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::MultiGit;
25 use PublicInbox::Search;
26 use PublicInbox::SearchIdx qw(prepare_stack is_ancestor is_bad_blob);
27 use PublicInbox::OverIdx;
28 use PublicInbox::MiscIdx;
29 use PublicInbox::MID qw(mids);
30 use PublicInbox::V2Writable;
31 use PublicInbox::InboxWritable;
32 use PublicInbox::ContentHash qw(content_hash);
34 use PublicInbox::DS qw(now add_timer);
35 use DBI qw(:sql_types); # SQL_BLOB
38 my (undef, $dir, $opt) = @_;
39 my $l = $opt->{indexlevel} // 'full';
40 $l !~ $PublicInbox::SearchIdx::INDEXLEVELS and
41 die "invalid indexlevel=$l\n";
42 $l eq 'basic' and die "E: indexlevel=basic not yet supported\n";
44 xpfx => "$dir/ei".PublicInbox::Search::SCHEMA_VERSION,
46 creat => $opt->{creat},
47 ibx_map => {}, # (newsgroup//inboxdir) => $ibx
48 ibx_active => [], # by config section order
49 ibx_known => [], # by config section order
55 lock_path => "$dir/ei.lock",
57 $self->{shards} = $self->count_shards ||
58 nproc_shards({ nproc => $opt->{jobs} });
59 my $oidx = PublicInbox::OverIdx->new("$self->{xpfx}/over.sqlite3");
60 $self->{-no_fsync} = $oidx->{-no_fsync} = 1 if !$opt->{fsync};
61 $self->{oidx} = $oidx;
66 my ($self, $ibx, $types) = @_;
67 $self->{ibx_map}->{$ibx->eidx_key} //= do {
68 delete $self->{-ibx_ary_known}; # invalidate cache
69 delete $self->{-ibx_ary_active}; # invalidate cache
70 $types //= [ qw(active known) ];
72 push @{$self->{"ibx_$t"}}, $ibx;
78 sub _ibx_attach { # each_inbox callback
79 my ($ibx, $self, $types) = @_;
80 attach_inbox($self, $ibx, $types);
84 my ($self, $cfg, $ibxs) = @_;
88 for my $ibx (@$ibxs) {
89 $self->{ibx_map}->{$ibx->eidx_key} //= do {
90 push @{$self->{ibx_active}}, $ibx;
91 push @{$self->{ibx_known}}, $ibx;
96 delete $self->{-ibx_ary_known};
97 delete $self->{-ibx_ary_active};
100 $types //= [ qw(known active) ];
101 $cfg->each_inbox(\&_ibx_attach, $self, $types);
104 sub check_batch_limit ($) {
106 my $self = $req->{self};
107 my $new_smsg = $req->{new_smsg};
108 my $n = $self->{transact_bytes} += $new_smsg->{bytes};
110 # set flag for PublicInbox::V2Writable::index_todo:
111 ${$req->{need_checkpoint}} = 1 if $n >= $self->{batch_bytes};
114 sub apply_boost ($$) {
115 my ($req, $smsg) = @_;
116 my $id2pos = $req->{id2pos}; # index in ibx_sorted
117 my $xr3 = $req->{self}->{oidx}->get_xref3($smsg->{num}, 1);
119 $id2pos->{$a->[0]} <=> $id2pos->{$b->[0]}
121 $a->[1] <=> $b->[1] # break ties with {xnum}
123 my $new_smsg = $req->{new_smsg};
124 return if $xr3->[0]->[2] ne pack('H*', $new_smsg->{blob}); # loser
126 # replace the old smsg with the more boosted one
127 $new_smsg->{num} = $smsg->{num};
128 $new_smsg->populate($req->{eml}, $req);
129 $req->{self}->{oidx}->add_overview($req->{eml}, $new_smsg);
132 sub remove_doc ($$) {
133 my ($self, $docid) = @_;
134 $self->{oidx}->delete_by_num($docid);
135 $self->{oidx}->eidxq_del($docid);
136 $self->idx_shard($docid)->ipc_do('xdb_remove', $docid);
139 sub _unref_doc ($$$$$;$) {
140 my ($sync, $docid, $ibx, $xnum, $oidbin, $eml) = @_;
144 $docid = $smsg->{num};
146 my $s = 'DELETE FROM xref3 WHERE oidbin = ?';
147 $s .= ' AND ibx_id = ?' if defined($ibx);
148 $s .= ' AND xnum = ?' if defined($xnum);
149 my $del = $sync->{self}->{oidx}->dbh->prepare_cached($s);
151 $del->bind_param(++$col, $oidbin, SQL_BLOB);
152 $del->bind_param(++$col, $ibx->{-ibx_id}) if $ibx;
153 $del->bind_param(++$col, $xnum) if defined($xnum);
155 my $xr3 = $sync->{self}->{oidx}->get_xref3($docid);
156 if (scalar(@$xr3) == 0) { # all gone
157 remove_doc($sync->{self}, $docid);
158 } else { # enqueue for reindex of remaining messages
160 my $ekey = $ibx->{-gc_eidx_key} // $ibx->eidx_key;
161 my $idx = $sync->{self}->idx_shard($docid);
162 $idx->ipc_do('remove_eidx_info', $docid, $ekey, $eml);
163 } # else: we can't remove_eidx_info in reindex-only path
165 # replace invalidated blob ASAP with something which should be
166 # readable since we may commit the transaction on checkpoint.
167 # eidxq processing will re-apply boost
168 $smsg //= $sync->{self}->{oidx}->get_art($docid);
169 my $hex = unpack('H*', $oidbin);
170 if ($smsg && $smsg->{blob} eq $hex) {
171 $xr3->[0] =~ /:([a-f0-9]{40,}+)\z/ or
172 die "BUG: xref $xr3->[0] has no OID";
173 $sync->{self}->{oidx}->update_blob($smsg, $1);
175 # yes, add, we'll need to re-apply boost
176 $sync->{self}->{oidx}->eidxq_add($docid);
182 my ($req, $smsg) = @_;
183 my $self = $req->{self};
184 my $docid = $smsg->{num};
185 my $oid = $req->{oid};
186 my $xibx = $req->{ibx};
187 my $eml = $req->{eml};
188 if (my $new_smsg = $req->{new_smsg}) { # 'm' on cross-posted message
189 my $eidx_key = $xibx->eidx_key;
190 my $xnum = $req->{xnum};
191 $self->{oidx}->add_xref3($docid, $xnum, $oid, $eidx_key);
192 my $idx = $self->idx_shard($docid);
193 $idx->ipc_do('add_eidx_info', $docid, $eidx_key, $eml);
194 apply_boost($req, $smsg) if $req->{boost_in_use};
195 } else { # 'd' no {xnum}
196 $oid = pack('H*', $oid);
197 _unref_doc($req, $docid, $xibx, undef, $oid, $eml);
201 # called by V2Writable::sync_prepare
202 sub artnum_max { $_[0]->{oidx}->eidx_max }
204 sub index_unseen ($) {
206 my $new_smsg = $req->{new_smsg} or die 'BUG: {new_smsg} unset';
207 my $eml = delete $req->{eml};
208 $new_smsg->populate($eml, $req);
209 my $self = $req->{self};
210 my $docid = $self->{oidx}->adj_counter('eidx_docid', '+');
211 $new_smsg->{num} = $docid;
212 my $idx = $self->idx_shard($docid);
213 $self->{oidx}->add_overview($eml, $new_smsg);
214 my $oid = $new_smsg->{blob};
215 my $ibx = delete $req->{ibx} or die 'BUG: {ibx} unset';
216 $self->{oidx}->add_xref3($docid, $req->{xnum}, $oid, $ibx->eidx_key);
217 $idx->index_eml($eml, $new_smsg, $ibx->eidx_key);
218 check_batch_limit($req);
221 sub do_finalize ($) {
223 if (my $indexed = $req->{indexed}) { # duplicated messages
224 do_xpost($req, $_) for @$indexed;
225 } elsif (exists $req->{new_smsg}) { # totally unseen messsage
228 # `d' message was already unindexed in the v1/v2 inboxes,
229 # so it's too noisy to warn, here.
