X-Git-Url: http://www.git.stargrave.org/?p=public-inbox.git;a=blobdiff_plain;f=lib%2FPublicInbox%2FOver.pm;h=081123868795fd9b22997488331b0bc39dc51800;hp=1faeff418f4d9caadc30eb6e6c9d19541b246caf;hb=6c853f5256f3a324c018a37a9e8a6b9fc8fdc063;hpb=77fe488ce5432a49ea8499605e0d8e155834575f diff --git a/lib/PublicInbox/Over.pm b/lib/PublicInbox/Over.pm index 1faeff41..08112386 100644 --- a/lib/PublicInbox/Over.pm +++ b/lib/PublicInbox/Over.pm @@ -6,7 +6,7 @@ # tweaked/updated over time and rebuilt. package PublicInbox::Over; use strict; -use warnings; +use v5.10.1; use DBI; use DBD::SQLite; use PublicInbox::Smsg; @@ -15,18 +15,55 @@ use constant DEFAULT_LIMIT => 1000; sub dbh_new { my ($self, $rw) = @_; - my $f = $self->{filename}; - if ($rw && !-f $f) { # SQLite defaults mode to 0644, we want 0666 - open my $fh, '+>>', $f or die "failed to open $f: $!"; + my $f = delete $self->{filename}; + if (!-f $f) { # SQLite defaults mode to 0644, we want 0666 + if ($rw) { + require PublicInbox::Spawn; + open my $fh, '+>>', $f or die "failed to open $f: $!"; + PublicInbox::Spawn::nodatacow_fd(fileno($fh)); + } else { + $self->{filename} = $f; # die on stat() below: + } + } + my (@st, $st, $dbh); + my $tries = 0; + do { + @st = stat($f) or die "failed to stat $f: $!"; + $st = pack('dd', $st[0], $st[1]); # 0: dev, 1: inode + $dbh = DBI->connect("dbi:SQLite:dbname=$f",'','', { + AutoCommit => 1, + RaiseError => 1, + PrintError => 0, + ReadOnly => !$rw, + sqlite_use_immediate_transaction => 1, + }); + $self->{st} = $st; + @st = stat($f) or die "failed to stat $f: $!"; + $st = pack('dd', $st[0], $st[1]); + } while ($st ne $self->{st} && $tries++ < 3); + warn "W: $f: .st_dev, .st_ino unstable\n" if $st ne $self->{st}; + + if ($rw) { + # TRUNCATE reduces I/O compared to the default (DELETE). + # + # Do not use WAL by default since we expect the case + # where any users may read via read-only daemons + # (-httpd/-imapd/-nntpd); but only a single user has + # write permissions for -watch/-mda. + # + # Read-only WAL support in SQLite 3.22.0 (2018-01-22) + # doesn't do what we need: it is only intended for + # immutable read-only media (e.g. CD-ROM) and not + # usable for our use case described above. + # + # If an admin is willing to give read-only daemons R/W + # permissions; they can enable WAL manually and we will + # respect that by not clobbering it. + my $jm = $dbh->selectrow_array('PRAGMA journal_mode'); + $dbh->do('PRAGMA journal_mode = TRUNCATE') if $jm ne 'wal'; + + $dbh->do('PRAGMA synchronous = OFF') if $rw > 1; } - my $dbh = DBI->connect("dbi:SQLite:dbname=$f",'','', { - AutoCommit => 1, - RaiseError => 1, - PrintError => 0, - ReadOnly => !$rw, - sqlite_use_immediate_transaction => 1, - }); - $dbh->{sqlite_unicode} = 1; $dbh; } @@ -35,16 +72,21 @@ sub new { bless { filename => $f }, $class; } -sub disconnect { $_[0]->{dbh} = undef } +sub dbh_close { + my ($self) = @_; + if (my $dbh = delete $self->{dbh}) { + delete $self->{-get_art}; + $self->{filename} = $dbh->sqlite_db_filename; + } +} -sub connect { $_[0]->{dbh} ||= $_[0]->dbh_new } +sub dbh ($) { $_[0]->{dbh} //= $_[0]->dbh_new } # dbh_new may be subclassed sub load_from_row ($;$) { my ($smsg, $cull) = @_; bless $smsg, 'PublicInbox::Smsg'; if (defined(my $data = delete $smsg->{ddd})) { $data = uncompress($data); - utf8::decode($data); PublicInbox::Smsg::load_from_data($smsg, $data); # saves over 600K for 1000+ message threads @@ -55,10 +97,9 @@ sub load_from_row ($;$) { sub do_get { my ($self, $sql, $opts, @args) = @_; - my $dbh = $self->connect; my $lim = (($opts->{limit} || 0) + 0) || DEFAULT_LIMIT; $sql .