-# Copyright (C) 2013, Eric Wong <normalperson@yhbt.net> and all contributors
+# Copyright (C) 2013-2015 all contributors <meta@public-inbox.org>
# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
+#
+# Used for generating Atom feeds for web-accessible mailing list archives.
package PublicInbox::Feed;
use strict;
use warnings;
-use Email::Address;
use Email::MIME;
-use Date::Parse qw(strptime str2time);
-use PublicInbox::Hval;
-use PublicInbox::GitCatFile;
+use Date::Parse qw(strptime);
+use PublicInbox::Hval qw/ascii_html/;
+use PublicInbox::Git;
+use PublicInbox::View;
+use PublicInbox::MID qw/mid_clean mid2path/;
+use PublicInbox::Address;
+use POSIX qw/strftime/;
use constant {
- DATEFMT => '%Y-%m-%dT%H:%M:%SZ',
- MAX_PER_PAGE => 25,
+ DATEFMT => '%Y-%m-%dT%H:%M:%SZ', # Atom standard
+ MAX_PER_PAGE => 25, # this needs to be tunable
};
# main function
sub generate {
- my ($class, $args) = @_;
- require XML::Atom::SimpleFeed;
- require PublicInbox::View;
- require POSIX;
- my $max = $args->{max} || MAX_PER_PAGE;
-
- my $feed_opts = get_feedopts($args);
- my $addr = $feed_opts->{address};
- $addr = $addr->[0] if ref($addr);
- my $feed = XML::Atom::SimpleFeed->new(
- title => $feed_opts->{description} || "unnamed feed",
- link => $feed_opts->{url} || "http://example.com/",
- link => {
- rel => 'self',
- href => $feed_opts->{atomurl} ||
- "http://example.com/atom.xml",
- },
- id => 'mailto:' . ($addr || 'public-inbox@example.com'),
- updated => POSIX::strftime(DATEFMT, gmtime),
- );
-
- my $git = PublicInbox::GitCatFile->new($args->{git_dir});
- each_recent_blob($args, sub {
- my ($add) = @_;
- add_to_feed($feed_opts, $feed, $add, $git);
- });
- $git = undef; # destroy pipes
- Email::Address->purge_cache;
- $feed->as_string;
+ my ($ctx) = @_;
+ sub { emit_atom($_[0], $ctx) };
+}
+
+sub generate_thread_atom {
+ my ($ctx) = @_;
+ sub { emit_atom_thread($_[0], $ctx) };
}
sub generate_html_index {
- my ($class, $args) = @_;
- require PublicInbox::Thread;
+ my ($ctx) = @_;
+ # if the 'r' query parameter is given, it is a legacy permalink
+ # which we must continue supporting:
+ my $qp = $ctx->{qp};
+ if ($qp && !$qp->{r} && $ctx->{srch}) {
+ return PublicInbox::View::index_topics($ctx);
+ }
+
+ my $env = $ctx->{env};
+ my $url = $ctx->{-inbox}->base_url($env) . 'new.html';
+ my $qs = $env->{QUERY_STRING};
+ $url .= "?$qs" if $qs ne '';
+ [302, [ 'Location', $url, 'Content-Type', 'text/plain'],
+ [ "Redirecting to $url\n" ] ];
+}
+
+sub new_html {
+ my ($ctx) = @_;
+ my @paths;
+ my (undef, $last) = each_recent_blob($ctx, sub {
+ my ($path, $commit, $ts, $u, $subj) = @_;
+ $ctx->{first} ||= $commit;
+ push @paths, $path;
+ });
+ if (!@paths) {
+ return [404, ['Content-Type', 'text/plain'],
+ ["No messages, yet\n"] ];
+ }
+ $ctx->{-html_tip} = '<pre>';
+ $ctx->{-upfx} = '';
+ PublicInbox::WwwStream->response($ctx, 200, sub {
+ while (my $path = shift @paths) {
+ my $m = do_cat_mail($ctx->{-inbox}, $path) or next;
+ my $more = scalar @paths;
+ my $s = PublicInbox::View::index_entry($m, $ctx, $more);
+ $s .= '</pre>' unless $more;
+ return $s;
+ }
+ new_html_footer($ctx, $last);
+ });
+}
+
+# private subs
+
+sub title_tag {
+ my ($title) = @_;
+ $title =~ tr/\t\n / /s; # squeeze spaces
+ # try to avoid the type attribute in title:
+ $title = ascii_html($title);
+ my $type = index($title, '&') >= 0 ? "\ntype=\"html\"" : '';
+ "<title$type>$title</title>";
+}
- my $max = $args->{max} || MAX_PER_PAGE;
- my $feed_opts = get_feedopts($args);
+sub atom_header {
+ my ($feed_opts, $title) = @_;
- my $title = $feed_opts->{description} || '';
- $title = PublicInbox::Hval->new_oneline($title)->as_html;
+ $title = title_tag($feed_opts->{description}) unless (defined $title);
- my @messages;
- my $git = PublicInbox::GitCatFile->new($args->{git_dir});
- my $last = each_recent_blob($args, sub {
- my $mime = do_cat_mail($git, $_[0]) or return 0;
- $mime->body_set(''); # save some memory
+ qq(<?xml version="1.0" encoding="us-ascii"?>\n) .
