diff options
Diffstat (limited to 'lib/PublicInbox/Mbox.pm')
| -rw-r--r-- | lib/PublicInbox/Mbox.pm | 192 |
1 files changed, 93 insertions, 99 deletions
diff --git a/lib/PublicInbox/Mbox.pm b/lib/PublicInbox/Mbox.pm index 0d67981f..2565ea58 100644 --- a/lib/PublicInbox/Mbox.pm +++ b/lib/PublicInbox/Mbox.pm @@ -6,32 +6,18 @@ package PublicInbox::Mbox; use strict; use warnings; -use PublicInbox::MID qw/mid2path mid_clean/; -use URI::Escape qw/uri_escape_utf8/; +use PublicInbox::MID qw/mid_clean mid_escape/; require Email::Simple; -sub thread_mbox { - my ($ctx, $srch, $sfx) = @_; - sub { - my ($response) = @_; # Plack callback - emit_mbox($response, $ctx, $srch, $sfx); - } -} - 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); - $fh->close; - } + my ($ctx, $msg) = @_; + $msg = Email::Simple->new($msg); + # single message should be easily renderable in browsers + [200, ['Content-Type', 'text/plain'], [ msg_str($ctx, $msg)] ] } -sub emit_msg { - my ($ctx, $fh, $simple) = @_; # Email::Simple object +sub msg_str { + my ($ctx, $simple) = @_; # Email::Simple object my $header_obj = $simple->header_obj; # drop potentially confusing headers, ssoma already should've dropped @@ -39,74 +25,67 @@ 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 $ibx = $ctx->{-inbox}; + my $base = $ibx->base_url($ctx->{env}); my $mid = mid_clean($header_obj->header('Message-ID')); - $mid = uri_escape_utf8($mid); - my @archived_at = $header_obj->header('Archived-At'); - push @archived_at, "<$base$mid/>"; - $header_obj->header_set('Archived-At', @archived_at); - $header_obj->header_set('List-Archive', "<$base>"); - $header_obj->header_set('List-Post', "<mailto:$feed_opts->{id_addr}>"); - - my $buf = $header_obj->as_string; - unless ($buf =~ /\AFrom /) { - $fh->write("From mboxrd\@z Thu Jan 1 00:00:00 1970\n"); + $mid = mid_escape($mid); + my @append = ( + 'Archived-At', "<$base$mid/>", + 'List-Archive', "<$base>", + 'List-Post', "<mailto:$ibx->{-primary_address}>", + ); + 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]; + my @v = $header_obj->header($k); + foreach (@v) { + if ($v eq $_) { + $v = undef; + last; + } + } + $buf .= "$k: $v$crlf" if defined $v; } - $fh->write($buf .= $simple->crlf); - - $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; + my $body = $simple->body; + $body =~ s/^(>*From )/>$1/gm; + $buf .= $body; + $buf .= "\n"; +} - $buf .= "\n" unless $buf =~ /\n\z/s; +sub thread_mbox { + my ($ctx, $srch, $sfx) = @_; + eval { require IO::Compress::Gzip }; + return sub { need_gzip(@_) } if $@; - $fh->write($buf); + my $cb = sub { $srch->get_thread($ctx->{mid}, @_) }; + # http://www.iana.org/assignments/media-types/application/gzip + [200, ['Content-Type' => 'application/gzip'], + PublicInbox::MboxGz->new($ctx, $cb) ]; } -sub emit_mbox { - my ($response, $ctx, $srch, $sfx) = @_; - my $type = 'mbox'; - if ($sfx) { - eval { require IO::Compress::Gzip }; - return need_gzip($response) if $@; - $type = 'gzip'; +sub emit_range { + my ($ctx, $range) = @_; + + eval { require IO::Compress::Gzip }; + return sub { need_gzip(@_) } if $@; + my $query; + if ($range eq 'all') { # TODO: YYYY[-MM] + $query = ''; + } else { + return [404, [qw(Content-Type text/plain)], []]; } + my $cb = sub { $ctx->{srch}->query($query, @_) }; # 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); - 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); - - $fh->close; + [200, [qw(Content-Type application/gzip)], + PublicInbox::MboxGz->new($ctx, $cb) ]; } sub need_gzip { @@ -123,40 +102,55 @@ EOF 1; -# fh may not be a proper IO, so we wrap the write and close methods -# to prevent IO::Compress::Gzip from complaining package PublicInbox::MboxGz; use strict; use warnings; sub new { - my ($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 => [], + opts => { offset => 0 }, }, $class; } -sub _flush_buf { +# called by Plack::Util::foreach or similar +sub getline { my ($self) = @_; - if (defined ${$self->{buf}}) { - $self->{fh}->write(${$self->{buf}}); - ${$self->{buf}} = undef; - } -} - -sub write { - $_[0]->{gz}->write($_[1]); - _flush_buf($_[0]); + my $ctx = $self->{ctx} or return; + my $res; + my $ibx = $ctx->{-inbox}; + my $gz = $self->{gz}; + do { + while (defined(my $smsg = shift @{$self->{msgs}})) { + my $msg = eval { $ibx->msg_by_smsg($smsg) } or next; + $msg = Email::Simple->new($msg); + $gz->write(PublicInbox::Mbox::msg_str($ctx, $msg)); + 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 ''; + } + $res = $self->{cb}->($self->{opts}); + $self->{msgs} = $res->{msgs}; + $res = scalar @{$self->{msgs}}; + $self->{opts}->{offset} += $res; + } while ($res); + $gz->close; + delete $self->{ctx}; + ${delete $self->{buf}}; } -sub close { - my ($self) = @_; - $self->{gz}->close; - _flush_buf($self); - $self->{fh}->close; -} +sub close {} # noop 1; |