1 # Copyright (C) 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 Scalar::Util qw(blessed);
22 use Sys::Hostname qw(hostname);
23 use POSIX qw(strftime);
24 use File::Glob qw(bsd_glob GLOB_NOSORT);
25 use PublicInbox::MultiGit;
26 use PublicInbox::Search;
27 use PublicInbox::SearchIdx qw(prepare_stack is_ancestor is_bad_blob);
28 use PublicInbox::OverIdx;
29 use PublicInbox::MiscIdx;
30 use PublicInbox::MID qw(mids);
31 use PublicInbox::V2Writable;
32 use PublicInbox::InboxWritable;
33 use PublicInbox::ContentHash qw(content_hash);
35 use PublicInbox::DS qw(now add_timer);
36 use DBI qw(:sql_types); # SQL_BLOB
39 my (undef, $dir, $opt) = @_;
40 my $l = $opt->{indexlevel} // 'full';
41 $l !~ $PublicInbox::SearchIdx::INDEXLEVELS and
42 die "invalid indexlevel=$l\n";
43 $l eq 'basic' and die "E: indexlevel=basic not yet supported\n";
45 xpfx => "$dir/ei".PublicInbox::Search::SCHEMA_VERSION,
47 creat => $opt->{creat},
48 ibx_map => {}, # (newsgroup//inboxdir) => $ibx
49 ibx_active => [], # by config section order
50 ibx_known => [], # by config section order
56 lock_path => "$dir/ei.lock",
58 $self->{shards} = $self->count_shards ||
59 nproc_shards({ nproc => $opt->{jobs} });
60 my $oidx = PublicInbox::OverIdx->new("$self->{xpfx}/over.sqlite3");
61 $self->{-no_fsync} = $oidx->{-no_fsync} = 1 if !$opt->{fsync};
62 $self->{-dangerous} = 1 if $opt->{dangerous};
63 $self->{oidx} = $oidx;
68 my ($self, $ibx, $types) = @_;
69 $self->{ibx_map}->{$ibx->eidx_key} //= do {
70 delete $self->{-ibx_ary_known}; # invalidate cache
71 delete $self->{-ibx_ary_active}; # invalidate cache
72 $types //= [ qw(active known) ];
74 push @{$self->{"ibx_$t"}}, $ibx;
80 sub _ibx_attach { # each_inbox callback
81 my ($ibx, $self, $types) = @_;
82 attach_inbox($self, $ibx, $types);
86 my ($self, $cfg, $ibxs) = @_;
90 for my $ibx (@$ibxs) {
91 $self->{ibx_map}->{$ibx->eidx_key} //= do {
92 push @{$self->{ibx_active}}, $ibx;
93 push @{$self->{ibx_known}}, $ibx;
98 delete $self->{-ibx_ary_known};
99 delete $self->{-ibx_ary_active};
100 $types = [ 'known' ];
102 $types //= [ qw(known active) ];
103 $cfg->each_inbox(\&_ibx_attach, $self, $types);
106 sub check_batch_limit ($) {
108 my $self = $req->{self};
109 my $new_smsg = $req->{new_smsg};
110 my $n = $self->{transact_bytes} += $new_smsg->{bytes};
112 # set flag for PublicInbox::V2Writable::index_todo:
113 ${$req->{need_checkpoint}} = 1 if $n >= $self->{batch_bytes};
116 sub apply_boost ($$) {
117 my ($req, $smsg) = @_;
118 my $id2pos = $req->{id2pos}; # index in ibx_sorted
119 my $xr3 = $req->{self}->{oidx}->get_xref3($smsg->{num}, 1);
121 $id2pos->{$a->[0]} <=> $id2pos->{$b->[0]}
123 $a->[1] <=> $b->[1] # break ties with {xnum}
125 my $new_smsg = $req->{new_smsg};
126 return if $xr3->[0]->[2] ne $new_smsg->oidbin; # loser
128 # replace the old smsg with the more boosted one
129 $new_smsg->{num} = $smsg->{num};
130 $new_smsg->populate($req->{eml}, $req);
131 $req->{self}->{oidx}->add_overview($req->{eml}, $new_smsg);
134 sub remove_doc ($$) {
135 my ($self, $docid) = @_;
136 $self->{oidx}->delete_by_num($docid);
137 $self->{oidx}->eidxq_del($docid);
138 $self->idx_shard($docid)->ipc_do('xdb_remove', $docid);
141 sub _unref_doc ($$$$$;$) {
142 my ($sync, $docid, $ibx, $xnum, $oidbin, $eml) = @_;
146 $docid = $smsg->{num};
148 if (defined($oidbin) && defined($xnum) && blessed($ibx) && $ibx->over) {
149 my $smsg = $ibx->over->get_art($xnum);
150 if ($smsg && $smsg->oidbin eq $oidbin) {
151 carp("BUG: (non-fatal) ".$ibx->eidx_key.
152 " #$xnum $smsg->{blob} still valid");
156 my $s = 'DELETE FROM xref3 WHERE oidbin = ?';
157 $s .= ' AND ibx_id = ?' if defined($ibx);
158 $s .= ' AND xnum = ?' if defined($xnum);
159 my $del = $sync->{self}->{oidx}->dbh->prepare_cached($s);
161 $del->bind_param(++$col, $oidbin, SQL_BLOB);
162 $del->bind_param(++$col, $ibx->{-ibx_id}) if $ibx;
163 $del->bind_param(++$col, $xnum) if defined($xnum);
165 my $xr3 = $sync->{self}->{oidx}->get_xref3($docid);
166 if (scalar(@$xr3) == 0) { # all gone
167 remove_doc($sync->{self}, $docid);
168 } else { # enqueue for reindex of remaining messages
170 my $ekey = $ibx->{-gc_eidx_key} // $ibx->eidx_key;
171 my $idx = $sync->{self}->idx_shard($docid);
172 $idx->ipc_do('remove_eidx_info', $docid, $ekey, $eml);
173 } # else: we can't remove_eidx_info in reindex-only path
175 # replace invalidated blob ASAP with something which should be
176 # readable since we may commit the transaction on checkpoint.
