X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FOverIdx.pm;h=d6d706f7fed04363d409eb7aa34564f08ad7dc2e;hb=ed0167d2a851b4f5128f57ad60309a0b76e62cfa;hp=07cca4e5bb30d9c2053d9bb8ab62788571f2289b;hpb=50ac81092ba1034f3055ddabb3d7cc7853edfa41;p=public-inbox.git diff --git a/lib/PublicInbox/OverIdx.pm b/lib/PublicInbox/OverIdx.pm index 07cca4e5..d6d706f7 100644 --- a/lib/PublicInbox/OverIdx.pm +++ b/lib/PublicInbox/OverIdx.pm @@ -1,4 +1,4 @@ -# Copyright (C) 2018-2020 all contributors +# Copyright (C) 2018-2021 all contributors # License: AGPL-3.0+ # for XOVER, OVER in NNTP, and feeds/homepage/threads in PSGI @@ -79,6 +79,11 @@ SELECT $id_col FROM $tbl WHERE $val_col = ? LIMIT 1 } } +sub ibx_id { + my ($self, $eidx_key) = @_; + id_for($self, 'inboxes', 'ibx_id', eidx_key => $eidx_key); +} + sub sid { my ($self, $path) = @_; return unless defined $path && $path ne ''; @@ -153,7 +158,8 @@ SELECT $cols FROM over WHERE over.num = ? LIMIT 1 foreach (@$nums) { $sth->execute($_->[0]); - my $smsg = $sth->fetchrow_hashref; + # $cb may delete rows and invalidate nums + my $smsg = $sth->fetchrow_hashref // next; $smsg = PublicInbox::Over::load_from_row($smsg); $cb->($self, $smsg, @arg) or return; } @@ -170,8 +176,14 @@ sub _resolve_mid_to_tid { $$tid = $cur_tid; } else { # rethreading, queue up dead ghosts $$tid = next_tid($self); - my $num = $smsg->{num}; - push(@{$self->{-ghosts_to_delete}}, $num) if $num < 0; + my $n = $smsg->{num}; + if ($n > 0) { + $self->{dbh}->prepare_cached(<<'')->execute($$tid, $n); +UPDATE over SET tid = ? WHERE num = ? + + } elsif ($n < 0) { + push(@{$self->{-ghosts_to_delete}}, $n); + } } 1; } @@ -232,32 +244,13 @@ sub link_refs { $tid; } -sub parse_references ($$$) { - my ($smsg, $hdr, $mids) = @_; - my $refs = references($hdr); - push(@$refs, @$mids) if scalar(@$mids) > 1; - return $refs if scalar(@$refs) == 0; - - # prevent circular references here: - my %seen = ( $smsg->{mid} => 1 ); - my @keep; - foreach my $ref (@$refs) { - if (length($ref) > PublicInbox::MID::MAX_MID_SIZE) { - warn "References: <$ref> too long, ignoring\n"; - next; - } - push(@keep, $ref) unless $seen{$ref}++; - } - $smsg->{references} = '<'.join('> <', @keep).'>' if @keep; - \@keep; -} - -# normalize subjects so they are suitable as pathnames for URLs -# XXX: consider for removal +# normalize subjects somewhat, they used to be ASCII-only but now +# we use \w for UTF-8 support. We may still drop it entirely and +# rely on Xapian for subject matches... sub subject_path ($) { my ($subj) = @_; $subj = subject_normalized($subj); - $subj =~ s![^a-zA-Z0-9_\.~/\-]+!_!g; + $subj =~ s![^\w\.~/\-]+!_!g; lc($subj); } @@ -272,7 +265,11 @@ sub add_overview { my ($self, $eml, $smsg) = @_; $smsg->{lines} = $eml->body_raw =~ tr!\n!