X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FLeiXSearch.pm;h=e15383915283715987b97d118066f64517b52f7c;hb=8ab43c1c27c725a8ef9307f5dba3e565169d48ca;hp=eb015978708024cfa1884ba9a98cf67727a6e20f;hpb=232f8e376fe2856c8d34092e33a2c0750734c285;p=public-inbox.git diff --git a/lib/PublicInbox/LeiXSearch.pm b/lib/PublicInbox/LeiXSearch.pm index eb015978..e1538391 100644 --- a/lib/PublicInbox/LeiXSearch.pm +++ b/lib/PublicInbox/LeiXSearch.pm @@ -18,6 +18,7 @@ use PublicInbox::MID qw(mids); use PublicInbox::Smsg; use PublicInbox::Eml; use Fcntl qw(SEEK_SET F_SETFL O_APPEND O_RDWR); +use PublicInbox::ContentHash qw(git_sha); sub new { my ($class) = @_; @@ -63,14 +64,18 @@ sub locals { @{$_[0]->{locals} // []} } sub remotes { @{$_[0]->{remotes} // []} } -# called by PublicInbox::Search::xdb +# called by PublicInbox::Search::xdb (usually via ->mset) sub xdb_shards_flat { @{$_[0]->{shards_flat} // []} } sub mitem_kw ($$;$) { my ($smsg, $mitem, $flagged) = @_; - my $kw = xap_terms('K', $mitem->get_document); + my $kw = xap_terms('K', my $doc = $mitem->get_document); $kw->{flagged} = 1 if $flagged; + # we keep the empty {kw} array here to prevent expensive work in + # ->xsmsg_vmd, _unbless_smsg will clobber it iff it's empty $smsg->{kw} = [ sort keys %$kw ]; + my $L = xap_terms('L', $doc); + $smsg->{L} = [ sort keys %$L ] if scalar(keys %$L); } # like over->get_art @@ -83,8 +88,10 @@ sub smsg_for { my $num = int(($docid - 1) / $nshard) + 1; my $ibx = $self->{shard2ibx}->[$shard]; my $smsg = $ibx->over->get_art($num); - mitem_kw($smsg, $mitem) if $ibx->can('msg_keywords'); - $smsg->{docid} = $docid; + return if $smsg->{bytes} == 0; # external message + if ($ibx->can('msg_keywords')) { + mitem_kw($smsg, $mitem); + } $smsg; } @@ -92,7 +99,7 @@ sub recent { my ($self, $qstr, $opt) = @_; $opt //= {}; $opt->{relevance} //= -2; - $self->mset($qstr //= 'bytes:1..', $opt); + $self->mset($qstr //= 'z:1..', $opt); } sub over {} @@ -104,7 +111,7 @@ sub _mset_more ($$) { } # $startq will EOF when do_augment is done augmenting and allow -# query_mset and query_thread_mset to proceed. +# query_combined_mset and query_thread_mset to proceed. sub wait_startq ($) { my ($lei) = @_; my $startq = delete $lei->{startq} or return; @@ -137,75 +144,115 @@ sub mset_progress { } } -sub query_thread_mset { # for --threads +sub query_one_mset { # for --threads and l2m w/o sort my ($self, $ibxish) = @_; - local $0 = "$0 query_thread_mset"; + local $0 = "$0 query_one_mset"; my $lei = $self->{lei}; my ($srch, $over) = ($ibxish->search, $ibxish->over); - my $desc = $ibxish->{inboxdir} // $ibxish->{topdir}; - return warn("$desc not indexed by Xapian\n") unless ($srch && $over); - my $mo = { %{$lei->{mset_opt}} }; + my $dir = $ibxish->{inboxdir} // $ibxish->{topdir}; + return warn("$dir not indexed by Xapian\n") unless ($srch && $over); + my $mo = { %{$lei->{mset_opt}} }; # copy my $mset; - my $each_smsg = $lei->{ovv}->ovv_each_smsg_cb($lei, $ibxish); + my $each_smsg = $lei->{ovv}->ovv_each_smsg_cb($lei); my $can_kw = !!$ibxish->can('msg_keywords'); - my $fl = $lei->{opt}->{threads} > 1; + my $threads = $lei->{opt}->{threads} // 0; + my $fl = $threads > 1 ? 1 : undef; + my $lss = $lei->{dedupe}; + $lss = undef unless $lss && $lss->can('cfg_set'); # saved search + my $maxk = "external.