X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FSearchIdx.pm;h=6e2e614c5085c7e45fa33ad06375a9137af1ee31;hb=7ffa98cbbf5bff8225fdfedb0f2b7c2beb5d0a6f;hp=0458d9c3ff18ec1ff24fd2e84735d5ac2d334241;hpb=98c03415fed31cce80f8b2030aee9311890c79d6;p=public-inbox.git diff --git a/lib/PublicInbox/SearchIdx.pm b/lib/PublicInbox/SearchIdx.pm index 0458d9c3..6e2e614c 100644 --- a/lib/PublicInbox/SearchIdx.pm +++ b/lib/PublicInbox/SearchIdx.pm @@ -1,6 +1,6 @@ -# Copyright (C) 2015-2020 all contributors +# Copyright (C) 2015-2021 all contributors # License: AGPL-3.0+ -# based on notmuch, but with no concept of folders, files or flags +# based on notmuch, but with no concept of folders, files # # Indexes mail with Xapian and our (SQLite-based) ::Msgmap for use # with the web and NNTP interfaces. This index maintains thread @@ -11,27 +11,36 @@ use strict; use v5.10.1; use parent qw(PublicInbox::Search PublicInbox::Lock Exporter); use PublicInbox::Eml; +use PublicInbox::Search qw(xap_terms); use PublicInbox::InboxWritable; use PublicInbox::MID qw(mids_for_index mids); use PublicInbox::MsgIter; use PublicInbox::IdxStack; -use Carp qw(croak); +use Carp qw(croak carp); use POSIX qw(strftime); use Time::Local qw(timegm); use PublicInbox::OverIdx; use PublicInbox::Spawn qw(spawn nodatacow_dir); use PublicInbox::Git qw(git_unquote); use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp); -our @EXPORT_OK = qw(crlf_adjust log2stack is_ancestor check_size); +use PublicInbox::Address; +use Config; +our @EXPORT_OK = qw(log2stack is_ancestor check_size prepare_stack + index_text term_generator add_val is_bad_blob); my $X = \%PublicInbox::Search::X; -my ($DB_CREATE_OR_OPEN, $DB_OPEN); +our ($DB_CREATE_OR_OPEN, $DB_OPEN); our $DB_NO_SYNC = 0; -our $BATCH_BYTES = $ENV{XAPIAN_FLUSH_THRESHOLD} ? 0x7fffffff : 1_000_000; +our $BATCH_BYTES = $ENV{XAPIAN_FLUSH_THRESHOLD} ? 0x7fffffff : + # assume a typical 64-bit system has 8x more RAM than a + # typical 32-bit system: + (($Config{ptrsize} >= 8 ? 8192 : 1024) * 1024); + use constant DEBUG => !!$ENV{DEBUG}; my $xapianlevels = qr/\A(?:full|medium)\z/; my $hex = '[a-f0-9]'; my $OID = $hex .'{40,}'; +my @VMD_MAP = (kw => 'K', L => 'L'); our $INDEXLEVELS = qr/\A(?:full|medium|basic)\z/; sub new { @@ -53,20 +62,16 @@ sub new { } } $ibx = PublicInbox::InboxWritable->new($ibx); - my $self = bless { - ibx => $ibx, - xpfx => $inboxdir, # for xpfx_init - -altid => $altid, - ibx_ver => $version, - indexlevel => $indexlevel, - }, $class; - $self->xpfx_init; + my $self = PublicInbox::Search->new($ibx); + bless $self, $class; + $self->{ibx} = $ibx; + $self->{-altid} = $altid; + $self->{indexlevel} = $indexlevel; $self->{-set_indexlevel_once} = 1 if $indexlevel eq 'medium'; if ($ibx->{-skip_docdata}) { $self->{-set_skip_docdata_once} = 1; $self->{-skip_docdata} = 1; } - $ibx->umask_prepare; if ($version == 1) { $self->{lock_path} = "$inboxdir/ssoma.lock"; my $dir = $self->xdir; @@ -89,7 +94,7 @@ sub need_xapian ($) { $_[0]->{indexlevel} =~ $xapianlevels } sub idx_release { my ($self, $wake) = @_; if (need_xapian($self)) { - my $xdb = delete $self->{xdb} or croak 'not acquired'; + my $xdb = delete $self->{xdb} or croak '{xdb} not acquired'; $xdb->close; } $self->lock_release($wake) if $self->{creat}; @@ -98,19 +103,21 @@ sub idx_release { sub load_xapian_writable () { return 1 if $X->{WritableDatabase}; - PublicInbox::Search::load_xapian() or return; + PublicInbox::Search::load_xapian() or die "failed to load Xapian: $@\n"; my $xap = $PublicInbox::Search::Xap; for (qw(Document TermGenerator WritableDatabase)) { $X->{$_} = $xap.'::'.$_; } eval 'require '.