177 # eidxq processing will re-apply boost
178 $smsg //= $sync->{self}->{oidx}->get_art($docid);
179 my $hex = unpack('H*', $oidbin);
180 if ($smsg && $smsg->{blob} eq $hex) {
181 $xr3->[0] =~ /:([a-f0-9]{40,}+)\z/ or
182 die "BUG: xref $xr3->[0] has no OID";
183 $sync->{self}->{oidx}->update_blob($smsg, $1);
185 # yes, add, we'll need to re-apply boost
186 $sync->{self}->{oidx}->eidxq_add($docid);
192 my ($req, $smsg) = @_;
193 my $self = $req->{self};
194 my $docid = $smsg->{num};
195 my $oid = $req->{oid};
196 my $xibx = $req->{ibx};
197 my $eml = $req->{eml};
198 if (my $new_smsg = $req->{new_smsg}) { # 'm' on cross-posted message
199 my $eidx_key = $xibx->eidx_key;
200 my $xnum = $req->{xnum};
201 $self->{oidx}->add_xref3($docid, $xnum, $oid, $eidx_key);
202 my $idx = $self->idx_shard($docid);
203 $idx->ipc_do('add_eidx_info', $docid, $eidx_key, $eml);
204 apply_boost($req, $smsg) if $req->{boost_in_use};
205 } else { # 'd' no {xnum}
206 $self->git->async_wait_all;
207 $oid = pack('H*', $oid);
208 _unref_doc($req, $docid, $xibx, undef, $oid, $eml);
212 # called by V2Writable::sync_prepare
213 sub artnum_max { $_[0]->{oidx}->eidx_max }
215 sub index_unseen ($) {
217 my $new_smsg = $req->{new_smsg} or die 'BUG: {new_smsg} unset';
218 my $eml = delete $req->{eml};
219 $new_smsg->populate($eml, $req);
220 my $self = $req->{self};
221 my $docid = $self->{oidx}->adj_counter('eidx_docid', '+');
222 $new_smsg->{num} = $docid;
223 my $idx = $self->idx_shard($docid);
224 $self->{oidx}->add_overview($eml, $new_smsg);
225 my $oid = $new_smsg->{blob};
226 my $ibx = delete $req->{ibx} or die 'BUG: {ibx} unset';
227 $self->{oidx}->add_xref3($docid, $req->{xnum}, $oid, $ibx->eidx_key);
228 $idx->index_eml($eml, $new_smsg, $ibx->eidx_key);
229 check_batch_limit($req);
232 sub do_finalize ($) {
234 if (my $indexed = $req->{indexed}) { # duplicated messages
235 do_xpost($req, $_) for @$indexed;
236 } elsif (exists $req->{new_smsg}) { # totally unseen messsage
239 # `d' message was already unindexed in the v1/v2 inboxes,
240 # so it's too noisy to warn, here.
242 # cur_cmt may be undef for unindex_oid, set by V2Writable::index_todo
243 if (defined(my $cur_cmt = $req->{cur_cmt})) {
244 ${$req->{latest_cmt}} = $cur_cmt;
248 sub do_step ($) { # main iterator for adding messages to the index
250 my $self = $req->{self} // die 'BUG: {self} missing';
252 if (my $next_arg = $req->{next_arg}) {
253 if (my $smsg = $self->{oidx}->next_by_mid(@$next_arg)) {
254 $req->{cur_smsg} = $smsg;
255 $self->git->cat_async($smsg->{blob},
256 \&ck_existing, $req);
257 return; # ck_existing calls do_step
259 delete $req->{next_arg};
261 die "BUG: {cur_smsg} still set" if $req->{cur_smsg};
262 my $mid = shift(@{$req->{mids}}) // last;
264 $req->{next_arg} = [ $mid, \$id, \$prev ];
270 sub _blob_missing ($$) { # called when a known $smsg->{blob} is gone
271 my ($req, $smsg) = @_;
272 # xnum and ibx are unknown, we only call this when an entry from
273 # /ei*/over.sqlite3 is bad, not on entries from xap*/over.sqlite3
274 $req->{self}->git->async_wait_all;
275 _unref_doc($req, $smsg, undef, undef, $smsg->oidbin);
278 sub ck_existing { # git->cat_async callback
279 my ($bref, $oid, $type, $size, $req) = @_;
280 my $smsg = delete $req->{cur_smsg} or die 'BUG: {cur_smsg} missing';
281 if ($type eq 'missing') {
282 _blob_missing($req, $smsg);
283 } elsif (!is_bad_blob($oid, $type, $size, $smsg->{blob})) {
284 my $self = $req->{self} // die 'BUG: {self} missing';
285 local $self->{current_info} = "$self->{current_info} $oid";
286 my $cur = PublicInbox::Eml->new($bref);
287 if (content_hash($cur) eq $req->{chash}) {
288 push @{$req->{indexed}}, $smsg; # for do_xpost
289 } # else { index_unseen later }
294 # is the messages visible in the inbox currently being indexed?
295 # return the number if so
296 sub cur_ibx_xnum ($$;$) {
297 my ($req, $bref, $mismatch) = @_;
298 my $ibx = $req->{ibx} or die 'BUG: current {ibx} missing';
300 $req->{eml} = PublicInbox::Eml->new($bref);
301 $req->{chash} = content_hash($req->{eml});
302 $req->{mids} = mids($req->{eml});
303 for my $mid (@{$req->{mids}}) {
305 while (my $x = $ibx->over->next_by_mid($mid, \$id, \$prev)) {
306 return $x->{num} if $x->{blob} eq $req->{oid};
307 push @$mismatch, $x if $mismatch;
313 sub index_oid { # git->cat_async callback for 'm'
314 my ($bref, $oid, $type, $size, $req) = @_;
315 my $self = $req->{self};
316 local $self->{current_info} = "$self->{current_info} $oid";
317 return if is_bad_blob($oid, $type, $size, $req->{oid});
318 my $new_smsg = $req->{new_smsg} = bless {
320 }, 'PublicInbox::Smsg';
321 $new_smsg->set_bytes($$bref, $size);
324 $req->{xnum} = cur_ibx_xnum($req, $bref, $mismatch) // do {
326 warn "# mismatch $_->{blob}\n" for @$mismatch;
327 ${$req->{latest_cmt}} = $req->{cur_cmt} //
328 die "BUG: {cur_cmt} unset ($oid)\n";
334 sub unindex_oid { # git->cat_async callback for 'd'
335 my ($bref, $oid, $type, $size, $req) = @_;
336 my $self = $req->{self};
337 local $self->{current_info} = "$self->{current_info} $oid";
338 return if is_bad_blob($oid, $type, $size, $req->{oid});
339 return if defined(cur_ibx_xnum($req, $bref)); # was re-added
343 # overrides V2Writable::last_commits, called by sync_ranges via sync_prepare
345 my ($self, $sync) = @_;
347 my $ekey = $sync->{ibx}->eidx_key;
348 my $uv = $sync->{ibx}->uidvalidity;
349 for my $i (0..$sync->{epoch_max}) {
350 $heads->[$i] = $self->{oidx}->eidx_meta("lc-v2:$ekey//$uv;$i");
355 sub _ibx_index_reject ($) {
357 $ibx->mm // return 'unindexed, no msgmap.sqlite3';
358 $ibx->uidvalidity // return 'no UIDVALIDITY';
359 $ibx->over // return 'unindexed, no over.sqlite3';
363 sub _sync_inbox ($$$) {
364 my ($self, $sync, $ibx) = @_;
365 my $ekey = $ibx->eidx_key;
366 if (defined(my $err = _ibx_index_reject($ibx))) {
367 return "W: skipping $ekey ($err)";
370 $sync->{nr} = \(my $nr = 0);
371 my $v = $ibx->version;
373 $sync->{epoch_max} = $ibx->max_git_epoch // return;
374 sync_prepare($self, $sync); # or return # TODO: once MiscIdx is stable
376 my $uv = $ibx->uidvalidity;
377 my $lc = $self->{oidx}->eidx_meta("lc-v1:$ekey//$uv");
378 my $head = $ibx->mm->last_commit //
379 return "E: $ibx->{inboxdir} is not indexed";
380 my $stk = prepare_stack($sync, $lc ? "$lc..$head" : $head);
381 my $unit = { stack => $stk, git => $ibx->git };
382 push @{$sync->{todo}}, $unit;
384 return "E: $ekey unsupported inbox version (v$v)";
386 for my $unit (@{delete($sync->{todo}) // []}) {
387 last if $sync->{quit};
388 index_todo($self, $sync, $unit);
390 $self->{midx}->index_ibx($ibx) unless $sync->{quit};
391 $ibx->git->cleanup; # done with this inbox, now
395 sub eidx_gc_scan_inboxes ($$) {
396 my ($self, $sync) = @_;
397 my ($x3_doc, $ibx_ck);
399 $x3_doc = $self->{oidx}->dbh->prepare(<<EOM);
