1 # Copyright (C) 2020 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 PublicInbox::Search;
24 use PublicInbox::SearchIdx qw(crlf_adjust prepare_stack is_ancestor
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);
34 use PublicInbox::DS qw(now);
35 use DBI qw(:sql_types); # SQL_BLOB
38 my (undef, $dir, $opt) = @_;
39 $dir = File::Spec->canonpath($dir);
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) => $ibx
55 lock_path => "$dir/ei.lock",
57 $self->{shards} = $self->count_shards || nproc_shards($opt->{creat});
58 my $oidx = PublicInbox::OverIdx->new("$self->{xpfx}/over.sqlite3");
59 $oidx->{-no_fsync} = 1 if $opt->{-no_fsync};
60 $self->{oidx} = $oidx;
65 my ($self, $ibx) = @_;
66 $ibx = PublicInbox::InboxWritable->new($ibx);
67 my $key = $ibx->eidx_key;
68 if (!$ibx->over || !$ibx->mm) {
69 warn "W: skipping $key (unindexed)\n";
72 if (!defined($ibx->uidvalidity)) {
73 warn "W: skipping $key (no UIDVALIDITY)\n";
76 my $ibxdir = File::Spec->canonpath($ibx->{inboxdir});
77 if ($ibxdir ne $ibx->{inboxdir}) {
78 warn "W: `$ibx->{inboxdir}' canonicalized to `$ibxdir'\n";
79 $ibx->{inboxdir} = $ibxdir;
81 $ibx = PublicInbox::InboxWritable->new($ibx);
82 $self->{ibx_map}->{$key} //= do {
83 push @{$self->{ibx_list}}, $ibx;
88 sub _ibx_attach { # each_inbox callback
89 my ($ibx, $self) = @_;
90 attach_inbox($self, $ibx);
94 my ($self, $cfg) = @_;
96 $cfg->each_inbox(\&_ibx_attach, $self);
99 sub check_batch_limit ($) {
101 my $self = $req->{self};
102 my $new_smsg = $req->{new_smsg};
104 # {raw_bytes} may be unset, so just use {bytes}
105 my $n = $self->{transact_bytes} += $new_smsg->{bytes};
107 # set flag for PublicInbox::V2Writable::index_todo:
108 ${$req->{need_checkpoint}} = 1 if $n >= $self->{batch_bytes};
112 my ($req, $smsg) = @_;
113 my $self = $req->{self};
114 my $docid = $smsg->{num};
115 my $idx = $self->idx_shard($docid);
116 my $oid = $req->{oid};
117 my $xibx = $req->{ibx};
118 my $eml = $req->{eml};
119 my $eidx_key = $xibx->eidx_key;
120 if (my $new_smsg = $req->{new_smsg}) { # 'm' on cross-posted message
121 my $xnum = $req->{xnum};
122 $self->{oidx}->add_xref3($docid, $xnum, $oid, $eidx_key);
123 $idx->shard_add_eidx_info($docid, $eidx_key, $eml);
124 check_batch_limit($req);
127 my $nr = $self->{oidx}->remove_xref3($docid, $oid, $eidx_key,
130 $self->{oidx}->eidxq_del($docid);
131 $idx->shard_remove($docid);
132 } elsif ($rm_eidx_info) {
133 $idx->shard_remove_eidx_info($docid, $eidx_key, $eml);
134 $self->{oidx}->eidxq_add($docid); # yes, add
139 # called by V2Writable::sync_prepare
140 sub artnum_max { $_[0]->{oidx}->eidx_max }
142 sub index_unseen ($) {
144 my $new_smsg = $req->{new_smsg} or die 'BUG: {new_smsg} unset';
145 my $eml = delete $req->{eml};
146 $new_smsg->populate($eml, $req);
147 my $self = $req->{self};
148 my $docid = $self->{oidx}->adj_counter('eidx_docid', '+');
149 $new_smsg->{num} = $docid;
150 my $idx = $self->idx_shard($docid);
151 $self->{oidx}->add_overview($eml, $new_smsg);
152 my $oid = $new_smsg->{blob};
153 my $ibx = delete $req->{ibx} or die 'BUG: {ibx} unset';
154 $self->{oidx}->add_xref3($docid, $req->{xnum}, $oid, $ibx->eidx_key);
155 $idx->index_raw(undef, $eml, $new_smsg, $ibx->eidx_key);
156 check_batch_limit($req);
159 sub do_finalize ($) {
161 if (my $indexed = $req->{indexed}) {
162 do_xpost($req, $_) for @$indexed;
163 } elsif (exists $req->{new_smsg}) { # totally unseen messsage
166 # `d' message was already unindexed in the v1/v2 inboxes,
167 # so it's too noisy to warn, here.