\n!; my $mids = mids_for_index($eml); - my $refs = parse_references($smsg, $eml, $mids); + my $refs = $smsg->parse_references($eml, $mids); + $mids->[0] //= do { + $smsg->{mid} //= ''; + $eml->{-lei_fake_mid}; + }; my $subj = $smsg->{subject}; my $xpath; if ($subj ne '') { @@ -298,7 +295,7 @@ sub _add_over { } } elsif ($n < 0) { # ghost $$old_tid //= $cur_valid ? $cur_tid : next_tid($self); - link_refs($self, $refs, $$old_tid); + $$old_tid = link_refs($self, $refs, $$old_tid); delete_by_num($self, $n); $$v++; } @@ -383,13 +380,12 @@ sub create_tables { $dbh->do(<<''); CREATE TABLE IF NOT EXISTS over ( - num INTEGER NOT NULL, /* NNTP article number == IMAP UID */ + num INTEGER PRIMARY KEY NOT NULL, /* NNTP article number == IMAP UID */ tid INTEGER NOT NULL, /* THREADID (IMAP REFERENCES threading, JMAP) */ sid INTEGER, /* Subject ID (IMAP ORDEREDSUBJECT "threading") */ ts INTEGER, /* IMAP INTERNALDATE (Received: header, git commit time) */ ds INTEGER, /* RFC-2822 sent Date: header, git author time */ - ddd VARBINARY, /* doc-data-deflated (->to_doc_data, ->load_from_data) */ - UNIQUE (num) + ddd VARBINARY /* doc-data-deflated (->to_doc_data, ->load_from_data) */ ) $dbh->do('CREATE INDEX IF NOT EXISTS idx_tid ON over (tid)'); @@ -463,10 +459,14 @@ sub dbh_close { sub create { my ($self) = @_; - unless (-r $self->{filename}) { + my $fn = $self->{filename} // do { + croak('BUG: no {filename}') unless $self->{dbh}; + return; + }; + unless (-r $fn) { require File::Path; - require File::Basename; - File::Path::mkpath(File::Basename::dirname($self->{filename})); + my ($dir) = ($fn =~ m!(.*?/)[^/]+\z!); + File::Path::mkpath($dir); } # create the DB: PublicInbox::Over::dbh($self); @@ -521,7 +521,7 @@ sub eidx_prep ($) { my ($self) = @_; $self->{-eidx_prep} //= do { my $dbh = $self->dbh; - $dbh->do(<<""); + $dbh->do(<<''); INSERT OR IGNORE INTO counter (key) VALUES ('eidx_docid') $dbh->do(<<''); @@ -543,9 +543,13 @@ CREATE TABLE IF NOT EXISTS xref3 ( $dbh->do('CREATE INDEX IF NOT EXISTS idx_docid ON xref3 (docid)'); # performance critical, this is not UNIQUE since we may need to - # tolerate some old bugs from indexing mirrors - $dbh->do('CREATE INDEX IF NOT EXISTS idx_nntp ON '. - 'xref3 (oidbin,xnum,ibx_id)'); + # tolerate some old bugs from indexing mirrors. n.b. we used + # to index oidbin here, but leaving it out speeds up reindexing + # and "XHDR Xref <$MSGID>" isn't any slower w/o oidbin + $dbh->do('CREATE INDEX IF NOT EXISTS idx_reindex ON '. + 'xref3 (xnum,ibx_id)'); + + $dbh->do('CREATE INDEX IF NOT EXISTS idx_oidbin ON xref3 (oidbin)'); $dbh->do(<<''); CREATE TABLE IF NOT EXISTS eidx_meta ( @@ -553,7 +557,14 @@ CREATE TABLE IF NOT EXISTS eidx_meta ( val VARCHAR(255) NOT NULL ) - $dbh; + # A queue of current docids which need reindexing. + # eidxq persists across aborted -extindex invocations + # Currently used for "-extindex --reindex" for Xapian + # data, but may be used in more places down the line. + $dbh->do(<<''); +CREATE TABLE IF NOT EXISTS eidxq (docid INTEGER PRIMARY KEY NOT NULL) + + 1; }; } @@ -583,7 +594,7 @@ sub eidx_max { sub add_xref3 { my ($self, $docid, $xnum, $oidhex, $eidx_key) = @_; begin_lazy($self); - my $ibx_id = id_for($self, 'inboxes', 'ibx_id', eidx_key => $eidx_key); + my $ibx_id = ibx_id($self, $eidx_key); my $oidbin = pack('H*', $oidhex); my $sth = $self->{dbh}->prepare_cached(<<''); INSERT OR IGNORE INTO xref3 (docid, ibx_id, xnum, oidbin) VALUES (?, ?, ?, ?) @@ -595,51 +606,6 @@ INSERT OR IGNORE INTO xref3 (docid, ibx_id, xnum, oidbin) VALUES (?, ?, ?, ?) $sth->execute; } -# returns remaining reference count to $docid -sub remove_xref3 { - my ($self, $docid, $oidhex, $eidx_key, $rm_eidx_info) = @_; - begin_lazy($self); - my $oidbin = pack('H*', $oidhex); - my ($sth, $ibx_id); - if (defined $eidx_key) { - $ibx_id = id_for($self, 'inboxes', 'ibx_id', - eidx_key => $eidx_key); - $sth = $self->{dbh}->prepare_cached(<<''); -DELETE FROM xref3 WHERE docid = ? AND ibx_id = ? AND oidbin = ? - - $sth->bind_param(1, $docid); - $sth->bind_param(2, $ibx_id); - $sth->bind_param(3, $oidbin, SQL_BLOB); - } else { - $sth = $self->{dbh}->prepare_cached(<<''); -DELETE FROM xref3 WHERE docid = ? AND oidbin = ? - - $sth->bind_param(1, $docid); - $sth->bind_param(2, $oidbin, SQL_BLOB); - } - $sth->execute; - $sth = $self->{dbh}->prepare_cached(<<'', undef, 1); -SELECT COUNT(*) FROM xref3 WHERE docid = ? - - $sth->execute($docid); - my $nr = $sth->fetchrow_array; - if ($nr == 0) { - delete_by_num($self, $docid); - } elsif (defined($ibx_id) && $rm_eidx_info) { - # if deduplication rules in ContentHash change, it's - # possible a docid can have multiple rows with the - # same ibx_id. This governs whether or not we call - # ->shard_remove_eidx_info in ExtSearchIdx. - $sth = $self->{dbh}->prepare_cached(<<'', undef, 1); -SELECT COUNT(*) FROM xref3 WHERE docid = ? AND ibx_id = ? - - $sth->execute($docid, $ibx_id); - my $count = $sth->fetchrow_array; - $$rm_eidx_info = ($count == 0); - } - $nr; -} - # for when an xref3 goes missing, this does NOT update {ts} sub update_blob { my ($self, $smsg, $oidhex) = @_; @@ -652,4 +618,56 @@ UPDATE over SET ddd = ? WHERE num = ? $sth->execute; } +sub merge_xref3 { # used for "-extindex --dedupe" + my ($self, $keep_docid, $drop_docid, $oidhex) = @_; + my $oidbin = pack('H*', $oidhex); + my $sth = $self->{dbh}->prepare_cached(<<''); +UPDATE OR IGNORE xref3 SET docid = ? WHERE docid = ? AND oidbin = ? + + $sth->bind_param(1, $keep_docid); + $sth->bind_param(2, $drop_docid); + $sth->bind_param(3, $oidbin, SQL_BLOB); + $sth->execute; + + # drop anything that conflicted + $sth = $self->{dbh}->prepare_cached(<<''); +DELETE FROM xref3 WHERE docid = ? AND oidbin = ? + + $sth->bind_param(1, $drop_docid); + $sth->bind_param(2, $oidbin, SQL_BLOB); + $sth->execute; +} + +sub eidxq_add { + my ($self, $docid) = @_; + $self->dbh->prepare_cached(<<'')->execute($docid); +INSERT OR IGNORE INTO eidxq (docid) VALUES (?) + +} + +sub eidxq_del { + my ($self, $docid) = @_; + $self->dbh->prepare_cached(<<'')->execute($docid); +DELETE FROM eidxq WHERE docid = ? + +} + +# returns true if we're vivifying a message for lei/store that was +# previously external-metadata only +sub vivify_xvmd { + my ($self, $smsg) = @_; + my @docids = $self->blob_exists($smsg->{blob}); + my @vivify_xvmd; + for my $id (@docids) { + if (my $cur = $self->get_art($id)) { + # already indexed if bytes > 0 + return if $cur->{bytes} > 0; + push @vivify_xvmd, $id; + } else { + warn "W: $smsg->{blob} #$id gone (bug?)\n"; + } + } + $smsg->{-vivify_xvmd} = \@vivify_xvmd; +} + 1;