mbox: support /$INBOX/all.mbox.gz endpoint

[public-inbox.git] / lib / PublicInbox / Mbox.pm

diff --git a/lib/PublicInbox/Mbox.pm b/lib/PublicInbox/Mbox.pm
index b81ec0a40e785dff75796fd573dd2e1a699e94c8..4c4b74fbc50c671e9ec201b8852b03b0686a8206 100644 (file)
--- a/lib/PublicInbox/Mbox.pm
+++ b/lib/PublicInbox/Mbox.pm
@@ -48,12 +48,16 @@ sub emit_msg {
        my $base = $feed_opts->{url};
        my $mid = mid_clean($header_obj->header('Message-ID'));
        $mid = uri_escape_utf8($mid);
-       my %append = (
-               'Archived-At' => "<$base$mid/>",
-               'List-Archive' => "<$base>",
-               'List-Post' => "<mailto:$feed_opts->{id_addr}>",
+       my @append = (
+               'Archived-At', "<$base$mid/>",
+               'List-Archive', "<$base>",
+               'List-Post', "<mailto:$feed_opts->{id_addr}>",
        );
-       while (my ($k, $v) = each %append) {
+       my $append = '';
+       my $crlf = $simple->crlf;
+       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 $_) {
@@ -61,14 +65,14 @@ sub emit_msg {
                                last;
                        }
                }
-               $header_obj->header_set($k, @v, $v) if defined $v;
+               $append .= "$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");
        }
-       $fh->write($buf .= $simple->crlf);
+       $append .= $crlf;
+       $fh->write($buf .= $append);
 
        $buf = $simple->body;
        $simple->body_set('');
@@ -117,6 +121,69 @@ sub emit_mbox {
        $fh->close;
 }
 
+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;
+       }
+
+       # 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;
+       };
+
+       $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;
+                       }
+               };
+               $io->write($next); # Danga::Socket::write
+               return;
+       }
+       $cb->() while ($nr > 0);
+       $fh->close;
+}
+
 sub need_gzip {
        my $fh = $_[0]->([501, ['Content-Type' => 'text/html']]);
        my $title = 'gzipped mbox not available';