X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FSearchIdx.pm;h=2e0b9a43e24750fce0e083c37120a47f34d8da8a;hb=7503aeb540af5afd5cb1b554b3c29f35f5fc918d;hp=d63dd7c742b4c21ef9420d810954df3e8332f45e;hpb=3fc59df0d633a17e0c5e43d633d12e8772c06ec3;p=public-inbox.git diff --git a/lib/PublicInbox/SearchIdx.pm b/lib/PublicInbox/SearchIdx.pm index d63dd7c7..2e0b9a43 100644 --- a/lib/PublicInbox/SearchIdx.pm +++ b/lib/PublicInbox/SearchIdx.pm @@ -1,5 +1,5 @@ -# Copyright (C) 2015 all contributors -# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt) +# Copyright (C) 2015-2018 all contributors +# License: AGPL-3.0+ # based on notmuch, but with no concept of folders, files or flags # # Indexes mail with Xapian and our (SQLite-based) ::Msgmap for use @@ -9,48 +9,79 @@ package PublicInbox::SearchIdx; use strict; use warnings; -use Fcntl qw(:flock :DEFAULT); +use base qw(PublicInbox::Search PublicInbox::Lock); use PublicInbox::MIME; -use Email::MIME::ContentType; -$Email::MIME::ContentType::STRICT_PARAMS = 0; -use base qw(PublicInbox::Search); -use PublicInbox::MID qw/mid_clean id_compress mid_mime/; +use PublicInbox::InboxWritable; +use PublicInbox::MID qw/mid_clean id_compress mid_mime mids references/; use PublicInbox::MsgIter; use Carp qw(croak); use POSIX qw(strftime); require PublicInbox::Git; -*xpfx = *PublicInbox::Search::xpfx; -use constant MAX_MID_SIZE => 244; # max term size - 1 in Xapian use constant { - PERM_UMASK => 0, - OLD_PERM_GROUP => 1, - OLD_PERM_EVERYBODY => 2, - PERM_GROUP => 0660, - PERM_EVERYBODY => 0664, + BATCH_BYTES => 1_000_000, + DEBUG => !!$ENV{DEBUG}, }; +my %GIT_ESC = ( + a => "\a", + b => "\b", + f => "\f", + n => "\n", + r => "\r", + t => "\t", + v => "\013", +); + +sub git_unquote ($) { + my ($s) = @_; + return $s unless ($s =~ /\A"(.*)"\z/); + $s = $1; + $s =~ s/\\([abfnrtv])/$GIT_ESC{$1}/g; + $s =~ s/\\([0-7]{1,3})/chr(oct($1))/ge; + $s; +} + sub new { - my ($class, $inbox, $creat) = @_; - my $git_dir = $inbox; - my $altid; - if (ref $inbox) { - $git_dir = $inbox->{mainrepo}; - $altid = $inbox->{altid}; + my ($class, $ibx, $creat, $part) = @_; + my $mainrepo = $ibx; # for "public-inbox-index" w/o entry in config + my $git_dir = $mainrepo; + my ($altid, $git); + my $version = 1; + if (ref $ibx) { + $mainrepo = $ibx->{mainrepo}; + $altid = $ibx->{altid}; + $version = $ibx->{version} || 1; if ($altid) { require PublicInbox::AltId; $altid = [ map { - PublicInbox::AltId->new($inbox, $_); + PublicInbox::AltId->new($ibx, $_); } @$altid ]; } + } else { # v1 + $ibx = { mainrepo => $git_dir, version => 1 }; } + $ibx = PublicInbox::InboxWritable->new($ibx); require Search::Xapian::WritableDatabase; - my $self = bless { git_dir => $git_dir, -altid => $altid }, $class; - my $perm = $self->_git_config_perm; - my $umask = _umask_for($perm); - $self->{umask} = $umask; - $self->{lock_path} = "$git_dir/ssoma.lock"; - $self->{git} = PublicInbox::Git->new($git_dir); + my $self = bless { + mainrepo => $mainrepo, + -inbox => $ibx, + git => $ibx->git, + -altid => $altid, + version => $version, + }, $class; + $ibx->umask_prepare; + if ($version == 1) { + $self->{lock_path} = "$mainrepo/ssoma.lock"; + } elsif ($version == 2) { + defined $part or die "partition is required for v2\n"; + # partition is a number or "all" + $self->{partition} = $part; + $self->{lock_path} = undef; + $self->{msgmap_path} = "$mainrepo/msgmap.sqlite3"; + } else { + die "unsupported inbox version=$version\n"; + } $self->{creat} = ($creat || 0) == 1; $self; } @@ -59,65 +90,36 @@ sub _xdb_release { my ($self) = @_; my $xdb = delete $self->{xdb} or croak 'not acquired'; $xdb->close; - _lock_release($self) if $self->{creat}; + $self->lock_release if $self->{creat}; undef; } sub _xdb_acquire { my ($self) = @_; croak 'already acquired' if $self->{xdb}; - my $dir = PublicInbox::Search->xdir($self->{git_dir}); + my $dir = $self->xdir; my $flag = Search::Xapian::DB_OPEN; if ($self->{creat}) { require File::Path; - _lock_acquire($self); + $self->lock_acquire; File::Path::mkpath($dir); - $self->{batch_size} = 100; $flag = Search::Xapian::DB_CREATE_OR_OPEN; } $self->{xdb} = Search::Xapian::WritableDatabase->new($dir, $flag); } -# we only acquire the flock if creating or reindexing; -# PublicInbox::Import already has the lock on its own. -sub _lock_acquire { - my ($self) = @_; - croak 'already locked' if $self->{lockfh}; - sysopen(my $lockfh, $self->{lock_path}, O_WRONLY|O_CREAT) or - die "failed to open lock $self->{lock_path}: $!\n"; - flock($lockfh, LOCK_EX) or die "lock failed: $!\n"; - $self->{lockfh} = $lockfh; -} - -sub _lock_release { - my ($self) = @_; - my $lockfh = delete $self->{lockfh} or croak 'not locked'; - flock($lockfh, LOCK_UN) or die "unlock failed: $!\n"; - close $lockfh or die "close failed: $!\n"; -} - sub add_val ($$$) { my ($doc, $col, $num) = @_; $num = Search::Xapian::sortable_serialise($num); $doc->add_value($col, $num); } -sub add_values ($$$) { - my ($smsg, $bytes, $num) = @_; - - my $ts = $smsg->ts; - my $doc = $smsg->{doc}; - add_val($doc, &PublicInbox::Search::TS, $ts); - - defined($num) and add_val($doc, &PublicInbox::Search::NUM, $num); - - defined($bytes) and add_val($doc, &PublicInbox::Search::BYTES, $bytes); - - add_val($doc, &PublicInbox::Search::LINES, - $smsg->{mime}->body_raw =~ tr!\n!\n!); - - my $yyyymmdd = strftime('%Y%m%d', gmtime($ts)); +sub add_values { + my ($doc, $ts, $ds, $num) = @_; + add_val($doc, PublicInbox::Search::TS, $ts); + my $yyyymmdd = strftime('%Y%m%d', gmtime($ds)); add_val($doc, PublicInbox::Search::YYYYMMDD, $yyyymmdd); + defined($num) and add_val($doc, PublicInbox::Search::NUM, $num); } sub index_users ($$) { @@ -135,40 +137,155 @@ sub index_users ($$) { $tg->increase_termpos; } +sub index_diff_inc ($$$$) { + my ($tg, $text, $pfx, $xnq) = @_; + if (@$xnq) { + $tg->index_text(join("\n", @$xnq), 1, 'XNQ'); + $tg->increase_termpos; + @$xnq = (); + } + $tg->index_text($text, 1, $pfx); + $tg->increase_termpos; +} + +sub index_old_diff_fn { + my ($tg, $seen, $fa, $fb, $xnq) = @_; + + # no renames or space support for traditional diffs, + # find the number of leading common paths to strip: + my @fa = split('/', $fa); + my @fb = split('/', $fb); + while (scalar(@fa) && scalar(@fb)) { + $fa = join('/', @fa); + $fb = join('/', @fb); + if ($fa eq $fb) { + unless ($seen->{$fa}++) { + index_diff_inc($tg, $fa, 'XDFN', $xnq); + } + return 1; + } + shift @fa; + shift @fb; + } + 0; +} + +sub index_diff ($$$) { + my ($tg, $lines, $doc) = @_; + my %seen; + my $in_diff; + my @xnq; + my $xnq = \@xnq; + foreach (@$lines) { + if ($in_diff && s/^ //) { # diff context + index_diff_inc($tg, $_, 'XDFCTX', $xnq); + } elsif (/^-- $/) { # email signature begins + $in_diff = undef; + } elsif (m!