X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FOver.pm;h=a055b4cdbe6114061f3e20d3dfb782418521b304;hb=5a69dffc4155c8810dd8f2f3474f7a8a5c6d16a2;hp=07e54b64ca5de9448a4cf35399faaa37d7e6295a;hpb=cfb8d16578e7f2f2e300f9f436205e4a8fc7f322;p=public-inbox.git diff --git a/lib/PublicInbox/Over.pm b/lib/PublicInbox/Over.pm index 07e54b64..a055b4cd 100644 --- a/lib/PublicInbox/Over.pm +++ b/lib/PublicInbox/Over.pm @@ -1,4 +1,4 @@ -# Copyright (C) 2018 all contributors +# Copyright (C) 2018-2020 all contributors # License: AGPL-3.0+ # for XOVER, OVER in NNTP, and feeds/homepage/threads in PSGI @@ -9,20 +9,43 @@ use strict; use warnings; use DBI; use DBD::SQLite; -use PublicInbox::SearchMsg; +use PublicInbox::Smsg; use Compress::Zlib qw(uncompress); +use constant DEFAULT_LIMIT => 1000; sub dbh_new { - my ($self) = @_; - my $ro = ref($self) eq 'PublicInbox::Over'; - my $dbh = DBI->connect("dbi:SQLite:dbname=$self->{filename}",'','', { - AutoCommit => 1, - RaiseError => 1, - PrintError => 0, - ReadOnly => $ro, - sqlite_use_immediate_transaction => 1, - }); - $dbh->{sqlite_unicode} = 1; + my ($self, $rw) = @_; + 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)); + my $j = "$f-journal"; + open $fh, '+>>', $j or die "failed to open $j: $!"; + PublicInbox::Spawn::nodatacow_fd(fileno($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}; + $dbh->do('PRAGMA synchronous = OFF') if ($rw // 0) > 1; $dbh; } @@ -31,17 +54,25 @@ sub new { bless { filename => $f }, $class; } -sub disconnect { $_[0]->{dbh} = undef } +sub disconnect { + my ($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 connect { $_[0]->{dbh} //= $_[0]->dbh_new } -sub load_from_row { - my ($smsg) = @_; - bless $smsg, 'PublicInbox::SearchMsg'; +sub load_from_row ($;$) { + my ($smsg, $cull) = @_; + bless $smsg, 'PublicInbox::Smsg'; if (defined(my $data = delete $smsg->{ddd})) { $data = uncompress($data); - utf8::decode($data); - $smsg->load_from_data($data); + PublicInbox::Smsg::load_from_data($smsg, $data); + + # saves over 600K for 1000+ message threads + PublicInbox::Smsg::psgi_cull($smsg) if $cull; } $smsg } @@ -49,10 +80,11 @@ sub load_from_row { sub do_get { my ($self, $sql, $opts, @args) = @_; my $dbh = $self->connect; - my $lim = (($opts->{limit} || 0) + 0) || 1000; + my $lim = (($opts->{limit} || 0) + 0) || DEFAULT_LIMIT; $sql .= "LIMIT $lim"; my $msgs = $dbh->selectall_arrayref($sql, { Slice => {} }, @args); - load_from_row($_) for @$msgs; + my $cull = $opts->{cull}; + load_from_row($_, $cull) for @$msgs; $msgs } @@ -72,11 +104,21 @@ 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 => $nr }, @_); +SELECT num,ds,ddd FROM over WHERE num IN ($in) + +} + sub nothing () { wantarray ? (0, []) : [] }; sub get_thread { my ($self, $mid, $prev) = @_; my $dbh = $self->connect; + my $opts = { cull => 1 }; my $id = $dbh->selectrow_array(<<'', undef, $mid); SELECT id FROM msgid WHERE mid = ? LIMIT 1 @@ -93,21 +135,47 @@ ORDER BY num ASC LIMIT 1 SELECT tid,sid FROM over WHERE num = ? LIMIT 1 defined $tid or return nothing; # $sid may be undef + + my $cond_all = '(tid = ? OR sid = ?) AND num > ?'; my $sort_col = 'ds'; $num = 0; - if ($prev) { + if ($prev) { # mboxrd stream, only $num = $prev->{num} || 0; $sort_col = 'num'; } - my $cond = '(tid = ? OR sid = ?) AND num > ?'; - my $msgs = do_get($self, <<"", {}, $tid, $sid, $num); -SELECT num,ts,ds,ddd FROM over WHERE $cond ORDER BY $sort_col ASC - return $msgs unless wantarray; + my $cols = 'num,ts,ds,ddd'; + unless (wantarray) { + return