1 # Copyright (C) 2015-2018 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # Streaming interface for formatting messages as an mboxrd.
5 # Used by the web interface
6 package PublicInbox::Mbox;
9 use PublicInbox::MID qw/mid_clean mid_escape/;
10 use PublicInbox::Hval qw/to_filename/;
12 use Email::MIME::Encode;
16 my $fn = $simple->header('Subject');
17 return 'no-subject' unless defined($fn);
19 # no need for full Email::MIME, here
21 eval { $fn = Encode::decode('MIME-Header', $fn) };
22 $fn = 'no-subject' if $@;
25 $fn = to_filename($fn);
26 $fn eq '' ? 'no-subject' : $fn;
31 bless $more, 'PublicInbox::Mbox';
34 # called by PSGI server as body response
36 my ($more) = @_; # self
37 my ($ctx, $head, $tail, $db, $cur) = @$more;
40 return msg_str($ctx, $cur);
42 for (; !defined($cur) && $head != $tail; $head++) {
43 my $smsg = PublicInbox::SearchMsg->get($head, $db, $ctx->{mid});
44 my $mref = $ctx->{-inbox}->msg_by_smsg($smsg) or next;
45 $cur = Email::Simple->new($mref);
46 $cur = msg_str($ctx, $cur);
56 my $mid = $ctx->{mid};
57 my $ibx = $ctx->{-inbox};
60 my ($head, $tail, $db);
62 if (my $srch = $ibx->search) {
63 $srch->retry_reopen(sub {
64 ($head, $tail, $db) = $srch->each_smsg_by_mid($mid);
65 for (; !defined($first) && $head != $tail; $head++) {
66 my @args = ($head, $db, $mid);
67 my $smsg = PublicInbox::SearchMsg->get(@args);
68 my $mref = $ibx->msg_by_smsg($smsg) or next;
69 $first = Email::Simple->new($mref);
72 $more = [ $ctx, $head, $tail, $db, $first ];
76 my $mref = $ibx->msg_by_mid($mid) or return;
77 $first = Email::Simple->new($mref);
79 return unless defined $first;
80 my $fn = subject_fn($first);
81 my @hdr = ('Content-Type');
82 if ($ibx->{obfuscate}) {
83 # obfuscation is stupid, but maybe scrapers are, too...
84 push @hdr, 'application/mbox';
87 push @hdr, 'text/plain';
90 push @hdr, 'Content-Disposition', "inline; filename=$fn";
91 [ 200, \@hdr, $more ? mb_stream($more) : [ msg_str($ctx, $first) ] ];
95 my ($ctx, $simple, $mid) = @_; # Email::Simple object
96 my $header_obj = $simple->header_obj;
98 # drop potentially confusing headers, ssoma already should've dropped
99 # Lines and Content-Length
100 foreach my $d (qw(Lines Bytes Content-Length Status)) {
101 $header_obj->header_set($d);
103 my $ibx = $ctx->{-inbox};
104 my $base = $ibx->base_url($ctx->{env});
105 $mid = $ctx->{mid} unless defined $mid;
106 $mid = mid_escape($mid);
108 'Archived-At', "<$base$mid/>",
109 'List-Archive', "<$base>",
110 'List-Post', "<mailto:$ibx->{-primary_address}>",
112 my $crlf = $simple->crlf;
113 my $buf = "From mboxrd\@z Thu Jan 1 00:00:00 1970\n" .