$dir.maxuid"; + my $stop_at = $lss ? $lss->{-cfg}->{$maxk} : undef; + if (defined $stop_at) { + die "$maxk=$stop_at has multiple values" if ref $stop_at; + my @e; + local $SIG{__WARN__} = sub { push @e, @_ }; + $stop_at += 0; + return warn("$maxk=$stop_at: @e") if @e; + } + my $first_ids; do { $mset = $srch->mset($mo->{qstr}, $mo); - mset_progress($lei, $desc, $mset->size, + mset_progress($lei, $dir, $mset->size, $mset->get_matches_estimated); + wait_startq($lei); # wait for keyword updates my $ids = $srch->mset_to_artnums($mset, $mo); - my $ctx = { ids => $ids }; + @$ids = grep { $_ > $stop_at } @$ids if defined($stop_at); my $i = 0; - my %n2item = map { ($ids->[$i++], $_) } $mset->items; - while ($over->expand_thread($ctx)) { - for my $n (@{$ctx->{xids}}) { - my $smsg = $over->get_art($n) or next; - wait_startq($lei); - my $mitem = delete $n2item{$smsg->{num}}; - if ($mitem) { - if ($can_kw) { + if ($threads) { + # copy $ids if $lss since over->expand_thread + # shifts @{$ctx->{ids}} + $first_ids = [ @$ids ] if $lss; + my $ctx = { ids => $ids }; + my %n2item = map { ($ids->[$i++], $_) } $mset->items; + while ($over->expand_thread($ctx)) { + for my $n (@{$ctx->{xids}}) { + my $smsg = $over->get_art($n) or next; + my $mitem = delete $n2item{$n}; + next if $smsg->{bytes} == 0; + if ($mitem && $can_kw) { mitem_kw($smsg, $mitem, $fl); - } else { - $smsg->{kw} = [ 'flagged' ]; + } elsif ($mitem && $fl) { + # call ->xsmsg_vmd, later + $smsg->{lei_q_tt_flagged} = 1; } + $each_smsg->($smsg, $mitem); } + @{$ctx->{xids}} = (); + } + } else { + $first_ids = $ids; + my @items = $mset->items; + for my $n (@$ids) { + my $mitem = $items[$i++]; + my $smsg = $over->get_art($n) or next; + next if $smsg->{bytes} == 0; + mitem_kw($smsg, $mitem, $fl) if $can_kw; $each_smsg->($smsg, $mitem); } - @{$ctx->{xids}} = (); } } while (_mset_more($mset, $mo)); - undef $each_smsg; # drops @io for l2m->{each_smsg_done} + if ($lss && scalar(@$first_ids)) { + undef $stop_at; + my $max = $first_ids->[0]; + $lss->cfg_set($maxk, $max); + undef $lss; + } + undef $each_smsg; # may commit $lei->{ovv}->ovv_atexit_child($lei); } -sub query_mset { # non-parallel for non-"--threads" users +sub query_combined_mset { # non-parallel for non-"--threads" users my ($self) = @_; - local $0 = "$0 query_mset"; + local $0 = "$0 query_combined_mset"; my $lei = $self->{lei}; my $mo = { %{$lei->{mset_opt}} }; my $mset; for my $loc (locals($self)) { attach_external($self, $loc); } - my $each_smsg = $lei->{ovv}->ovv_each_smsg_cb($lei, $self); + my $each_smsg = $lei->{ovv}->ovv_each_smsg_cb($lei); do { $mset = $self->mset($mo->{qstr}, $mo); mset_progress($lei, 'xsearch', $mset->size, $mset->size, $mset->get_matches_estimated); + wait_startq($lei); # wait for keyword updates for my $mitem ($mset->items) { my $smsg = smsg_for($self, $mitem) or next; - wait_startq($lei); $each_smsg->($smsg, $mitem); } } while (_mset_more($mset, $mo)); - undef $each_smsg; # drops @io for l2m->{each_smsg_done} + undef $each_smsg; # may commit $lei->{ovv}->ovv_atexit_child($lei); } sub each_remote_eml { # callback for MboxReader->mboxrd my ($eml, $self, $lei, $each_smsg) = @_; - $lei->{sto}->ipc_do('set_eml', $eml) if $lei->{sto}; # --import-remote + my $xoids = $lei->{ale}->xoids_for($eml, 1); + if ($self->{import_sto} && !