400 SELECT docid,xnum,oidbin FROM xref3 WHERE ibx_id = ?
402 $ibx_ck = $self->{oidx}->dbh->prepare(<<EOM);
403 SELECT ibx_id,eidx_key FROM inboxes
406 while (my ($ibx_id, $eidx_key) = $ibx_ck->fetchrow_array) {
407 next if $self->{ibx_map}->{$eidx_key};
408 $self->{midx}->remove_eidx_key($eidx_key);
409 warn "I: deleting messages for $eidx_key...\n";
410 $x3_doc->execute($ibx_id);
411 my $ibx = { -ibx_id => $ibx_id, -gc_eidx_key => $eidx_key };
412 while (my ($docid, $xnum, $oid) = $x3_doc->fetchrow_array) {
413 my $r = _unref_doc($sync, $docid, $ibx, $xnum, $oid);
414 $oid = unpack('H*', $oid);
415 $r = $r ? 'unref' : 'remove';
416 warn "I: $r #$docid $eidx_key $oid\n";
417 if (checkpoint_due($sync)) {
418 $x3_doc = $ibx_ck = undef;
419 reindex_checkpoint($self, $sync);
423 $self->{oidx}->dbh->do(<<'', undef, $ibx_id);
424 DELETE FROM inboxes WHERE ibx_id = ?
426 # drop last_commit info
428 $pat =~ s/([_%\\])/\\$1/g;
429 $self->{oidx}->dbh->do('PRAGMA case_sensitive_like = ON');
430 my $lc_i = $self->{oidx}->dbh->prepare(<<'');
431 SELECT key FROM eidx_meta WHERE key LIKE ? ESCAPE ?
433 $lc_i->execute("lc-%:$pat//%", '\\');
434 while (my ($key) = $lc_i->fetchrow_array) {
435 next if $key !~ m!\Alc-v[1-9]+:\Q$eidx_key\E//!;
436 warn "I: removing $key\n";
437 $self->{oidx}->dbh->do(<<'', undef, $key);
438 DELETE FROM eidx_meta WHERE key = ?
441 warn "I: $eidx_key removed\n";
445 sub eidx_gc_scan_shards ($$) { # TODO: use for lei/store
446 my ($self, $sync) = @_;
447 my $nr = $self->{oidx}->dbh->do(<<'');
448 DELETE FROM xref3 WHERE docid NOT IN (SELECT num FROM over)
450 warn "I: eliminated $nr stale xref3 entries\n" if $nr != 0;
451 reindex_checkpoint($self, $sync) if checkpoint_due($sync);
453 # fixup from old bugs:
454 $nr = $self->{oidx}->dbh->do(<<'');
455 DELETE FROM over WHERE num > 0 AND num NOT IN (SELECT docid FROM xref3)
457 warn "I: eliminated $nr stale over entries\n" if $nr != 0;
458 reindex_checkpoint($self, $sync) if checkpoint_due($sync);
460 $nr = $self->{oidx}->dbh->do(<<'');
461 DELETE FROM eidxq WHERE docid NOT IN (SELECT num FROM over)
463 warn "I: eliminated $nr stale reindex queue entries\n" if $nr != 0;
464 reindex_checkpoint($self, $sync) if checkpoint_due($sync);
466 my ($cur) = $self->{oidx}->dbh->selectrow_array(<<EOM);
467 SELECT MIN(num) FROM over WHERE num > 0
469 $cur // return; # empty
470 my ($r, $n, %active_shards);
473 $r = $self->{oidx}->dbh->selectcol_arrayref(<<"", undef, $cur);
474 SELECT num FROM over WHERE num >= ? ORDER BY num ASC LIMIT 10000
476 last unless scalar(@$r);
477 while (defined($n = shift @$r)) {
478 for my $i ($cur..($n - 1)) {
479 my $idx = idx_shard($self, $i);
480 $idx->ipc_do('xdb_remove_quiet', $i);
481 $active_shards{$idx} = $idx;
485 if (checkpoint_due($sync)) {
486 for my $idx (values %active_shards) {
487 $nr += $idx->ipc_do('nr_quiet_rm')
490 reindex_checkpoint($self, $sync);
493 warn "I: eliminated $nr stale Xapian documents\n" if $nr != 0;
497 my ($self, $opt) = @_;
498 $self->{cfg} or die "E: GC requires ->attach_config\n";
501 need_checkpoint => \(my $need_checkpoint = 0),
503 next_check => now() + 10,
504 checkpoint_unlocks => 1,
508 $self->idx_init($opt); # acquire lock via V2Writable::_idx_init
509 eidx_gc_scan_inboxes($self, $sync);
510 eidx_gc_scan_shards($self, $sync);
515 my ($self, $sync, $smsg) = @_;
516 my $ibx_id = delete($smsg->{ibx_id}) // die '{ibx_id} unset';
517 my $pos = $sync->{id2pos}->{$ibx_id} // die "$ibx_id no pos";
518 $self->{-ibx_ary_known}->[$pos] //
519 die "BUG: ibx for $smsg->{blob} not mapped"
522 sub _fd_constrained ($) {
524 $self->{-fd_constrained} //= do {
526 if (eval { require BSD::Resource; 1 }) {
527 my $NOFILE = BSD::Resource::RLIMIT_NOFILE();
528 ($soft, undef) = BSD::Resource::getrlimit($NOFILE);
530 chomp($soft = `sh -c 'ulimit -n'`);
532 if (defined($soft)) {
533 # $want is an estimate
534 my $want = scalar(@{$self->{ibx_active}}) + 64;
535 my $ret = $want > $soft;
538 RLIMIT_NOFILE=$soft insufficient (want: $want), will close DB handles early
543 warn "Unable to determine RLIMIT_NOFILE: $@\n";
549 sub _reindex_finalize ($$$) {
550 my ($req, $smsg, $eml) = @_;
551 my $sync = $req->{sync};
552 my $self = $sync->{self};
553 my $by_chash = delete $req->{by_chash} or die 'BUG: no {by_chash}';
554 my $nr = scalar(keys(%$by_chash)) or die 'BUG: no content hashes';
555 my $orig_smsg = $req->{orig_smsg} // die 'BUG: no {orig_smsg}';
556 my $docid = $smsg->{num} = $orig_smsg->{num};