$X->{WritableDatabase} or die; *sortable_serialise = $xap.'::sortable_serialise'; - *sortable_unserialise = $xap.'::sortable_unserialise'; $DB_CREATE_OR_OPEN = eval($xap.'::DB_CREATE_OR_OPEN()'); $DB_OPEN = eval($xap.'::DB_OPEN()'); my $ver = (eval($xap.'::major_version()') << 16) | - (eval($xap.'::minor_version()') << 8); + (eval($xap.'::minor_version()') << 8) | + eval($xap.'::revision()'); $DB_NO_SYNC = 0x4 if $ver >= 0x10400; + # Xapian v1.2.21..v1.2.24 were missing close-on-exec on OFD locks + $X->{CLOEXEC_UNSET} = 1 if $ver >= 0x010215 && $ver <= 0x010218; 1; } @@ -137,7 +144,7 @@ sub idx_acquire { } } return unless defined $flag; - $flag |= $DB_NO_SYNC if $self->{ibx}->{-no_fsync}; + $flag |= $DB_NO_SYNC if ($self->{ibx} // $self->{eidx})->{-no_fsync}; my $xdb = eval { ($X->{WritableDatabase})->new($dir, $flag) }; croak "Failed opening $dir: $@" if $@; $self->{xdb} = $xdb; @@ -154,27 +161,49 @@ sub term_generator ($) { # write-only $self->{term_generator} //= do { my $tg = $X->{TermGenerator}->new; - $tg->set_stemmer($self->stemmer); + $tg->set_stemmer(PublicInbox::Search::stemmer($self)); $tg; } } +sub index_phrase ($$$$) { + my ($self, $text, $wdf_inc, $prefix) = @_; + + my $tg = term_generator($self); + $tg->index_text($text, $wdf_inc, $prefix); + $tg->increase_termpos; +} + sub index_text ($$$$) { my ($self, $text, $wdf_inc, $prefix) = @_; - my $tg = term_generator($self); # man Search::Xapian::TermGenerator if ($self->{indexlevel} eq 'full') { - $tg->index_text($text, $wdf_inc, $prefix); - $tg->increase_termpos; + index_phrase($self, $text, $wdf_inc, $prefix); } else { + my $tg = term_generator($self); $tg->index_text_without_positions($text, $wdf_inc, $prefix); } } sub index_headers ($$) { my ($self, $smsg) = @_; - my @x = (from => 'A', # Author - subject => 'S', to => 'XTO', cc => 'XCC'); + my @x = (from => 'A', to => 'XTO', cc => 'XCC'); # A: Author + while (my ($field, $pfx) = splice(@x, 0, 2)) { + my $val = $smsg->{$field}; + next if $val eq ''; + # include "(comments)" after the address, too, so not using + # PublicInbox::Address::names or pairs + index_text($self, $val, 1, $pfx); + + # we need positional info for email addresses since they + # can be considered phrases + if ($self->{indexlevel} eq 'medium') { + for my $addr (PublicInbox::Address::emails($val)) { + index_phrase($self, $addr, 1, $pfx); + } + } + } + @x = (subject => 'S'); while (my ($field, $pfx) = splice(@x, 0, 2)) { my $val = $smsg->{$field}; index_text($self, $val, 1, $pfx) if $val ne ''; @@ -187,7 +216,11 @@ sub index_diff_inc ($$$$) { index_text($self, join("\n", @$xnq), 1, 'XNQ'); @$xnq = (); } - index_text($self, $text, 1, $pfx); + if ($pfx eq 'XDFN') { + index_phrase($self, $text, 1, $pfx); + } else { + index_text($self, $text, 1, $pfx); + } } sub index_old_diff_fn { @@ -226,6 +259,7 @@ sub index_diff ($$$) { } elsif (m!^diff --git "?[^/]+/.+ "?[^/]+/.+\z!) { # wait until "---" and "+++" to capture filenames $in_diff = 1; + push @xnq, $_; # traditional diff: } elsif (m/^diff -(.+) (\S+) (\S+)$/) { my ($opt, $fa, $fb) = ($1, $2, $3); @@ -293,7 +327,7 @@ sub index_xapian { # msg_iter callback my $ct = $part->content_type || 'text/plain'; my $fn = $part->filename; if (defined $fn && $fn ne '') { - index_text($self, $fn, 1, 'XFN'); + index_phrase($self, $fn, 1, 'XFN'); } if ($part->{is_submsg}) { my $mids = mids_for_index($part); @@ -331,28 +365,29 @@ sub index_list_id ($$$) { $l =~ /<([^>]+)>/ or next; my $lid = lc $1; $doc->add_boolean_term('G' . $lid); - index_text($self, $lid, 1, 'XL'); # probabilistic + index_phrase($self, $lid, 1, 'XL'); # probabilistic } } sub index_ids ($$$$) { my ($self, $doc, $hdr, $mids) = @_; for my $mid (@$mids) { - index_text($self, $mid, 1, 'XM'); + index_phrase($self, $mid, 1, 'XM'); # because too many Message-IDs are prefixed with # "Pine.LNX."