$xoids) { + $self->{import_sto}->ipc_do('add_eml', $eml); + } my $smsg = bless {}, 'PublicInbox::Smsg'; + $smsg->{blob} = $xoids ? (keys(%$xoids))[0] + : git_sha(1, $eml)->hexdigest; $smsg->populate($eml); $smsg->parse_references($eml, mids($eml)); $smsg->{$_} //= '' for qw(from to cc ds subject references mid); @@ -218,7 +265,7 @@ sub each_remote_eml { # callback for MboxReader->mboxrd if ($now > $next) { $lei->{-next_progress} = $now + 1; my $nr = $lei->{-nr_remote_eml}; - $lei->err("# $lei->{-current_url} $nr/?"); + mset_progress($lei, $lei->{-current_url}, $nr, '?'); } } $each_smsg->($smsg, undef, $eml); @@ -249,6 +296,7 @@ sub query_remote_mboxrd { my $curl = PublicInbox::LeiCurl->new($lei, $self->{curl}) or return; push @$curl, '-s', '-d', ''; my $each_smsg = $lei->{ovv}->ovv_each_smsg_cb($lei); + $self->{import_sto} = $lei->{sto} if $lei->{opt}->{'import-remote'}; for my $uri (@$uris) { $lei->{-current_url} = $uri->as_string; $lei->{-nr_remote_eml} = 0; @@ -257,7 +305,7 @@ sub query_remote_mboxrd { $lei->qerr("# $cmd"); my ($fh, $pid) = popen_rd($cmd, undef, $rdr); $reap_curl = PublicInbox::OnDestroy->new($sigint_reap, $pid); - $fh = IO::Uncompress::Gunzip->new($fh); + $fh = IO::Uncompress::Gunzip->new($fh, MultiStream => 1); PublicInbox::MboxReader->mboxrd($fh, \&each_remote_eml, $self, $lei, $each_smsg); my $err = waitpid($pid, 0) == $pid ? undef @@ -289,27 +337,7 @@ sub query_remote_mboxrd { $lei->{ovv}->ovv_atexit_child($lei); } -# called by LeiOverview::each_smsg_cb -sub git { $_[0]->{git_tmp} // die 'BUG: caller did not set {git_tmp}' } - -sub git_tmp ($) { - my ($self) = @_; - my (%seen, @dirs); - my $tmp = File::Temp->newdir("lei_xsearch_git.$$-XXXX", TMPDIR => 1); - for my $ibxish (locals($self)) { - my $d = File::Spec->canonpath($ibxish->git->{git_dir}); - $seen{$d} //= push @dirs, "$d/objects\n" - } - my $git_dir = $tmp->dirname; - PublicInbox::Import::init_bare($git_dir); - my $f = "$git_dir/objects/info/alternates"; - open my $alt, '>', $f or die "open($f): $!"; - print $alt @dirs or die "print $f: $!"; - close $alt or die "close $f: $!"; - my $git = PublicInbox::Git->new($git_dir); - $git->{-tmp} = $tmp; - $git; -} +sub git { $_[0]->{git} // die 'BUG: git uninitialized' } sub xsearch_done_wait { # dwaitpid callback my ($arg, $pid) = @_; @@ -324,6 +352,7 @@ sub query_done { # EOF callback for main daemon if (my $lxs = delete $lei->{lxs}) { $lxs->wq_wait_old(\&xsearch_done_wait, $lei); } + my $wait = $lei->{sto} ? $lei->{sto}->ipc_do('done') : undef; $lei->{ovv}->ovv_end($lei); if ($l2m) { # close() calls LeiToMail reap_compress if (my $out = delete $lei->{old_1}) { @@ -338,6 +367,7 @@ Error closing $lei->{ovv}->{dst}: $! $l2m->poke_dst; $lei->poke_mua; } else { # mbox users + delete $l2m->{mbl}; # drop dotlock $lei->start_mua; } } @@ -348,7 +378,7 @@ Error closing $lei->{ovv}->{dst}: $! sub do_post_augment { my ($lei) = @_; - my $l2m = $lei->{l2m} or die 'BUG: unexpected do_post_augment'; + my $l2m = $lei->{l2m} or return; # client disconnected my $err; eval { $l2m->post_augment($lei) }; $err = $@; @@ -367,7 +397,7 @@ sub do_post_augment { sub incr_post_augment { # called whenever an l2m shard finishes