+ qq{<feed\nxmlns="http://www.w3.org/2005/Atom">} .
+ qq{$title} .
+ qq(<link\nrel="alternate"\ntype="text/html") .
+ qq(\nhref="$feed_opts->{url}"/>) .
+ qq(<link\nrel="self"\nhref="$feed_opts->{atomurl}"/>) .
+ qq(<id>mailto:$feed_opts->{id_addr}</id>);
+}
- my $t = eval { str2time($mime->header('Date')) };
- defined($t) or $t = 0;
- $mime->header_set('X-PI-Date', $t);
- push @messages, $mime;
+sub emit_atom {
+ my ($cb, $ctx) = @_;
+ my $feed_opts = get_feedopts($ctx);
+ my $fh = $cb->([ 200, ['Content-Type' => 'application/atom+xml']]);
+ my $max = $ctx->{max} || MAX_PER_PAGE;
+ my $x = atom_header($feed_opts);
+ my $ibx = $ctx->{-inbox};
+ each_recent_blob($ctx, sub {
+ my ($path, undef, $ts) = @_;
+ if (defined $x) {
+ $fh->write($x . feed_updated(undef, $ts));
+ $x = undef;
+ }
+ my $s = feed_entry($feed_opts, $path, $ibx) or return 0;
+ $fh->write($s);
1;
});
- $git = undef; # destroy pipes.
-
- my $th = PublicInbox::Thread->new(@messages);
- $th->thread;
- my $html = "<html><head><title>$title</title>" .
- '<link rel="alternate" title="Atom feed" href="' .
- $feed_opts->{atomurl} . '" type="application/atom+xml"/>' .
- '</head><body><pre>';
-
- # sort by date, most recent at top
- $th->order(sub {
- sort {
- $b->topmost->message->header('X-PI-Date') <=>
- $a->topmost->message->header('X-PI-Date')
- } @_;
- });
- dump_html_line($_, 0, \$html) for $th->rootset;
+ end_feed($fh);
+}
- Email::Address->purge_cache;
+sub _no_thread {
+ my ($cb) = @_;
+ $cb->([404, ['Content-Type', 'text/plain'],
+ ["No feed found for thread\n"]]);
+}
- my $footer = nav_footer($args->{cgi}, $last);
- $footer = "<hr /><pre>$footer</pre>" if $footer;
- $html . "</pre>$footer</html>";
+sub end_feed {
+ my ($fh) = @_;
+ $fh->write('</feed>');
+ $fh->close;
}
-# private subs
+sub emit_atom_thread {
+ my ($cb, $ctx) = @_;
+ my $mid = $ctx->{mid};
+ my $res = $ctx->{srch}->get_thread($mid);
+ return _no_thread($cb) unless $res->{total};
+ my $feed_opts = get_feedopts($ctx);
+ my $fh = $cb->([200, ['Content-Type' => 'application/atom+xml']]);
+ my $ibx = $ctx->{-inbox};
+ my $html_url = $ibx->base_url($ctx->{env});
+ $html_url .= PublicInbox::Hval->new_msgid($mid)->as_href;
+
+ $feed_opts->{url} = $html_url;
+ $feed_opts->{emit_header} = 1;
+
+ foreach my $msg (@{$res->{msgs}}) {
+ my $s = feed_entry($feed_opts, mid2path($msg->mid), $ibx);
+ $fh->write($s) if defined $s;
+ }
+ end_feed($fh);
+}
+
+sub _html_index_top {
+ my ($feed_opts, $srch) = @_;
-sub nav_footer {
- my ($cgi, $last) = @_;
- $cgi or return '';
- my $old_r = $cgi->param('r');
- my $head = ' ';
+ my $title = ascii_html($feed_opts->{description} || '');
+ my $top = "<b>$title</b> (<a\nhref=\"new.atom\">Atom feed</a>)";
+ if ($srch) {
+ $top = qq{<form\naction=""><pre>$top} .