... if ($mid =~ /\w{12,}/) { my @long = ($mid =~ /(\w{3,}+)/g); - index_text($self, join(' ', @long), 1, 'XM'); + index_phrase($self, join(' ', @long), 1, 'XM'); } } $doc->add_boolean_term('Q' . $_) for @$mids; index_list_id($self, $doc, $hdr); } -sub add_xapian ($$$$) { +sub eml2doc ($$$;$) { my ($self, $eml, $smsg, $mids) = @_; + $mids //= mids_for_index($eml); my $doc = $X->{Document}->new; add_val($doc, PublicInbox::Search::TS(), $smsg->{ts}); my @ds = gmtime($smsg->{ds}); @@ -369,9 +404,7 @@ sub add_xapian ($$$$) { index_headers($self, $smsg); if (defined(my $eidx_key = $smsg->{eidx_key})) { - $doc->add_boolean_term('O'.$eidx_key); - $doc->add_boolean_term('P'. - "$eidx_key:$smsg->{num}:$smsg->{blob}"); + $doc->add_boolean_term('O'.$eidx_key) if $eidx_key ne '.'; } msg_iter($eml, \&index_xapian, [ $self, $doc ]); index_ids($self, $doc, $eml, $mids); @@ -382,7 +415,7 @@ sub add_xapian ($$$$) { if (!$self->{-skip_docdata}) { # WWW doesn't need {to} or {cc}, only NNTP $smsg->{to} = $smsg->{cc} = ''; - PublicInbox::OverIdx::parse_references($smsg, $eml, $mids); + $smsg->parse_references($eml, $mids); my $data = $smsg->to_doc_data; $doc->set_data($data); } @@ -397,22 +430,38 @@ sub add_xapian ($$$$) { } } } + $doc; +} + +sub add_xapian ($$$$) { + my ($self, $eml, $smsg, $mids) = @_; + begin_txn_lazy($self); + my $merge_vmd = delete $smsg->{-merge_vmd}; + my $doc = eml2doc($self, $eml, $smsg, $mids); + if (my $old = $merge_vmd ? _get_doc($self, $smsg->{num}) : undef) { + my @x = @VMD_MAP; + while (my ($field, $pfx) = splice(@x, 0, 2)) { + for my $term (xap_terms($pfx, $old)) { + $doc->add_boolean_term($pfx.$term); + } + } + } $self->{xdb}->replace_document($smsg->{num}, $doc); } sub _msgmap_init ($) { my ($self) = @_; - die "BUG: _msgmap_init is only for v1\n" if $self->{ibx_ver} != 1; - $self->{mm} //= eval { + die "BUG: _msgmap_init is only for v1\n" if $self->{ibx}->version != 1; + $self->{mm} //= do { require PublicInbox::Msgmap; - my $rw = $self->{ibx}->{-no_fsync} ? 2 : 1; - PublicInbox::Msgmap->new($self->{ibx}->{inboxdir}, $rw); + PublicInbox::Msgmap->new_file($self->{ibx}, 1); }; } sub add_message { # mime = PublicInbox::Eml or Email::MIME object my ($self, $mime, $smsg, $sync) = @_; + begin_txn_lazy($self); my $mids = mids_for_index($mime); $smsg //= bless { blob => '' }, 'PublicInbox::Smsg'; # test-only compat $smsg->{mid} //= $mids->[0]; # v1 compatibility @@ -446,108 +495,172 @@ sub add_message { $smsg->{num}; } -sub _get_doc ($$$) { - my ($self, $docid, $oid) = @_; +sub _get_doc ($$) { + my ($self, $docid) = @_; my $doc = eval { $self->{xdb}->get_document($docid) }; $doc // do { warn "E: $@\n" if $@; - warn "E: #$docid $oid missing in Xapian\n"; + warn "E: #$docid missing in Xapian\n"; undef; } } -sub add_xref3 { - my ($self, $docid, $xnum, $oid, $eidx_key, $eml) = @_; +sub add_eidx_info { + my ($self, $docid, $eidx_key, $eml) = @_; begin_txn_lazy($self); - my $doc = _get_doc($self, $docid, $oid) or return; + my $doc = _get_doc($self, $docid) or return; term_generator($self)->set_document($doc); - $doc->add_boolean_term('O'.$eidx_key); - $doc->add_boolean_term('P'."$eidx_key:$xnum:$oid"); + + # '.' is special for lei_store + $doc->add_boolean_term('O'.