231 # cur_cmt may be undef for unindex_oid, set by V2Writable::index_todo
232 if (defined(my $cur_cmt = $req->{cur_cmt})) {
233 ${$req->{latest_cmt}} = $cur_cmt;
237 sub do_step ($) { # main iterator for adding messages to the index
239 my $self = $req->{self} // die 'BUG: {self} missing';
241 if (my $next_arg = $req->{next_arg}) {
242 if (my $smsg = $self->{oidx}->next_by_mid(@$next_arg)) {
243 $req->{cur_smsg} = $smsg;
244 $self->git->cat_async($smsg->{blob},
245 \&ck_existing, $req);
246 return; # ck_existing calls do_step
248 delete $req->{next_arg};
250 die "BUG: {cur_smsg} still set" if $req->{cur_smsg};
251 my $mid = shift(@{$req->{mids}}) // last;
253 $req->{next_arg} = [ $mid, \$id, \$prev ];
259 sub _blob_missing ($$) { # called when a known $smsg->{blob} is gone
260 my ($req, $smsg) = @_;
261 # xnum and ibx are unknown, we only call this when an entry from
262 # /ei*/over.sqlite3 is bad, not on entries from xap*/over.sqlite3
263 my $oidbin = pack('H*', $smsg->{blob});
264 _unref_doc($req, $smsg, undef, undef, $oidbin);
267 sub ck_existing { # git->cat_async callback
268 my ($bref, $oid, $type, $size, $req) = @_;
269 my $smsg = delete $req->{cur_smsg} or die 'BUG: {cur_smsg} missing';
270 if ($type eq 'missing') {
271 _blob_missing($req, $smsg);
272 } elsif (!is_bad_blob($oid, $type, $size, $smsg->{blob})) {
273 my $self = $req->{self} // die 'BUG: {self} missing';
274 local $self->{current_info} = "$self->{current_info} $oid";
275 my $cur = PublicInbox::Eml->new($bref);
276 if (content_hash($cur) eq $req->{chash}) {
277 push @{$req->{indexed}}, $smsg; # for do_xpost
278 } # else { index_unseen later }
283 # is the messages visible in the inbox currently being indexed?
284 # return the number if so
285 sub cur_ibx_xnum ($$) {
286 my ($req, $bref) = @_;
287 my $ibx = $req->{ibx} or die 'BUG: current {ibx} missing';
289 $req->{eml} = PublicInbox::Eml->new($bref);
290 $req->{chash} = content_hash($req->{eml});
291 $req->{mids} = mids($req->{eml});
292 for my $mid (@{$req->{mids}}) {
294 while (my $x = $ibx->over->next_by_mid($mid, \$id, \$prev)) {
295 return $x->{num} if $x->{blob} eq $req->{oid};
301 sub index_oid { # git->cat_async callback for 'm'
302 my ($bref, $oid, $type, $size, $req) = @_;
303 my $self = $req->{self};
304 local $self->{current_info} = "$self->{current_info} $oid";
305 return if is_bad_blob($oid, $type, $size, $req->{oid});
306 my $new_smsg = $req->{new_smsg} = bless {
308 }, 'PublicInbox::Smsg';
309 $new_smsg->set_bytes($$bref, $size);
310 defined($req->{xnum} = cur_ibx_xnum($req, $bref)) or return;
315 sub unindex_oid { # git->cat_async callback for 'd'
316 my ($bref, $oid, $type, $size, $req) = @_;
317 my $self = $req->{self};
318 local $self->{current_info} = "$self->{current_info} $oid";
319 return if is_bad_blob($oid, $type, $size, $req->{oid});
320 return if defined(cur_ibx_xnum($req, $bref)); # was re-added
324 # overrides V2Writable::last_commits, called by sync_ranges via sync_prepare
326 my ($self, $sync) = @_;
328 my $ekey = $sync->{ibx}->eidx_key;
329 my $uv = $sync->{ibx}->uidvalidity;
330 for my $i (0..$sync->{epoch_max}) {
331 $heads->[$i] = $self->{oidx}->eidx_meta("lc-v2:$ekey//$uv;$i");
336 sub _ibx_index_reject ($) {
338 $ibx->mm // return 'unindexed, no msgmap.sqlite3';
339 $ibx->uidvalidity // return 'no UIDVALIDITY';
340 $ibx->over // return 'unindexed, no over.sqlite3';
344 sub _sync_inbox ($$$) {
345 my ($self, $sync, $ibx) = @_;
346 my $ekey = $ibx->eidx_key;
347 if (defined(my $err = _ibx_index_reject($ibx))) {
348 return "W: skipping $ekey ($err)";
351 $sync->{nr} = \(my $nr = 0);
352 my $v = $ibx->version;
354 $sync->{epoch_max} = $ibx->max_git_epoch // return;
355 sync_prepare($self, $sync); # or return # TODO: once MiscIdx is stable
357 my $uv = $ibx->uidvalidity;
358 my $lc = $self->{oidx}->eidx_meta("lc-v1:$ekey//$uv");
359 my $head = $ibx->mm->last_commit //
360 return "E: $ibx->{inboxdir} is not indexed";
361 my $stk = prepare_stack($sync, $lc ? "$lc..$head" : $head);
362 my $unit = { stack => $stk, git => $ibx->git };
363 push @{$sync->{todo}}, $unit;
365 return "E: $ekey unsupported inbox version (v$v)";
367 for my $unit (@{delete($sync->{todo}) // []}) {
368 last if $sync->{quit};
369 index_todo($self, $sync, $unit);
371 $self->{midx}->index_ibx($ibx) unless $sync->{quit};
372 $ibx->git->cleanup; # done with this inbox, now
376 sub eidx_gc_scan_inboxes ($$) {
377 my ($self, $sync) = @_;
378 my ($x3_doc, $ibx_ck);
380 $x3_doc = $self->{oidx}->dbh->prepare(<<EOM);