+ qq{ <input\nname=q\ntype=text />} .
+ qq{<input\ntype=submit\nvalue=search />} .
+ q{</pre></form><pre>}
+ } else {
+ $top = '<pre>' . $top . "\n";
+ }
+
+ "<html><head><title>$title</title>" .
+ "<link\nrel=alternate\ntitle=\"Atom feed\"\n".
+ "href=\"new.atom\"\ntype=\"application/atom+xml\"/>" .
+ PublicInbox::Hval::STYLE .
+ "</head><body>$top";
+}
+
+sub new_html_footer {
+ my ($ctx, $last) = @_;
+ my $qp = delete $ctx->{qp} or return;
+ my $old_r = $qp->{r};
+ my $latest = '';
my $next = ' ';
if ($last) {
- $next = qq!<a href="?r=$last">next</a>!;
+ $next = qq!<a\nhref="?r=$last"\nrel=next>next</a>!;
}
if ($old_r) {
- $head = $cgi->path_info;
- $head = qq!<a href="$head">head</a>!;
+ $latest = qq! <a\nhref='./new.html'>latest</a>!;
}
- "$next $head";
+ "<hr><pre>page: $next$latest</pre>";
}
sub each_recent_blob {
- my ($args, $cb) = @_;
- my $max = $args->{max} || MAX_PER_PAGE;
+ my ($ctx, $cb) = @_;
+ my $max = $ctx->{max} || MAX_PER_PAGE;
my $hex = '[a-f0-9]';
my $addmsg = qr!^:000000 100644 \S+ \S+ A\t(${hex}{2}/${hex}{38})$!;
my $delmsg = qr!^:100644 000000 \S+ \S+ D\t(${hex}{2}/${hex}{38})$!;
- my $refhex = qr/${hex}{4,40}(?:~\d+)?/;
- my $cgi = $args->{cgi};
+ my $refhex = qr/(?:HEAD|${hex}{4,40})(?:~\d+)?/;
+ my $qp = $ctx->{qp};
# revision ranges may be specified
my $range = 'HEAD';
- my $r = $cgi->param('r') if $cgi;
+ my $r = $qp->{r} if $qp;
if ($r && ($r =~ /\A(?:$refhex\.\.)?$refhex\z/o)) {
$range = $r;