^diff --git ("?a/.+) ("?b/.+)\z!) { + my ($fa, $fb) = ($1, $2); + my $fn = (split('/', git_unquote($fa), 2))[1]; + $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq); + $fn = (split('/', git_unquote($fb), 2))[1]; + $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq); + $in_diff = 1; + # traditional diff: + } elsif (m/^diff -(.+) (\S+) (\S+)$/) { + my ($opt, $fa, $fb) = ($1, $2, $3); + push @xnq, $_; + # only support unified: + next unless $opt =~ /[uU]/; + $in_diff = index_old_diff_fn($tg, \%seen, $fa, $fb, + $xnq); + } elsif (m!^--- ("?a/.+)!) { + my $fn = (split('/', git_unquote($1), 2))[1]; + $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq); + $in_diff = 1; + } elsif (m!^\+\+\+ ("?b/.+)!) { + my $fn = (split('/', git_unquote($1), 2))[1]; + $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq); + $in_diff = 1; + } elsif (/^--- (\S+)/) { + $in_diff = $1; + push @xnq, $_; + } elsif (defined $in_diff && /^\+\+\+ (\S+)/) { + $in_diff = index_old_diff_fn($tg, \%seen, $in_diff, $1, + $xnq); + } elsif ($in_diff && s/^\+//) { # diff added + index_diff_inc($tg, $_, 'XDFB', $xnq); + } elsif ($in_diff && s/^-//) { # diff removed + index_diff_inc($tg, $_, 'XDFA', $xnq); + } elsif (m!^index ([a-f0-9]+)\.\.([a-f0-9]+)!) { + my ($ba, $bb) = ($1, $2); + index_git_blob_id($doc, 'XDFPRE', $ba); + index_git_blob_id($doc, 'XDFPOST', $bb); + $in_diff = 1; + } elsif (/^@@ (?:\S+) (?:\S+) @@\s*$/) { + # traditional diff w/o -p + } elsif (/^@@ (?:\S+) (?:\S+) @@\s*(\S+.*)$/) { + # hunk header context + index_diff_inc($tg, $1, 'XDFHH', $xnq); + # ignore the following lines: + } elsif (/^(?:dis)similarity index/ || + /^(?:old|new) mode/ || + /^(?:deleted|new) file mode/ || + /^(?:copy|rename) (?:from|to) / || + /^(?:dis)?similarity index / || + /^\\ No newline at end of file/ || + /^Binary files .* differ/) { + push @xnq, $_; + } elsif ($_ eq '') { + $in_diff = undef; + } else { + push @xnq, $_; + warn "non-diff line: $_\n" if DEBUG && $_ ne ''; + $in_diff = undef; + } + } + + $tg->index_text(join("\n", @xnq), 1, 'XNQ'); + $tg->increase_termpos; +} + sub index_body ($$$) { - my ($tg, $lines, $inc) = @_; - $tg->index_text(join("\n", @$lines), $inc, $inc ? 'XNQ' : 'XQUOT'); - @$lines = (); + my ($tg, $lines, $doc) = @_; + my $txt = join("\n", @$lines); + if ($doc) { + # does it look like a diff? + if ($txt =~ /^(?:diff|---|\+\+\+) /ms) { + $txt = undef; + index_diff($tg, $lines, $doc); + } else { + $tg->index_text($txt, 1, 'XNQ'); + } + } else { + $tg->index_text($txt, 0, 'XQUOT'); + } $tg->increase_termpos; + @$lines = (); } sub add_message { - my ($self, $mime, $bytes, $num, $blob) = @_; # mime = Email::MIME object - my $db = $self->{xdb}; - - my ($doc_id, $old_tid); - my $mid = mid_clean(mid_mime($mime)); + # mime = Email::MIME object + my ($self, $mime, $bytes, $num, $oid, $mid0) = @_; + my $doc_id; + my $mids = mids($mime->header_obj); + my $skel = $self->{skeleton}; eval { - die 'Message-ID too long' if length($mid) > MAX_MID_SIZE; - my $smsg = $self->lookup_message($mid); - if ($smsg) { - # convert a ghost to a regular message - # it will also clobber any existing regular message - $doc_id = $smsg->{doc_id}; - $old_tid = $smsg->thread_id; - } - $smsg = PublicInbox::SearchMsg->new($mime); + my $smsg = PublicInbox::SearchMsg->new($mime); my $doc = $smsg->{doc}; - $doc->add_term(xpfx('mid') . $mid); - my $subj = $smsg->subject; + my $xpath; if ($subj ne '') { - my $path = $self->subject_path($subj); - $doc->add_term(xpfx('path') . id_compress($path)); + $xpath = $self->subject_path($subj); + $xpath = id_compress($xpath); } - add_values($smsg, $bytes, $num); + my $lines = $mime->body_raw =~ tr!