557 $self->{oidx}->add_overview($eml, $smsg); # may rethread
558 check_batch_limit({ %$sync, new_smsg => $smsg });
559 my $chash0 = $smsg->{chash} // die "BUG: $smsg->{blob} no {chash}";
560 my $stable = delete($by_chash->{$chash0}) //
561 die "BUG: $smsg->{blob} chash missing";
562 my $idx = $self->idx_shard($docid);
563 my $top_smsg = pop @$stable;
564 $top_smsg == $smsg or die 'BUG: top_smsg != smsg';
565 my $ibx = _ibx_for($self, $sync, $smsg);
566 $idx->index_eml($eml, $smsg, $ibx->eidx_key);
567 for my $x (reverse @$stable) {
568 $ibx = _ibx_for($self, $sync, $x);
569 my $hdr = delete $x->{hdr} // die 'BUG: no {hdr}';
570 $idx->ipc_do('add_eidx_info', $docid, $ibx->eidx_key, $hdr);
572 return if $nr == 1; # likely, all good
574 $self->git->async_wait_all;
575 warn "W: #$docid split into $nr due to deduplication change\n";
577 for my $ary (values %$by_chash) {
578 for my $x (reverse @$ary) {
579 warn "removing #$docid xref3 $x->{blob}\n";
580 my $bin = $x->oidbin;
581 my $n = _unref_doc($sync, $docid, undef, undef, $bin);
582 die "BUG: $x->{blob} invalidated #$docid" if $n == 0;
584 my $x = pop(@$ary) // die "BUG: #$docid {by_chash} empty";
585 $x->{num} = delete($x->{xnum}) // die '{xnum} unset';
586 $ibx = _ibx_for($self, $sync, $x);
587 if (my $over = $ibx->over) {
588 my $e = $over->get_art($x->{num});
589 $e->{blob} eq $x->{blob} or die <<EOF;
590 $x->{blob} != $e->{blob} (${\$ibx->eidx_key}:$e->{num});
592 push @todo, $ibx, $e;
593 $over->dbh_close if _fd_constrained($self);
595 die "$ibx->{inboxdir}: over.sqlite3 unusable: $!\n";
599 while (my ($ibx, $e) = splice(@todo, 0, 2)) {
600 reindex_unseen($self, $sync, $ibx, $e);
604 sub _reindex_oid { # git->cat_async callback
605 my ($bref, $oid, $type, $size, $req) = @_;
606 my $sync = $req->{sync};
607 my $self = $sync->{self};
608 my $orig_smsg = $req->{orig_smsg} // die 'BUG: no {orig_smsg}';
609 my $expect_oid = $req->{xr3r}->[$req->{ix}]->[2];
610 my $docid = $orig_smsg->{num};
611 if (is_bad_blob($oid, $type, $size, $expect_oid)) {
612 my $oidbin = pack('H*', $expect_oid);
613 my $remain = _unref_doc($sync, $docid, undef, undef, $oidbin);
615 warn "W: #$docid ($oid) gone or corrupt\n";
616 } elsif (my $next_oid = $req->{xr3r}->[++$req->{ix}]->[2]) {
617 $self->git->cat_async($next_oid, \&_reindex_oid, $req);
619 warn "BUG: #$docid ($oid) gone (UNEXPECTED)\n";
623 my $ci = $self->{current_info};
624 local $self->{current_info} = "$ci #$docid $oid";
625 my $re_smsg = bless { blob => $oid }, 'PublicInbox::Smsg';
626 $re_smsg->set_bytes($$bref, $size);
627 my $eml = PublicInbox::Eml->new($bref);
628 $re_smsg->populate($eml, { autime => $orig_smsg->{ds},
629 cotime => $orig_smsg->{ts} });
630 my $chash = content_hash($eml);
631 $re_smsg->{chash} = $chash;
632 $re_smsg->{xnum} = $req->{xr3r}->[$req->{ix}]->[1];
633 $re_smsg->{ibx_id} = $req->{xr3r}->[$req->{ix}]->[0];
634 $re_smsg->{hdr} = $eml->header_obj;
635 push @{$req->{by_chash}->{$chash}}, $re_smsg;
636 if (my $next_oid = $req->{xr3r}->[++$req->{ix}]->[2]) {
637 $self->git->cat_async($next_oid, \&_reindex_oid, $req);
638 } else { # last $re_smsg is the highest priority xref3
639 local $self->{current_info} = "$ci #$docid";
640 _reindex_finalize($req, $re_smsg, $eml);
644 sub _reindex_smsg ($$$) {
645 my ($self, $sync, $smsg) = @_;
646 my $docid = $smsg->{num};
647 my $xr3 = $self->{oidx}->get_xref3($docid, 1);
648 if (scalar(@$xr3) == 0) { # _reindex_check_stale should've covered this
650 BUG? #$docid $smsg->{blob} is not referenced by inboxes during reindex
652 remove_doc($self, $docid);
656 # we sort {xr3r} in the reverse order of ibx_sorted so we can
657 # hit the common case in _reindex_finalize without rereading
658 # from git (or holding multiple messages in memory).
659 my $id2pos = $sync->{id2pos}; # index in ibx_sorted
661 $id2pos->{$b->[0]} <=> $id2pos->{$a->[0]}
663 $b->[1] <=> $a->[1] # break ties with {xnum}
665 @$xr3 = map { [ $_->[0], $_->[1], unpack('H*', $_->[2]) ] } @$xr3;
666 my $req = { orig_smsg => $smsg, sync => $sync, xr3r => $xr3, ix => 0 };
667 $self->git->cat_async($xr3->[$req->{ix}]->[2], \&_reindex_oid, $req);
670 sub checkpoint_due ($) {
672 ${$sync->{need_checkpoint}} || (now() > $sync->{next_check});
676 # I've copied FS images and only changed the hostname before,
677 # so prepend hostname. Use `state' since these a BOFH can change
678 # these while this process is running and we always want to be
679 # able to release locks taken by this process.