381 SELECT docid,xnum,oidbin FROM xref3 WHERE ibx_id = ?
383 $ibx_ck = $self->{oidx}->dbh->prepare(<<EOM);
384 SELECT ibx_id,eidx_key FROM inboxes
387 while (my ($ibx_id, $eidx_key) = $ibx_ck->fetchrow_array) {
388 next if $self->{ibx_map}->{$eidx_key};
389 $self->{midx}->remove_eidx_key($eidx_key);
390 warn "I: deleting messages for $eidx_key...\n";
391 $x3_doc->execute($ibx_id);
392 my $ibx = { -ibx_id => $ibx_id, -gc_eidx_key => $eidx_key };
393 while (my ($docid, $xnum, $oid) = $x3_doc->fetchrow_array) {
394 my $r = _unref_doc($sync, $docid, $ibx, $xnum, $oid);
395 $oid = unpack('H*', $oid);
396 $r = $r ? 'unref' : 'remove';
397 warn "I: $r #$docid $eidx_key $oid\n";
398 if (checkpoint_due($sync)) {
399 $x3_doc = $ibx_ck = undef;
400 reindex_checkpoint($self, $sync);
404 $self->{oidx}->dbh->do(<<'', undef, $ibx_id);
405 DELETE FROM inboxes WHERE ibx_id = ?
407 # drop last_commit info
409 $pat =~ s/([_%\\])/\\$1/g;
410 $self->{oidx}->dbh->do('PRAGMA case_sensitive_like = ON');
411 my $lc_i = $self->{oidx}->dbh->prepare(<<'');
412 SELECT key FROM eidx_meta WHERE key LIKE ? ESCAPE ?
414 $lc_i->execute("lc-%:$pat//%", '\\');
415 while (my ($key) = $lc_i->fetchrow_array) {
416 next if $key !~ m!\Alc-v[1-9]+:\Q$eidx_key\E//!;
417 warn "I: removing $key\n";
418 $self->{oidx}->dbh->do(<<'', undef, $key);
419 DELETE FROM eidx_meta WHERE key = ?
422 warn "I: $eidx_key removed\n";
426 sub eidx_gc_scan_shards ($$) { # TODO: use for lei/store
427 my ($self, $sync) = @_;
428 my $nr = $self->{oidx}->dbh->do(<<'');
429 DELETE FROM xref3 WHERE docid NOT IN (SELECT num FROM over)
431 warn "I: eliminated $nr stale xref3 entries\n" if $nr != 0;
432 reindex_checkpoint($self, $sync) if checkpoint_due($sync);
434 # fixup from old bugs:
435 $nr = $self->{oidx}->dbh->do(<<'');
436 DELETE FROM over WHERE num > 0 AND num NOT IN (SELECT docid FROM xref3)
438 warn "I: eliminated $nr stale over entries\n" if $nr != 0;
439 reindex_checkpoint($self, $sync) if checkpoint_due($sync);
441 $nr = $self->{oidx}->dbh->do(<<'');
442 DELETE FROM eidxq WHERE docid NOT IN (SELECT num FROM over)
444 warn "I: eliminated $nr stale reindex queue entries\n" if $nr != 0;
445 reindex_checkpoint($self, $sync) if checkpoint_due($sync);
447 my ($cur) = $self->{oidx}->dbh->selectrow_array(<<EOM);
448 SELECT MIN(num) FROM over WHERE num > 0
450 $cur // return; # empty
451 my ($r, $n, %active_shards);
454 $r = $self->{oidx}->dbh->selectcol_arrayref(<<"", undef, $cur);
455 SELECT num FROM over WHERE num >= ? ORDER BY num ASC LIMIT 10000
457 last unless scalar(@$r);
458 while (defined($n = shift @$r)) {
459 for my $i ($cur..($n - 1)) {
460 my $idx = idx_shard($self, $i);
461 $idx->ipc_do('xdb_remove_quiet', $i);
462 $active_shards{$idx} = $idx;
466 if (checkpoint_due($sync)) {
467 for my $idx (values %active_shards) {
468 $nr += $idx->ipc_do('nr_quiet_rm')
471 reindex_checkpoint($self, $sync);
474 warn "I: eliminated $nr stale Xapian documents\n" if $nr != 0;
478 my ($self, $opt) = @_;
479 $self->{cfg} or die "E: GC requires ->attach_config\n";
482 need_checkpoint => \(my $need_checkpoint = 0),
484 next_check => now() + 10,
485 checkpoint_unlocks => 1,
489 $self->idx_init($opt); # acquire lock via V2Writable::_idx_init
490 eidx_gc_scan_inboxes($self, $sync);
491 eidx_gc_scan_shards($self, $sync);
496 my ($self, $sync, $smsg) = @_;
497 my $ibx_id = delete($smsg->{ibx_id}) // die '{ibx_id} unset';
498 my $pos = $sync->{id2pos}->{$ibx_id} // die "$ibx_id no pos";
499 $self->{-ibx_ary_known}->[$pos] //
500 die "BUG: ibx for $smsg->{blob} not mapped"
503 sub _fd_constrained ($) {
505 $self->{-fd_constrained} //= do {
507 if (eval { require BSD::Resource; 1 }) {
508 my $NOFILE = BSD::Resource::RLIMIT_NOFILE();
509 ($soft, undef) = BSD::Resource::getrlimit($NOFILE);
511 chomp($soft = `sh -c 'ulimit -n'`);
513 if (defined($soft)) {
514 # $want is an estimate
515 my $want = scalar(@{$self->{ibx_active}}) + 64;
516 my $ret = $want > $soft;
519 RLIMIT_NOFILE=$soft insufficient (want: $want), will close DB handles early
524 warn "Unable to determine RLIMIT_NOFILE: $@\n";
530 sub _reindex_finalize ($$$) {
531 my ($req, $smsg, $eml) = @_;
532 my $sync = $req->{sync};
533 my $self = $sync->{self};
534 my $by_chash = delete $req->{by_chash} or die 'BUG: no {by_chash}';
535 my $nr = scalar(keys(%$by_chash)) or die 'BUG: no content hashes';
536 my $orig_smsg = $req->{orig_smsg} // die 'BUG: no {orig_smsg}';
537 my $docid = $smsg->{num} = $orig_smsg->{num};
