X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FMbox.pm;h=0be1968584d9c68a7a76a72b98f82afea13bc564;hb=0dceebd0a85774c92af247e6da5e2f5a0ee8417c;hp=4c4b74fbc50c671e9ec201b8852b03b0686a8206;hpb=1269631ae596b910b7c26890c3efa9f85301a89b;p=public-inbox.git diff --git a/lib/PublicInbox/Mbox.pm b/lib/PublicInbox/Mbox.pm index 4c4b74fb..0be19685 100644 --- a/lib/PublicInbox/Mbox.pm +++ b/lib/PublicInbox/Mbox.pm @@ -1,37 +1,98 @@ -# Copyright (C) 2015 all contributors -# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt) +# Copyright (C) 2015-2018 all contributors +# License: AGPL-3.0+ # Streaming interface for formatting messages as an mboxrd. # Used by the web interface package PublicInbox::Mbox; use strict; use warnings; -use PublicInbox::MID qw/mid2path mid_clean/; -use URI::Escape qw/uri_escape_utf8/; -require Email::Simple; +use PublicInbox::MID qw/mid_clean mid_escape/; +use PublicInbox::Hval qw/to_filename/; +use Email::Simple; +use Email::MIME::Encode; -sub thread_mbox { - my ($ctx, $srch, $sfx) = @_; - sub { - my ($response) = @_; # Plack callback - emit_mbox($response, $ctx, $srch, $sfx); +sub subject_fn ($) { + my ($simple) = @_; + my $fn = $simple->header('Subject'); + return 'no-subject' unless defined($fn); + + # no need for full Email::MIME, here + if ($fn =~ /=\?/) { + eval { $fn = Encode::decode('MIME-Header', $fn) }; + $fn = 'no-subject' if $@; } + $fn =~ s/^re:\s+//i; + $fn = to_filename($fn); + $fn eq '' ? 'no-subject' : $fn; +} + +sub mb_stream { + my ($more) = @_; + bless $more, 'PublicInbox::Mbox'; } -sub emit1 { - my $simple = Email::Simple->new(pop); - my $ctx = pop; - sub { - my ($response) = @_; - # single message should be easily renderable in browsers - my $fh = $response->([200, ['Content-Type'=>'text/plain']]); - emit_msg($ctx, $fh, $simple); - $fh->close; +# called by PSGI server as body response +sub getline { + my ($more) = @_; # self + my ($ctx, $head, $tail, $db, $cur) = @$more; + if ($cur) { + pop @$more; + return msg_str($ctx, $cur); } + for (; !defined($cur) && $head != $tail; $head++) { + my $smsg = PublicInbox::SearchMsg->get($head, $db, $ctx->{mid}); + my $mref = $ctx->{-inbox}->msg_by_smsg($smsg) or next; + $cur = Email::Simple->new($mref); + $cur = msg_str($ctx, $cur); + } + $more->[1] = $head; + $cur; } -sub emit_msg { - my ($ctx, $fh, $simple) = @_; # Email::Simple object +sub close {} # noop + +sub emit_raw { + my ($ctx) = @_; + my $mid = $ctx->{mid}; + my $ibx = $ctx->{-inbox}; + my $first; + my $more; + my ($head, $tail, $db); + my %seen; + if (my $srch = $ibx->search) { + $srch->retry_reopen(sub { + ($head, $tail, $db) = $srch->each_smsg_by_mid($mid); + for (; !defined($first) && $head != $tail; $head++) { + my @args = ($head, $db, $mid); + my $smsg = PublicInbox::SearchMsg->get(@args); + my $mref = $ibx->msg_by_smsg($smsg) or next; + $first = Email::Simple->new($mref); + } + if ($head != $tail) { + $more = [ $ctx, $head, $tail, $db, $first ]; + } + }); + } else { + my $mref = $ibx->msg_by_mid($mid) or return; + $first = Email::Simple->new($mref); + } + return unless defined $first; + my $fn = subject_fn($first); + my @hdr = ('Content-Type'); + if ($ibx->{obfuscate}) { + # obfuscation is stupid, but maybe scrapers are, too... + push @hdr, 'application/mbox'; + $fn .= '.mbox'; + } else { + push @hdr, 'text/plain'; + $fn .