680 state $retval = hostname . '-' . do {
681 my $m; # machine-id(5) is systemd
682 if (open(my $fh, '<', '/etc/machine-id')) { $m = <$fh> }
683 # (g)hostid(1) is in GNU coreutils, kern.hostid is most BSDs
684 chomp($m ||= `{ sysctl -n kern.hostid ||
685 hostid || ghostid; } 2>/dev/null`
686 || "no-machine-id-or-hostid-on-$^O");
693 my $expect = delete($self->{-eidxq_locked}) or return;
694 my ($owner_pid, undef) = split(/-/, $expect);
695 return if $owner_pid != $$; # shards may fork
696 my $oidx = $self->{oidx};
698 my $cur = $oidx->eidx_meta('eidxq_lock') // '';
699 if ($cur eq $expect) {
700 $oidx->eidx_meta('eidxq_lock', '');
702 } elsif ($cur ne '') {
703 warn "E: eidxq_lock($expect) stolen by $cur\n";
705 warn "E: eidxq_lock($expect) released by another process\n";
712 eidxq_release($self) and $self->{oidx}->commit_lazy;
715 sub _eidxq_take ($) {
717 my $val = "$$-${\time}-$>-".host_ident;
718 $self->{oidx}->eidx_meta('eidxq_lock', $val);
719 $self->{-eidxq_locked} = $val;
722 sub eidxq_lock_acquire ($) {
724 my $oidx = $self->{oidx};
726 my $cur = $oidx->eidx_meta('eidxq_lock') || return _eidxq_take($self);
727 if (my $locked = $self->{-eidxq_locked}) { # be lazy
728 return $locked if $locked eq $cur;
730 my ($pid, $time, $euid, $ident) = split(/-/, $cur, 4);
731 my $t = strftime('%Y-%m-%d %k:%M %z', localtime($time));
732 local $self->{current_info} = 'eidxq';
733 if ($euid == $> && $ident eq host_ident) {
736 I: PID:$pid (re)indexing since $t, it will continue our work
740 warn "I: eidxq_lock is stale ($cur), clobbering\n";
741 return _eidxq_take($self);
743 warn "E: kill(0, $pid) failed: $!\n"; # fall-through:
745 my $fn = $oidx->dbh->sqlite_db_filename;
747 W: PID:$pid, UID:$euid on $ident is indexing Xapian since $t
748 W: If this is unexpected, delete `eidxq_lock' from the `eidx_meta' table:
749 W: sqlite3 $fn 'DELETE FROM eidx_meta WHERE key = "eidxq_lock"'
754 sub ibx_sorted ($$) {
755 my ($self, $type) = @_;
756 $self->{"-ibx_ary_$type"} //= do {
757 # highest boost first, stable for config-ordering tiebreaker
760 ($b->{boost} // 0) <=> ($a->{boost} // 0)
761 } @{$self->{'ibx_'.$type} // die "BUG: $type unknown"} ];
765 sub prep_id2pos ($) {
769 $id2pos{$_->{-ibx_id}} = $pos++ for (@{ibx_sorted($self, 'known')});
773 sub eidxq_process ($$) { # for reindexing
774 my ($self, $sync) = @_;
775 local $self->{current_info} = 'eidxq process';
776 return unless ($self->{cfg} && eidxq_lock_acquire($self));
777 my $dbh = $self->{oidx}->dbh;
778 my $tot = $dbh->selectrow_array('SELECT COUNT(*) FROM eidxq') or return;
780 local $sync->{-regen_fmt} = "%u/$tot\n";
781 my $pr = $sync->{-opt}->{-progress};
783 my $min = $dbh->selectrow_array('SELECT MIN(docid) FROM eidxq');
784 my $max = $dbh->selectrow_array('SELECT MAX(docid) FROM eidxq');
785 $pr->("Xapian indexing $min..$max (total=$tot)\n");
787 $sync->{id2pos} //= prep_id2pos($self);
790 $del = $dbh->prepare('DELETE FROM eidxq WHERE docid = ?');
791 $iter = $dbh->prepare('SELECT docid FROM eidxq ORDER BY docid ASC');
793 while (defined(my $docid = $iter->fetchrow_array)) {
794 last if $sync->{quit};
795 if (my $smsg = $self->{oidx}->get_art($docid)) {
796 _reindex_smsg($self, $sync, $smsg);
798 warn "E: #$docid does not exist in over\n";
800 $del->execute($docid);
803 if (checkpoint_due($sync)) {
804 $dbh = $del = $iter = undef;
805 reindex_checkpoint($self, $sync); # release lock
806 $dbh = $self->{oidx}->dbh;
810 $self->git->async_wait_all;
811 $pr->("reindexed ${$sync->{nr}}/$tot\n") if $pr;
814 sub _reindex_unseen { # git->cat_async callback
815 my ($bref, $oid, $type, $size, $req) = @_;
816 return if is_bad_blob($oid, $type, $size, $req->{oid});
817 my $self = $req->{self} // die 'BUG: {self} unset';
818 local $self->{current_info} = "$self->{current_info} $oid";
819 my $new_smsg = bless { blob => $oid, }, 'PublicInbox::Smsg';
820 $new_smsg->set_bytes($$bref, $size);
821 my $eml = $req->{eml} = PublicInbox::Eml->new($bref);
822 $req->{new_smsg} = $new_smsg;
823 $req->{chash} = content_hash($eml);
824 $req->{mids} = mids($eml); # do_step iterates through this
825 do_step($req); # enter the normal indexing flow
828 # --reindex may catch totally unseen messages, this handles them
829 sub reindex_unseen ($$$$) {
830 my ($self, $sync, $ibx, $xsmsg) = @_;
833 autime => $xsmsg->{ds},
834 cotime => $xsmsg->{ts},
835 oid => $xsmsg->{blob},
837 xnum => $xsmsg->{num},
838 # {mids} and {chash} will be filled in at _reindex_unseen
840 warn "I: reindex_unseen ${\$ibx->eidx_key}:$req->{xnum}:$req->{oid}\n";
841 $self->git->cat_async($xsmsg->{blob}, \&_reindex_unseen, $req);
844 sub _unref_stale_range ($$$) {
845 my ($sync, $ibx, $lt_or_gt) = @_;
849 $r = $sync->{self}->{oidx}->dbh->selectall_arrayref(
850 <<EOS, undef, $ibx->{-ibx_id});
851 SELECT docid,xnum,oidbin FROM xref3
852 WHERE ibx_id = ? AND $lt_or_gt LIMIT $lim
854 return if $sync->{quit};
855 for (@$r) { # hopefully rare, not worth optimizing:
856 my ($docid, $xnum, $oidbin) = @$_;
857 my $hex = unpack('H*', $oidbin);
858 warn("# $xnum:$hex (#$docid): stale\n");
859 _unref_doc($sync, $docid, $ibx, $xnum, $oidbin);
861 } while (scalar(@$r) == $lim);
865 sub _reindex_check_ibx ($$$) {
866 my ($self, $sync, $ibx) = @_;
867 my $ibx_id = $ibx->{-ibx_id};
869 my $opt = { limit => $slice };
870 my ($beg, $end) = (1, $slice);
871 my $ekey = $ibx->eidx_key;
874 $max0 = $ibx->mm->num_highwater;
875 sync_inbox($self, $sync, $ibx) and return; # warned
876 $max = $ibx->mm->num_highwater;
877 return if $sync->{quit};
878 } while ($max > $max0 &&
879 warn("# $ekey moved $max0..$max, resyncing..\n"));
880 $end = $max if $end > $max;
882 # first, check if we missed any messages in target $ibx
884 my $pr = $sync->{-opt}->{-progress};
885 local $sync->{-regen_fmt} = "$ekey checking %u/$max\n";
887 my $fast = $sync->{-opt}->{fast};
888 my $usr; # _unref_stale_range (< $lo) called
890 while (scalar(@{$msgs = $ibx->over->query_xover($beg, $end, $opt)})) {
891 ${$sync->{nr}} = $beg;
892 $beg = $msgs->[-1]->{num} + 1;
893 $end = $beg + $slice;
894 $end = $max if $end > $max;
895 if (checkpoint_due($sync)) {
896 reindex_checkpoint($self, $sync); # release lock
898 ($lo, $hi) = ($msgs->[0]->{num}, $msgs->[-1]->{num});
899 $usr //= _unref_stale_range($sync, $ibx, "xnum < $lo");
900 my $x3a = $self->{oidx}->dbh->selectall_arrayref(
901 <<"", undef, $ibx_id, $lo, $hi);
902 SELECT xnum,oidbin,docid FROM xref3 WHERE
903 ibx_id = ? AND xnum >= ? AND xnum <= ?