538 $self->{oidx}->add_overview($eml, $smsg); # may rethread
539 check_batch_limit({ %$sync, new_smsg => $smsg });
540 my $chash0 = $smsg->{chash} // die "BUG: $smsg->{blob} no {chash}";
541 my $stable = delete($by_chash->{$chash0}) //
542 die "BUG: $smsg->{blob} chash missing";
543 my $idx = $self->idx_shard($docid);
544 my $top_smsg = pop @$stable;
545 $top_smsg == $smsg or die 'BUG: top_smsg != smsg';
546 my $ibx = _ibx_for($self, $sync, $smsg);
547 $idx->index_eml($eml, $smsg, $ibx->eidx_key);
548 for my $x (reverse @$stable) {
549 $ibx = _ibx_for($self, $sync, $x);
550 my $hdr = delete $x->{hdr} // die 'BUG: no {hdr}';
551 $idx->ipc_do('add_eidx_info', $docid, $ibx->eidx_key, $hdr);
553 return if $nr == 1; # likely, all good
555 warn "W: #$docid split into $nr due to deduplication change\n";
557 for my $ary (values %$by_chash) {
558 for my $x (reverse @$ary) {
559 warn "removing #$docid xref3 $x->{blob}\n";
560 my $bin = pack('H*', $x->{blob});
561 my $n = _unref_doc($sync, $docid, undef, undef, $bin);
562 die "BUG: $x->{blob} invalidated #$docid" if $n == 0;
564 my $x = pop(@$ary) // die "BUG: #$docid {by_chash} empty";
565 $x->{num} = delete($x->{xnum}) // die '{xnum} unset';
566 $ibx = _ibx_for($self, $sync, $x);
567 if (my $over = $ibx->over) {
568 my $e = $over->get_art($x->{num});
569 $e->{blob} eq $x->{blob} or die <<EOF;
570 $x->{blob} != $e->{blob} (${\$ibx->eidx_key}:$e->{num});
572 push @todo, $ibx, $e;
573 $over->dbh_close if _fd_constrained($self);
575 die "$ibx->{inboxdir}: over.sqlite3 unusable: $!\n";
579 while (my ($ibx, $e) = splice(@todo, 0, 2)) {
580 reindex_unseen($self, $sync, $ibx, $e);
584 sub _reindex_oid { # git->cat_async callback
585 my ($bref, $oid, $type, $size, $req) = @_;
586 my $sync = $req->{sync};
587 my $self = $sync->{self};
588 my $orig_smsg = $req->{orig_smsg} // die 'BUG: no {orig_smsg}';
589 my $expect_oid = $req->{xr3r}->[$req->{ix}]->[2];
590 my $docid = $orig_smsg->{num};
591 if (is_bad_blob($oid, $type, $size, $expect_oid)) {
592 my $oidbin = pack('H*', $expect_oid);
593 my $remain = _unref_doc($sync, $docid, undef, undef, $oidbin);
595 warn "W: #$docid gone or corrupted\n";
596 } elsif (my $next_oid = $req->{xr3r}->[++$req->{ix}]->[2]) {
597 $self->git->cat_async($next_oid, \&_reindex_oid, $req);
599 warn "BUG: #$docid gone (UNEXPECTED)\n";
603 my $ci = $self->{current_info};
604 local $self->{current_info} = "$ci #$docid $oid";
605 my $re_smsg = bless { blob => $oid }, 'PublicInbox::Smsg';
606 $re_smsg->set_bytes($$bref, $size);
607 my $eml = PublicInbox::Eml->new($bref);
608 $re_smsg->populate($eml, { autime => $orig_smsg->{ds},
609 cotime => $orig_smsg->{ts} });
610 my $chash = content_hash($eml);
611 $re_smsg->{chash} = $chash;
612 $re_smsg->{xnum} = $req->{xr3r}->[$req->{ix}]->[1];
613 $re_smsg->{ibx_id} = $req->{xr3r}->[$req->{ix}]->[0];
614 $re_smsg->{hdr} = $eml->header_obj;
615 push @{$req->{by_chash}->{$chash}}, $re_smsg;
616 if (my $next_oid = $req->{xr3r}->[++$req->{ix}]->[2]) {
617 $self->git->cat_async($next_oid, \&_reindex_oid, $req);
618 } else { # last $re_smsg is the highest priority xref3
619 local $self->{current_info} = "$ci #$docid";
620 _reindex_finalize($req, $re_smsg, $eml);
624 sub _reindex_smsg ($$$) {
625 my ($self, $sync, $smsg) = @_;
626 my $docid = $smsg->{num};
627 my $xr3 = $self->{oidx}->get_xref3($docid, 1);
628 if (scalar(@$xr3) == 0) { # _reindex_check_stale should've covered this
630 BUG? #$docid $smsg->{blob} is not referenced by inboxes during reindex
632 remove_doc($self, $docid);
636 # we sort {xr3r} in the reverse order of ibx_sorted so we can
637 # hit the common case in _reindex_finalize without rereading
638 # from git (or holding multiple messages in memory).
639 my $id2pos = $sync->{id2pos}; # index in ibx_sorted
641 $id2pos->{$b->[0]} <=> $id2pos->{$a->[0]}
643 $b->[1] <=> $a->[1] # break ties with {xnum}
645 @$xr3 = map { [ $_->[0], $_->[1], unpack('H*', $_->[2]) ] } @$xr3;
646 my $req = { orig_smsg => $smsg, sync => $sync, xr3r => $xr3, ix => 0 };
647 $self->git->cat_async($xr3->[$req->{ix}]->[2], \&_reindex_oid, $req);
650 sub checkpoint_due ($) {
652 ${$sync->{need_checkpoint}} || (now() > $sync->{next_check});
656 # I've copied FS images and only changed the hostname before,
657 # so prepend hostname. Use `state' since these a BOFH can change
658 # these while this process is running and we always want to be
659 # able to release locks taken by this process.
