X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FMsgmap.pm;h=d28e96c8588b002e62a09b64dc6eaaad51346309;hb=5f6a0d2342323541e44ff2f1e7329053d0263800;hp=2f64d90c71f741cb100c3beacd124e2698260d71;hpb=7ee29d11b25b09b8919b9a125732e4c2c5784d28;p=public-inbox.git diff --git a/lib/PublicInbox/Msgmap.pm b/lib/PublicInbox/Msgmap.pm index 2f64d90c..d28e96c8 100644 --- a/lib/PublicInbox/Msgmap.pm +++ b/lib/PublicInbox/Msgmap.pm @@ -1,12 +1,18 @@ -# Copyright (C) 2015 all contributors -# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt) -# bidirectional Message-ID <-> Article Number mapping +# Copyright (C) 2015-2020 all contributors +# License: AGPL-3.0+ + +# bidirectional Message-ID <-> Article Number mapping for the NNTP +# and web interfaces. This is required for implementing stable article +# numbers for NNTP and allows prefix lookups for partial Message-IDs +# in case URLs get truncated from copy-n-paste errors by users. +# +# This is maintained by ::SearchIdx package PublicInbox::Msgmap; use strict; -use warnings; -use fields qw(dbh mid_insert mid_for num_for num_minmax); use DBI; use DBD::SQLite; +use PublicInbox::Over; +use PublicInbox::Spawn; sub new { my ($class, $git_dir, $writable) = @_; @@ -15,53 +21,60 @@ sub new { my $err = $!; -d $d or die "$d not created: $err"; } - my $f = "$d/msgmap.sqlite3"; - my $dbh = DBI->connect("dbi:SQLite:dbname=$f",'','', { - AutoCommit => 1, - RaiseError => 1, - PrintError => 0, - sqlite_use_immediate_transaction => 1, - }); - $dbh->do('PRAGMA case_sensitive_like = ON'); - my $self = fields::new($class); - $self->{dbh} = $dbh; - - if ($writable) { + new_file($class, "$d/msgmap.sqlite3", $writable); +} + +sub new_file { + my ($class, $f, $rw) = @_; + return if !$rw && !-r $f; + + my $self = bless { filename => $f }, $class; + my $dbh = $self->{dbh} = PublicInbox::Over::dbh_new($self, $rw); + if ($rw) { + $dbh->begin_work; create_tables($dbh); $self->created_at(time) unless $self->created_at; + + my $max = $self->max // 0; + $self->num_highwater($max); + $dbh->commit; } $self; } +# used to keep track of used numeric mappings for v2 reindex +sub tmp_clone { + my ($self, $dir) = @_; + require File::Temp; + my $tmp = "mm_tmp-$$-XXXXXX"; + my ($fh, $fn) = File::Temp::tempfile($tmp, EXLOCK => 0, DIR => $dir); + PublicInbox::Spawn::nodatacow_fd(fileno($fh)); + $self->{dbh}->sqlite_backup_to_file($fn); + $tmp = ref($self)->new_file($fn, 2); + $tmp->{dbh}->do('PRAGMA journal_mode = MEMORY'); + $tmp->{pid} = $$; + $tmp; +} + +# n.b. invoked directly by scripts/xhdr-num2mid sub meta_accessor { my ($self, $key, $value) = @_; - use constant { - meta_select => 'SELECT val FROM meta WHERE key = ? LIMIT 1', - meta_update => 'UPDATE meta SET val = ? WHERE key = ? LIMIT 1', - meta_insert => 'INSERT INTO meta (key,val) VALUES (?,?)', - }; + my $sql = 'SELECT val FROM meta WHERE key = ? LIMIT 1'; my $dbh = $self->{dbh}; my $prev; - defined $value or - return $dbh->selectrow_array(meta_select, undef, $key); + defined $value or return $dbh->selectrow_array($sql, undef, $key); - $dbh->begin_work; - eval { - $prev = $dbh->selectrow_array(meta_select, undef, $key); + $prev = $dbh->selectrow_array($sql, undef, $key); - if (defined $prev) { - $dbh->do(meta_update, undef, $value, $key); - } else { - $dbh->do(meta_insert, undef, $key, $value); - } - $dbh->commit; - }; - my $err = $@; - return $prev unless $err; - - $dbh->rollback; - die $err; + if (defined $prev) { + $sql = 'UPDATE meta SET val = ? WHERE key = ?'; + $dbh->do($sql, undef, $value, $key); + } else { + $sql = 'INSERT INTO meta (key,val) VALUES (?,?)'; + $dbh->do($sql, undef, $key, $value); + } + $prev; } sub last_commit { @@ -69,26 +82,47 @@ sub last_commit { $self->meta_accessor('last_commit', $commit); } +# v2 uses this to keep track of how up-to-date Xapian is +# old versions may be automatically GC'ed away in the future, +# but it's a trivial amount of storage. +sub last_commit_xap { + my ($self, $version, $i, $commit) = @_; + $self->meta_accessor("last_xap$version-$i", $commit); +} + sub created_at { my ($self, $second) = @_; $self->meta_accessor('created_at', $second); } +sub num_highwater { + my ($self, $num) = @_; + my $high = $self->{num_highwater} ||= + $self->meta_accessor('num_highwater'); + if (defined($num) && (!defined($high) || ($num > $high))) { + $self->{num_highwater} = $num; + $self->meta_accessor('num_highwater', $num); + } + $self->{num_highwater}; +} + sub mid_insert { my ($self, $mid) = @_; my $dbh = $self->{dbh}; - use constant MID_INSERT => 'INSERT INTO msgmap (mid) VALUES (?)'; - my $sth = $self->{mid_insert} ||= $dbh->prepare(MID_INSERT); - $sth->bind_param(1, $mid); - $sth->execute; - $dbh->last_insert_id(undef, undef, 'msgmap', 'num'); + my $sth = $dbh->prepare_cached(<<''); +INSERT INTO msgmap (mid) VALUES (?) + + return unless eval { $sth->execute($mid) }; + my $num = $dbh->last_insert_id(undef, undef, 'msgmap', 'num'); + $self->num_highwater($num) if defined($num); + $num; } -use constant MID_FOR => 'SELECT mid FROM msgmap WHERE num = ? LIMIT 1'; sub mid_for { my ($self, $num) = @_; my $dbh = $self->{dbh}; - my $sth = $self->{mid_for} ||= $dbh->prepare(MID_FOR); + my $sth = $self->{mid_for} ||= + $dbh->prepare('SELECT mid FROM msgmap WHERE num = ? LIMIT 1'); $sth->bind_param(1, $num); $sth->execute; $sth->fetchrow_array; @@ -97,75 +131,171 @@ sub mid_for { sub num_for { my ($self, $mid) = @_; my $dbh = $self->{dbh}; - use constant NUM_FOR => 'SELECT num FROM msgmap WHERE mid = ? LIMIT 1'; - my $sth = $self->{num_for} ||= $dbh->prepare(NUM_FOR); + my $sth = $self->{num_for} ||= + $dbh->prepare('SELECT num FROM msgmap WHERE mid = ? LIMIT 1'); $sth->bind_param(1, $mid); $sth->execute; $sth->fetchrow_array; } -sub minmax { - my ($self) = @_; - my $dbh = $self->{dbh}; - use constant NUM_MINMAX => 'SELECT MIN(num),MAX(num) FROM msgmap'; - my $sth = $self->{num_minmax} ||= $dbh->prepare(NUM_MINMAX); +sub max { + my $sth = $_[0]->{dbh}->prepare_cached('SELECT MAX(num) FROM msgmap', + undef, 1); $sth->execute; - $sth->fetchrow_array; + $sth->fetchrow_array; } -sub mid_prefixes { - my ($self, $pfx, $limit) = @_; - - die "No prefix given" unless (defined $pfx && $pfx ne ''); - $pfx =~ s/([%_])/\\$1/g; - $pfx .= '%'; - - $limit ||= 100; - $limit += 0; # force to integer - $limit ||= 100; - - $self->{dbh}->selectcol_arrayref('SELECT mid FROM msgmap ' . - 'WHERE mid LIKE ? ESCAPE ? ' . - "ORDER BY num DESC LIMIT $limit", - undef, $pfx, '\\'); +sub minmax { + # breaking MIN and MAX into separate queries speeds up from 250ms + # to around 700us with 2.7million messages. + my $sth = $_[0]->{dbh}->prepare_cached('SELECT MIN(num) FROM msgmap', + undef, 1); + $sth->execute; + ($sth->fetchrow_array, max($_[0])); } sub mid_delete { my ($self, $mid) = @_; my $dbh = $self->{dbh}; - use constant MID_DELETE => 'DELETE FROM msgmap WHERE mid = ?'; - my $sth = $dbh->prepare(MID_DELETE); + my $sth = $dbh->prepare('DELETE FROM msgmap WHERE mid = ?'); $sth->bind_param(1, $mid); $sth->execute; } +sub num_delete { + my ($self, $num) = @_; + my $dbh = $self->{dbh}; + my $sth = $dbh->prepare('DELETE FROM msgmap WHERE num = ?'); + $sth->bind_param(1, $num); + $sth->execute; +} + sub create_tables { my ($dbh) = @_; - my $e; - $e = eval { $dbh->selectrow_array('EXPLAIN SELECT * FROM msgmap;') }; - defined $e or $dbh->do('CREATE TABLE msgmap (' . - 'num INTEGER PRIMARY KEY AUTOINCREMENT, '. - 'mid VARCHAR(1000) NOT NULL, ' . - 'UNIQUE (mid) )'); + $dbh->do(<<''); +CREATE TABLE IF NOT EXISTS msgmap ( + num INTEGER PRIMARY KEY AUTOINCREMENT, + mid VARCHAR(1000) NOT NULL, + UNIQUE (mid) +) + + $dbh->do(<<''); +CREATE TABLE IF NOT EXISTS meta ( + key VARCHAR(32) PRIMARY KEY, + val VARCHAR(255) NOT NULL +) - $e = eval { $dbh->selectrow_array('EXPLAIN SELECT * FROM meta') }; - defined $e or $dbh->do('CREATE TABLE meta (' . - 'key VARCHAR(32) PRIMARY KEY, '. - 'val VARCHAR(255) NOT NULL)'); } -sub id_batch { - my ($self, $num, $cb) = @_; +# used by NNTP.pm +sub ids_after { + my ($self, $num) = @_; + my $ids = $self->{dbh}->selectcol_arrayref(<<'', undef, $$num); +SELECT num FROM msgmap WHERE num > ? +ORDER BY num ASC LIMIT 1000 + + $$num = $ids->[-1] if @$ids; + $ids; +} + +sub msg_range { + my ($self, $beg, $end, $cols) = @_; + $cols //= 'num,mid'; my $dbh = $self->{dbh}; - my $sth = $dbh->prepare('SELECT num FROM msgmap WHERE num > ? '. - 'ORDER BY num ASC LIMIT 1000'); - $sth->execute($num); - my $ary = $sth->fetchall_arrayref; - @$ary = map { $_->[0] } @$ary; - my $nr = scalar @$ary; - $cb->($ary) if $nr; - $nr; + my $attr = { Columns => [] }; + my $mids = $dbh->selectall_arrayref(<<"", $attr, $$beg, $end); +SELECT $cols FROM msgmap WHERE num >= ? AND num <= ? +ORDER BY num ASC LIMIT 1000 + + $$beg = $mids->[-1]->[0] + 1 if @$mids; + $mids +} + +# only used for mapping external serial numbers (e.g. articles from gmane) +# see scripts/xhdr-num2mid or PublicInbox::Filter::RubyLang for usage +sub mid_set { + my ($self, $num, $mid) = @_; + my $sth = $self->{mid_set} ||= do { + $self->{dbh}->prepare( + 'INSERT OR IGNORE INTO msgmap (num,mid) VALUES (?,?)'); + }; + my $result = $sth->execute($num, $mid); + $self->num_highwater($num) if (defined($result) && $result == 1); + $result; +} + +sub DESTROY { + my ($self) = @_; + my $dbh = $self->{dbh} or return; + if (($self->{pid} // 0) == $$) { + my $f = $dbh->sqlite_db_filename; + unlink $f or warn "failed to unlink $f: $!\n"; + } +} + +sub atfork_parent { + my ($self) = @_; + $self->{pid} or die 'BUG: not a temporary clone'; + $self->{dbh} and die 'BUG: tmp_clone dbh not prepared for parent'; + defined($self->{filename}) or die 'BUG: {filename} not defined'; + $self->{dbh} = PublicInbox::Over::dbh_new($self, 2); + $self->{dbh}->do('PRAGMA journal_mode = MEMORY'); +} + +sub atfork_prepare { + my ($self) = @_; + my $pid = $self->{pid} or die 'BUG: not a temporary clone'; + $pid == $$ or die "BUG: atfork_prepare not called by $pid"; + my $dbh = $self->{dbh} or die 'BUG: temporary clone not open'; + + # must clobber prepared statements + %$self = (filename => $dbh->sqlite_db_filename, pid => $pid); +} + +sub skip_artnum { + my ($self, $skip_artnum) = @_; + return meta_accessor($self, 'skip_artnum') if !defined($skip_artnum); + + my $cur = num_highwater($self) // 0; + if ($skip_artnum < $cur) { + die "E: current article number $cur ", + "exceeds --skip-artnum=$skip_artnum\n"; + } else { + my $ok; + for (1..10) { + my $mid = 'skip'.rand.'@'.rand.'.example.com'; + $ok = mid_set($self, $skip_artnum, $mid); + if ($ok) { + mid_delete($self, $mid); + last; + } + } + $ok or die '--skip-artnum failed'; + + # in the future, the indexer may use this value for + # new messages in old epochs + meta_accessor($self, 'skip_artnum', $skip_artnum); + } +} + +sub check_inodes { + my ($self) = @_; + # no filename if in-:memory: + my $f = $self->{dbh}->sqlite_db_filename // return; + if (my @st = stat($f)) { # did st_dev, st_ino change? + my $st = pack('dd', $st[0], $st[1]); + if ($st ne ($self->{st} // $st)) { + my $tmp = eval { ref($self)->new_file($f) }; + if ($@) { + warn "E: DBI->connect($f): $@\n"; + } else { + %$self = %$tmp; + } + } + } else { + warn "W: stat $f: $!\n"; + } } 1;