114 $header_obj->as_string;
115 for (my $i = 0; $i < @append; $i += 2) {
117 my $v = $append[$i + 1];
118 my @v = $header_obj->header($k);
125 $buf .= "$k: $v$crlf" if defined $v;
129 # mboxrd quoting style
130 # ref: http://www.qmail.org/man/man5/mbox.html
131 my $body = $simple->body;
132 $body =~ s/^(>*From )/>$1/gm;
138 my ($ctx, $srch, $sfx) = @_;
139 eval { require IO::Compress::Gzip };
140 return sub { need_gzip(@_) } if $@;
141 my $mid = $ctx->{mid};
142 my $msgs = $srch->get_thread($mid, 0);
143 return [404, [qw(Content-Type text/plain)], []] if !@$msgs;
144 my $prev = $msgs->[-1]->{num};
148 if (my $smsg = $msgs->[$i++]) {
152 $msgs = $srch->get_thread($mid, $prev);
153 return unless @$msgs;
154 $prev = $msgs->[-1]->{num};
158 PublicInbox::MboxGz->response($ctx, $cb, $msgs->[0]->subject);
162 my ($ctx, $range) = @_;
165 if ($range eq 'all') { # TODO: YYYY[-MM]
168 return [404, [qw(Content-Type text/plain)], []];
170 mbox_all($ctx, $query);
176 my $ids = $ctx->{-inbox}->mm->ids_after(\$prev) or return
177 [404, [qw(Content-Type text/plain)], ["No results found\n"]];
179 my $over = $ctx->{srch}->{over_ro};
182 while ((my $num = $ids->[$i++])) {
183 my $smsg = $over->get_art($num) or next;
186 $ids = $ctx->{-inbox}->mm->ids_after(\$prev);
191 return PublicInbox::MboxGz->response($ctx, $cb, 'all');
195 my ($ctx, $query) = @_;
197 eval { require IO::Compress::Gzip };
198 return sub { need_gzip(@_) } if $@;
199 return mbox_all_ids($ctx) if $query eq '';
200 my $opts = { mset => 2 };
201 my $srch = $ctx->{srch};
202 my $mset = $srch->query($query, $opts);
203 $opts->{offset} = $mset->size or
204 return [404, [qw(Content-Type text/plain)],
205 ["No results found\n"]];
207 my $cb = sub { # called by MboxGz->getline
209 while (my $mi = (($mset->items)[$i++])) {
210 my $doc = $mi->get_document;
211 my $smsg = $srch->retry_reopen(sub {
212 PublicInbox::SearchMsg->load_doc($doc);
217 $mset = $srch->query($query, $opts);
218 my $size = $mset->size or return;
219 $opts->{offset} += $size;
223 PublicInbox::MboxGz->response($ctx, $cb, 'results-'.$query);
227 my $fh = $_[0]->([501, ['Content-Type' => 'text/html']]);
228 my $title = 'gzipped mbox not available';
230 <html><head><title>$title</title><body><pre>$title
231 The administrator needs to install the IO::Compress::Gzip Perl module
232 to support gzipped mboxes.
233 <a href="../">Return to index</a></pre></body></html>
240 package PublicInbox::MboxGz;
243 use PublicInbox::Hval qw/to_filename/;
246 my ($class, $ctx, $cb) = @_;
250 gz => IO::Compress::Gzip->new(\$buf, Time => 0),
257 my ($class, $ctx, $cb, $fn) = @_;
258 my $body = $class->new($ctx, $cb);
259 # http://www.iana.org/assignments/media-types/application/gzip
260 my @h = qw(Content-Type application/gzip);
262 $fn = to_filename($fn);
263 push @h, 'Content-Disposition', "inline; filename=$fn.mbox.gz";
268 # called by Plack::Util::foreach or similar
271 my $ctx = $self->{ctx} or return;
272 while (my $smsg = $self->{cb}->()) {
273 my $msg = $ctx->{-inbox}->msg_by_smsg($smsg) or next;
274 $msg = Email::Simple->new($msg);
275 $self->{gz}->write(PublicInbox::Mbox::msg_str($ctx, $msg,
277 my $bref = $self->{buf};
278 if (length($$bref) >= 8192) {
279 my $ret = $$bref; # copy :<
280 ${$self->{buf}} = '';
284 # be fair to other clients on public-inbox-httpd:
287 delete($self->{gz})->close;
288 # signal that we're done and can return undef next call:
290 ${delete $self->{buf}};