660 state $retval = hostname . '-' . do {
661 my $m; # machine-id(5) is systemd
662 if (open(my $fh, '<', '/etc/machine-id')) { $m = <$fh> }
663 # (g)hostid(1) is in GNU coreutils, kern.hostid is most BSDs
664 chomp($m ||= `{ sysctl -n kern.hostid ||
665 hostid || ghostid; } 2>/dev/null`
666 || "no-machine-id-or-hostid-on-$^O");
673 my $expect = delete($self->{-eidxq_locked}) or return;
674 my ($owner_pid, undef) = split(/-/, $expect);
675 return if $owner_pid != $$; # shards may fork
676 my $oidx = $self->{oidx};
678 my $cur = $oidx->eidx_meta('eidxq_lock') // '';
679 if ($cur eq $expect) {
680 $oidx->eidx_meta('eidxq_lock', '');
682 } elsif ($cur ne '') {
683 warn "E: eidxq_lock($expect) stolen by $cur\n";
685 warn "E: eidxq_lock($expect) released by another process\n";
692 eidxq_release($self) and $self->{oidx}->commit_lazy;
695 sub _eidxq_take ($) {
697 my $val = "$$-${\time}-$>-".host_ident;
698 $self->{oidx}->eidx_meta('eidxq_lock', $val);
699 $self->{-eidxq_locked} = $val;
702 sub eidxq_lock_acquire ($) {
704 my $oidx = $self->{oidx};
706 my $cur = $oidx->eidx_meta('eidxq_lock') || return _eidxq_take($self);
707 if (my $locked = $self->{-eidxq_locked}) { # be lazy
708 return $locked if $locked eq $cur;
710 my ($pid, $time, $euid, $ident) = split(/-/, $cur, 4);
711 my $t = strftime('%Y-%m-%d %k:%M:%S', gmtime($time));
712 if ($euid == $> && $ident eq host_ident) {
715 I: PID:$pid (re)indexing Xapian since $t, it will continue our work
719 warn "I: eidxq_lock is stale ($cur), clobbering\n";
720 return _eidxq_take($self);
722 warn "E: kill(0, $pid) failed: $!\n"; # fall-through:
724 my $fn = $oidx->dbh->sqlite_db_filename;
726 W: PID:$pid, UID:$euid on $ident is indexing Xapian since $t
727 W: If this is unexpected, delete `eidxq_lock' from the `eidx_meta' table:
728 W: sqlite3 $fn 'DELETE FROM eidx_meta WHERE key = "eidxq_lock"'
733 sub ibx_sorted ($$) {
734 my ($self, $type) = @_;
735 $self->{"-ibx_ary_$type"} //= do {
736 # highest boost first, stable for config-ordering tiebreaker
739 ($b->{boost} // 0) <=> ($a->{boost} // 0)
740 } @{$self->{'ibx_'.$type} // die "BUG: $type unknown"} ];
744 sub prep_id2pos ($) {
748 $id2pos{$_->{-ibx_id}} = $pos++ for (@{ibx_sorted($self, 'known')});
752 sub eidxq_process ($$) { # for reindexing
753 my ($self, $sync) = @_;
754 return unless $self->{cfg};
756 return unless eidxq_lock_acquire($self);
757 my $dbh = $self->{oidx}->dbh;
758 my $tot = $dbh->selectrow_array('SELECT COUNT(*) FROM eidxq') or return;
760 local $sync->{-regen_fmt} = "%u/$tot\n";
761 my $pr = $sync->{-opt}->{-progress};
763 my $min = $dbh->selectrow_array('SELECT MIN(docid) FROM eidxq');
764 my $max = $dbh->selectrow_array('SELECT MAX(docid) FROM eidxq');
765 $pr->("Xapian indexing $min..$max (total=$tot)\n");
767 $sync->{id2pos} //= prep_id2pos($self);
770 $del = $dbh->prepare('DELETE FROM eidxq WHERE docid = ?');
771 $iter = $dbh->prepare('SELECT docid FROM eidxq ORDER BY docid ASC');
773 while (defined(my $docid = $iter->fetchrow_array)) {
774 last if $sync->{quit};
775 if (my $smsg = $self->{oidx}->get_art($docid)) {
776 _reindex_smsg($self, $sync, $smsg);
778 warn "E: #$docid does not exist in over\n";
780 $del->execute($docid);
783 if (checkpoint_due($sync)) {
784 $dbh = $del = $iter = undef;
785 reindex_checkpoint($self, $sync); # release lock
786 $dbh = $self->{oidx}->dbh;
790 $self->git->async_wait_all;
791 $pr->("reindexed ${$sync->{nr}}/$tot\n") if $pr;
794 sub _reindex_unseen { # git->cat_async callback
795 my ($bref, $oid, $type, $size, $req) = @_;
796 return if is_bad_blob($oid, $type, $size, $req->{oid});
797 my $self = $req->{self} // die 'BUG: {self} unset';
798 local $self->{current_info} = "$self->{current_info} $oid";
799 my $new_smsg = bless { blob => $oid, }, 'PublicInbox::Smsg';
800 $new_smsg->set_bytes($$bref, $size);
801 my $eml = $req->{eml} = PublicInbox::Eml->new($bref);
802 $req->{new_smsg} = $new_smsg;
803 $req->{chash} = content_hash($eml);
804 $req->{mids} = mids($eml); # do_step iterates through this
805 do_step($req); # enter the normal indexing flow
808 # --reindex may catch totally unseen messages, this handles them
809 sub reindex_unseen ($$$$) {
810 my ($self, $sync, $ibx, $xsmsg) = @_;
813 autime => $xsmsg->{ds},
814 cotime => $xsmsg->{ts},
815 oid => $xsmsg->{blob},
817 xnum => $xsmsg->{num},
818 # {mids} and {chash} will be filled in at _reindex_unseen
820 warn "I: reindex_unseen ${\$ibx->eidx_key}:$req->{xnum}:$req->{oid}\n";
821 $self->git->cat_async($xsmsg->{blob}, \&_reindex_unseen, $req);
824 sub _unref_stale_range ($$$) {
825 my ($sync, $ibx, $lt_or_gt) = @_;
829 $r = $sync->{self}->{oidx}->dbh->selectall_arrayref(
830 <<EOS, undef, $ibx->{-ibx_id});
831 SELECT docid,xnum,oidbin FROM xref3
832 WHERE ibx_id = ? AND xnum $lt_or_gt LIMIT $lim
834 return if $sync->{quit};
835 for (@$r) { # hopefully rare, not worth optimizing:
836 my ($docid, $xnum, $oidbin) = @$_;
837 my $hex = unpack('H*', $oidbin);
838 warn("# $xnum:$hex (#$docid): stale\n");
839 _unref_doc($sync, $docid, $ibx, $xnum, $oidbin);
841 } while (scalar(@$r) == $lim);
845 sub _reindex_check_ibx ($$$) {
846 my ($self, $sync, $ibx) = @_;
847 my $ibx_id = $ibx->{-ibx_id};
849 my $opt = { limit => $slice };
850 my ($beg, $end) = (1, $slice);
851 my $err = sync_inbox($self, $sync, $ibx) and return;
852 my $max = $ibx->over->max;
853 $end = $max if $end > $max;
855 # first, check if we missed any messages in target $ibx
857 my $pr = $sync->{-opt}->{-progress};
858 my $ekey = $ibx->eidx_key;
859 local $sync->{-regen_fmt} = "$ekey checking %u/$max\n";
861 my $fast = $sync->{-opt}->{fast};
862 my $dsu; # _unref_stale_range (< $lo) called
864 while (scalar(@{$msgs = $ibx->over->query_xover($beg, $end, $opt)})) {
865 ${$sync->{nr}} = $beg;
866 $beg = $msgs->[-1]->{num} + 1;
867 $end = $beg + $slice;
868 $end = $max if $end > $max;
869 if (checkpoint_due($sync)) {
870 reindex_checkpoint($self, $sync); # release lock
872 ($lo, $hi) = ($msgs->[0]->{num}, $msgs->[-1]->{num});
873 $dsu //= _unref_stale_range($sync, $ibx, "< $lo");
874 my $x3a = $self->{oidx}->dbh->selectall_arrayref(
875 <<"", undef, $ibx_id, $lo, $hi);
876 SELECT xnum,oidbin,docid FROM xref3 WHERE
877 ibx_id = ? AND xnum >= ? AND xnum <= ?