$eidx_key) if $eidx_key ne '.'; + index_list_id($self, $doc, $eml); $self->{xdb}->replace_document($docid, $doc); } -sub remove_xref3 { - my ($self, $docid, $oid, $eidx_key, $eml) = @_; +sub get_terms { + my ($self, $pfx, $docid) = @_; + begin_txn_lazy($self); + xap_terms($pfx, $self->{xdb}, $docid); +} + +sub remove_eidx_info { + my ($self, $docid, $eidx_key, $eml) = @_; + begin_txn_lazy($self); + my $doc = _get_doc($self, $docid) or return; + eval { $doc->remove_term('O'.$eidx_key) }; + warn "W: ->remove_term O$eidx_key: $@\n" if $@; + for my $l ($eml ? $eml->header_raw('List-Id') : ()) { + $l =~ /<([^>]+)>/ or next; + my $lid = lc $1; + eval { $doc->remove_term('G' . $lid) }; + warn "W: ->remove_term G$lid: $@\n" if $@; + + # nb: we don't remove the XL probabilistic terms + # since terms may overlap if cross-posted. + # + # IOW, a message which has both + # and would have overlapping + # "XLexample" and "XLcom" as terms and which we + # wouldn't know if they're safe to remove if we just + # unindex while preserving + # . + # + # In any case, this entire sub is will likely never + # be needed and users using the "l:" prefix are probably + # rarer. + } + $self->{xdb}->replace_document($docid, $doc); +} + +sub set_vmd { + my ($self, $docid, $vmd) = @_; begin_txn_lazy($self); - my $doc = _get_doc($self, $docid, $oid) or return; - my $xref3 = PublicInbox::Smsg::xref3(undef, $doc); - my %x3 = map { $_ => undef } @$xref3; - for (grep(/\A\Q$eidx_key\E:[0-9]+:\Q$oid\E\z/, @$xref3)) { - delete $x3{$_}; - $doc->remove_term('P' . $_); - } - if (scalar(keys(%x3)) == 0) { - $self->{xdb}->delete_document($docid); - if (my $del_fh = $self->{del_fh}) { # TODO - print $del_fh $docid, "\n" or die "E: print $!"; + my $doc = _get_doc($self, $docid) or return; + my ($end, @rm, @add); + my @x = @VMD_MAP; + while (my ($field, $pfx) = splice(@x, 0, 2)) { + my $set = $vmd->{$field} // next; + my %keep = map { $_ => 1 } @$set; + my %add = %keep; + $end //= $doc->termlist_end; + for (my $cur = $doc->termlist_begin; $cur != $end; $cur++) { + $cur->skip_to($pfx); + last if $cur == $end; + my $v = $cur->get_termname; + $v =~ s/\A$pfx//s or next; + $keep{$v} ? delete($add{$v}) : push(@rm, $pfx.$v); } - } else { - if (!grep(/\A\Q$eidx_key\E:/, keys %x3)) { - $doc->remove_term('O'.$eidx_key); + push(@add, map { $pfx.$_ } keys %add); + } + return unless scalar(@rm) || scalar(@add); + $doc->remove_term($_) for @rm; + $doc->add_boolean_term($_) for @add; + $self->{xdb}->replace_document($docid, $doc); +} + +sub apply_vmd_mod ($$) { + my ($doc, $vmd_mod) = @_; + my $updated = 0; + my @x = @VMD_MAP; + while (my ($field, $pfx) = splice(@x, 0, 2)) { + # field: "L" or "kw" + for my $val (@{$vmd_mod->{"-$field"} // []}) { + eval { + $doc->remove_term($pfx . $val); + ++$updated; + }; } - for my $l ($eml->header_raw('List-Id')) { - $l =~ /<([^>]+)>/ or next; - my $lid = lc $1; - $doc->remove_term('G' . $lid); - - # nb: we don't remove the XL probabilistic terms - # since terms may overlap if cross-posted. - # - # IOW, a message which has both - # and would have overlapping - # "XLexample" and "XLcom" as terms and which we - # wouldn't know if they're safe to remove if we just - # unindex while preserving - # . - # - # In any case, this entire sub is will likely never - # be needed and users using the "l:" prefix are probably - # rarer. + for my $val (@{$vmd_mod->{"+$field"} // []}) { + $doc->add_boolean_term($pfx . $val); + ++$updated; } - $self->{xdb}->replace_document($docid, $doc); } + $updated; } -sub get_val ($$) { - my ($doc, $col) = @_; - sortable_unserialise($doc->get_value($col)); +sub add_vmd { + my ($self, $docid, $vmd) = @_; + begin_txn_lazy($self); + my $doc = _get_doc($self, $docid) or return; + my @x = @VMD_MAP; + my $updated = 0; + while (my ($field, $pfx) = splice(@x, 0, 2)) { + my $add = $vmd->{$field} // next; + $doc->add_boolean_term($pfx . $_) for @$add; + $updated += scalar(@$add); + } + $updated += apply_vmd_mod($doc, $vmd); + $self->{xdb}->replace_document($docid, $doc) if $updated; } -sub smsg_from_doc ($) { - my ($doc) = @_; - my $data = $doc->get_data or return; - my $smsg = bless {}, 'PublicInbox::Smsg'; - $smsg->{ts} = get_val($doc, PublicInbox::Search::TS()); - my $dt = get_val($doc, PublicInbox::Search::DT()); - my ($yyyy, $mon, $dd, $hh, $mm, $ss) = unpack('A4A2A2A2A2A2', $dt); - $smsg->{ds} = timegm($ss, $mm, $hh, $dd, $mon - 1, $yyyy); - $smsg->load_from_data($data); - $smsg; +sub remove_vmd { + my ($self, $docid, $vmd) = @_; + begin_txn_lazy($self); + my $doc = _get_doc($self, $docid) or return; + my $replace; + my @x = @VMD_MAP; + while (my ($field, $pfx) = splice(@x, 0, 2)) { + my $rm = $vmd->{$field} // next; + for (@$rm) { + eval { + $doc->remove_term($pfx . $_); + $replace = 1; + }; + } + } + $self->{xdb}->replace_document($docid, $doc) if $replace; +} + +sub update_vmd { + my ($self, $docid, $vmd_mod) = @_; + begin_txn_lazy($self); + my $doc = _get_doc($self, $docid) or return; + my $updated = apply_vmd_mod($doc, $vmd_mod); + $self->{xdb}->replace_document($docid, $doc) if $updated; + $updated; } sub xdb_remove { - my ($self, $oid, @removed) = @_; - my $xdb = $self->{xdb} or return; - for my $num (@removed) { - my $doc = _get_doc($self, $num, $oid) or next; - my $smsg = smsg_from_doc($doc); - my $blob = $smsg->{blob}; # may be undef if --skip-docdata - if (!defined($blob) || $blob eq $oid) { - $xdb->delete_document($num); - } else { - warn "E: #$num $oid != $blob in Xapian\n"; - } + my ($self, @docids) = @_; + begin_txn_lazy($self); + my $xdb = $self->{xdb} // die 'BUG: missing {xdb}'; + for my $docid (@docids) { + eval { $xdb->delete_document($docid) }; + warn "E: #$docid not in in Xapian? $@\n" if $@; } } -sub remove_by_oid { - my ($self, $oid, $num) = @_; - die "BUG: remove_by_oid is v2-only\n" if $self->{oidx}; - $self->begin_txn_lazy; - xdb_remove($self, $oid, $num) if need_xapian($self); +sub xdb_remove_quiet { + my ($self, $docid) = @_; + begin_txn_lazy($self); + my $xdb = $self->{xdb} // die 'BUG: missing {xdb}'; + eval { $xdb->delete_document($docid) }; + ++$self->{-quiet_rm} unless $@; } +sub nr_quiet_rm { delete($_[0]->{-quiet_rm}) // 0 } + sub index_git_blob_id { my ($doc, $pfx, $objid) = @_; @@ -570,8 +683,8 @@ sub unindex_eml { $tmp{$_}++ for @removed; } if (!$nr) { - $mids = join('> <', @$mids); - warn "W: <$mids> missing for removal from overview\n"; + my $m = join('> <', @$mids); + warn "W: <$m> missing for removal from overview\n"; } while (my ($num, $nr) = each %tmp) { warn "BUG: $num appears >1 times ($nr) for $oid\n" if $nr != 1; @@ -581,7 +694,7 @@ sub unindex_eml { } else { # just in case msgmap and over.sqlite3 become desynched: $self->{mm}->mid_delete($mids->[0]); } - xdb_remove($self, $oid, keys %tmp) if need_xapian($self); + xdb_remove($self, keys %tmp) if need_xapian($self); } sub index_mm { @@ -601,45 +714,64 @@ sub index_mm { } } -# returns the number of bytes to add if given a non-CRLF arg -sub crlf_adjust ($) { - if (index($_[0], "\r\n") < 0) { - # common case is LF-only, every \n needs an \r; - # so favor a cheap tr// over an expensive m//g - $_[0] =~ tr/\n/\n/; - } else { # count number of '\n' w/o '\r', expensive: - scalar(my @n = ($_[0] =~ m/(?cat_async callback my ($bref, $oid, $type, $size, $sync) = @_; + return if is_bad_blob($oid, $type, $size, $sync->{oid}); my ($nr, $max) = @$sync{qw(nr max)}; ++$$nr; $$max -= $size; - $size += crlf_adjust($$bref); - my $smsg = bless { bytes => $size, blob => $oid }, 'PublicInbox::Smsg'; + my $smsg = bless { blob => $oid }, 'PublicInbox::Smsg'; + $smsg->set_bytes($$bref, $size); my $self = $sync->{sidx}; + local $self->{current_info} = "$self->{current_info}: $oid"; my $eml = PublicInbox::Eml->new($bref); $smsg->{num} = index_mm($self, $eml, $oid, $sync) or die "E: could not generate NNTP article number for $oid"; add_message($self, $eml, $smsg, $sync); + ++$self->{nidx}; + my $cur_cmt = $sync->{cur_cmt} // die 'BUG: {cur_cmt} missing'; + ${$sync->{latest_cmt}} = $cur_cmt; } sub unindex_both { # git->cat_async callback - my ($bref, $oid, $type, $size, $self) = @_; + my ($bref, $oid, $type, $size, $sync) = @_; + return if is_bad_blob($oid, $type, $size, $sync->{oid}); + my $self = $sync->{sidx}; + local $self->{current_info} = "$self->{current_info}: $oid"; unindex_eml($self, $oid, PublicInbox::Eml->new($bref)); + # may be undef if leftover + if (defined(my $cur_cmt = $sync->{cur_cmt})) { + ${$sync->{latest_cmt}} = $cur_cmt; + } + ++$self->{nidx}; +} + +sub with_umask { + my $self = shift; + ($self->{ibx} // $self->{eidx})->with_umask(@_); } # called by public-inbox-index sub index_sync { my ($self, $opt) = @_; delete $self->{lock_path} if $opt->{-skip_lock}; - $self->{ibx}->with_umask(\&_index_sync, $self, $opt); - if ($opt->{reindex}) { + $self->with_umask(\&_index_sync, $self, $opt); + if ($opt->{reindex} && !$opt->{quit} && + !grep(defined, @$opt{qw(since until)})) { my %again = %$opt; delete @again{qw(rethread reindex)}; index_sync($self, \%again); + $opt->{quit} = $again{quit}; # propagate to caller } } @@ -657,43 +789,43 @@ sub v1_checkpoint ($$;$) { my ($self, $sync, $stk) = @_; $self->{ibx}->git->async_wait_all; - # latest_cmt may be undef - my $newest = $stk ? $stk->{latest_cmt} : undef; - if ($newest) { - my $cur = $self->{mm}->last_commit || ''; - if (need_update($self, $cur, $newest)) { + # $newest may be undef + my $newest = $stk ? $stk->{latest_cmt} : ${$sync->{latest_cmt}}; + if (defined($newest)) { + my $cur = $self->{mm}->last_commit; + if (need_update($self, $sync, $cur, $newest)) { $self->{mm}->last_commit($newest); } - } else { - ${$sync->{max}} = $self->{batch_bytes}; } + ${$sync->{max}} = $self->{batch_bytes}; $self->{mm}->{dbh}->commit; - if ($newest && need_xapian($self)) { - my $xdb = $self->{xdb}; + eval { $self->{mm}->{dbh}->do('PRAGMA optimize') }; + my $xdb = $self->{xdb}; + if ($newest && $xdb) { my $cur = $xdb->get_metadata('last_commit'); - if (need_update($self, $cur, $newest)) { + if (need_update($self, $sync, $cur, $newest)) { $xdb->set_metadata('last_commit', $newest); } - + } + if ($stk) { # all done if $stk is passed # let SearchView know a full --reindex was done so it can # generate ->has_threadid-dependent links - if ($sync->{reindex} && !ref($sync->{reindex})) { + if ($xdb && $sync->{reindex} && !ref($sync->{reindex})) { my $n = $xdb->get_metadata('has_threadid'); $xdb->set_metadata('has_threadid', '1') if $n ne '1'; } + $self->{oidx}->rethread_done($sync->{-opt}); # all done } - - $self->{oidx}->rethread_done($sync->{-opt}) if $newest; # all done commit_txn_lazy($self); - $self->{ibx}->git->cleanup; + $sync->{ibx}->git->cleanup; my $nr = ${$sync->{nr}}; idx_release($self, $nr); # let another process do some work... if (my $pr = $sync->{-opt}->{-progress}) { $pr->("indexed $nr/$sync->{ntodo}\n") if $nr; } - if (!$stk) { # more to come + if (!$stk && !