169 # cur_cmt may be undef for unindex_oid, set by V2Writable::index_todo
170 if (defined(my $cur_cmt = $req->{cur_cmt})) {
171 ${$req->{latest_cmt}} = $cur_cmt;
175 sub do_step ($) { # main iterator for adding messages to the index
177 my $self = $req->{self} // die 'BUG: {self} missing';
179 if (my $next_arg = $req->{next_arg}) {
180 if (my $smsg = $self->{oidx}->next_by_mid(@$next_arg)) {
181 $req->{cur_smsg} = $smsg;
182 $self->git->cat_async($smsg->{blob},
183 \&ck_existing, $req);
184 return; # ck_existing calls do_step
186 delete $req->{cur_smsg};
187 delete $req->{next_arg};
189 my $mid = shift(@{$req->{mids}});
190 last unless defined $mid;
192 $req->{next_arg} = [ $mid, \$id, \$prev ];
198 sub _blob_missing ($) { # called when req->{cur_smsg}->{blob} is bad
200 my $smsg = $req->{cur_smsg} or die 'BUG: {cur_smsg} missing';
201 my $self = $req->{self};
202 my $xref3 = $self->{oidx}->get_xref3($smsg->{num});
203 my @keep = grep(!/:$smsg->{blob}\z/, @$xref3);
205 $keep[0] =~ /:([a-f0-9]{40,}+)\z/ or
206 die "BUG: xref $keep[0] has no OID";
208 $self->{oidx}->remove_xref3($smsg->{num}, $smsg->{blob});
209 my $upd = $self->{oidx}->update_blob($smsg, $oidhex);
210 my $saved = $self->{oidx}->get_art($smsg->{num});
212 $self->{oidx}->delete_by_num($smsg->{num});
216 sub ck_existing { # git->cat_async callback
217 my ($bref, $oid, $type, $size, $req) = @_;
218 my $smsg = $req->{cur_smsg} or die 'BUG: {cur_smsg} missing';
219 if ($type eq 'missing') {
221 } elsif (!is_bad_blob($oid, $type, $size, $smsg->{blob})) {
222 my $self = $req->{self} // die 'BUG: {self} missing';
223 local $self->{current_info} = "$self->{current_info} $oid";
224 my $cur = PublicInbox::Eml->new($bref);
225 if (content_hash($cur) eq $req->{chash}) {
226 push @{$req->{indexed}}, $smsg; # for do_xpost
227 } # else { index_unseen later }
232 # is the messages visible in the inbox currently being indexed?
233 # return the number if so
234 sub cur_ibx_xnum ($$) {
235 my ($req, $bref) = @_;
236 my $ibx = $req->{ibx} or die 'BUG: current {ibx} missing';
238 $req->{eml} = PublicInbox::Eml->new($bref);
239 $req->{chash} = content_hash($req->{eml});
240 $req->{mids} = mids($req->{eml});
241 my @q = @{$req->{mids}}; # copy
242 while (defined(my $mid = shift @q)) {
244 while (my $x = $ibx->over->next_by_mid($mid, \$id, \$prev)) {
245 return $x->{num} if $x->{blob} eq $req->{oid};
251 sub index_oid { # git->cat_async callback for 'm'
252 my ($bref, $oid, $type, $size, $req) = @_;
253 my $self = $req->{self};
254 local $self->{current_info} = "$self->{current_info} $oid";
255 return if is_bad_blob($oid, $type, $size, $req->{oid});
256 my $new_smsg = $req->{new_smsg} = bless {
258 }, 'PublicInbox::Smsg';
259 $new_smsg->{bytes} = $size + crlf_adjust($$bref);
260 defined($req->{xnum} = cur_ibx_xnum($req, $bref)) or return;
265 sub unindex_oid { # git->cat_async callback for 'd'
266 my ($bref, $oid, $type, $size, $req) = @_;
267 my $self = $req->{self};
268 local $self->{current_info} = "$self->{current_info} $oid";
269 return if is_bad_blob($oid, $type, $size, $req->{oid});
270 return if defined(cur_ibx_xnum($req, $bref)); # was re-added
