1 # Copyright (C) 2020-2021 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
3 package PublicInbox::LeiDedupe;
6 use PublicInbox::SharedKV;
7 use PublicInbox::ContentHash qw(content_hash);
10 # n.b. mutt sets most of these headers not sure about Bytes
11 our @OID_IGNORE = qw(Status X-Status Content-Length Lines Bytes);
13 # best-effort regeneration of OID when augmenting existing results
16 my @stash; # stash away headers we shouldn't have in git
17 for my $k (@OID_IGNORE) {
18 my @v = $eml->header_raw($k) or next;
19 push @stash, [ $k, \@v ];
20 $eml->header_set($k); # restore below
22 my $dig = Digest::SHA->new(1); # XXX SHA256 later
23 my $buf = $eml->as_string;
24 $dig->add('blob '.length($buf)."\0");
28 for my $kv (@stash) { # restore stashed headers
30 $eml->header_set($k, @v);
35 sub _oidbin ($) { defined($_[0]) ? pack('H*', $_[0]) : undef }
39 my $dig = Digest::SHA->new(256);
40 my $x = join("\0", @$smsg{qw(from to cc ds subject references mid)});
49 (sub { # may be called in a child process
51 $skv->set_maybe(_oidbin($oid) // _regen_oid($eml), '');
54 $skv->set_maybe(_oidbin($smsg->{blob}), '');
58 # dangerous if there's duplicate messages with different Message-IDs
61 (sub { # may be called in a child process
63 # TODO: lei will support non-public messages w/o Message-ID
64 my $mid = $eml->header_raw('Message-ID') // _oidbin($oid) //
66 $skv->set_maybe($mid, '');
69 my $mid = $smsg->{mid};
70 $mid = undef if $mid eq '';
71 $mid //= smsg_hash($smsg) // _oidbin($smsg->{blob});
72 $skv->set_maybe($mid, '');
76 # our default deduplication strategy (used by v2, also)
77 sub dedupe_content ($) {
79 (sub { # may be called in a child process
80 my ($eml) = @_; # oid = $_[1], ignored
81 $skv->set_maybe(content_hash($eml), '');
84 $skv->set_maybe(smsg_hash($smsg), '');
88 # no deduplication at all
90 sub dedupe_none ($) { (\&true, \&true) }
94 my $dd = $lei->{opt}->{dedupe} // 'content';
95 my $dst = $lei->{ovv}->{dst};
97 # allow "none" to bypass Eml->new if writing to directory:
98 return if ($dd eq 'none' && substr($dst // '', -1) eq '/');
100 $cls->can($m) or die "unsupported dedupe strategy: $dd\n";
101 my $skv = $dd eq 'none' ? undef : PublicInbox::SharedKV->new;
103 # [ $skv, $eml_cb, $smsg_cb, "dedupe_$dd" ]
104 bless [ $skv, undef, undef, $m ], $cls;
107 # returns true on seen messages according to the deduplication strategy,
108 # returns false if unseen
110 my ($self, $eml, $oid) = @_;
111 !$self->[1]->($eml, $oid);
115 my ($self, $smsg) = @_;
116 !$self->[2]->($smsg);
121 my $skv = $self->[0];
122 $self->[1] or @$self[1,2] = $self->can($self->[3])->($skv);
123 $skv ? $skv->dbh : undef;
128 my $skv = $self->[0];
129 delete($skv->{dbh}) if $skv;