}
# get recent messages
# we could use git log -z, but, we already know ssoma will not
# leave us with filenames with spaces in them..
- my @cmd = ('git', "--git-dir=$args->{git_dir}",
- qw/log --no-notes --no-color --raw -r/);
- push @cmd, $range;
-
- my $pid = open(my $log, '-|', @cmd) or
- die('open `'.join(' ', @cmd) . " pipe failed: $!\n");
+ my $log = $ctx->{-inbox}->git->popen(qw/log
+ --no-notes --no-color --raw -r
+ --abbrev=16 --abbrev-commit/,
+ "--format=%h%x00%ct%x00%an%x00%s%x00",
+ $range);
my %deleted; # only an optimization at this point
my $last;
my $nr = 0;
- my @commits = ();
- while (my $line = <$log>) {
+ my ($cur_commit, $first_commit, $last_commit);
+ my ($ts, $subj, $u);
+ local $/ = "\n";
+ while (defined(my $line = <$log>)) {
if ($line =~ /$addmsg/o) {
my $add = $1;
- next if $deleted{$add};
- $nr += $cb->($add);
+ next if $deleted{$add}; # optimization-only
+ $nr += $cb->($add, $cur_commit, $ts, $u, $subj);
if ($nr >= $max) {
$last = 1;
last;
}
} elsif ($line =~ /$delmsg/o) {
$deleted{$1} = 1;
- } elsif ($line =~ /^commit (${hex}{40})/) {
- push @commits, $1;
+ } elsif ($line =~ /^${hex}{7,40}/o) {
+ ($cur_commit, $ts, $u, $subj) = split("\0", $line);
+ unless (defined $first_commit) {
+ $first_commit = $cur_commit;
+ }
}
}
if ($last) {
+ local $/ = "\n";
while (my $line = <$log>) {
- if ($line =~ /^commit (${hex}{40})/) {
- push @commits, $1;
+ if ($line =~ /^(${hex}{7,40})/o) {
+ $last_commit = $1;
last;
}
}
- } else {
- push @commits, undef;
}
- close $log; # we may EPIPE here
# for pagination
- $commits[-1];
+ ($first_commit, $last_commit);
}
# private functions below
sub get_feedopts {
- my ($args) = @_;
- my $pi_config = $args->{pi_config};
- my $listname = $args->{listname};
- my $cgi = $args->{cgi};
- my %rv;
- if (open my $fh, '<', "$args->{git_dir}/description") {
- chomp($rv{description} = <$fh>);
- close $fh;
- }
-
- if ($pi_config && defined $listname && length $listname) {
- foreach my $key (qw(address)) {
- $rv{$key} = $pi_config->get($listname, $key) || "";
- }
- }
- my $url_base;
- if ($cgi) {
- my $path_info = $cgi->path_info;
- my $base;
- if (ref($cgi) eq 'CGI') {
- $base = $cgi->url(-base);
- } else {
- $base = $cgi->base->as_string;
- $base =~ s!/\z!!;
- }
- $url_base = $path_info;
- if ($url_base =~ s!/(?:|index\.html)?\z!!) {
- $rv{atomurl} = "$base$url_base/atom.xml";
- } else {
- $url_base =~ s!/atom\.xml\z!!;
- $rv{atomurl} = $base . $path_info;
- $url_base = $base . $url_base; # XXX is this needed?
- }
+ my ($ctx) = @_;
+ my $inbox = $ctx->{inbox};
+ my $obj = $ctx->{-inbox};
+ my %rv = ( description => $obj->description );
+
+ $rv{address} = $obj->{address};
+ $rv{id_addr} = $obj->{-primary_address};
+ my $url_base = $obj->base_url($ctx->{env});
+ if (my $mid = $ctx->{mid}) { # per-thread feed:
+ $rv{atomurl} = "$url_base$mid/t.atom";
} else {
- $url_base = "http://example.com";
- $rv{atomurl} = "$url_base/atom.xml";
+ $rv{atomurl} = $url_base."new.atom";
}
- $rv{url} ||= "$url_base/";
- $rv{midurl} = "$url_base/m/";
- $rv{fullurl} = "$url_base/f/";
+ $rv{url} ||= $url_base;
+ $rv{midurl} = $url_base;
\%rv;
}
-sub mime_header {
- my ($mime, $name) = @_;
- PublicInbox::Hval->new_oneline($mime->header($name))->raw;
-}
-
-sub feed_date {
- my ($date) = @_;
- my @t = eval { strptime($date) };
+sub feed_updated {
+ my ($date, $ts) = @_;
+ my @t = eval { strptime($date) } if defined $date;
+ @t = gmtime($ts || time) unless scalar @t;
- scalar(@t) ? POSIX::strftime(DATEFMT, @t) : 0;
+ '<updated>' . strftime(DATEFMT, @t) . '</updated>';
}
-# returns 0 (skipped) or 1 (added)
-sub add_to_feed {
- my ($feed_opts, $feed, $add, $git) = @_;