274 # overrides V2Writable::last_commits, called by sync_ranges via sync_prepare
276 my ($self, $sync) = @_;
278 my $ekey = $sync->{ibx}->eidx_key;
279 my $uv = $sync->{ibx}->uidvalidity;
280 for my $i (0..$sync->{epoch_max}) {
281 $heads->[$i] = $self->{oidx}->eidx_meta("lc-v2:$ekey//$uv;$i");
286 sub _sync_inbox ($$$) {
287 my ($self, $sync, $ibx) = @_;
289 $sync->{nr} = \(my $nr = 0);
290 my $v = $ibx->version;
291 my $ekey = $ibx->eidx_key;
294 defined($ibx->git_dir_latest(\$epoch_max)) or return;
295 $sync->{epoch_max} = $epoch_max;
296 sync_prepare($self, $sync); # or return # TODO: once MiscIdx is stable
298 my $uv = $ibx->uidvalidity;
299 my $lc = $self->{oidx}->eidx_meta("lc-v1:$ekey//$uv");
300 my $head = $ibx->mm->last_commit;
301 unless (defined $head) {
302 warn "E: $ibx->{inboxdir} is not indexed\n";
305 my $stk = prepare_stack($sync, $lc ? "$lc..$head" : $head);
306 my $unit = { stack => $stk, git => $ibx->git };
307 push @{$sync->{todo}}, $unit;
309 warn "E: $ekey unsupported inbox version (v$v)\n";
312 for my $unit (@{delete($sync->{todo}) // []}) {
313 last if $sync->{quit};
314 index_todo($self, $sync, $unit);
316 $self->{midx}->index_ibx($ibx) unless $sync->{quit};
317 $ibx->git->cleanup; # done with this inbox, now
320 sub gc_unref_doc ($$$$) {
321 my ($self, $ibx_id, $eidx_key, $docid) = @_;
322 my $dbh = $self->{oidx}->dbh;
324 # for debug/info purposes, oids may no longer be accessible
325 my $sth = $dbh->prepare_cached(<<'', undef, 1);
326 SELECT oidbin FROM xref3 WHERE docid = ? AND ibx_id = ?
328 $sth->execute($docid, $ibx_id);
329 my @oid = map { unpack('H*', $_->[0]) } @{$sth->fetchall_arrayref};
331 $dbh->prepare_cached(<<'')->execute($docid, $ibx_id);
332 DELETE FROM xref3 WHERE docid = ? AND ibx_id = ?
334 my $remain = $self->{oidx}->get_xref3($docid);
335 if (scalar(@$remain)) {
336 $self->{oidx}->eidxq_add($docid); # enqueue for reindex
338 warn "I: unref #$docid $eidx_key $oid\n";
341 warn "I: remove #$docid $eidx_key @oid\n";
342 $self->idx_shard($docid)->shard_remove($docid);
347 my ($self, $opt) = @_;
348 $self->{cfg} or die "E: GC requires ->attach_config\n";
350 $self->idx_init($opt); # acquire lock via V2Writable::_idx_init
352 my $dbh = $self->{oidx}->dbh;
353 my $x3_doc = $dbh->prepare('SELECT docid FROM xref3 WHERE ibx_id = ?');
354 my $ibx_ck = $dbh->prepare('SELECT ibx_id,eidx_key FROM inboxes');
355 my $lc_i = $dbh->prepare('SELECT key FROM eidx_meta WHERE key LIKE ?');
358 while (my ($ibx_id, $eidx_key) = $ibx_ck->fetchrow_array) {
359 next if $self->{ibx_map}->{$eidx_key};
360 $self->{midx}->remove_eidx_key($eidx_key);
361 warn "I: deleting messages for $eidx_key...\n";
362 $x3_doc->execute($ibx_id);
363 while (defined(my $docid = $x3_doc->fetchrow_array)) {
364 gc_unref_doc($self, $ibx_id, $eidx_key, $docid);
366 $dbh->prepare_cached(<<'')->execute($ibx_id);
367 DELETE FROM inboxes WHERE ibx_id = ?
369 # drop last_commit info
371 $pat =~ s/([_%])/\\$1/g;
372 $lc_i->execute("lc-%:$pat//%");
373 while (my ($key) = $lc_i->fetchrow_array) {
374 next if $key !~ m!\Alc-v[1-9]+:\Q$eidx_key\E//!;
375 warn "I: removing $key\n";
376 $dbh->prepare_cached(<<'')->execute($key);