do_get($self, <<"", $opts, $tid, $sid, $num); +SELECT $cols FROM over WHERE $cond_all +ORDER BY $sort_col ASC - my $nr = $dbh->selectrow_array(<<"", undef, $tid, $sid, $num); -SELECT COUNT(num) FROM over WHERE $cond + } + # HTML view always wants an array and never uses $prev, + # but the mbox stream never wants an array and always has $prev + die '$prev not supported with wantarray' if $prev; + my $nr = $dbh->selectrow_array(<<"", undef, $tid, $sid, $num); +SELECT COUNT(num) FROM over WHERE $cond_all + + # giant thread, prioritize strict (tid) matches and throw + # in the loose (sid) matches at the end + my $msgs = do_get($self, <<"", $opts, $tid, $num); +SELECT $cols FROM over WHERE tid = ? AND num > ? +ORDER BY $sort_col ASC + + # do we have room for loose matches? get the most recent ones, first: + my $lim = DEFAULT_LIMIT - scalar(@$msgs); + if ($lim > 0) { + $opts->{limit} = $lim; + my $loose = do_get($self, <<"", $opts, $tid, $sid, $num); +SELECT $cols FROM over WHERE tid != ? AND sid = ? AND num > ? +ORDER BY $sort_col DESC + + # TODO separate strict and loose matches here once --reindex + # is fixed to preserve `tid' properly + push @$msgs, @$loose; + } ($nr, $msgs); } @@ -116,18 +184,18 @@ sub recent { my ($s, @v); if (defined($before)) { if (defined($after)) { - $s = 'num > 0 AND ts >= ? AND ts <= ? ORDER BY ts DESC'; + $s = '+num > 0 AND ts >= ? AND ts <= ? ORDER BY ts DESC'; @v = ($after, $before); } else { - $s = 'num > 0 AND ts <= ? ORDER BY ts DESC'; + $s = '+num > 0 AND ts <= ? ORDER BY ts DESC'; @v = ($before); } } else { if (defined($after)) { - $s = 'num > 0 AND ts >= ? ORDER BY ts ASC'; + $s = '+num > 0 AND ts >= ? ORDER BY ts ASC'; @v = ($after); } else { - $s = 'num > 0 ORDER BY ts DESC'; + $s = '+num > 0 ORDER BY ts DESC'; } } my $msgs = do_get($self, <<"", $opts, @v); @@ -143,12 +211,13 @@ SELECT COUNT(num) FROM over WHERE num > 0 sub get_art { my ($self, $num) = @_; - my $dbh = $self->connect; - my $smsg = $dbh->selectrow_hashref(<<'', undef, $num); + # caching $sth ourselves is faster than prepare_cached + my $sth = $self->{-get_art} //= $self->connect->prepare(<<''); SELECT num,ds,ts,ddd FROM over WHERE num = ? LIMIT 1 - return load_from_row($smsg) if $smsg; - undef; + $sth->execute($num); + my $smsg = $sth->fetchrow_hashref; + $smsg ? load_from_row($smsg) : undef; } sub next_by_mid { @@ -171,13 +240,52 @@ ORDER BY num ASC LIMIT 1 $sth->execute($$id, $$prev); my $num = $sth->fetchrow_array or return; $$prev = $num; + get_art($self, $num); +} - $sth = $dbh->prepare_cached(<<"", undef, 1); -SELECT num,ts,ds,ddd FROM over WHERE num = ? LIMIT 1 +# IMAP search, this is limited by callers to UID_SLICE size (50K) +sub uid_range { + my ($self, $beg, $end, $sql) = @_; + my $dbh = $self->connect; + my $q = 'SELECT num FROM over WHERE num >= ? AND num <= ?'; - $sth->execute($num); - my $smsg = $sth->fetchrow_hashref or return; - load_from_row($smsg); + # This is read-only, anyways; but caller should verify it's + # only sending \A[0-9]+\z for ds and ts column ranges + $q .= $$sql if $sql; + $q .= ' ORDER BY num ASC'; + $dbh->selectcol_arrayref($q, undef, $beg, $end); +} + +sub max { + my ($self) = @_; + my $sth = $self->connect->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 = $self->connect->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], $st[1]); + + # don't actually reopen, just let {dbh} be recreated later + disconnect($self) if $st ne ($self->{st} // $st); + } else { + warn "W: stat $f: $!\n"; + } } 1;