-progress};
1197 my $ekey = $ibx->eidx_key;
1198 local $0 = "sync $ekey";
1199 $pr->("indexing $ekey\n") if $pr;
1200 $self->idx_init($opt);
1201 sync_inbox($self, $self->{-watch_sync}, $ibx);
1202 $self->{-commit_timer} //= add_timer($opt->{'commit-interval'} // 10,
1203 \&_watch_commit, $self);
1206 sub eidx_reload { # -extindex --watch SIGHUP handler
1207 my ($self, $idler) = @_;
1209 my $pr = $self->{-watch_sync}->{-opt}->{-progress};
1210 $pr->('reloading ...') if $pr;
1211 delete $self->{-resync_queue};
1212 delete $self->{-ibx_ary};
1213 $self->{ibx_cfg} = [];
1214 %{$self->{ibx_map}} = ();
1215 delete $self->{-watch_sync}->{id2pos};
1216 my $cfg = PublicInbox::Config->new;
1217 attach_config($self, $cfg);
1218 $idler->refresh($cfg);
1219 $pr->(" done\n") if $pr;
1221 warn "reload not supported without --all\n";
1225 sub eidx_resync_start ($) { # -extindex --watch SIGUSR1 handler
1227 $self->{-resync_queue} //= [ @{ibx_sorted($self)} ];
1228 PublicInbox::DS::requeue($self); # trigger our ->event_step
1231 sub event_step { # PublicInbox::DS::requeue callback
1233 if (my $resync_queue = $self->{-resync_queue}) {
1234 if (my $ibx = shift(@$resync_queue)) {
1235 on_inbox_unlock($self, $ibx);
1236 PublicInbox::DS::requeue($self);
1238 delete $self->{-resync_queue};
1239 _watch_commit($self);
1242 done($self) unless $self->{-commit_timer};
1246 sub eidx_watch { # public-inbox-extindex --watch main loop
1247 my ($self, $opt) = @_;
1249 for my $sig (qw(HUP USR1 TSTP QUIT INT TERM)) {
1250 $SIG{$sig} = sub { warn "SIG$sig ignored while scanning\n" };
1252 require PublicInbox::InboxIdle;
1253 require PublicInbox::DS;
1254 require PublicInbox::Syscall;
1255 require PublicInbox::Sigfd;
1256 my $idler = PublicInbox::InboxIdle->new($self->{cfg});
1257 if (!$self->{cfg}) {
1258 $idler->watch_inbox($_) for (@{ibx_sorted($self)});
1260 $_->subscribe_unlock(__PACKAGE__, $self) for (@{ibx_sorted($self)});
1261 my $pr = $opt->{-progress};
1262 $pr->("performing initial scan ...\n") if $pr;
1263 my $sync = eidx_sync($self, $opt); # initial sync
1264 return if $sync->{quit};
1265 my $oldset = PublicInbox::DS::block_signals();
1266 local $self->{current_info} = '';
1267 my $cb = $SIG{__WARN__} || \&CORE::warn;
1268 local $SIG{__WARN__} = sub {
1269 return if PublicInbox::Eml::warn_ignore(@_);
1270 $cb->($self->{current_info}, ': ', @_);
1273 HUP => sub { eidx_reload($self, $idler) },
1274 USR1 => sub { eidx_resync_start($self) },
1275 TSTP => sub { kill('STOP', $$) },
1277 my $quit = PublicInbox::SearchIdx::quit_cb($sync);
1278 $sig->{QUIT} = $sig->{INT} = $sig->{TERM} = $quit;
1279 my $sigfd = PublicInbox::Sigfd->new($sig,
1280 $PublicInbox::Syscall::SFD_NONBLOCK);
1281 %SIG = (%SIG, %$sig) if !$sigfd;
1282 local $self->{-watch_sync} = $sync; # for ->on_inbox_unlock
1284 # wake up every second to accept signals if we don't
1285 # have signalfd or IO::KQueue:
1286 PublicInbox::DS::sig_setmask($oldset);
1287 PublicInbox::DS->SetLoopTimeout(1000);
1289 PublicInbox::DS->SetPostLoopCallback(sub { !$sync->{quit} });
1290 $pr->("initial scan complete, entering event loop\n") if $pr;
1291 PublicInbox::DS->EventLoop; # calls InboxIdle->event_step
1296 *done = \&PublicInbox::V2Writable::done;
1297 *with_umask = \&PublicInbox::InboxWritable::with_umask;
1298 *parallel_init = \&PublicInbox::V2Writable::parallel_init;
1299 *nproc_shards = \&PublicInbox::V2Writable::nproc_shards;
1300 *sync_prepare = \&PublicInbox::V2Writable::sync_prepare;
1301 *index_todo = \&PublicInbox::V2Writable::index_todo;
1302 *count_shards = \&PublicInbox::V2Writable::count_shards;
1303 *atfork_child = \&PublicInbox::V2Writable::atfork_child;
1304 *idx_shard = \&PublicInbox::V2Writable::idx_shard;
1305 *reindex_checkpoint = \&PublicInbox::V2Writable::reindex_checkpoint;
1306 *checkpoint = \&PublicInbox::V2Writable::checkpoint;