377 DELETE FROM eidx_meta WHERE key = ?
381 warn "I: $eidx_key removed\n";
384 # it's not real unless it's in `over', we use parallelism here,
385 # shards will be reading directly from over, so commit
386 $self->{oidx}->commit_lazy;
387 $self->{oidx}->begin_lazy;
389 for my $idx (@{$self->{idx_shards}}) {
390 warn "I: cleaning up shard #$idx->{shard}\n";
391 $idx->shard_over_check($self->{oidx});
393 my $nr = $dbh->do(<<'');
394 DELETE FROM xref3 WHERE docid NOT IN (SELECT num FROM over)
396 warn "I: eliminated $nr stale xref3 entries\n" if $nr != 0;
402 my ($self, $sync, $smsg) = @_;
403 my $ibx_id = delete($smsg->{ibx_id}) // die '{ibx_id} unset';
404 my $pos = $sync->{id2pos}->{$ibx_id} // die "$ibx_id no pos";
405 $self->{ibx_list}->[$pos] // die "BUG: ibx for $smsg->{blob} not mapped"
408 sub _reindex_finalize ($$$) {
409 my ($req, $smsg, $eml) = @_;
410 my $sync = $req->{sync};
411 my $self = $sync->{self};
412 my $by_chash = delete $req->{by_chash} or die 'BUG: no {by_chash}';
413 my $nr = scalar(keys(%$by_chash)) or die 'BUG: no content hashes';
414 my $orig_smsg = $req->{orig_smsg} // die 'BUG: no {orig_smsg}';
415 my $docid = $smsg->{num} = $orig_smsg->{num};
416 $self->{oidx}->add_overview($eml, $smsg); # may rethread
417 check_batch_limit({ %$sync, new_smsg => $smsg });
418 my $chash0 = $smsg->{chash} // die "BUG: $smsg->{blob} no {chash}";
419 my $stable = delete($by_chash->{$chash0}) //
420 die "BUG: $smsg->{blob} chash missing";
421 my $idx = $self->idx_shard($docid);
422 my $top_smsg = pop @$stable;
423 $top_smsg == $smsg or die 'BUG: top_smsg != smsg';
424 my $ibx = _ibx_for($self, $sync, $smsg);
425 $idx->index_raw(undef, $eml, $smsg, $ibx->eidx_key);
426 for my $x (reverse @$stable) {
427 $ibx = _ibx_for($self, $sync, $x);
428 my $hdr = delete $x->{hdr} // die 'BUG: no {hdr}';
429 $idx->shard_add_eidx_info($docid, $ibx->eidx_key, $hdr);
431 return if $nr == 1; # likely, all good
433 warn "W: #$docid split into $nr due to deduplication change\n";
435 for my $ary (values %$by_chash) {
436 for my $x (reverse @$ary) {
437 warn "removing #$docid xref3 $x->{blob}\n";
438 my $n = $self->{oidx}->remove_xref3($docid, $x->{blob});
439 die "BUG: $x->{blob} invalidated #$docid" if $n == 0;
441 my $x = pop(@$ary) // die "BUG: #$docid {by_chash} empty";
442 $x->{num} = delete($x->{xnum}) // die '{xnum} unset';
443 $ibx = _ibx_for($self, $sync, $x);
444 my $e = $ibx->over->get_art($x->{num});
445 $e->{blob} eq $x->{blob} or die <<EOF;
446 $x->{blob} != $e->{blob} (${\$ibx->eidx_key}:$e->{num});
448 push @todo, $ibx, $e;
451 while (my ($ibx, $e) = splice(@todo, 0, 2)) {
452 reindex_unseen($self, $sync, $ibx, $e);
456 sub _reindex_oid { # git->cat_async callback
457 my ($bref, $oid, $type, $size, $req) = @_;
458 my $sync = $req->{sync};
459 my $self = $sync->{self};
460 my $orig_smsg = $req->{orig_smsg} // die 'BUG: no {orig_smsg}';
461 my $expect_oid = $req->{xr3r}->[$req->{ix}]->[2];
462 my $docid = $orig_smsg->{num};
463 if (is_bad_blob($oid, $type, $size, $expect_oid)) {
464 my $remain = $self->{oidx}->remove_xref3($docid, $expect_oid);
466 warn "W: #$docid gone or corrupted\n";
467 $self->idx_shard($docid)->shard_remove($docid);
468 } elsif (my $next_oid = $req->{xr3r}->[++$req->{ix}]->[2]) {
469 $self->git->cat_async($next_oid, \&_reindex_oid, $req);
471 warn "BUG: #$docid gone (UNEXPECTED)\n";
472 $self->idx_shard($docid)->shard_remove($docid);
476 my $ci = $self->{current_info};
477 local $self->{current_info} = "$ci #$docid $oid";
478 my $re_smsg = bless { blob => $oid }, 'PublicInbox::Smsg';
479 $re_smsg->{bytes} = $size + crlf_adjust($$bref);
480 my $eml = PublicInbox::Eml->new($bref);
481 $re_smsg->populate($eml, { autime => $orig_smsg->{ds},
482 cotime => $orig_smsg->{ts} });
483 my $chash = content_hash($eml);
484 $re_smsg->{chash} = $chash;
485 $re_smsg->{xnum} = $req->{xr3r}->[$req->{ix}]->[1];
486 $re_smsg->{ibx_id} = $req->{xr3r}->[$req->{ix}]->[0];
487 $re_smsg->{hdr} = $eml->header_obj;
488 push @{$req->{by_chash}->{$chash}}, $re_smsg;
489 if (my $next_oid = $req->{xr3r}->[++$req->{ix}]->[2]) {
490 $self->git->cat_async($next_oid, \&_reindex_oid, $req);
491 } else { # last $re_smsg is the highest priority xref3
492 local $self->{current_info} = "$ci #$docid";
493 _reindex_finalize($req, $re_smsg, $eml);
497 sub _reindex_smsg ($$$) {
498 my ($self, $sync, $smsg) = @_;
499 my $docid = $smsg->{num};
500 my $xr3 = $self->{oidx}->get_xref3($docid, 1);
501 if (scalar(@$xr3) == 0) { # _reindex_check_stale should've covered this
503 BUG? #$docid $smsg->{blob} is not referenced by inboxes during reindex
505 $self->{oidx}->delete_by_num($docid);
506 $self->idx_shard($docid)->shard_remove($docid);
510 # we sort {xr3r} in the reverse order of {ibx_list} so we can
511 # hit the common case in _reindex_finalize without rereading
512 # from git (or holding multiple messages in memory).