881 my $k = pack('J', $_->[0]) . $_->[1];
882 push @{$x3m{$k}}, $_->[2];
885 for my $xsmsg (@$msgs) {
886 my $k = pack('JH*', $xsmsg->{num}, $xsmsg->{blob});
887 my $docids = delete($x3m{$k});
888 if (!defined($docids)) {
889 reindex_unseen($self, $sync, $ibx, $xsmsg);
891 for my $num (@$docids) {
892 $self->{oidx}->eidxq_add($num);
894 return if $sync->{quit};
897 return if $sync->{quit};
898 next unless scalar keys %x3m;
900 # eliminate stale/mismatched entries
901 my %mismatch = map { $_->{num} => $_->{blob} } @$msgs;
902 while (my ($k, $docids) = each %x3m) {
903 my ($xnum, $hex) = unpack('JH*', $k);
904 my $bin = pack('H*', $hex);
905 my $exp = $mismatch{$xnum};
906 my $m = defined($exp) ? "mismatch (!= $exp)" : 'stale';
907 warn("# $xnum:$hex (#@$docids): $m\n");
908 for my $i (@$docids) {
909 _unref_doc($sync, $i, $ibx, $xnum, $bin);
913 _unref_stale_range($sync, $ibx, "> $hi") if defined($hi);
916 sub _reindex_inbox ($$$) {
917 my ($self, $sync, $ibx) = @_;
918 my $ekey = $ibx->eidx_key;
919 local $self->{current_info} = $ekey;
920 if (defined(my $err = _ibx_index_reject($ibx))) {
921 warn "W: cannot reindex $ekey ($err)\n";
923 _reindex_check_ibx($self, $sync, $ibx);
925 delete @$ibx{qw(over mm search git)}; # won't need these for a bit
929 my ($self, $sync) = @_;
930 return unless $self->{cfg};
932 # acquire eidxq_lock early because full reindex takes forever
933 # and incremental -extindex processes can run during our checkpoints
934 if (!eidxq_lock_acquire($self)) {
935 warn "E: aborting --reindex\n";
938 for my $ibx (@{ibx_sorted($self, 'active')}) {
939 _reindex_inbox($self, $sync, $ibx);
940 last if $sync->{quit};
942 $self->git->async_wait_all; # ensure eidxq gets filled completely
943 eidxq_process($self, $sync) unless $sync->{quit};
947 my ($self, $sync, $ibx) = @_;
948 my $err = _sync_inbox($self, $sync, $ibx);
949 delete @$ibx{qw(mm over)};
950 warn $err, "\n" if defined($err);
954 sub dd_smsg { # git->cat_async callback
955 my ($bref, $oid, $type, $size, $dd) = @_;
956 my $smsg = $dd->{smsg} // die 'BUG: dd->{smsg} missing';
957 my $self = $dd->{self} // die 'BUG: {self} missing';
958 my $per_mid = $dd->{per_mid} // die 'BUG: {per_mid} missing';
959 if ($type eq 'missing') {
960 _blob_missing($dd, $smsg);
961 } elsif (!is_bad_blob($oid, $type, $size, $smsg->{blob})) {
962 local $self->{current_info} = "$self->{current_info} $oid";
963 my $chash = content_hash(PublicInbox::Eml->new($bref));
964 push(@{$per_mid->{dd_chash}->{$chash}}, $smsg);
966 return if $per_mid->{last_smsg} != $smsg;
967 while (my ($chash, $ary) = each %{$per_mid->{dd_chash}}) {
968 my $keep = shift @$ary;
969 next if !scalar(@$ary);
970 $per_mid->{sync}->{dedupe_cull} += scalar(@$ary);
972 "# <$keep->{mid}> keeping #$keep->{num}, dropping ",
973 join(', ', map { "#$_->{num}" } @$ary),"\n";
974 next if $per_mid->{sync}->{-opt}->{'dry-run'};
975 my $oidx = $self->{oidx};
976 for my $smsg (@$ary) {
977 my $gone = $smsg->{num};
978 $oidx->merge_xref3($keep->{num}, $gone, $smsg->{blob});
979 remove_doc($self, $gone);
984 sub eidx_dedupe ($$$) {
985 my ($self, $sync, $msgids) = @_;
986 $sync->{dedupe_cull} = 0;
989 return unless eidxq_lock_acquire($self);
990 my ($iter, $cur_mid);
993 my ($max_id) = $self->{oidx}->dbh->selectrow_array(<<EOS);
994 SELECT MAX(id) FROM msgid
996 local $sync->{-regen_fmt} = "dedupe %u/$max_id\n";
998 # note: we could write this query more intelligently,
999 # but that causes lock contention with read-only processes
1001 $cur_mid = $msgids->[$idx];
1002 if ($cur_mid eq '') { # all Message-IDs
1003 $iter = $self->{oidx}->dbh->prepare(<<EOS);
1004 SELECT mid,id FROM msgid WHERE id > ? ORDER BY id ASC
1006 $iter->execute($min_id);
1008 $iter = $self->{oidx}->dbh->prepare(<<EOS);
1009 SELECT mid,id FROM msgid WHERE mid = ? AND id > ? ORDER BY id ASC
1011 $iter->execute($cur_mid, $min_id);
1013 while (my ($mid, $id) = $iter->fetchrow_array) {
1014 last if $sync->{quit};
1015 $self->{current_info} = "dedupe $mid";
1016 ${$sync->{nr}} = $min_id = $id;
1018 while (my $x = $self->{oidx}->next_by_mid($mid, \$id, \$prv)) {
1021 next if scalar(@smsg) < 2;
1023 dd_chash => {}, # chash => [ary of smsgs]
1024 last_smsg => $smsg[-1],
1028 $candidates += scalar(@smsg) - 1;
1029 for my $smsg (@smsg) {
1031 per_mid => $per_mid,
1035 $self->git->cat_async($smsg->{blob}, \&dd_smsg, $dd);
1037 # need to wait on every single one @smsg contents can get
1038 # invalidated inside dd_smsg for messages with multiple
1040 $self->git->async_wait_all;
1042 if (checkpoint_due($sync)) {
1044 reindex_checkpoint($self, $sync);
1045 goto dedupe_restart;
1048 goto dedupe_restart if defined($msgids->[++$idx]);
1050 my $n = delete $sync->{dedupe_cull};
1051 if (my $pr = $sync->{-opt}->{-progress}) {
1052 $pr->("culled $n/$candidates candidates ($nr_mid msgids)\n");
1057 sub eidx_sync { # main entry point
1058 my ($self, $opt) = @_;
1060 my $warn_cb = $SIG{__WARN__} || \&CORE::warn;
1061 local $self->{current_info} = '';
1062 local $SIG{__WARN__} = sub {