= "LIMIT $lim"; - my $msgs = $dbh->selectall_arrayref($sql, { Slice => {} }, @args); + my $msgs = dbh($self)->selectall_arrayref($sql, { Slice => {} }, @args); my $cull = $opts->{cull}; load_from_row($_, $cull) for @$msgs; $msgs @@ -80,11 +121,20 @@ ORDER BY num ASC } +sub get_all { + my $self = shift; + my $nr = scalar(@_) or return []; + my $in = '?' . (',?' x ($nr - 1)); + do_get($self, <<"", { cull => 1, limit => $nr }, @_); +SELECT num,ts,ds,ddd FROM over WHERE num IN ($in) + +} + sub nothing () { wantarray ? (0, []) : [] }; sub get_thread { my ($self, $mid, $prev) = @_; - my $dbh = $self->connect; + my $dbh = dbh($self); my $opts = { cull => 1 }; my $id = $dbh->selectrow_array(<<'', undef, $mid); @@ -146,6 +196,35 @@ ORDER BY $sort_col DESC ($nr, $msgs); } +# strict `tid' matches, only, for thread-expanded mbox.gz search results +# and future CLI interface +# returns true if we have IDs, undef if not +sub expand_thread { + my ($self, $ctx) = @_; + my $dbh = dbh($self); + do { + defined(my $num = $ctx->{ids}->[0]) or return; + my ($tid) = $dbh->selectrow_array(<<'', undef, $num); +SELECT tid FROM over WHERE num = ? + + if (defined($tid)) { + my $sql = <<''; +SELECT num FROM over WHERE tid = ? AND num > ? +ORDER BY num ASC LIMIT 1000 + + my $xids = $dbh->selectcol_arrayref($sql, undef, $tid, + $ctx->{prev} // 0); + if (scalar(@$xids)) { + $ctx->{prev} = $xids->[-1]; + $ctx->{xids} = $xids; + return 1; # success + } + } + $ctx->{prev} = 0; + shift @{$ctx->{ids}}; + } while (1); +} + sub recent { my ($self, $opts, $after, $before) = @_; my ($s, @v); @@ -165,22 +244,16 @@ sub recent { $s = '+num > 0 ORDER BY ts DESC'; } } - my $msgs = do_get($self, <<"", $opts, @v); + do_get($self, <<"", $opts, @v); SELECT ts,ds,ddd FROM over WHERE $s - return $msgs unless wantarray; - - my $nr = $self->{dbh}->selectrow_array(<<''); -SELECT COUNT(num) FROM over WHERE num > 0 - - ($nr, $msgs); } sub get_art { my ($self, $num) = @_; - my $dbh = $self->connect; - my $sth = $dbh->prepare_cached(<<'', undef, 1); -SELECT num,ds,ts,ddd FROM over WHERE num = ? LIMIT 1 + # caching $sth ourselves is faster than prepare_cached + my $sth = $self->{-get_art} //= dbh($self)->prepare(<<''); +SELECT num,tid,ds,ts,ddd FROM over WHERE num = ? LIMIT 1 $sth->execute($num); my $smsg = $sth->fetchrow_hashref; @@ -189,7 +262,7 @@ SELECT num,ds,ts,ddd FROM over WHERE num = ? LIMIT 1 sub next_by_mid { my ($self, $mid, $id, $prev) = @_; - my $dbh = $self->connect; + my $dbh = dbh($self); unless (defined $$id) { my $sth = $dbh->prepare_cached(<<'', undef, 1); @@ -207,26 +280,52 @@ ORDER BY num ASC LIMIT 1 $sth->execute($$id, $$prev); my $num = $sth->fetchrow_array or return; $$prev = $num; - - $sth = $dbh->prepare_cached(<<"", undef, 1); -SELECT num,ts,ds,ddd FROM over WHERE num = ? LIMIT 1 - - $sth->execute($num); - my $smsg = $sth->fetchrow_hashref or return; - load_from_row($smsg); + get_art($self, $num); } -# IMAP search +# IMAP search, this is limited by callers to UID_SLICE size (50K) sub uid_range { my ($self, $beg, $end, $sql) = @_; - my $dbh = $self->connect; + my $dbh = dbh($self); my $q = 'SELECT num FROM over WHERE num >= ? AND num <= ?'; # This is read-only, anyways; but caller should verify it's # only sending \A[0-9]+\z for ds and ts column ranges $q .= $$sql if $sql; - $q .= ' ORDER BY num ASC LIMIT ' . DEFAULT_LIMIT; + $q .= ' ORDER BY num ASC'; $dbh->selectcol_arrayref($q, undef, $beg, $end); } +sub max { + my ($self) = @_; + my $sth = dbh($self)->prepare_cached(<<'', undef, 1); +SELECT MAX(num) FROM over WHERE num > 0 + + $sth->execute; + $sth->fetchrow_array // 0; +} + +sub imap_exists { + my ($self, $uid_base, $uid_end) = @_; + my $sth = dbh($self)->prepare_cached(<<'', undef, 1); +SELECT COUNT(num) FROM over WHERE num > ? AND num <= ? + + $sth->execute($uid_base, $uid_end); + $sth->fetchrow_array; +} + +sub check_inodes { + my ($self) = @_; + my $dbh = $self->{dbh} or return; + my $f = $dbh->sqlite_db_filename; + if (my @st = stat($f)) { # did st_dev, st_ino change? + my $st = pack('dd', $st[0], $st[1]); + + # don't actually reopen, just let {dbh} be recreated later + dbh_close($self) if $st ne ($self->{st} // $st); + } else { + warn "W: stat $f: $!\n"; + } +} + 1;