513 my $id2pos = $sync->{id2pos}; # index in {ibx_list}
515 $id2pos->{$b->[0]} <=> $id2pos->{$a->[0]}
517 $b->[1] <=> $a->[1] # break ties with {xnum}
519 @$xr3 = map { [ $_->[0], $_->[1], unpack('H*', $_->[2]) ] } @$xr3;
520 my $req = { orig_smsg => $smsg, sync => $sync, xr3r => $xr3, ix => 0 };
521 $self->git->cat_async($xr3->[$req->{ix}]->[2], \&_reindex_oid, $req);
524 sub checkpoint_due ($) {
526 ${$sync->{need_checkpoint}} || (now() > $sync->{next_check});
530 # I've copied FS images and only changed the hostname before,
531 # so prepend hostname. Use `state' since these a BOFH can change
532 # these while this process is running and we always want to be
533 # able to release locks taken by this process.
534 state $retval = hostname . '-' . do {
535 my $m; # machine-id(5) is systemd
536 if (open(my $fh, '<', '/etc/machine-id')) { $m = <$fh> }
537 # hostid(1) is in GNU coreutils, kern.hostid is FreeBSD
538 chomp($m ||= `hostid` || `sysctl -n kern.hostid`);
545 my $expect = delete($self->{-eidxq_locked}) or return;
546 my ($owner_pid, undef) = split(/-/, $expect);
547 return if $owner_pid != $$; # shards may fork
548 my $oidx = $self->{oidx};
550 my $cur = $oidx->eidx_meta('eidxq_lock') // '';
551 if ($cur eq $expect) {
552 $oidx->eidx_meta('eidxq_lock', '');
554 } elsif ($cur ne '') {
555 warn "E: eidxq_lock($expect) stolen by $cur\n";
557 warn "E: eidxq_lock($expect) released by another process\n";
564 eidxq_release($self) and $self->{oidx}->commit_lazy;
567 sub _eidxq_take ($) {
569 my $val = "$$-${\time}-$>-".host_ident;
570 $self->{oidx}->eidx_meta('eidxq_lock', $val);
571 $self->{-eidxq_locked} = $val;
574 sub eidxq_lock_acquire ($) {
576 my $oidx = $self->{oidx};
578 my $cur = $oidx->eidx_meta('eidxq_lock') || return _eidxq_take($self);
579 if (my $locked = $self->{-eidxq_locked}) { # be lazy
580 return $locked if $locked eq $cur;
582 my ($pid, $time, $euid, $ident) = split(/-/, $cur, 4);
583 my $t = strftime('%Y-%m-%d %k:%M:%S', gmtime($time));
584 if ($euid == $> && $ident eq host_ident) {
587 I: PID:$pid (re)indexing Xapian since $t, it will continue our work
591 warn "I: eidxq_lock is stale ($cur), clobbering\n";
592 return _eidxq_take($self);
594 warn "E: kill(0, $pid) failed: $!\n"; # fall-through:
596 my $fn = $oidx->dbh->sqlite_db_filename;
598 W: PID:$pid, UID:$euid on $ident is indexing Xapian since $t
599 W: If this is unexpected, delete `eidxq_lock' from the `eidx_meta' table:
600 W: sqlite3 $fn 'DELETE FROM eidx_meta WHERE key = "eidxq_lock"'
605 sub eidxq_process ($$) { # for reindexing
606 my ($self, $sync) = @_;
608 return unless eidxq_lock_acquire($self);
609 my $dbh = $self->{oidx}->dbh;
610 my $tot = $dbh->selectrow_array('SELECT COUNT(*) FROM eidxq') or return;
612 $sync->{-regen_fmt} = "%u/$tot\n";