\n!\n!; + $smsg->{lines} = $mime->body_raw =~ tr!\n!\n!; + defined $bytes or $bytes = length($mime->as_string); + $smsg->{bytes} = $bytes; + add_values($doc, $smsg->ts, $smsg->ds, $num); my $tg = $self->term_generator; @@ -206,7 +323,7 @@ sub add_message { my @lines = split(/\n/, $body); while (defined(my $l = shift @lines)) { if ($l =~ /^>/) { - index_body($tg, \@orig, 1) if @orig; + index_body($tg, \@orig, $doc) if @orig; push @quot, $l; } else { index_body($tg, \@quot, 0) if @quot; @@ -214,51 +331,123 @@ sub add_message { } } index_body($tg, \@quot, 0) if @quot; - index_body($tg, \@orig, 1) if @orig; + index_body($tg, \@orig, $doc) if @orig; }); - link_message($self, $smsg, $old_tid); - $tg->index_text($mid, 1, 'XMID'); - $doc->set_data($smsg->to_doc_data($blob)); - + # populates smsg->references for smsg->to_doc_data + my $refs = parse_references($smsg); + $mid0 = $mids->[0] unless defined $mid0; # v1 compatibility + my $data = $smsg->to_doc_data($oid, $mid0); + foreach my $mid (@$mids) { + $tg->index_text($mid, 1, 'XM'); + } + $doc->set_data($data); if (my $altid = $self->{-altid}) { foreach my $alt (@$altid) { - my $id = $alt->mid2alt($mid); - next unless defined $id; - $doc->add_term($alt->{xprefix} . $id); + my $pfx = $alt->{xprefix}; + foreach my $mid (@$mids) { + my $id = $alt->mid2alt($mid); + next unless defined $id; + $doc->add_boolean_term($pfx . $id); + } } } - if (defined $doc_id) { - $db->replace_document($doc_id, $doc); + + $self->delete_article($num) if defined $num; # for reindexing + if ($skel) { + my @vals = ($smsg->ts, $num, $mids, $xpath, $data); + $skel->index_skeleton(\@vals); + $doc->add_boolean_term('Q' . $_) foreach @$mids; + $doc->add_boolean_term('XNUM' . $num) if defined $num; + $doc_id = $self->{xdb}->add_document($doc); } else { - $doc_id = $db->add_document($doc); + $doc_id = link_and_save($self, $doc, $mids, $refs, + $num, $xpath); } }; if ($@) { - warn "failed to index message <$mid>: $@\n"; + warn "failed to index message <".join('> <',@$mids).">: $@\n"; return undef; } $doc_id; } -# returns deleted doc_id on success, undef on missing +# returns begin and end PostingIterator +sub find_doc_ids { + my ($self, $termval) = @_; + my $db = $self->{xdb}; + + ($db->postlist_begin($termval), $db->postlist_end($termval)); +} + +sub batch_do { + my ($self, $termval, $cb) = @_; + my $batch_size = 1000; # don't let @ids grow too large to avoid OOM + while (1) { + my ($head, $tail) = $self->find_doc_ids($termval); + return if $head == $tail; + my @ids; + for (; $head != $tail && @ids < $batch_size; $head->inc) { + push @ids, $head->get_docid; + } + $cb->(\@ids); + } +} + sub remove_message { my ($self, $mid) = @_; my $db = $self->{xdb}; - my $doc_id; + my $called; $mid = mid_clean($mid); eval { - $doc_id = $self->find_unique_doc_id('mid', $mid); - $db->delete_document($doc_id) if defined $doc_id; + batch_do($self, 'Q' . $mid, sub { + my ($ids) = @_; + $db->delete_document($_) for @$ids; + $called = 1; + }); }; - if ($@) { warn "failed to remove message <$mid>: $@\n"; - return undef; + } elsif (!