augment my ($lei) = @_; - my $l2m = $lei->{l2m} or die 'BUG: unexpected incr_post_augment'; + my $l2m = $lei->{l2m} or return; # client disconnected return if ++$lei->{nr_post_augment} != $l2m->{-wq_nr_workers}; do_post_augment($lei); } @@ -382,14 +412,14 @@ sub concurrency { $nl + $nr; } -sub start_query { # always runs in main (lei-daemon) process - my ($self) = @_; - if ($self->{threads}) { +sub start_query ($;$) { # always runs in main (lei-daemon) process + my ($self, $l2m) = @_; + if ($self->{opt_threads} || ($l2m && !$self->{opt_sort})) { for my $ibxish (locals($self)) { - $self->wq_io_do('query_thread_mset', [], $ibxish); + $self->wq_io_do('query_one_mset', [], $ibxish); } } elsif (locals($self)) { - $self->wq_io_do('query_mset', []); + $self->wq_io_do('query_combined_mset', []); } my $i = 0; my $q = []; @@ -405,12 +435,12 @@ sub start_query { # always runs in main (lei-daemon) process sub incr_start_query { # called whenever an l2m shard starts do_post_auth my ($self, $l2m) = @_; return if ++$self->{nr_start_query} != $l2m->{-wq_nr_workers}; - start_query($self); + start_query($self, $l2m); } sub ipc_atfork_child { my ($self) = @_; - $self->{lei}->lei_atfork_child; + $self->{lei}->_lei_atfork_child; $SIG{__WARN__} = PublicInbox::Eml::warn_ignore_cb(); $self->SUPER::ipc_atfork_child; } @@ -430,10 +460,11 @@ sub do_query { 'incr_start_query' => [ \&incr_start_query, $self, $l2m ], }; $lei->{auth}->op_merge($ops, $l2m) if $l2m && $lei->{auth}; - ($lei->{pkt_op_c}, $lei->{pkt_op_p}) = PublicInbox::PktOp->pair($ops); + my $end = $lei->pkt_op_pair; $lei->{1}->autoflush(1); $lei->start_pager if delete $lei->{need_pager}; $lei->{ovv}->ovv_begin($lei); + die 'BUG: xdb|over open' if $lei->{lse}->{xdb} || $lei->{lse}->{over}; if ($l2m) { $l2m->pre_augment($lei); if ($lei->{opt}->{augment} && delete $lei->{early_mua}) { @@ -445,25 +476,20 @@ sub do_query { # 1031: F_SETPIPE_SZ fcntl($lei->{startq}, 1031, 4096) if $^O eq 'linux'; } - if (!$lei->{opt}->{threads} && locals($self)) { # for query_mset - # lei->{git_tmp} is set for wq_wait_old so we don't - # delete until all lei2mail + lei_xsearch workers are reaped - $lei->{git_tmp} = $self->{git_tmp} = git_tmp($self); - } $self->wq_workers_start('lei_xsearch', undef, $lei->oldset, { lei => $lei }); - my $op = delete $lei->{pkt_op_c}; + my $op_c = delete $lei->{pkt_op_c}; delete $lei->{pkt_op_p}; - $self->{threads} = $lei->{opt}->{threads}; + @$end = (); + $self->{opt_threads} = $lei->{opt}->{threads}; + $self->{opt_sort} = $lei->{opt}->{'sort'}; if ($l2m) { $l2m->net_merge_complete unless $lei->{auth}; } else { start_query($self); } $lei->event_step_init; # wait for shutdowns - if ($lei->{oneshot}) { - while ($op->{sock}) { $op->event_step } - } + $op_c->op_wait_event($ops); } sub add_uri { @@ -487,8 +513,10 @@ sub prepare_external { return add_uri($self, URI->new($loc)); } elsif (-f "$loc/ei.lock") { require PublicInbox::ExtSearch; + die "`\\n' not allowed in `$loc'\n" if index($loc, "\n") >= 0; $loc = PublicInbox::ExtSearch->new($loc); } elsif (-f "$loc/inbox.lock" || -d "$loc/public-inbox") { + die "`\\n' not allowed in `$loc'\n" if index($loc, "\n") >= 0; require PublicInbox::Inbox; # v2, v1 $loc = bless { inboxdir => $loc }, 'PublicInbox::Inbox'; } else {