613 my $pr = $sync->{-opt}->{-progress};
615 my $min = $dbh->selectrow_array('SELECT MIN(docid) FROM eidxq');
616 my $max = $dbh->selectrow_array('SELECT MAX(docid) FROM eidxq');
617 $pr->("Xapian indexing $min..$max (total=$tot)\n");
619 $sync->{id2pos} //= do {
622 $id2pos{$_->{-ibx_id}} = $pos++ for @{$self->{ibx_list}};
627 $del = $dbh->prepare('DELETE FROM eidxq WHERE docid = ?');
628 $iter = $dbh->prepare('SELECT docid FROM eidxq ORDER BY docid ASC');
630 while (defined(my $docid = $iter->fetchrow_array)) {
631 last if $sync->{quit};
632 if (my $smsg = $self->{oidx}->get_art($docid)) {
633 _reindex_smsg($self, $sync, $smsg);
635 warn "E: #$docid does not exist in over\n";
637 $del->execute($docid);
640 if (checkpoint_due($sync)) {
641 $dbh = $del = $iter = undef;
642 reindex_checkpoint($self, $sync); # release lock
643 $dbh = $self->{oidx}->dbh;
647 $self->git->async_wait_all;
648 $pr->("reindexed ${$sync->{nr}}/$tot\n") if $pr;
651 sub _reindex_unseen { # git->cat_async callback
652 my ($bref, $oid, $type, $size, $req) = @_;
653 return if is_bad_blob($oid, $type, $size, $req->{oid});
654 my $self = $req->{self} // die 'BUG: {self} unset';
655 local $self->{current_info} = "$self->{current_info} $oid";
656 my $new_smsg = bless { blob => $oid, }, 'PublicInbox::Smsg';
657 $new_smsg->{bytes} = $size + crlf_adjust($$bref);
658 my $eml = $req->{eml} = PublicInbox::Eml->new($bref);
659 $req->{new_smsg} = $new_smsg;
660 $req->{chash} = content_hash($eml);
661 $req->{mids} = mids($eml); # do_step iterates through this
662 do_step($req); # enter the normal indexing flow
665 # --reindex may catch totally unseen messages, this handles them
666 sub reindex_unseen ($$$$) {
667 my ($self, $sync, $ibx, $xsmsg) = @_;
670 autime => $xsmsg->{ds},
671 cotime => $xsmsg->{ts},
672 oid => $xsmsg->{blob},
674 xnum => $xsmsg->{num},
675 # {mids} and {chash} will be filled in at _reindex_unseen
677 warn "I: reindex_unseen ${\$ibx->eidx_key}:$req->{xnum}:$req->{oid}\n";
678 $self->git->cat_async($xsmsg->{blob}, \&_reindex_unseen, $req);
681 sub _reindex_check_unseen ($$$) {
682 my ($self, $sync, $ibx) = @_;
683 my $ibx_id = $ibx->{-ibx_id};
685 my ($beg, $end) = (1, $slice);
687 # first, check if we missed any messages in target $ibx
689 my $pr = $sync->{-opt}->{-progress};
690 my $ekey = $ibx->eidx_key;
691 $sync->{-regen_fmt} = "$ekey checking unseen %u/".$ibx->over->max."\n";
694 while (scalar(@{$msgs = $ibx->over->query_xover($beg, $end)})) {
695 ${$sync->{nr}} = $beg;
696 $beg = $msgs->[-1]->{num} + 1;
697 $end = $beg + $slice;
698 if (checkpoint_due($sync)) {
699 reindex_checkpoint($self, $sync); # release lock
702 my $inx3 = $self->{oidx}->dbh->prepare_cached(<<'', undef, 1);
703 SELECT DISTINCT(docid) FROM xref3 WHERE
704 ibx_id = ? AND xnum = ? AND oidbin = ?