1063 return if PublicInbox::Eml::warn_ignore(@_);
1064 $warn_cb->($self->{current_info}, ': ', @_);
1066 $self->idx_init($opt); # acquire lock via V2Writable::_idx_init
1067 $self->{oidx}->rethread_prepare($opt);
1069 need_checkpoint => \(my $need_checkpoint = 0),
1071 next_check => now() + 10,
1073 # DO NOT SET {reindex} here, it's incompatible with reused
1074 # V2Writable code, reindex is totally different here
1075 # compared to v1/v2 inboxes because we have multiple histories
1077 -regen_fmt => "%u/?\n",
1079 local $SIG{USR1} = sub { $need_checkpoint = 1 };
1080 my $quit = PublicInbox::SearchIdx::quit_cb($sync);
1081 local $SIG{QUIT} = $quit;
1082 local $SIG{INT} = $quit;
1083 local $SIG{TERM} = $quit;
1084 for my $ibx (@{ibx_sorted($self, 'known')}) {
1085 $ibx->{-ibx_id} //= $self->{oidx}->ibx_id($ibx->eidx_key);
1088 if (scalar(grep { defined($_->{boost}) } @{$self->{ibx_known}})) {
1089 $sync->{id2pos} //= prep_id2pos($self);
1090 $sync->{boost_in_use} = 1;
1093 if (my $msgids = delete($opt->{dedupe})) {
1094 local $sync->{checkpoint_unlocks} = 1;
1095 eidx_dedupe($self, $sync, $msgids);
1097 if (delete($opt->{reindex})) {
1098 local $sync->{checkpoint_unlocks} = 1;
1099 eidx_reindex($self, $sync);
1102 # don't use $_ here, it'll get clobbered by reindex_checkpoint
1103 if ($opt->{scan} // 1) {
1104 for my $ibx (@{ibx_sorted($self, 'active')}) {
1105 last if $sync->{quit};
1106 sync_inbox($self, $sync, $ibx);
1109 $self->{oidx}->rethread_done($opt) unless $sync->{quit};
1110 eidxq_process($self, $sync) unless $sync->{quit};
1112 eidxq_release($self);
1114 $sync; # for eidx_watch
1117 sub update_last_commit { # overrides V2Writable
1118 my ($self, $sync, $stk) = @_;
1119 my $unit = $sync->{unit} // return;
1120 my $latest_cmt = $stk ? $stk->{latest_cmt} : ${$sync->{latest_cmt}};
1121 defined($latest_cmt) or return;
1122 my $ibx = $sync->{ibx} or die 'BUG: {ibx} missing';
1123 my $ekey = $ibx->eidx_key;
1124 my $uv = $ibx->uidvalidity;
1125 my $epoch = $unit->{epoch};
1127 my $v = $ibx->version;
1129 die 'No {epoch} for v2 unit' unless defined $epoch;
1130 $meta_key = "lc-v2:$ekey//$uv;$epoch";
1132 die 'Unexpected {epoch} for v1 unit' if defined $epoch;
1133 $meta_key = "lc-v1:$ekey//$uv";
1135 die "Unsupported inbox version: $v";
1137 my $last = $self->{oidx}->eidx_meta($meta_key);
1138 if (defined $last && is_ancestor($self->git, $last, $latest_cmt)) {
1139 my @cmd = (qw(rev-list --count), "$last..$latest_cmt");
1140 chomp(my $n = $unit->{git}->qx(@cmd));
1141 return if $n ne '' && $n == 0;
1143 $self->{oidx}->eidx_meta($meta_key, $latest_cmt);
1146 sub _idx_init { # with_umask callback
1147 my ($self, $opt) = @_;
1148 PublicInbox::V2Writable::_idx_init($self, $opt); # acquires ei.lock
1149 $self->{midx} = PublicInbox::MiscIdx->new($self);
1152 sub symlink_packs ($$) {
1153 my ($ibx, $pd) = @_;
1155 my $glob = "$ibx->{inboxdir}/git/*.git/objects/pack/*.idx";
1156 for my $idx (bsd_glob($glob, GLOB_NOSORT)) {
1157 my $src = substr($idx, 0, -length('.idx'));
1158 my $dst = $pd . substr($src, rindex($src, '/'));
1159 if (-f "$src.pack" and
1160 symlink("$src.pack", "$dst.pack") and
1161 symlink($idx, "$dst.idx") and
1164 # .promisor, .bitmap, .rev and .keep are optional
1165 # XXX should we symlink .keep here?
1166 for my $s (qw(promisor bitmap rev)) {
1167 symlink("$src.$s", "$dst.$s") if -f "$src.$s";
1169 } elsif (!$!{EEXIST}) {
1170 warn "W: ln -s $src.{pack,idx} => $dst.*: $!\n";
1171 unlink "$dst.pack", "$dst.idx";
1177 sub idx_init { # similar to V2Writable
1178 my ($self, $opt) = @_;
1179 return if $self->{idx_shards};
1181 $self->git->cleanup;
1183 my $ALL = $self->git->{git_dir}; # topdir/ALL.git
1184 my ($has_new, $alt, $seen);
1185 if ($opt->{-private}) { # LeiStore
1186 my $local = "$self->{topdir}/local"; # lei/store
1187 $self->{mg} //= PublicInbox::MultiGit->new($self->{topdir},
1188 'ALL.git', 'local');
1191 umask 077; # don't bother restoring for lei
1192 PublicInbox::Import::init_bare($ALL);
1193 $self->git->qx(qw(config core.sharedRepository 0600));
1195 ($alt, $seen) = $self->{mg}->read_alternates(\$mode);
1196 $has_new = $self->{mg}->merge_epochs($alt, $seen);
1197 } else { # extindex has no epochs
1198 $self->{mg} //= PublicInbox::MultiGit->new($self->{topdir},
1200 ($alt, $seen) = $self->{mg}->read_alternates(\$mode,
1202 PublicInbox::Import::init_bare($ALL);
1205 # git-multi-pack-index(1) can speed up "git cat-file" startup slightly
1207 my $pd = "$ALL/objects/pack";
1208 if (opendir(my $dh, $pd)) { # drop stale symlinks
1209 while (defined(my $dn = readdir($dh))) {
1210 if ($dn =~ /\.(?:idx|pack|promisor|bitmap|rev)\z/) {
1212 unlink($f) if -l $f && !-e $f;
1215 } elsif ($!{ENOENT}) {
1216 mkdir($pd) or die "mkdir($pd): $!";
1218 die "opendir($pd): $!";
1221 for my $ibx (@{ibx_sorted($self, 'active')}) {
1222 # create symlinks for multi-pack-index
1223 $git_midx += symlink_packs($ibx, $pd);
1224 # add new lines to our alternates file
1225 my $d = $ibx->git->{git_dir} . '/objects';
1226 next if exists $alt->{$d};
1227 if (my @st = stat($d)) {