907 my $k = pack('J', $_->[0]) . $_->[1];
908 push @{$x3m{$k}}, $_->[2];
911 for my $xsmsg (@$msgs) {
912 my $k = pack('JH*', $xsmsg->{num}, $xsmsg->{blob});
913 my $docids = delete($x3m{$k});
914 if (!defined($docids)) {
915 reindex_unseen($self, $sync, $ibx, $xsmsg);
917 for my $num (@$docids) {
918 $self->{oidx}->eidxq_add($num);
921 return if $sync->{quit};
923 next unless scalar keys %x3m;
924 $self->git->async_wait_all; # wait for reindex_unseen
926 # eliminate stale/mismatched entries
927 my %mismatch = map { $_->{num} => $_->{blob} } @$msgs;
928 while (my ($k, $docids) = each %x3m) {
929 my ($xnum, $hex) = unpack('JH*', $k);
930 my $bin = pack('H*', $hex);
931 my $exp = $mismatch{$xnum};
933 my $smsg = $ibx->over->get_art($xnum) // next;
934 # $xnum may be expired by another process
935 if ($smsg->{blob} eq $hex) {
937 BUG: (non-fatal) $ekey #$xnum $smsg->{blob} still matches (old exp: $exp)
940 } # else: continue to unref
942 my $m = defined($exp) ? "mismatch (!= $exp)" : 'stale';
943 warn("# $xnum:$hex (#@$docids): $m\n");
944 for my $i (@$docids) {
945 _unref_doc($sync, $i, $ibx, $xnum, $bin);
947 return if $sync->{quit};
950 defined($hi) and ($hi < $max) and
951 _unref_stale_range($sync, $ibx, "xnum > $hi AND xnum <= $max");
954 sub _reindex_inbox ($$$) {
955 my ($self, $sync, $ibx) = @_;
956 my $ekey = $ibx->eidx_key;
957 local $self->{current_info} = $ekey;
958 if (defined(my $err = _ibx_index_reject($ibx))) {
959 warn "W: cannot reindex $ekey ($err)\n";
961 _reindex_check_ibx($self, $sync, $ibx);
963 delete @$ibx{qw(over mm search git)}; # won't need these for a bit
967 my ($self, $sync) = @_;
968 return unless $self->{cfg};
970 # acquire eidxq_lock early because full reindex takes forever
971 # and incremental -extindex processes can run during our checkpoints
972 if (!eidxq_lock_acquire($self)) {
973 warn "E: aborting --reindex\n";
976 for my $ibx (@{ibx_sorted($self, 'active')}) {
977 _reindex_inbox($self, $sync, $ibx);
978 last if $sync->{quit};
980 $self->git->async_wait_all; # ensure eidxq gets filled completely
981 eidxq_process($self, $sync) unless $sync->{quit};
985 my ($self, $sync, $ibx) = @_;
986 my $err = _sync_inbox($self, $sync, $ibx);
987 delete @$ibx{qw(mm over)};
988 warn $err, "\n" if defined($err);
992 sub dd_smsg { # git->cat_async callback
993 my ($bref, $oid, $type, $size, $dd) = @_;
994 my $smsg = $dd->{smsg} // die 'BUG: dd->{smsg} missing';
995 my $self = $dd->{self} // die 'BUG: {self} missing';
996 my $per_mid = $dd->{per_mid} // die 'BUG: {per_mid} missing';
997 if ($type eq 'missing') {
998 _blob_missing($dd, $smsg);
999 } elsif (!is_bad_blob($oid, $type, $size, $smsg->{blob})) {
1000 local $self->{current_info} = "$self->{current_info} $oid";
1001 my $chash = content_hash(PublicInbox::Eml->new($bref));
1002 push(@{$per_mid->{dd_chash}->{$chash}}, $smsg);
1004 return if $per_mid->{last_smsg} != $smsg;
1005 while (my ($chash, $ary) = each %{$per_mid->{dd_chash}}) {
1006 my $keep = shift @$ary;
1007 next if !scalar(@$ary);
1008 $per_mid->{sync}->{dedupe_cull} += scalar(@$ary);
1010 "# <$keep->{mid}> keeping #$keep->{num}, dropping ",
1011 join(', ', map { "#$_->{num}" } @$ary),"\n";
1012 next if $per_mid->{sync}->{-opt}->{'dry-run'};
1013 my $oidx = $self->{oidx};
1014 for my $smsg (@$ary) {
1015 my $gone = $smsg->{num};
1016 $oidx->merge_xref3($keep->{num}, $gone, $smsg->oidbin);
1017 remove_doc($self, $gone);
1022 sub eidx_dedupe ($$$) {
1023 my ($self, $sync, $msgids) = @_;
1024 $sync->{dedupe_cull} = 0;
1027 return unless eidxq_lock_acquire($self);
1028 my ($iter, $cur_mid);
1031 my ($max_id) = $self->{oidx}->dbh->selectrow_array(<<EOS);
1032 SELECT MAX(id) FROM msgid
1034 local $sync->{-regen_fmt} = "dedupe %u/$max_id\n";
1036 # note: we could write this query more intelligently,
1037 # but that causes lock contention with read-only processes
1039 $cur_mid = $msgids->[$idx];
1040 if ($cur_mid eq '') { # all Message-IDs
1041 $iter = $self->{oidx}->dbh->prepare(<<EOS);
1042 SELECT mid,id FROM msgid WHERE id > ? ORDER BY id ASC
1044 $iter->execute($min_id);
1046 $iter = $self->{oidx}->dbh->prepare(<<EOS);
1047 SELECT mid,id FROM msgid WHERE mid = ? AND id > ? ORDER BY id ASC
1049 $iter->execute($cur_mid, $min_id);
1051 while (my ($mid, $id) = $iter->fetchrow_array) {
1052 last if $sync->{quit};