$sync->{quit}) { # more to come begin_txn_lazy($self); $self->{mm}->{dbh}->begin_work; } @@ -702,27 +834,32 @@ sub v1_checkpoint ($$;$) { # only for v1 sub process_stack { my ($self, $sync, $stk) = @_; - my $git = $self->{ibx}->git; + my $git = $sync->{ibx}->git; my $max = $self->{batch_bytes}; my $nr = 0; $sync->{nr} = \$nr; $sync->{max} = \$max; $sync->{sidx} = $self; + $sync->{latest_cmt} = \(my $latest_cmt); $self->{mm}->{dbh}->begin_work; if (my @leftovers = keys %{delete($sync->{D}) // {}}) { warn('W: unindexing '.scalar(@leftovers)." leftovers\n"); for my $oid (@leftovers) { + last if $sync->{quit}; $oid = unpack('H*', $oid); - $git->cat_async($oid, \&unindex_both, $self); + $git->cat_async($oid, \&unindex_both, $sync); } } if ($sync->{max_size} = $sync->{-opt}->{max_size}) { $sync->{index_oid} = \&index_both; } - while (my ($f, $at, $ct, $oid) = $stk->pop_rec) { + while (my ($f, $at, $ct, $oid, $cur_cmt) = $stk->pop_rec) { + my $arg = { %$sync, cur_cmt => $cur_cmt, oid => $oid }; + last if $sync->{quit}; if ($f eq 'm') { - my $arg = { %$sync, autime => $at, cotime => $ct }; + $arg->{autime} = $at; + $arg->{cotime} = $ct; if ($sync->{max_size}) { $git->check_async($oid, \&check_size, $arg); } else { @@ -730,17 +867,17 @@ sub process_stack { } v1_checkpoint($self, $sync) if $max <= 0; } elsif ($f eq 'd') { - $git->cat_async($oid, \&unindex_both, $self); + $git->cat_async($oid, \&unindex_both, $arg); } } - v1_checkpoint($self, $sync, $stk); + v1_checkpoint($self, $sync, $sync->{quit} ? undef : $stk); } -sub log2stack ($$$$) { - my ($sync, $git, $range, $ibx) = @_; +sub log2stack ($$$) { + my ($sync, $git, $range) = @_; my $D = $sync->{D}; # OID_BIN => NR (if reindexing, undef otherwise) my ($add, $del); - if ($ibx->version == 1) { + if ($sync->{ibx}->version == 1) { my $path = $hex.'{2}/'.$hex.'{38}'; $add = qr!\A:000000 100644 \S+ ($OID) A\t$path$!; $del = qr!\A:100644 000000 ($OID) \S+ D\t$path$!; @@ -751,20 +888,26 @@ sub log2stack ($$$$) { # Count the new files so they can be added newest to oldest # and still have numbers increasing from oldest to newest - my $fh = $git->popen(qw(log --raw -r --pretty=tformat:%at-%ct-%H - --no-notes --no-color --no-renames --no-abbrev), - $range); - my ($at, $ct, $stk); + my @cmd = qw(log --raw -r --pretty=tformat:%at-%ct-%H + --no-notes --no-color --no-renames --no-abbrev); + for my $k (qw(since until)) { + my $v = $sync->{-opt}->{$k} // next; + next if !$sync->{-opt}->{reindex}; + push @cmd, "--$k=$v"; + } + my $fh = $git->popen(@cmd, $range); + my ($at, $ct, $stk, $cmt); while (<$fh>) { + return if $sync->{quit}; if (/\A([0-9]+)-([0-9]+)-($OID)$/o) { - ($at, $ct) = ($1 + 0, $2 + 0); - $stk //= PublicInbox::IdxStack->new($3); + ($at, $ct, $cmt) = ($1 + 0, $2 + 0, $3); + $stk //= PublicInbox::IdxStack->new($cmt); } elsif (/$del/) { my $oid = $1; if ($D) { # reindex case $D->{pack('H*', $oid)}++; } else { # non-reindex case: - $stk->push_rec('d', $at, $ct, $oid); + $stk->push_rec('d', $at, $ct, $oid, $cmt); } } elsif (/$add/) { my $oid = $1; @@ -772,12 +915,10 @@ sub log2stack ($$$$) { my $oid_bin = pack('H*', $oid); my $nr = --$D->{$oid_bin}; delete($D->{$oid_bin}) if $nr <= 0; - # nr < 0 (-1) means it never existed - $stk->push_rec('m', $at, $ct, $oid) if $nr < 0; - } else { - $stk->push_rec('m', $at, $ct, $oid); + next if $nr >= 0; } + $stk->push_rec('m', $at, $ct, $oid, $cmt); } } close $fh or die "git log failed: \$?