1228 next if $seen->{"$st[0]\0$st[1]"}++;
1230 warn "W: stat($d) failed (from $ibx->{inboxdir}): $!\n";
1231 next if $opt->{-idx_gc};
1235 ($has_new || $new ne '') and
1236 $self->{mg}->write_alternates($mode, $alt, $new);
1237 $git_midx and $self->with_umask(sub {
1238 my @cmd = ('multi-pack-index');
1239 push @cmd, '--no-progress' if ($opt->{quiet}//0) > 1;
1240 my $lk = $self->lock_for_scope;
1241 system('git', "--git-dir=$ALL", @cmd, 'write');
1242 # ignore errors, fairly new command, may not exist
1244 $self->parallel_init($self->{indexlevel});
1245 $self->with_umask(\&_idx_init, $self, $opt);
1246 $self->{oidx}->begin_lazy;
1247 $self->{oidx}->eidx_prep;
1248 $self->{midx}->create_xdb if $new ne '';
1251 sub _watch_commit { # PublicInbox::DS::add_timer callback
1253 delete $self->{-commit_timer};
1254 eidxq_process($self, $self->{-watch_sync});
1255 eidxq_release($self);
1256 my $fmt = delete $self->{-watch_sync}->{-regen_fmt};
1257 reindex_checkpoint($self, $self->{-watch_sync});
1258 $self->{-watch_sync}->{-regen_fmt} = $fmt;
1260 # call event_step => done unless commit_timer is armed
1261 PublicInbox::DS::requeue($self);
1264 sub on_inbox_unlock { # called by PublicInbox::InboxIdle
1265 my ($self, $ibx) = @_;
1266 my $opt = $self->{-watch_sync}->{-opt};
1267 my $pr = $opt->{-progress};
1268 my $ekey = $ibx->eidx_key;
1269 local $0 = "sync $ekey";
1270 $pr->("indexing $ekey\n") if $pr;
1271 $self->idx_init($opt);
1272 sync_inbox($self, $self->{-watch_sync}, $ibx);
1273 $self->{-commit_timer} //= add_timer($opt->{'commit-interval'} // 10,
1274 \&_watch_commit, $self);
1277 sub eidx_reload { # -extindex --watch SIGHUP handler
1278 my ($self, $idler) = @_;
1280 my $pr = $self->{-watch_sync}->{-opt}->{-progress};
1281 $pr->('reloading ...') if $pr;
1282 delete $self->{-resync_queue};
1283 delete $self->{-ibx_ary_known};
1284 delete $self->{-ibx_ary_active};
1285 $self->{ibx_known} = [];
1286 $self->{ibx_active} = [];
1287 %{$self->{ibx_map}} = ();
1288 delete $self->{-watch_sync}->{id2pos};
1289 my $cfg = PublicInbox::Config->new;
1290 attach_config($self, $cfg);
1291 $idler->refresh($cfg);
1292 $pr->(" done\n") if $pr;
1294 warn "reload not supported without --all\n";
1298 sub eidx_resync_start ($) { # -extindex --watch SIGUSR1 handler
1300 $self->{-resync_queue} //= [ @{ibx_sorted($self, 'active')} ];
1301 PublicInbox::DS::requeue($self); # trigger our ->event_step
1304 sub event_step { # PublicInbox::DS::requeue callback
1306 if (my $resync_queue = $self->{-resync_queue}) {
1307 if (my $ibx = shift(@$resync_queue)) {
1308 on_inbox_unlock($self, $ibx);
1309 PublicInbox::DS::requeue($self);
1311 delete $self->{-resync_queue};
1312 _watch_commit($self);
1315 done($self) unless $self->{-commit_timer};
1319 sub eidx_watch { # public-inbox-extindex --watch main loop
1320 my ($self, $opt) = @_;
1321 local @SIG{keys %SIG} = values %SIG;
1322 for my $sig (qw(HUP USR1 TSTP QUIT INT TERM)) {
1323 $SIG{$sig} = sub { warn "SIG$sig ignored while scanning\n" };
1325 require PublicInbox::InboxIdle;
1326 require PublicInbox::DS;
1327 require PublicInbox::Syscall;
1328 require PublicInbox::Sigfd;
1329 my $idler = PublicInbox::InboxIdle->new($self->{cfg});
1330 if (!$self->{cfg}) {
1331 $idler->watch_inbox($_) for (@{ibx_sorted($self, 'active')});
1333 for my $ibx (@{ibx_sorted($self, 'active')}) {
1334 $ibx->subscribe_unlock(__PACKAGE__, $self)
1336 my $pr = $opt->{-progress};
1337 $pr->("performing initial scan ...\n") if $pr;
1338 my $sync = eidx_sync($self, $opt); # initial sync
1339 return if $sync->{quit};
1340 my $oldset = PublicInbox::DS::block_signals();
1341 local $self->{current_info} = '';
1342 my $cb = $SIG{__WARN__} || \&CORE::warn;
1343 local $SIG{__WARN__} = sub {
1344 return if PublicInbox::Eml::warn_ignore(@_);
1345 $cb->($self->{current_info}, ': ', @_);
1348 HUP => sub { eidx_reload($self, $idler) },
1349 USR1 => sub { eidx_resync_start($self) },
1350 TSTP => sub { kill('STOP', $$) },
1352 my $quit = PublicInbox::SearchIdx::quit_cb($sync);
1353 $sig->{QUIT} = $sig->{INT} = $sig->{TERM} = $quit;
1354 local $self->{-watch_sync} = $sync; # for ->on_inbox_unlock
1355 PublicInbox::DS->SetPostLoopCallback(sub { !$sync->{quit} });
1356 $pr->("initial scan complete, entering event loop\n") if $pr;
1357 # calls InboxIdle->event_step:
1358 PublicInbox::DS::event_loop($sig, $oldset);
1363 *done = \&PublicInbox::V2Writable::done;
1364 *with_umask = \&PublicInbox::InboxWritable::with_umask;
1365 *parallel_init = \&PublicInbox::V2Writable::parallel_init;
1366 *nproc_shards = \&PublicInbox::V2Writable::nproc_shards;
1367 *sync_prepare = \&PublicInbox::V2Writable::sync_prepare;
1368 *index_todo = \&PublicInbox::V2Writable::index_todo;
1369 *count_shards = \&PublicInbox::V2Writable::count_shards;
1370 *atfork_child = \&PublicInbox::V2Writable::atfork_child;
1371 *idx_shard = \&PublicInbox::V2Writable::idx_shard;
1372 *reindex_checkpoint = \&PublicInbox::V2Writable::reindex_checkpoint;
1373 *checkpoint = \&PublicInbox::V2Writable::checkpoint;