1053 $self->{current_info} = "dedupe $mid";
1054 ${$sync->{nr}} = $min_id = $id;
1056 while (my $x = $self->{oidx}->next_by_mid($mid, \$id, \$prv)) {
1059 next if scalar(@smsg) < 2;
1061 dd_chash => {}, # chash => [ary of smsgs]
1062 last_smsg => $smsg[-1],
1066 $candidates += scalar(@smsg) - 1;
1067 for my $smsg (@smsg) {
1069 per_mid => $per_mid,
1073 $self->git->cat_async($smsg->{blob}, \&dd_smsg, $dd);
1075 # need to wait on every single one @smsg contents can get
1076 # invalidated inside dd_smsg for messages with multiple
1078 $self->git->async_wait_all;
1080 if (checkpoint_due($sync)) {
1082 reindex_checkpoint($self, $sync);
1083 goto dedupe_restart;
1086 goto dedupe_restart if defined($msgids->[++$idx]);
1088 my $n = delete $sync->{dedupe_cull};
1089 if (my $pr = $sync->{-opt}->{-progress}) {
1090 $pr->("culled $n/$candidates candidates ($nr_mid msgids)\n");
1095 sub eidx_sync { # main entry point
1096 my ($self, $opt) = @_;
1098 my $warn_cb = $SIG{__WARN__} || \&CORE::warn;
1099 local $self->{current_info} = '';
1100 local $SIG{__WARN__} = sub {
1101 return if PublicInbox::Eml::warn_ignore(@_);
1102 $warn_cb->($self->{current_info}, ': ', @_);
1104 $self->idx_init($opt); # acquire lock via V2Writable::_idx_init
1105 $self->{oidx}->rethread_prepare($opt);
1107 need_checkpoint => \(my $need_checkpoint = 0),
1109 next_check => now() + 10,
1111 # DO NOT SET {reindex} here, it's incompatible with reused
1112 # V2Writable code, reindex is totally different here
1113 # compared to v1/v2 inboxes because we have multiple histories
1115 -regen_fmt => "%u/?\n",
1117 local $SIG{USR1} = sub { $need_checkpoint = 1 };
1118 my $quit = PublicInbox::SearchIdx::quit_cb($sync);
1119 local $SIG{QUIT} = $quit;
1120 local $SIG{INT} = $quit;
1121 local $SIG{TERM} = $quit;
1122 for my $ibx (@{ibx_sorted($self, 'known')}) {
1123 $ibx->{-ibx_id} //= $self->{oidx}->ibx_id($ibx->eidx_key);
1126 if (scalar(grep { defined($_->{boost}) } @{$self->{ibx_known}})) {
1127 $sync->{id2pos} //= prep_id2pos($self);
1128 $sync->{boost_in_use} = 1;
1131 if (my $msgids = delete($opt->{dedupe})) {
1132 local $sync->{checkpoint_unlocks} = 1;
1133 eidx_dedupe($self, $sync, $msgids);
1135 if (delete($opt->{reindex})) {
1136 local $sync->{checkpoint_unlocks} = 1;
1137 eidx_reindex($self, $sync);
1140 # don't use $_ here, it'll get clobbered by reindex_checkpoint
1141 if ($opt->{scan} // 1) {
1142 for my $ibx (@{ibx_sorted($self, 'active')}) {
1143 last if $sync->{quit};
1144 sync_inbox($self, $sync, $ibx);
1147 $self->{oidx}->rethread_done($opt) unless $sync->{quit};
1148 eidxq_process($self, $sync) unless $sync->{quit};
1150 eidxq_release($self);
1152 $sync; # for eidx_watch
1155 sub update_last_commit { # overrides V2Writable
1156 my ($self, $sync, $stk) = @_;
1157 my $unit = $sync->{unit} // return;
1158 my $latest_cmt = $stk ? $stk->{latest_cmt} : ${$sync->{latest_cmt}};
1159 defined($latest_cmt) or return;
1160 my $ibx = $sync->{ibx} or die 'BUG: {ibx} missing';
1161 my $ekey = $ibx->eidx_key;
1162 my $uv = $ibx->uidvalidity;
1163 my $epoch = $unit->{epoch};
1165 my $v = $ibx->version;
1167 die 'No {epoch} for v2 unit' unless defined $epoch;
1168 $meta_key = "lc-v2:$ekey//$uv;$epoch";
1170 die 'Unexpected {epoch} for v1 unit' if defined $epoch;
1171 $meta_key = "lc-v1:$ekey//$uv";
1173 die "Unsupported inbox version: $v";
1175 my $last = $self->{oidx}->eidx_meta($meta_key);
1176 if (defined $last && is_ancestor($self->git, $last, $latest_cmt)) {
1177 my @cmd = (qw(rev-list --count), "$last..$latest_cmt");
1178 chomp(my $n = $unit->{git}->qx(@cmd));
1179 return if $n ne '' && $n == 0;
1181 $self->{oidx}->eidx_meta($meta_key, $latest_cmt);
1184 sub _idx_init { # with_umask callback
1185 my ($self, $opt) = @_;
1186 PublicInbox::V2Writable::_idx_init($self, $opt); # acquires ei.lock
1187 $self->{midx} = PublicInbox::MiscIdx->new($self);
1190 sub symlink_packs ($$) {
1191 my ($ibx, $pd) = @_;
1193 my $glob = "$ibx->{inboxdir}/git/*.git/objects/pack/*.idx";
1194 for my $idx (bsd_glob($glob, GLOB_NOSORT)) {
1195 my $src = substr($idx, 0, -length('.idx'));
1196 my $dst = $pd . substr($src, rindex($src, '/'));
1197 if (-f "$src.pack" and
1198 symlink("$src.pack", "$dst.pack") and
1199 symlink($idx, "$dst.idx") and
1202 # .promisor, .bitmap, .rev and .keep are optional
1203 # XXX should we symlink .keep here?