=$?"; @@ -785,9 +926,9 @@ sub log2stack ($$$$) { $stk->read_prepare; } -sub prepare_stack ($$$) { - my ($self, $sync, $range) = @_; - my $git = $self->{ibx}->git; +sub prepare_stack ($$) { + my ($sync, $range) = @_; + my $git = $sync->{ibx}->git; if (index($range, '..') < 0) { # don't show annoying git errors to users who run -index @@ -796,7 +937,7 @@ sub prepare_stack ($$$) { return PublicInbox::IdxStack->new->read_prepare if $?; } $sync->{D} = $sync->{reindex} ? {} : undef; # OID_BIN => NR - log2stack($sync, $git, $range, $self->{ibx}); + log2stack($sync, $git, $range); } # --is-ancestor requires git 1.8.0+ @@ -810,10 +951,17 @@ sub is_ancestor ($$$) { $? == 0; } -sub need_update ($$$) { - my ($self, $cur, $new) = @_; +sub need_update ($$$$) { + my ($self, $sync, $cur, $new) = @_; my $git = $self->{ibx}->git; - return 1 if $cur && !is_ancestor($git, $cur, $new); + $cur //= ''; # XS Search::Xapian ->get_metadata doesn't give undef + + # don't rewind if --{since,until,before,after} are in use + return if $cur ne '' && + grep(defined, @{$sync->{-opt}}{qw(since until)}) && + is_ancestor($git, $new, $cur); + + return 1 if $cur ne '' && !is_ancestor($git, $cur, $new); my $range = $cur eq '' ? $new : "$cur..$new"; chomp(my $n = $git->qx(qw(rev-list --count), $range)); ($n eq '' || $n > 0); @@ -844,15 +992,30 @@ sub reindex_from ($$) { ref($reindex) eq 'HASH' ? $reindex->{from} : ''; } +sub quit_cb ($) { + my ($sync) = @_; + sub { + # we set {-opt}->{quit} too, so ->index_sync callers + # can abort multi-inbox loops this way + $sync->{quit} = $sync->{-opt}->{quit} = 1; + warn "gracefully quitting\n"; + } +} + # indexes all unindexed messages (v1 only) sub _index_sync { my ($self, $opt) = @_; my $tip = $opt->{ref} || 'HEAD'; - my $git = $self->{ibx}->git; + my $ibx = $self->{ibx}; + local $self->{current_info} = "$ibx->{inboxdir}"; $self->{batch_bytes} = $opt->{batch_size} // $BATCH_BYTES; - $git->batch_prepare; + $ibx->git->batch_prepare; my $pr = $opt->{-progress}; - my $sync = { reindex => $opt->{reindex}, -opt => $opt }; + my $sync = { reindex => $opt->{reindex}, -opt => $opt, ibx => $ibx }; + my $quit = quit_cb($sync); + local $SIG{QUIT} = $quit; + local $SIG{INT} = $quit; + local $SIG{TERM} = $quit; my $xdb = $self->begin_txn_lazy; $self->{oidx}->rethread_prepare($opt); my $mm = _msgmap_init($self); @@ -870,10 +1033,10 @@ sub _index_sync { my $lx = reindex_from($sync->{reindex}, $last_commit); my $range = $lx eq '' ? $tip : "$lx..$tip"; $pr->("counting changes\n\t$range ... ") if $pr; - my $stk = prepare_stack($self, $sync, $range); + my $stk = prepare_stack($sync, $range); $sync->{ntodo} = $stk ? $stk->num_records : 0; $pr->("$sync->{ntodo}\n") if $pr; # continue previous line - process_stack($self, $sync, $stk); + process_stack($self, $sync, $stk) if !$sync->{quit}; } sub DESTROY { @@ -893,7 +1056,7 @@ sub _begin_txn { sub begin_txn_lazy { my ($self) = @_; - $self->{ibx}->with_umask(\&_begin_txn, $self) if !$self->{txn}; + $self->with_umask(\&_begin_txn, $self) if !$self->{txn}; } # store 'indexlevel=medium' in v2 shard=0 and v1 (only one shard) @@ -921,6 +1084,10 @@ sub set_metadata_once { sub _commit_txn { my ($self) = @_; + if (my $eidx = $self->{eidx}) { + $eidx->git->async_wait_all; + $eidx->{transact_bytes} = 0; + } if (my $xdb = $self->{xdb}) { set_metadata_once($self); $xdb->commit_transaction; @@ -931,15 +1098,21 @@ sub _commit_txn { sub commit_txn_lazy { my ($self) = @_; delete($self->{txn}) and - $self->{ibx}->with_umask(\&_commit_txn, $self); + $self->with_umask(\&_commit_txn, $self); } -sub worker_done { - my ($self) = @_; - if (need_xapian($self)) { - die "$$ $0 xdb not released\n" if $self->{xdb}; - } - die "$$ $0 still in transaction\n" if $self->{txn}; +sub eidx_shard_new { + my ($class, $eidx, $shard) = @_; + my $self = bless { + eidx => $eidx, + xpfx => $eidx->{xpfx}, + indexlevel => $eidx->{indexlevel}, + -skip_docdata => 1, + shard => $shard, + creat => 1, + }, $class; + $self->{-set_indexlevel_once} = 1 if $self->{indexlevel} eq 'medium'; + $self; } 1;