$called) { + warn "cannot remove non-existent <$mid>\n"; } - $doc_id; +} + +sub delete_article { + my ($self, $num) = @_; + my $ndel = 0; + batch_do($self, 'XNUM' . $num, sub { + my ($ids) = @_; + $ndel += scalar @$ids; + $self->{xdb}->delete_document($_) for @$ids; + }); +} + +# MID is a hint in V2 +sub remove_by_oid { + my ($self, $oid, $mid) = @_; + my $db = $self->{xdb}; + + # XXX careful, we cannot use batch_do here since we conditionally + # delete documents based on other factors, so we cannot call + # find_doc_ids twice. + my ($head, $tail) = $self->find_doc_ids('Q' . $mid); + return if $head == $tail; + + # there is only ONE element in @delete unless we + # have bugs in our v2writable deduplication check + my @delete; + for (; $head != $tail; $head->inc) { + my $docid = $head->get_docid; + my $doc = $db->get_document($docid); + my $smsg = PublicInbox::SearchMsg->wrap($doc, $mid); + $smsg->load_expand; + push(@delete, $docid) if $smsg->{blob} eq $oid; + } + $db->delete_document($_) foreach @delete; + scalar(@delete); } sub term_generator { # write-only @@ -274,7 +463,7 @@ sub term_generator { # write-only } # increments last_thread_id counter -# returns a 64-bit integer represented as a hex string +# returns a 64-bit integer represented as a decimal string sub next_thread_id { my ($self) = @_; my $db = $self->{xdb}; @@ -285,73 +474,101 @@ sub next_thread_id { $last_thread_id; } -sub link_message { - my ($self, $smsg, $old_tid) = @_; - my $doc = $smsg->{doc}; - my $mid = $smsg->mid; +sub parse_references ($) { + my ($smsg) = @_; my $mime = $smsg->{mime}; my $hdr = $mime->header_obj; - my $refs = $hdr->header_raw('References'); - my @refs = $refs ? ($refs =~ /<([^>]+)>/g) : (); - my $irt = $hdr->header_raw('In-Reply-To'); - if (defined $irt) { - $irt = mid_clean($irt); - $irt = undef if $mid eq $irt; - } - - my $tid; - if (@refs) { - my %uniq = ($mid => 1); - my @orig_refs = @refs; - @refs = (); - - if (defined $irt) { - # to check MAX_MID_SIZE - push @orig_refs, $irt; - - # below, we will ensure IRT (if specified) - # is the last References - $uniq{$irt} = 1; - } - - # prevent circular references via References: here: - foreach my $ref (@orig_refs) { - if (length($ref) > MAX_MID_SIZE) { - warn "References: <$ref> too long, ignoring\n"; - } - next if $uniq{$ref}; - $uniq{$ref} = 1; - push @refs, $ref; + my $refs = references($hdr); + return $refs if scalar(@$refs) == 0; + + # prevent circular references via References here: + my %mids = map { $_ => 1 } @{mids($hdr)}; + my @keep; + foreach my $ref (@$refs) { + if (length($ref) > PublicInbox::MID::MAX_MID_SIZE) { + warn "References: <$ref> too long, ignoring\n"; + next; } + next if $mids{$ref}; + push @keep, $ref; } + $smsg->{references} = '<'.join('> <', @keep).'>' if @keep; + \@keep; +} - # last References should be IRT, but some mail clients do things - # out of order, so trust IRT over References iff IRT exists - push @refs, $irt if defined $irt; - - if (@refs) { - $smsg->{references} = '<'.join('> <', @refs).'