706 for my $xsmsg (@$msgs) {
707 my $oidbin = pack('H*', $xsmsg->{blob});
708 $inx3->bind_param(1, $ibx_id);
709 $inx3->bind_param(2, $xsmsg->{num});
710 $inx3->bind_param(3, $oidbin, SQL_BLOB);
712 my $docids = $inx3->fetchall_arrayref;
713 # index messages which were totally missed
714 # the first time around ASAP:
715 if (scalar(@$docids) == 0) {
716 reindex_unseen($self, $sync, $ibx, $xsmsg);
717 } else { # already seen, reindex later
718 for my $r (@$docids) {
719 $self->{oidx}->eidxq_add($r->[0]);
722 last if $sync->{quit};
724 last if $sync->{quit};
728 sub _reindex_check_stale ($$$) {
729 my ($self, $sync, $ibx) = @_;
731 my $pr = $sync->{-opt}->{-progress};
733 my $ekey = $ibx->eidx_key;
734 $sync->{-regen_fmt} =
735 "$ekey check stale/missing %u/".$ibx->over->max."\n";
738 if (checkpoint_due($sync)) {
739 reindex_checkpoint($self, $sync); # release lock
741 # now, check if there's stale xrefs
742 my $iter = $self->{oidx}->dbh->prepare_cached(<<'', undef, 1);
743 SELECT docid,xnum,oidbin FROM xref3 WHERE ibx_id = ? AND docid > ?
744 ORDER BY docid,xnum ASC LIMIT 10000
746 $iter->execute($ibx->{-ibx_id}, $min);
749 while (my ($docid, $xnum, $oidbin) = $iter->fetchrow_array) {
750 return if $sync->{quit};
751 ${$sync->{nr}} = $xnum;
753 $fetching = $min = $docid;
754 my $smsg = $ibx->over->get_art($xnum);
755 my $oidhex = unpack('H*', $oidbin);
759 } elsif ($smsg->{blob} ne $oidhex) {
760 $err = "mismatch (!= $smsg->{blob})";
762 next; # likely, all good
764 # current_info already has eidx_key
765 warn "$xnum:$oidhex (#$docid): $err\n";
766 my $del = $self->{oidx}->dbh->prepare_cached(<<'');
767 DELETE FROM xref3 WHERE ibx_id = ? AND xnum = ? AND oidbin = ?
769 $del->bind_param(1, $ibx->{-ibx_id});
770 $del->bind_param(2, $xnum);
771 $del->bind_param(3, $oidbin, SQL_BLOB);
774 # get_xref3 over-fetches, but this is a rare path:
775 my $xr3 = $self->{oidx}->get_xref3($docid);
776 my $idx = $self->idx_shard($docid);
777 if (scalar(@$xr3) == 0) { # all gone
778 $self->{oidx}->delete_by_num($docid);
779 $self->{oidx}->eidxq_del($docid);
780 $idx->shard_remove($docid);
781 } else { # enqueue for reindex of remaining messages
782 $idx->shard_remove_eidx_info($docid,
784 $self->{oidx}->eidxq_add($docid); # yes, add
787 } while (defined $fetching);
790 sub _reindex_inbox ($$$) {
791 my ($self, $sync, $ibx) = @_;
792 local $self->{current_info} = $ibx->eidx_key;
793 _reindex_check_unseen($self, $sync, $ibx);
794 _reindex_check_stale($self, $sync, $ibx) unless $sync->{quit};
795 delete @$ibx{qw(over mm search git)}; # won't need these for a bit
799 my ($self, $sync) = @_;
801 # acquire eidxq_lock early because full reindex takes forever
802 # and incremental -extindex processes can run during our checkpoints
803 if (!eidxq_lock_acquire($self)) {
804 warn "E: aborting --reindex\n";
807 for my $ibx (@{$self->{ibx_list}}) {
808 _reindex_inbox($self, $sync, $ibx);
809 last if $sync->{quit};
811 $self->git->async_wait_all; # ensure eidxq gets filled completely
812 eidxq_process($self, $sync) unless $sync->{quit};
815 sub eidx_sync { # main entry point
816 my ($self, $opt) = @_;
818 my $warn_cb = $SIG{__WARN__} || sub { print STDERR @_ };
819 local $self->{current_info} = '';
820 local $SIG{__WARN__} = sub {
821 $warn_cb->($self->{current_info}, ': ', @_);
823 $self->idx_init($opt); # acquire lock via V2Writable::_idx_init
824 $self->{oidx}->rethread_prepare($opt);
826 need_checkpoint => \(my $need_checkpoint = 0),
828 next_check => now() + 10,
830 # DO NOT SET {reindex} here, it's incompatible with reused
831 # V2Writable code, reindex is totally different here
832 # compared to v1/v2 inboxes because we have multiple histories
834 -regen_fmt => "%u/?\n",