1204 for my $s (qw(promisor bitmap rev)) {
1205 symlink("$src.$s", "$dst.$s") if -f "$src.$s";
1207 } elsif (!$!{EEXIST}) {
1208 warn "W: ln -s $src.{pack,idx} => $dst.*: $!\n";
1209 unlink "$dst.pack", "$dst.idx";
1215 sub idx_init { # similar to V2Writable
1216 my ($self, $opt) = @_;
1217 return if $self->{idx_shards};
1219 $self->git->cleanup;
1221 my $ALL = $self->git->{git_dir}; # topdir/ALL.git
1222 my ($has_new, $alt, $seen, $prune, $prune_nr);
1223 if ($opt->{-private}) { # LeiStore
1224 my $local = "$self->{topdir}/local"; # lei/store
1225 $self->{mg} //= PublicInbox::MultiGit->new($self->{topdir},
1226 'ALL.git', 'local');
1229 umask 077; # don't bother restoring for lei
1230 PublicInbox::Import::init_bare($ALL);
1231 $self->git->qx(qw(config core.sharedRepository 0600));
1233 ($alt, $seen) = $self->{mg}->read_alternates(\$mode);
1234 $has_new = $self->{mg}->merge_epochs($alt, $seen);
1235 } else { # extindex has no epochs
1236 $self->{mg} //= PublicInbox::MultiGit->new($self->{topdir},
1238 $prune = $opt->{-idx_gc} ? \$prune_nr : undef;
1239 ($alt, $seen) = $self->{mg}->read_alternates(\$mode, $prune);
1240 PublicInbox::Import::init_bare($ALL);
1243 # git-multi-pack-index(1) can speed up "git cat-file" startup slightly
1245 my $pd = "$ALL/objects/pack";
1246 if (opendir(my $dh, $pd)) { # drop stale symlinks
1247 while (defined(my $dn = readdir($dh))) {
1248 if ($dn =~ /\.(?:idx|pack|promisor|bitmap|rev)\z/) {
1250 unlink($f) if -l $f && !-e $f;
1253 } elsif ($!{ENOENT}) {
1254 mkdir($pd) or die "mkdir($pd): $!";
1256 die "opendir($pd): $!";
1259 for my $ibx (@{ibx_sorted($self, 'active')}) {
1260 # create symlinks for multi-pack-index
1261 $git_midx += symlink_packs($ibx, $pd);
1262 # add new lines to our alternates file
1263 my $d = $ibx->git->{git_dir} . '/objects';
1264 next if exists $alt->{$d};
1265 if (my @st = stat($d)) {
1266 next if $seen->{"$st[0]\0$st[1]"}++;
1268 warn "W: stat($d) failed (from $ibx->{inboxdir}): $!\n";
1269 next if $opt->{-idx_gc};
1273 ($has_new || $prune_nr || $new ne '') and
1274 $self->{mg}->write_alternates($mode, $alt, $new);
1275 $git_midx and $self->with_umask(sub {
1276 my @cmd = ('multi-pack-index');
1277 push @cmd, '--no-progress' if ($opt->{quiet}//0) > 1;
1278 my $lk = $self->lock_for_scope;
1279 system('git', "--git-dir=$ALL", @cmd, 'write');
1280 # ignore errors, fairly new command, may not exist
1282 $self->parallel_init($self->{indexlevel});
1283 $self->with_umask(\&_idx_init, $self, $opt);
1284 $self->{oidx}->begin_lazy;
1285 $self->{oidx}->eidx_prep;
1286 $self->{midx}->create_xdb if $new ne '';
1289 sub _watch_commit { # PublicInbox::DS::add_timer callback
1291 delete $self->{-commit_timer};
1292 eidxq_process($self, $self->{-watch_sync});
1293 eidxq_release($self);
1294 my $fmt = delete $self->{-watch_sync}->{-regen_fmt};
1295 reindex_checkpoint($self, $self->{-watch_sync});
1296 $self->{-watch_sync}->{-regen_fmt} = $fmt;
1298 # call event_step => done unless commit_timer is armed
1299 PublicInbox::DS::requeue($self);
1302 sub on_inbox_unlock { # called by PublicInbox::InboxIdle
1303 my ($self, $ibx) = @_;
1304 my $opt = $self->{-watch_sync}->{-opt};
1305 my $pr = $opt->{-progress};
1306 my $ekey = $ibx->eidx_key;
1307 local $0 = "sync $ekey";
1308 $pr->("indexing $ekey\n") if $pr;
1309 $self->idx_init($opt);
1310 sync_inbox($self, $self->{-watch_sync}, $ibx);
1311 $self->{-commit_timer} //= add_timer($opt->{'commit-interval'} // 10,
1312 \&_watch_commit, $self);
1315 sub eidx_reload { # -extindex --watch SIGHUP handler
1316 my ($self, $idler) = @_;
1318 my $pr = $self->{-watch_sync}->{-opt}->{-progress};
1319 $pr->('reloading ...') if $pr;
1320 delete $self->{-resync_queue};
1321 delete $self->{-ibx_ary_known};
1322 delete $self->{-ibx_ary_active};
1323 $self->{ibx_known} = [];
1324 $self->{ibx_active} = [];
1325 %{$self->{ibx_map}} = ();
1326 delete $self->{-watch_sync}->{id2pos};
1327 my $cfg = PublicInbox::Config->new;
1328 attach_config($self, $cfg);
1329 $idler->refresh($cfg);
1330 $pr->(" done\n") if $pr;
1332 warn "reload not supported without --all\n";
1336 sub eidx_resync_start ($) { # -extindex --watch SIGUSR1 handler
1338 $self->{-resync_queue} //= [ @{ibx_sorted($self, 'active')} ];
1339 PublicInbox::DS::requeue($self); # trigger our ->event_step
1342 sub event_step { # PublicInbox::DS::requeue callback
1344 if (my $resync_queue = $self->{-resync_queue}) {
1345 if (my $ibx = shift(@$resync_queue)) {
1346 on_inbox_unlock($self, $ibx);
1347 PublicInbox::DS::requeue($self);
1349 delete $self->{-resync_queue};
1350 _watch_commit($self);
1353 done($self) unless $self->{-commit_timer};
1357 sub eidx_watch { # public-inbox-extindex --watch main loop
1358 my ($self, $opt) = @_;
1359 local @SIG{keys %SIG} = values %SIG;
1360 for my $sig (qw(HUP USR1 TSTP QUIT INT TERM)) {
1361 $SIG{$sig} = sub { warn "SIG$sig ignored while scanning\n" };
1363 require PublicInbox::InboxIdle;
1364 require PublicInbox::DS;
1365 require PublicInbox::Syscall;
1366 require PublicInbox::Sigfd;
1367 my $idler = PublicInbox::InboxIdle->new($self->{cfg});
1368 if (!$self->{cfg}) {
1369 $idler->watch_inbox($_) for (@{ibx_sorted($self, 'active')});
1371 for my $ibx (@{ibx_sorted($self, 'active')}) {
1372 $ibx->subscribe_unlock(__PACKAGE__, $self)
1374 my $pr = $opt->{-progress};
1375 $pr->("performing initial scan ...\n") if $pr;
1376 my $sync = eidx_sync($self, $opt); # initial sync
1377 return if $sync->{quit};
1378 my $oldset = PublicInbox::DS::block_signals();
1379 local $self->{current_info} = '';
1380 my $cb = $SIG{__WARN__} || \&CORE::warn;
1381 local $SIG{__WARN__} = sub {
1382 return if PublicInbox::Eml::warn_ignore(@_);
1383 $cb->($self->{current_info}, ': ', @_);
1386 HUP => sub { eidx_reload($self, $idler) },
1387 USR1 => sub { eidx_resync_start($self) },
1388 TSTP => sub { kill('STOP', $$) },
1390 my $quit = PublicInbox::SearchIdx::quit_cb($sync);
1391 $sig->{QUIT} = $sig->{INT} = $sig->{TERM} = $quit;
1392 local $self->{-watch_sync} = $sync; # for ->on_inbox_unlock
1393 PublicInbox::DS->SetPostLoopCallback(sub { !$sync->{quit} });
1394 $pr->("initial scan complete, entering event loop\n") if $pr;
1395 # calls InboxIdle->event_step:
1396 PublicInbox::DS::event_loop($sig, $oldset);
1401 *done = \&PublicInbox::V2Writable::done;
1402 *with_umask = \&PublicInbox::InboxWritable::with_umask;
1403 *parallel_init = \&PublicInbox::V2Writable::parallel_init;
1404 *nproc_shards = \&PublicInbox::V2Writable::nproc_shards;
1405 *sync_prepare = \&PublicInbox::V2Writable::sync_prepare;
1406 *index_todo = \&PublicInbox::V2Writable::index_todo;
1407 *count_shards = \&PublicInbox::V2Writable::count_shards;
1408 *atfork_child = \&PublicInbox::V2Writable::atfork_child;
1409 *idx_shard = \&PublicInbox::V2Writable::idx_shard;
1410 *reindex_checkpoint = \&PublicInbox::V2Writable::reindex_checkpoint;
1411 *checkpoint = \&PublicInbox::V2Writable::checkpoint;