1 # Copyright (C) 2016 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # ref: https://cr.yp.to/proto/maildir.html
5 # http://wiki2.dovecot.org/MailboxFormat/Maildir
6 package PublicInbox::WatchMaildir;
10 use Email::MIME::ContentType;
11 $Email::MIME::ContentType::STRICT_PARAMS = 0; # user input is imperfect
13 use PublicInbox::Import;
15 use PublicInbox::Spawn qw(spawn);
18 my ($class, $config) = @_;
19 my (%mdmap, @mdir, $spamc);
21 # XXX is "publicinboxlearn" really a good namespace for this?
22 my $k = 'publicinboxlearn.watchspam';
23 if (my $spamdir = $config->{$k}) {
24 if ($spamdir =~ s/\Amaildir://) {
26 # skip "new", no MUA has seen it, yet.
27 my $cur = "$spamdir/cur";
29 $mdmap{$cur} = 'watchspam';
31 warn "unsupported $k=$spamdir\n";
35 $k = 'publicinboxwatch.spamcheck';
36 my $spamcheck = $config->{$k};
38 if ($spamcheck eq 'spamc') {
39 $spamcheck = 'PublicInbox::Spamcheck::Spamc';
41 if ($spamcheck =~ /::/) {
42 eval "require $spamcheck";
43 $spamcheck = _spamcheck_cb($spamcheck->new);
45 warn "unsupported $k=$spamcheck\n";
49 foreach $k (keys %$config) {
50 $k =~ /\Apublicinbox\.([^\.]+)\.watch\z/ or next;
52 my $watch = $config->{$k};
53 if ($watch =~ s/\Amaildir://) {
55 my $inbox = $config->lookup_name($name);
56 if (my $wm = $inbox->{watchheader}) {
57 my ($k, $v) = split(/:/, $wm, 2);
58 $inbox->{-watchheader} = [ $k, qr/\Q$v\E/ ];
60 my $new = "$watch/new";
61 my $cur = "$watch/cur";
62 push @mdir, $new, $cur;
63 die "$new already in use\n" if $mdmap{$new};
64 die "$cur already in use\n" if $mdmap{$cur};
65 $mdmap{$new} = $mdmap{$cur} = $inbox;
67 warn "watch unsupported: $k=$watch\n";
72 my $mdre = join('|', map { quotemeta($_) } @mdir);
73 $mdre = qr!\A($mdre)/!;
75 spamcheck => $spamcheck,
84 $_->done foreach values %{$_[0]->{importers}};
88 my ($self, $paths) = @_;
89 _try_path($self, $_->{path}) foreach @$paths;
94 my ($self, $path) = @_;
95 $path =~ /:2,[A-R]*S[T-Z]*\z/i or return;
96 my $mime = _path_to_mime($path) or return;
98 foreach my $inbox (values %{$self->{mdmap}}) {
99 next unless ref $inbox;
100 my $im = _importer_for($self, $inbox);
102 if (my $scrub = _scrubber_for($inbox)) {
103 my $scrubbed = $scrub->scrub($mime) or next;
104 $im->remove($scrubbed);
109 # used to hash the relevant portions of a message when there are conflicts
113 my $dig = Digest::SHA->new('SHA-1');
114 $dig->add($mime->header_obj->header_raw('Subject'));
115 $dig->add($mime->body_raw);
121 # probably a bad idea, but we inject a Message-Id if
122 # one is missing, here..
123 my $mid = $mime->header_obj->header_raw('Message-Id');
124 if (!defined $mid || $mid =~ /\A\s*\z/) {
125 $mid = '<' . _hash_mime2($mime) . '@generated>';
126 $mime->header_set('Message-Id', $mid);
131 my ($self, $path) = @_;
132 my @p = split(m!/+!, $path);
133 return if $p[-1] !~ /\A[a-zA-Z0-9][\w:,=\.]+\z/;
134 if ($p[-1] =~ /:2,([A-Z]+)\z/i) {
136 return if $flags =~ /[DT]/; # no [D]rafts or [T]rashed mail
138 return unless -f $path;
139 if ($path !~ $self->{mdre}) {
140 warn "unrecognized path: $path\n";
143 my $inbox = $self->{mdmap}->{$1};
145 warn "unmappable dir: $1\n";
148 if (!ref($inbox) && $inbox eq 'watchspam') {
149 return _remove_spam($self, $path);
151 my $im = _importer_for($self, $inbox);
152 my $mime = _path_to_mime($path) or return;
153 $mime->header_set($_) foreach @PublicInbox::MDA::BAD_HEADERS;
154 my $wm = $inbox->{-watchheader};
156 my $v = $mime->header_obj->header_raw($wm->[0]);
157 return unless ($v && $v =~ $wm->[1]);
159 if (my $scrub = _scrubber_for($inbox)) {
160 $mime = $scrub->scrub($mime) or return;
164 $im->add($mime, $self->{spamcheck});
169 my $cb = sub { _try_fsn_paths($self, \@_) };
170 my $mdir = $self->{mdir};
172 # lazy load here, we may support watching via IMAP IDLE
174 require Filesys::Notify::Simple;
175 my $watcher = Filesys::Notify::Simple->new($mdir);
176 $watcher->wait($cb) while (1);
181 my $mdir = $self->{mdir};
182 foreach my $dir (@$mdir) {
183 my $ok = opendir(my $dh, $dir);
185 warn "failed to open $dir: $!\n";
188 while (my $fn = readdir($dh)) {
189 _try_path($self, "$dir/$fn");
193 _done_for_now($self);
198 if (open my $fh, '<', $path) {
202 return Email::MIME->new(\$str);
203 } elsif ($!{ENOENT}) {
206 warn "failed to open $path: $!\n";
212 my ($self, $inbox) = @_;
213 my $im = $inbox->{-import} ||= eval {
214 my $git = $inbox->git;
215 my $name = $inbox->{name};
216 my $addr = $inbox->{-primary_address};
217 PublicInbox::Import->new($git, $name, $addr);
219 $self->{importers}->{"$im"} = $im;
224 my $f = $inbox->{filter};
225 if ($f && $f =~ /::/) {
241 if ($sc->spamcheck($mime, \$tmp)) {
242 return Email::MIME->new(\$tmp);
244 warn $mime->header('Message-ID')." failed spam check\n";