1 # Copyright (C) 2015-2021 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # bidirectional Message-ID <-> Article Number mapping for the NNTP
5 # and web interfaces. This is required for implementing stable article
6 # numbers for NNTP and allows prefix lookups for partial Message-IDs
7 # in case URLs get truncated from copy-n-paste errors by users.
9 # This is maintained by ::SearchIdx (v1) and ::V2Writable (v2)
10 package PublicInbox::Msgmap;
15 use PublicInbox::Over;
16 use PublicInbox::Spawn;
19 my ($class, $git_dir, $writable) = @_;
20 my $d = "$git_dir/public-inbox";
21 if ($writable && !-d $d && !mkdir $d) {
23 -d $d or die "$d not created: $err";
25 new_file($class, "$d/msgmap.sqlite3", $writable);
29 my ($class, $f, $rw) = @_;
30 return if !$rw && !-r $f;
32 my $self = bless { filename => $f }, $class;
33 my $dbh = $self->{dbh} = PublicInbox::Over::dbh_new($self, $rw);
37 $self->created_at(time) unless $self->created_at;
39 $self->num_highwater(max($self));
45 # used to keep track of used numeric mappings for v2 reindex
47 my ($self, $dir) = @_;
49 my $tmp = "mm_tmp-$$-XXXX";
50 my ($fh, $fn) = File::Temp::tempfile($tmp, EXLOCK => 0, DIR => $dir);
51 PublicInbox::Spawn::nodatacow_fd(fileno($fh));
52 $self->{dbh}->sqlite_backup_to_file($fn);
53 $tmp = ref($self)->new_file($fn, 2);
54 $tmp->{dbh}->do('PRAGMA journal_mode = MEMORY');
59 # n.b. invoked directly by scripts/xhdr-num2mid
61 my ($self, $key, $value) = @_;
63 my $sql = 'SELECT val FROM meta WHERE key = ? LIMIT 1';
64 my $prev = $self->{dbh}->selectrow_array($sql, undef, $key);
65 $value // return $prev;
68 $sql = 'UPDATE meta SET val = ? WHERE key = ?';
69 $self->{dbh}->do($sql, undef, $value, $key);
71 $sql = 'INSERT INTO meta (key,val) VALUES (?,?)';
72 $self->{dbh}->do($sql, undef, $key, $value);
78 my ($self, $commit) = @_;
79 $self->meta_accessor('last_commit', $commit);
82 # v2 uses this to keep track of how up-to-date Xapian is
83 # old versions may be automatically GC'ed away in the future,
84 # but it's a trivial amount of storage.
86 my ($self, $version, $i, $commit) = @_;
87 $self->meta_accessor("last_xap$version-$i", $commit);
90 # this is the UIDVALIDITY for IMAP (cf. RFC 3501 sec 2.3.1.1. item 3)
92 my ($self, $second) = @_;
93 $self->meta_accessor('created_at', $second);
97 my ($self, $num) = @_;
98 my $high = $self->{num_highwater} ||=
99 $self->meta_accessor('num_highwater');
100 if (defined($num) && (!defined($high) || ($num > $high))) {
101 $self->{num_highwater} = $num;
102 $self->meta_accessor('num_highwater', $num);
104 $self->{num_highwater};
108 my ($self, $mid) = @_;
109 my $sth = $self->{dbh}->prepare_cached(<<'');
110 INSERT INTO msgmap (mid) VALUES (?)
112 return unless eval { $sth->execute($mid) };
113 my $num = $self->{dbh}->last_insert_id(undef, undef, 'msgmap', 'num');
114 $self->num_highwater($num) if defined($num);
119 my ($self, $num) = @_;
120 my $sth = $self->{dbh}->prepare_cached(<<"", undef, 1);
121 SELECT mid FROM msgmap WHERE num = ? LIMIT 1
124 $sth->fetchrow_array;
128 my ($self, $mid) = @_;
129 my $sth = $self->{dbh}->prepare_cached(<<"", undef, 1);
130 SELECT num FROM msgmap WHERE mid = ? LIMIT 1
133 $sth->fetchrow_array;
137 my $sth = $_[0]->{dbh}->prepare_cached('SELECT MAX(num) FROM msgmap',
140 $sth->fetchrow_array // 0;
144 # breaking MIN and MAX into separate queries speeds up from 250ms
145 # to around 700us with 2.7million messages.
