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>
4 # Local storage (cache/memo) for lei(1), suitable for personal/private
5 # mail iff on encrypted device/FS. Based on v2, but only deduplicates
8 # for xref3, the following are constant: $eidx_key = '.', $xnum = -1
9 package PublicInbox::LeiStore;
12 use parent qw(PublicInbox::Lock PublicInbox::IPC);
13 use PublicInbox::ExtSearchIdx;
14 use PublicInbox::Import;
15 use PublicInbox::InboxWritable qw(eml_from_path);
16 use PublicInbox::V2Writable;
17 use PublicInbox::ContentHash qw(content_hash content_digest);
18 use PublicInbox::MID qw(mids mids_in);
19 use PublicInbox::LeiSearch;
21 use List::Util qw(max);
24 my (undef, $dir, $opt) = @_;
25 my $eidx = PublicInbox::ExtSearchIdx->new($dir, $opt);
26 my $self = bless { priv_eidx => $eidx }, __PACKAGE__;
27 eidx_init($self)->done if $opt->{creat};
31 sub git { $_[0]->{priv_eidx}->git } # read-only
33 sub packing_factor { $PublicInbox::V2Writable::PACKING_FACTOR }
36 $_[0]->{rotate_bytes} // ((1024 * 1024 * 1024) / $_[0]->packing_factor)
39 sub git_pfx { "$_[0]->{priv_eidx}->{topdir}/local" };
43 if (opendir(my $dh, $self->git_pfx)) {
45 substr($_, 0, -4) + 0; # drop ".git" suffix
46 } grep(/\A[0-9]+\.git\z/, readdir($dh))) // 0;
48 $!{ENOENT} ? 0 : die("opendir ${\$self->git_pfx}: $!\n");
54 chomp(my $i = $git->qx(qw(var GIT_COMMITTER_IDENT)));
55 warn "$git->{git_dir} GIT_COMMITTER_IDENT failed\n" if $?;
56 $i =~ /\A(.+) <([^>]+)> [0-9]+ [-\+]?[0-9]+$/ ? ($1, $2) :
57 ('lei user', 'x@example.com')
65 return $im if $im->{bytes_added} < $self->rotate_bytes;
71 $max = $self->git_epoch_max + 1;
73 my $pfx = $self->git_pfx;
74 $max //= $self->git_epoch_max;
76 my $latest = "$pfx/$max.git";
78 PublicInbox::Import::init_bare($latest);
79 my $git = PublicInbox::Git->new($latest);
80 $git->qx(qw(config core.sharedRepository 0600)) if !$old;
81 my $packed_bytes = $git->packed_bytes;
82 my $unpacked_bytes = $packed_bytes / $self->packing_factor;
83 if ($unpacked_bytes >= $self->rotate_bytes) {
87 my ($n, $e) = git_ident($git);
88 $self->{im} = $im = PublicInbox::Import->new($git, $n, $e);
89 $im->{bytes_added} = int($packed_bytes / $self->packing_factor);
90 $im->{lock_path} = undef;
91 $im->{path_type} = 'v2';
97 PublicInbox::LeiSearch->new($_[0]->{priv_eidx}->{topdir});
102 my $eidx = $self->{priv_eidx};
103 $eidx->idx_init({-private => 1});
107 # when a message has no Message-IDs at all, this is needed for
108 # unsent Draft messages, at least
109 sub _fake_mid_for ($$) {
110 my ($eml, $dig) = @_;
111 my $mids = mids_in($eml, qw(X-Alt-Message-ID Resent-Message-ID));
112 $eml->{-lei_fake_mid} =
113 $mids->[0] // PublicInbox::Import::digest2mid($dig, $eml);
116 sub _docids_for ($$) {
117 my ($self, $eml) = @_;
119 my $dig = content_digest($eml);
120 my $chash = $dig->clone->digest;
121 my $eidx = eidx_init($self);
122 my $oidx = $eidx->{oidx};
123 my $im = $self->{im};
124 my $mids = mids($eml);
125 $mids->[0] //= _fake_mid_for($eml, $dig);
126 for my $mid (@$mids) {
128 while (my $cur = $oidx->next_by_mid($mid, \$id, \$prev)) {
129 my $oid = $cur->{blob};
130 my $docid = $cur->{num};
131 my $bref = $im ? $im->cat_blob($oid) : undef;
132 $bref //= $eidx->git->cat_file($oid) // do {
133 warn "W: $oid (#$docid) <$mid> not found\n";