= '.txt'; + } + push @hdr, 'Content-Disposition', "inline; filename=$fn"; + [ 200, \@hdr, $more ? mb_stream($more) : [ msg_str($ctx, $first) ] ]; +} + +sub msg_str { + my ($ctx, $simple, $mid) = @_; # Email::Simple object my $header_obj = $simple->header_obj; # drop potentially confusing headers, ssoma already should've dropped @@ -39,22 +100,18 @@ sub emit_msg { foreach my $d (qw(Lines Bytes Content-Length Status)) { $header_obj->header_set($d); } - my $feed_opts = $ctx->{feed_opts}; - unless ($feed_opts) { - require PublicInbox::Feed; # FIXME: gross - $feed_opts = PublicInbox::Feed::get_feedopts($ctx); - $ctx->{feed_opts} = $feed_opts; - } - my $base = $feed_opts->{url}; - my $mid = mid_clean($header_obj->header('Message-ID')); - $mid = uri_escape_utf8($mid); + my $ibx = $ctx->{-inbox}; + my $base = $ibx->base_url($ctx->{env}); + $mid = $ctx->{mid} unless defined $mid; + $mid = mid_escape($mid); my @append = ( 'Archived-At', "<$base$mid/>", 'List-Archive', "<$base>", - 'List-Post', "{id_addr}>", + 'List-Post', "{-primary_address}>", ); - my $append = ''; my $crlf = $simple->crlf; + my $buf = "From mboxrd\@z Thu Jan 1 00:00:00 1970\n" . + $header_obj->as_string; for (my $i = 0; $i < @append; $i += 2) { my $k = $append[$i]; my $v = $append[$i + 1]; @@ -65,123 +122,68 @@ sub emit_msg { last; } } - $append .= "$k: $v$crlf" if defined $v; + $buf .= "$k: $v$crlf" if defined $v; } - my $buf = $header_obj->as_string; - unless ($buf =~ /\AFrom /) { - $fh->write("From mboxrd\@z Thu Jan 1 00:00:00 1970\n"); - } - $append .= $crlf; - $fh->write($buf .= $append); - - $buf = $simple->body; - $simple->body_set(''); + $buf .= $crlf; # mboxrd quoting style # ref: http://www.qmail.org/man/man5/mbox.html - $buf =~ s/^(>*From )/>$1/gm; - - $fh->write($buf .= "\n"); + my $body = $simple->body; + $body =~ s/^(>*From )/>$1/gm; + $buf .= $body; + $buf .= "\n"; } -sub emit_mbox { - my ($response, $ctx, $srch, $sfx) = @_; - my $type = 'mbox'; - if ($sfx) { - eval { require IO::Compress::Gzip }; - return need_gzip($response) if $@; - $type = 'gzip'; - } - - # http://www.iana.org/assignments/media-types/application/gzip - # http://www.iana.org/assignments/media-types/application/mbox - my $fh = $response->([200, ['Content-Type' => "application/$type"]]); - $fh = PublicInbox::MboxGz->new($fh) if $sfx; - - require PublicInbox::Git; - my $mid = $ctx->{mid}; - my $git = $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir}); - my %opts = (offset => 0, asc => 1); - my $nr; - do { - my $res = $srch->get_thread($mid, \%opts); - my $msgs = $res->{msgs}; - $nr = scalar @$msgs; - while (defined(my $smsg = shift @$msgs)) { - my $msg = eval { - my $p = 'HEAD:'.mid2path($smsg->mid); - Email::Simple->new($git->cat_file($p)); - }; - emit_msg($ctx, $fh, $msg) if $msg; - } - - $opts{offset} += $nr; - } while ($nr > 0); - - $fh->close; +sub thread_mbox { + my ($ctx, $srch, $sfx) = @_; + eval { require IO::Compress::Gzip }; + return sub { need_gzip(@_) } if $@; + my $prev = 0; + my $cb = sub { + my $msgs = $srch->get_thread($ctx->{mid}, $prev); + $prev = $msgs->[-1]->{num} if scalar(@$msgs); + $msgs; + }; + PublicInbox::MboxGz->response($ctx, $cb); } sub emit_range { my ($ctx, $range) = @_; - sub { _emit_range($_[0], $ctx, $range) }; -} - -sub _emit_range { - my ($res, $ctx, $range) = @_; - eval { require IO::Compress::Gzip }; - return need_gzip($res) if $@; my $query; if ($range eq 'all') { # TODO: YYYY[-MM] $query = ''; } else { - $res->([404, [qw(Content-Type text/plain)], []]); - return; + return [404, [qw(Content-Type text/plain)], []]; } + mbox_all($ctx, $query); +} - # http://www.iana.org/assignments/media-types/application/gzip - my $fh = $res->([200, [qw(Content-Type application/gzip)]]); - $fh = PublicInbox::MboxGz->new($fh); - my $env = $ctx->{cgi}->env; - my $srch = $ctx->{srch}; - my $git = $ctx->{git}; - my %opts = (offset => 0, asc => 1); - my $nr; - my $cb = sub { - my $res = $srch->query($query, \%opts); - my $msgs = $res->{msgs}; - $nr = scalar @$msgs; - while (defined(my $smsg = shift @$msgs)) { - my $msg = eval { - my $p = 'HEAD:'.mid2path($smsg->mid); - Email::Simple->new($git->cat_file($p)); - }; - emit_msg($ctx, $fh, $msg) if $msg; - } - - $opts{offset} += $nr; - }; +sub mbox_all { + my ($ctx, $query) = @_; - $cb->(); # first part is free - return $fh->close if $nr == 0; - - if ($env->{'pi-httpd.async'}) { - my $io = $env->{'psgix.io'} or die "no IO"; - my $next; - $next = sub { - $cb->(); - if ($nr > 0) { - $io->write($next); - } else { - $next = undef; - $fh->close; - } + eval { require IO::Compress::Gzip }; + return sub { need_gzip(@_) } if $@; + if ($query eq '') { + my $prev = 0; + my $msgs = []; + my $cb = sub { + $ctx->{-inbox}->mm->id_batch($prev, sub { + $msgs = $_[0]; + }); + $prev = $msgs->[-1] if @$msgs; + $msgs; }; - $io->write($next); # Danga::Socket::write - return; + return PublicInbox::MboxGz->response($ctx, $cb, 'all'); } - $cb->() while ($nr > 0); - $fh->close; + my $opts = { offset => 0 }; + my $srch = $ctx->{srch}; + my $cb = sub { # called by MboxGz->getline + my $msgs = $srch->query($query, $opts); + $opts->{offset} += scalar @$msgs; + $msgs; + }; + PublicInbox::MboxGz->response($ctx, $cb, 'results-'.$query); } sub need_gzip { @@ -198,40 +200,86 @@ EOF 1; -# fh may not be a proper IO, so we wrap the write and close methods -# to prevent IO::Compress::Gzip from complaining package PublicInbox::MboxGz; use strict; use warnings; +use PublicInbox::Hval qw/to_filename/; sub new { - my ($class, $fh) = @_; - my $buf; + my ($class, $ctx, $cb) = @_; + my $buf = ''; bless { buf => \$buf, - gz => IO::Compress::Gzip->new(\$buf), - fh => $fh, + gz => IO::Compress::Gzip->new(\$buf, Time => 0), + cb => $cb, + ctx => $ctx, + msgs => [], }, $class; } -sub _flush_buf { - my ($self) = @_; - if (defined ${$self->{buf}}) { - $self->{fh}->write(${$self->{buf}}); - ${$self->{buf}} = undef; - } +sub response { + my ($class, $ctx, $cb, $fn) = @_; + my $body = $class->new($ctx, $cb); + # http://www.iana.org/assignments/media-types/application/gzip + $body->{hdr} = [ 'Content-Type', 'application/gzip' ]; + $body->{fn} = $fn; + my $hdr = $body->getline; # fill in Content-Disposition filename + [ 200, $hdr, $body ]; } -sub write { - $_[0]->{gz}->write($_[1]); - _flush_buf($_[0]); +sub set_filename ($$) { + my ($fn, $msg) = @_; + return to_filename($fn) if defined($fn); + + PublicInbox::Mbox::subject_fn($msg); } -sub close { +# called by Plack::Util::foreach or similar +sub getline { my ($self) = @_; - $self->{gz}->close; - _flush_buf($self); - $self->{fh}->close; + my $ctx = $self->{ctx} or return; + my $ibx = $ctx->{-inbox}; + my $gz = $self->{gz}; + my $msgs = $self->{msgs}; + do { + # work on existing result set + while (defined(my $smsg = shift @$msgs)) { + # id_batch may return integers + ref($smsg) or + $smsg = $ctx->{srch}->{over_ro}->get_art($smsg); + + my $msg = eval { $ibx->msg_by_smsg($smsg) } or next; + $msg = Email::Simple->new($msg); + $gz->write(PublicInbox::Mbox::msg_str($ctx, $msg, + $smsg->mid)); + + # use subject of first message as subject + if (my $hdr = delete $self->{hdr}) { + my $fn = set_filename($self->{fn}, $msg); + push @$hdr, 'Content-Disposition', + "inline; filename=$fn.mbox.gz"; + return $hdr; + } + my $bref = $self->{buf}; + if (length($$bref) >= 8192) { + my $ret = $$bref; # copy :< + ${$self->{buf}} = ''; + return $ret; + } + + # be fair to other clients on public-inbox-httpd: + return ''; + } + + # refill result set + $msgs = $self->{msgs} = $self->{cb}->(); + } while (@$msgs); + $gz->close; + # signal that we're done and can return undef next call: + delete $self->{ctx}; + ${delete $self->{buf}}; } +sub close {} # noop + 1;