>'; +sub link_doc { + my ($self, $doc, $refs, $old_tid) = @_; + my $tid; + if (@$refs) { # first ref *should* be the thread root, # but we can never trust clients to do the right thing - my $ref = shift @refs; - $tid = $self->_resolve_mid_to_tid($ref); - $self->merge_threads($tid, $old_tid) if defined $old_tid; + my $ref = shift @$refs; + $tid = resolve_mid_to_tid($self, $ref); + merge_threads($self, $tid, $old_tid) if defined $old_tid; # the rest of the refs should point to this tid: - foreach $ref (@refs) { - my $ptid = $self->_resolve_mid_to_tid($ref); + foreach $ref (@$refs) { + my $ptid = resolve_mid_to_tid($self, $ref); merge_threads($self, $tid, $ptid); } } else { - $tid = $self->next_thread_id; + $tid = defined $old_tid ? $old_tid : $self->next_thread_id; } - $doc->add_term(xpfx('thread') . $tid); + $doc->add_boolean_term('G' . $tid); + $tid; } -sub index_blob { - my ($self, $mime, $bytes, $num, $blob) = @_; - $self->add_message($mime, $bytes, $num, $blob); +sub link_and_save { + my ($self, $doc, $mids, $refs, $num, $xpath) = @_; + my $db = $self->{xdb}; + my $old_tid; + my $doc_id; + $doc->add_boolean_term('XNUM' . $num) if defined $num; + $doc->add_boolean_term('XPATH' . $xpath) if defined $xpath; + $doc->add_boolean_term('Q' . $_) foreach @$mids; + + $self->{skel} and die "Should not have read-only skel here\n";; + foreach my $mid (@$mids) { + my $vivified = 0; + $self->each_smsg_by_mid($mid, sub { + my ($cur) = @_; + my $type = $cur->type; + my $cur_tid = $cur->thread_id; + $old_tid = $cur_tid unless defined $old_tid; + if ($type eq 'mail') { + # do not break existing mail messages, + # just merge the threads + merge_threads($self, $old_tid, $cur_tid); + return 1; + } + if ($type ne 'ghost') { + die "<$mid> has a bad type: $type\n"; + } + my $tid = link_doc($self, $doc, $refs, $old_tid); + $old_tid = $tid unless defined $old_tid; + $doc_id = $cur->{doc_id}; + $self->{xdb}->replace_document($doc_id, $doc); + ++$vivified; + 1; + }); + $vivified > 1 and warn + "BUG: vivified multiple ($vivified) ghosts for $mid\n"; + } + # not really important, but we return any vivified ghost docid, here: + return $doc_id if defined $doc_id; + link_doc($self, $doc, $refs, $old_tid); + $self->{xdb}->add_document($doc); +} + +sub index_git_blob_id { + my ($doc, $pfx, $objid) = @_; + + my $len = length($objid); + for (my $len = length($objid); $len >= 7; ) { + $doc->add_term($pfx.$objid); + $objid = substr($objid, 0, --$len); + } } sub unindex_blob { @@ -362,7 +579,13 @@ sub unindex_blob { sub index_mm { my ($self, $mime) = @_; - $self->{mm}->mid_insert(mid_clean(mid_mime($mime))); + my $mid = mid_clean(mid_mime($mime)); + my $mm = $self->{mm}; + my $num = $mm->mid_insert($mid); + return $num if defined $num; + + # fallback to num_for since filters like RubyLang set the number + $mm->num_for($mid); } sub unindex_mm { @@ -373,7 +596,7 @@ sub unindex_mm { sub index_mm2 { my ($self, $mime, $bytes, $blob) = @_; my $num = $self->{mm}->num_for(mid_clean(mid_mime($mime))); - index_blob($self, $mime, $bytes, $num, $blob); + add_message($self, $mime, $bytes, $num, $blob); } sub unindex_mm2 { @@ -385,7 +608,7 @@ sub unindex_mm2 { sub index_both { my ($self, $mime, $bytes, $blob) = @_; my $num = index_mm($self, $mime); - index_blob($self, $mime, $bytes, $num, $blob); + add_message($self, $mime, $bytes, $num, $blob); } sub unindex_both { @@ -407,9 +630,19 @@ sub do_cat_mail { sub index_sync { my ($self, $opts) = @_; - with_umask($self, sub { $self->_index_sync($opts) }); + $self->{-inbox}->with_umask(sub { $self->_index_sync($opts) }) +} + +sub batch_adjust ($$$$) { + my ($max, $bytes, $batch_cb, $latest) = @_; + $$max -= $bytes; + if ($$max <= 0) { + $$max = BATCH_BYTES; + $batch_cb->($latest, 1); + } } +# only for v1 sub rlog { my ($self, $log, $add_cb, $del_cb, $batch_cb) = @_; my $hex = '[a-f0-9]'; @@ -419,23 +652,21 @@ sub rlog { my $git = $self->{git}; my $latest; my $bytes; - my $max = $self->{batch_size}; # may be undef + my $max = BATCH_BYTES; local $/ = "\n"; my $line; while (defined($line = <$log>)) { if ($line =~ /$addmsg/o) { my $blob = $1; my $mime = do_cat_mail($git, $blob, \$bytes) or next; + batch_adjust(\$max, $bytes, $batch_cb, $latest); $add_cb->($self, $mime, $bytes, $blob); } elsif ($line =~ /$delmsg/o) { my $blob = $1; - my $mime = do_cat_mail($git, $blob) or next; + my $mime = do_cat_mail($git, $blob, \$bytes) or next; + batch_adjust(\$max, $bytes, $batch_cb, $latest); $del_cb->($self, $mime); } elsif ($line =~ /^commit ($h40)/o) { - if (defined $max && --$max <= 0) { - $max = $self->{batch_size}; - $batch_cb->($latest, 1); - } $latest = $1; } } @@ -444,9 +675,14 @@ sub rlog { sub _msgmap_init { my ($self) = @_; - $self->{mm} = eval { + $self->{mm} ||= eval { require PublicInbox::Msgmap; - PublicInbox::Msgmap->new($self->{git_dir}, 1); + my $msgmap_path = $self->{msgmap_path}; + if (defined $msgmap_path) { # v2 + PublicInbox::Msgmap->new_file($msgmap_path, 1); + } else { + PublicInbox::Msgmap->new($self->{mainrepo}, 1); + } }; } @@ -539,16 +775,27 @@ sub _index_sync { } } else { # user didn't install DBD::SQLite and DBI - rlog($self, $xlog, *index_blob, *unindex_blob, $cb); + rlog($self, $xlog, *add_message, *unindex_blob, $cb); } } # this will create a ghost as necessary -sub _resolve_mid_to_tid { +sub resolve_mid_to_tid { my ($self, $mid) = @_; + my $tid; + $self->each_smsg_by_mid($mid, sub { + my ($smsg) = @_; + my $cur_tid = $smsg->thread_id; + if (defined $tid) { + merge_threads($self, $tid, $cur_tid); + } else { + $tid = $smsg->thread_id; + } + 1; + }); + return $tid if defined $tid; - my $smsg = $self->lookup_message($mid) || $self->create_ghost($mid); - $smsg->thread_id; + $self->create_ghost($mid)->thread_id; } sub create_ghost { @@ -556,9 +803,9 @@ sub create_ghost { my $tid = $self->next_thread_id; my $doc = Search::Xapian::Document->new; - $doc->add_term(xpfx('mid') . $mid); - $doc->add_term(xpfx('thread') . $tid); - $doc->add_term(xpfx('type') . 'ghost'); + $doc->add_boolean_term('Q' . $mid); + $doc->add_boolean_term('G' . $tid); + $doc->add_boolean_term('T' . 'ghost'); my $smsg = PublicInbox::SearchMsg->wrap($doc, $mid); $self->{xdb}->add_document($doc); @@ -569,79 +816,80 @@ sub create_ghost { sub merge_threads { my ($self, $winner_tid, $loser_tid) = @_; return if $winner_tid == $loser_tid; - my ($head, $tail) = $self->find_doc_ids('thread', $loser_tid); - my $thread_pfx = xpfx('thread'); my $db = $self->{xdb}; + batch_do($self, 'G' . $loser_tid, sub { + my ($ids) = @_; + foreach my $docid (@$ids) { + my $doc = $db->get_document($docid); + $doc->remove_term('G' . $loser_tid); + $doc->add_boolean_term('G' . $winner_tid); + $db->replace_document($docid, $doc); + } + }); +} - for (; $head != $tail; $head->inc) { - my $docid = $head->get_docid; - my $doc = $db->get_document($docid); - $doc->remove_term($thread_pfx . $loser_tid); - $doc->add_term($thread_pfx . $winner_tid); - $db->replace_document($docid, $doc); - } +sub DESTROY { + # order matters for unlocking + $_[0]->{xdb} = undef; + $_[0]->{lockfh} = undef; } -sub _read_git_config_perm { +# remote_* subs are only used by SearchIdxPart and SearchIdxSkeleton +sub remote_commit { my ($self) = @_; - my @cmd = qw(config core.sharedRepository); - my $fh = PublicInbox::Git->new($self->{git_dir})->popen(@cmd); - local $/ = "\n"; - my $perm = <$fh>; - chomp $perm if defined $perm; - $perm; -} - -sub _git_config_perm { - my $self = shift; - my $perm = scalar @_ ? $_[0] : _read_git_config_perm($self); - return PERM_GROUP if (!defined($perm) || $perm eq ''); - return PERM_UMASK if ($perm eq 'umask'); - return PERM_GROUP if ($perm eq 'group'); - if ($perm =~ /\A(?:all|world|everybody)\z/) { - return PERM_EVERYBODY; + if (my $w = $self->{w}) { + print $w "commit\n" or die "failed to write commit: $!"; + } else { + $self->commit_txn_lazy; + if (my $skel = $self->{skeleton}) { + $skel->commit_txn_lazy; + } } - return PERM_GROUP if ($perm =~ /\A(?:true|yes|on|1)\z/); - return PERM_UMASK if ($perm =~ /\A(?:false|no|off|0)\z/); - - my $i = oct($perm); - return PERM_UMASK if ($i == PERM_UMASK); - return PERM_GROUP if ($i == OLD_PERM_GROUP); - return PERM_EVERYBODY if ($i == OLD_PERM_EVERYBODY); +} - if (($i & 0600) != 0600) { - die "core.sharedRepository mode invalid: ". - sprintf('%.3o', $i) . "\nOwner must have permissions\n"; +sub remote_close { + my ($self) = @_; + if (my $w = delete $self->{w}) { + my $pid = delete $self->{pid} or die "no process to wait on\n"; + print $w "close\n" or die "failed to write to pid:$pid: $!\n"; + close $w or die "failed to close pipe for pid:$pid: $!\n"; + waitpid($pid, 0) == $pid or die "remote process did not finish"; + $? == 0 or die ref($self)." pid:$pid exited with: $?"; + } else { + die "transaction in progress $self\n" if $self->{txn}; + $self->_xdb_release if $self->{xdb}; } - ($i & 0666); } -sub _umask_for { - my ($perm) = @_; # _git_config_perm return value - my $rv = $perm; - return umask if $rv == 0; +sub remote_remove { + my ($self, $oid, $mid) = @_; + if (my $w = $self->{w}) { + # triggers remove_by_oid in partition or skeleton + print $w "D $oid $mid\n" or die "failed to write remove $!"; + } else { + $self->begin_txn_lazy; + $self->remove_by_oid($oid, $mid); + } +} - # set +x bit if +r or +w were set - $rv |= 0100 if ($rv & 0600); - $rv |= 0010 if ($rv & 0060); - $rv |= 0001 if ($rv & 0006); - (~$rv & 0777); +sub begin_txn_lazy { + my ($self) = @_; + return if $self->{txn}; + my $xdb = $self->{xdb} || $self->_xdb_acquire; + $xdb->begin_transaction; + $self->{txn} = 1; } -sub with_umask { - my ($self, $cb) = @_; - my $old = umask $self->{umask}; - my $rv = eval { $cb->() }; - my $err = $@; - umask $old; - die $err if $@; - $rv; +sub commit_txn_lazy { + my ($self) = @_; + delete $self->{txn} or return; + $self->{xdb}->commit_transaction; } -sub DESTROY { - # order matters for unlocking - $_[0]->{xdb} = undef; - $_[0]->{lockfh} = undef; +sub worker_done { + my ($self) = @_; + die "$$ $0 xdb not released\n" if $self->{xdb}; + die "$$ $0 still in transaction\n" if $self->{txn}; } 1;