136 local $self->{current_info} = $oid;
137 my $x = PublicInbox::Eml->new($bref);
138 $docids{$docid} = $docid if content_hash($x) eq $chash;
141 sort { $a <=> $b } values %docids;
144 sub set_eml_keywords {
145 my ($self, $eml, @kw) = @_;
146 my $eidx = eidx_init($self);
147 my @docids = _docids_for($self, $eml);
148 for my $docid (@docids) {
149 $eidx->idx_shard($docid)->ipc_do('set_keywords', $docid, @kw);
154 sub add_eml_keywords {
155 my ($self, $eml, @kw) = @_;
156 my $eidx = eidx_init($self);
157 my @docids = _docids_for($self, $eml);
158 for my $docid (@docids) {
159 $eidx->idx_shard($docid)->ipc_do('add_keywords', $docid, @kw);
164 sub remove_eml_keywords {
165 my ($self, $eml, @kw) = @_;
166 my $eidx = eidx_init($self);
167 my @docids = _docids_for($self, $eml);
168 for my $docid (@docids) {
169 $eidx->idx_shard($docid)->ipc_do('remove_keywords', $docid, @kw)
174 # cf: https://doc.dovecot.org/configuration_manual/mail_location/mbox/
175 my %status2kw = (F => 'flagged', A => 'answered', R => 'seen', T => 'draft');
176 # O (old/non-recent), and D (deleted) aren't in JMAP,
177 # so probably won't be supported by us.
180 my $s = "@{[$eml->header_raw('X-Status'),$eml->header_raw('Status')]}";
182 $s =~ s/([FART])/$kw{$status2kw{$1}} = 1/sge;
186 # cf: https://cr.yp.to/proto/maildir.html
187 my %c2kw = ('D' => 'draft', F => 'flagged', R => 'answered', S => 'seen');
188 sub maildir_keywords {
189 $_[-1] =~ /:2,([A-Z]+)\z/i ?
190 sort(map { $c2kw{$_} // () } split(//, $1)) : ();
194 my ($self, $eml, @kw) = @_;
195 my $im = $self->importer; # may create new epoch
196 my $eidx = eidx_init($self); # writes ALL.git/objects/info/alternates
197 my $oidx = $eidx->{oidx};
198 my $smsg = bless { -oidx => $oidx }, 'PublicInbox::Smsg';
199 $im->add($eml, undef, $smsg) or return; # duplicate returns undef
201 local $self->{current_info} = $smsg->{blob};
202 if (my @docids = _docids_for($self, $eml)) {
203 for my $docid (@docids) {
204 my $idx = $eidx->idx_shard($docid);
205 $oidx->add_xref3($docid, -1, $smsg->{blob}, '.');
206 # add_eidx_info for List-Id
207 $idx->ipc_do('add_eidx_info', $docid, '.', $eml);
208 $idx->ipc_do('add_keywords', $docid, @kw) if @kw;
212 $smsg->{num} = $oidx->adj_counter('eidx_docid', '+');
213 $oidx->add_overview($eml, $smsg);
214 $oidx->add_xref3($smsg->{num}, -1, $smsg->{blob}, '.');
215 my $idx = $eidx->idx_shard($smsg->{num});
216 $idx->index_eml($eml, $smsg);
217 $idx->ipc_do('add_keywords', $smsg->{num}, @kw) if @kw;
223 my ($self, $eml, @kw) = @_;
224 add_eml($self, $eml, @kw) // set_eml_keywords($self, $eml, @kw);
227 sub set_eml_from_maildir {
228 my ($self, $f, $set_kw) = @_;
229 my $eml = eml_from_path($f) or return;
230 set_eml($self, $eml, $set_kw ? maildir_keywords($f) : ());
236 if (my $im = delete($self->{im})) {
239 $err .= "import done: $@\n";
243 $self->{priv_eidx}->done;
247 sub ipc_atfork_child {
249 my $lei = delete $self->{lei};
250 $lei->lei_atfork_child(1) if $lei;
251 $self->SUPER::ipc_atfork_child;
255 my ($self, $lei) = @_;
256 $self->ipc_lock_init;
257 # Mail we import into lei are private, so headers filtered out
258 # by -mda for public mail are not appropriate
259 local @PublicInbox::MDA::BAD_HEADERS = ();
260 $self->ipc_worker_spawn('lei_store', $lei->oldset, { lei => $lei });