+# returns undef or string
+sub feed_entry {
+ my ($feed_opts, $add, $ibx) = @_;
- my $mime = do_cat_mail($git, $add) or return 0;
- my $midurl = $feed_opts->{midurl} || 'http://example.com/m/';
- my $fullurl = $feed_opts->{fullurl} || 'http://example.com/f/';
+ my $mime = do_cat_mail($ibx, $add) or return;
+ my $url = $feed_opts->{url};
+ my $midurl = $feed_opts->{midurl};
- my $mid = $mime->header_obj->header_raw('Message-ID');
- defined $mid or return 0;
+ my $header_obj = $mime->header_obj;
+ my $mid = $header_obj->header_raw('Message-ID');
+ defined $mid or return;
$mid = PublicInbox::Hval->new_msgid($mid);
- my $href = $mid->as_href . '.html';
- my $content = PublicInbox::View->as_feed_entry($mime, $fullurl . $href);
- defined($content) or return 0;
-
- my $subject = mime_header($mime, 'Subject') or return 0;
-
- my $from = mime_header($mime, 'From') or return 0;
- my @from = Email::Address->parse($from);
- my $name = $from[0]->name;
- defined $name or $name = "";
- my $email = $from[0]->address;
- defined $email or $email = "";
-
- my $date = $mime->header('Date');
- $date = PublicInbox::Hval->new_oneline($date);
- $date = feed_date($date->raw) or return 0;
- $add =~ tr!/!!d;
- my $h = '[a-f0-9]';
- my (@uuid5) = ($add =~ m!\A($h{8})($h{4})($h{4})($h{4})($h{12})!o);
+ my $href = $midurl . $mid->as_href . '/';
- $feed->add_entry(
- author => { name => $name, email => $email },
- title => $subject,
- updated => $date,
- content => { type => 'xhtml', content => $content },
- link => $midurl . $href,
- id => 'urn:uuid:' . join('-', @uuid5),
- );
- 1;
-}
+ my $date = $header_obj->header('Date');
+ my $updated = feed_updated($date);
-sub dump_html_line {
- my ($self, $level, $html) = @_;
- if ($self->message) {
- $$html .= (' ' x $level);
- my $mime = $self->message;
- my $subj = $mime->header('Subject');
- my $mid = $mime->header_obj->header_raw('Message-ID');
- $mid = PublicInbox::Hval->new_msgid($mid);
- my $href = 'm/' . $mid->as_href . '.html';
- my $from = mime_header($mime, 'From');
-
- my @from = Email::Address->parse($from);
- $from = $from[0]->name;
- (defined($from) && length($from)) or $from = $from[0]->address;
-
- $from = PublicInbox::Hval->new_oneline($from)->as_html;
- $subj = PublicInbox::Hval->new_oneline($subj)->as_html;
- $$html .= "<a href=\"$href\">$subj</a> $from\n";
+ my $title = $header_obj->header('Subject');
+ defined $title or return;
+ $title = title_tag($title);
+
+ my $from = $header_obj->header('From') or return;
+ my ($email) = PublicInbox::Address::emails($from);
+ my $name = join(', ',PublicInbox::Address::names($from));
+ $name = ascii_html($name);
+ $email = ascii_html($email);
+
+ my $s = '';
+ if (delete $feed_opts->{emit_header}) {
+ $s .= atom_header($feed_opts, $title) . $updated;
}
- dump_html_line($self->child, $level+1, $html) if $self->child;
- dump_html_line($self->next, $level, $html) if $self->next;
+ $s .= "<entry><author><name>$name</name><email>$email</email>" .
+ "</author>$title$updated" .
+ qq{<content\ntype="xhtml">} .
+ qq{<div\nxmlns="http://www.w3.org/1999/xhtml">} .
+ qq(<pre\nstyle="white-space:pre-wrap">) .
+ PublicInbox::View::multipart_text_as_html($mime, $href) .
+ '</pre>';
+
+ $add =~ tr!/!!d;
+ my $h = '[a-f0-9]';
+ my (@uuid5) = ($add =~ m!\A($h{8})($h{4})($h{4})($h{4})($h{12})!o);
+ my $id = 'urn:uuid:' . join('-', @uuid5);
+ $s .= qq!</div></content><link\nhref="$href"/>!.
+ "<id>$id</id></entry>";
}
sub do_cat_mail {
- my ($git, $path) = @_;
- my $str = $git->cat_file("HEAD:$path");
- Email::MIME->new($str);
+ my ($ibx, $path) = @_;
+ my $mime = eval { $ibx->msg_by_path($path) } or return;
+ Email::MIME->new($mime);
}
1;