836 local $SIG{USR1} = sub { $need_checkpoint = 1 };
837 my $quit = PublicInbox::SearchIdx::quit_cb($sync);
838 local $SIG{QUIT} = $quit;
839 local $SIG{INT} = $quit;
840 local $SIG{TERM} = $quit;
841 for my $ibx (@{$self->{ibx_list}}) {
842 $ibx->{-ibx_id} //= $self->{oidx}->ibx_id($ibx->eidx_key);
844 if (delete($opt->{reindex})) {
845 $sync->{checkpoint_unlocks} = 1;
846 eidx_reindex($self, $sync);
849 # don't use $_ here, it'll get clobbered by reindex_checkpoint
850 for my $ibx (@{$self->{ibx_list}}) {
851 last if $sync->{quit};
852 _sync_inbox($self, $sync, $ibx);
854 $self->{oidx}->rethread_done($opt) unless $sync->{quit};
855 eidxq_process($self, $sync) unless $sync->{quit};
857 eidxq_release($self);
858 PublicInbox::V2Writable::done($self);
861 sub update_last_commit { # overrides V2Writable
862 my ($self, $sync, $stk) = @_;
863 my $unit = $sync->{unit} // return;
864 my $latest_cmt = $stk ? $stk->{latest_cmt} : ${$sync->{latest_cmt}};
865 defined($latest_cmt) or return;
866 my $ibx = $sync->{ibx} or die 'BUG: {ibx} missing';
867 my $ekey = $ibx->eidx_key;
868 my $uv = $ibx->uidvalidity;
869 my $epoch = $unit->{epoch};
871 my $v = $ibx->version;
873 die 'No {epoch} for v2 unit' unless defined $epoch;
874 $meta_key = "lc-v2:$ekey//$uv;$epoch";
876 die 'Unexpected {epoch} for v1 unit' if defined $epoch;
877 $meta_key = "lc-v1:$ekey//$uv";
879 die "Unsupported inbox version: $v";
881 my $last = $self->{oidx}->eidx_meta($meta_key);
882 if (defined $last && is_ancestor($self->git, $last, $latest_cmt)) {
883 my @cmd = (qw(rev-list --count), "$last..$latest_cmt");
884 chomp(my $n = $unit->{git}->qx(@cmd));
885 return if $n ne '' && $n == 0;
887 $self->{oidx}->eidx_meta($meta_key, $latest_cmt);
890 sub _idx_init { # with_umask callback
891 my ($self, $opt) = @_;
892 PublicInbox::V2Writable::_idx_init($self, $opt);
893 $self->{midx} = PublicInbox::MiscIdx->new($self);
896 sub idx_init { # similar to V2Writable
897 my ($self, $opt) = @_;
898 return if $self->{idx_shards};
902 my $ALL = $self->git->{git_dir}; # ALL.git
903 PublicInbox::Import::init_bare($ALL) unless -d $ALL;
904 my $info_dir = "$ALL/objects/info";
905 my $alt = "$info_dir/alternates";
907 my (@old, @new, %seen); # seen: st_dev + st_ino
909 open(my $fh, '<', $alt) or die "open $alt: $!";
910 $mode = (stat($fh))[2] & 07777;
911 while (my $line = <$fh>) {
912 chomp(my $d = $line);
913 if (my @st = stat($d)) {
914 next if $seen{"$st[0]\0$st[1]"}++;
916 warn "W: stat($d) failed (from $alt): $!\n";
917 next if $opt->{-idx_gc};
922 for my $ibx (@{$self->{ibx_list}}) {
923 my $line = $ibx->git->{git_dir} . "/objects\n";
924 chomp(my $d = $line);
925 if (my @st = stat($d)) {
926 next if $seen{"$st[0]\0$st[1]"}++;
928 warn "W: stat($d) failed (from $ibx->{inboxdir}): $!\n";
929 next if $opt->{-idx_gc};
936 PublicInbox::V2Writable::write_alternates($info_dir, $mode, $o);
938 $self->parallel_init($self->{indexlevel});
939 $self->umask_prepare;
940 $self->with_umask(\&_idx_init, $self, $opt);
941 $self->{oidx}->begin_lazy;
942 $self->{oidx}->eidx_prep;
943 $self->{midx}->begin_txn;
947 *done = \&PublicInbox::V2Writable::done;
948 *umask_prepare = \&PublicInbox::InboxWritable::umask_prepare;
949 *with_umask = \&PublicInbox::InboxWritable::with_umask;
950 *parallel_init = \&PublicInbox::V2Writable::parallel_init;
951 *nproc_shards = \&PublicInbox::V2Writable::nproc_shards;
952 *sync_prepare = \&PublicInbox::V2Writable::sync_prepare;
953 *index_todo = \&PublicInbox::V2Writable::index_todo;
954 *count_shards = \&PublicInbox::V2Writable::count_shards;
955 *atfork_child = \&PublicInbox::V2Writable::atfork_child;
956 *idx_shard = \&PublicInbox::V2Writable::idx_shard;
957 *reindex_checkpoint = \&PublicInbox::V2Writable::reindex_checkpoint;