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);
9 # n.b. mutt sets most of these headers not sure about Bytes
10 our @OID_IGNORE = qw(Status X-Status Content-Length Lines Bytes);
12 # best-effort regeneration of OID when augmenting existing results
15 my @stash; # stash away headers we shouldn't have in git
16 for my $k (@OID_IGNORE) {
17 my @v = $eml->header_raw($k) or next;
18 push @stash, [ $k, \@v ];
19 $eml->header_set($k); # restore below
21 my $dig = Digest::SHA->new(1); # XXX SHA256 later
22 my $buf = $eml->as_string;
23 $dig->add('blob '.length($buf)."\0");
27 for my $kv (@stash) { # restore stashed headers
29 $eml->header_set($k, @v);
34 sub _oidbin ($) { defined($_[0]) ? pack('H*', $_[0]) : undef }
38 my $dig = Digest::SHA->new(256);
39 my $x = join("\0", @$smsg{qw(from to cc ds subject references mid)});
48 (sub { # may be called in a child process
50 $skv->set_maybe(_oidbin($oid) // _regen_oid($eml), '');
53 $skv->set_maybe(_oidbin($smsg->{blob}), '');
57 # dangerous if there's duplicate messages with different Message-IDs
60 (sub { # may be called in a child process
62 # TODO: lei will support non-public messages w/o Message-ID
63 my $mid = $eml->header_raw('Message-ID') // _oidbin($oid) //
65 $skv->set_maybe($mid, '');
68 my $mid = $smsg->{mid};
69 $mid = undef if $mid eq '';
70 $mid //= smsg_hash($smsg) // _oidbin($smsg->{blob});
71 $skv->set_maybe($mid, '');
75 # our default deduplication strategy (used by v2, also)
76 sub dedupe_content ($) {
78 (sub { # may be called in a child process
79 my ($eml) = @_; # oid = $_[1], ignored
80 $skv->set_maybe(content_hash($eml), '');
83 $skv->set_maybe(smsg_hash($smsg), '');
87 # no deduplication at all
89 sub dedupe_none ($) { (\&true, \&true) }
93 my $dd = $lei->{opt}->{dedupe} // 'content';
94 my $dst = $lei->{ovv}->{dst};
96 # allow "none" to bypass Eml->new if writing to directory:
97 return if ($dd eq 'none' && substr($dst // '', -1) eq '/');
99 $cls->can($m) or die "unsupported dedupe strategy: $dd\n";
100 my $skv = $dd eq 'none' ? undef : PublicInbox::SharedKV->new;
102 # [ $skv, $eml_cb, $smsg_cb, "dedupe_$dd" ]
103 bless [ $skv, undef, undef, $m ], $cls;
106 # returns true on seen messages according to the deduplication strategy,
107 # returns false if unseen
109 my ($self, $eml, $oid) = @_;
110 !$self->[1]->($eml, $oid);
114 my ($self, $smsg) = @_;
115 !$self->[2]->($smsg);
120 my $skv = $self->[0];
121 $self->[1] or @$self[1,2] = $self->can($self->[3])->($skv);
122 $skv ? $skv->dbh : undef;
127 my $skv = $self->[0];
128 delete($skv->{dbh}) if $skv;