146 my $sth = $_[0]->{dbh}->prepare_cached('SELECT MIN(num) FROM msgmap',
149 ($sth->fetchrow_array // 0, max($_[0]));
153 my ($self, $mid) = @_;
154 $self->{dbh}->do('DELETE FROM msgmap WHERE mid = ?', undef, $mid);
158 my ($self, $num) = @_;
159 $self->{dbh}->do('DELETE FROM msgmap WHERE num = ?', undef, $num);
166 CREATE TABLE IF NOT EXISTS msgmap (
167 num INTEGER PRIMARY KEY AUTOINCREMENT,
168 mid VARCHAR(1000) NOT NULL,
173 CREATE TABLE IF NOT EXISTS meta (
174 key VARCHAR(32) PRIMARY KEY,
175 val VARCHAR(255) NOT NULL
181 my ($self, $beg, $end, $cols) = @_;
183 my $attr = { Columns => [] };
184 my $mids = $self->{dbh}->selectall_arrayref(<<"", $attr, $$beg, $end);
185 SELECT $cols FROM msgmap WHERE num >= ? AND num <= ?
186 ORDER BY num ASC LIMIT 1000
188 $$beg = $mids->[-1]->[0] + 1 if @$mids;
192 # only used for mapping external serial numbers (e.g. articles from gmane)
193 # see scripts/xhdr-num2mid or PublicInbox::Filter::RubyLang for usage
195 my ($self, $num, $mid) = @_;
196 my $sth = $self->{dbh}->prepare_cached(<<"");
197 INSERT OR IGNORE INTO msgmap (num,mid) VALUES (?,?)
199 my $result = $sth->execute($num, $mid);
200 $self->num_highwater($num) if (defined($result) && $result == 1);
206 my $dbh = $self->{dbh} or return;
207 if (($self->{pid} // 0) == $$) {
208 my $f = $dbh->sqlite_db_filename;
209 unlink $f or warn "failed to unlink $f: $!\n";
215 $self->{pid} or die 'BUG: not a temporary clone';
216 $self->{dbh} and die 'BUG: tmp_clone dbh not prepared for parent';
217 defined($self->{filename}) or die 'BUG: {filename} not defined';
218 $self->{dbh} = PublicInbox::Over::dbh_new($self, 2);
219 $self->{dbh}->do('PRAGMA journal_mode = MEMORY');
224 my $pid = $self->{pid} or die 'BUG: not a temporary clone';
225 $pid == $$ or die "BUG: atfork_prepare not called by $pid";
226 my $dbh = $self->{dbh} or die 'BUG: temporary clone not open';
228 # must clobber prepared statements
229 %$self = (filename => $dbh->sqlite_db_filename, pid => $pid);
233 my ($self, $skip_artnum) = @_;
234 return meta_accessor($self, 'skip_artnum') if !defined($skip_artnum);
236 my $cur = num_highwater($self) // 0;
237 if ($skip_artnum < $cur) {
238 die "E: current article number $cur ",
239 "exceeds --skip-artnum=$skip_artnum\n";
243 my $mid = 'skip'.rand.'@'.rand.'.example.com';
244 $ok = mid_set($self, $skip_artnum, $mid);
246 mid_delete($self, $mid);
250 $ok or die '--skip-artnum failed';
252 # in the future, the indexer may use this value for
253 # new messages in old epochs
254 meta_accessor($self, 'skip_artnum', $skip_artnum);
260 $self->{dbh} // return;
261 my $rw = !$self->{dbh}->{ReadOnly};
262 PublicInbox::Over::check_inodes($self);
263 $self->{dbh} //= PublicInbox::Over::dbh_new($self, !$rw);