-# 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 Encode qw/find_encoding/;
-use Encode::MIME::Header;
-use CGI qw(escapeHTML);
-use Date::Parse qw(strptime str2time);
-use PublicInbox::Hval;
-eval { require Git }; # this is GPLv2+, so we are OK to use it
+use Email::MIME;
+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 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
};
-my $enc_utf8 = find_encoding('utf8');
-my $enc_ascii = find_encoding('us-ascii');
-my $enc_mime = find_encoding('MIME-Header');
-
-# FIXME: workaround https://rt.cpan.org/Public/Bug/Display.html?id=22817
# main function
sub generate {
- my ($class, $args) = @_;
- require XML::Atom::SimpleFeed;
- require PublicInbox::View;
- require Email::MIME;
- require POSIX;
- my $max = $args->{max} || MAX_PER_PAGE;
- my $top = $args->{top}; # bool
-
- local $ENV{GIT_DIR} = $args->{git_dir};
- 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 => $addr || 'public-inbox@example.com',
- updated => POSIX::strftime(DATEFMT, gmtime),
- );
-
- my $git = try_git_pm($args->{git_dir});
- each_recent_blob($args, sub {
- my ($add) = @_;
- add_to_feed($feed_opts, $feed, $add, $top, $git);
- });
- $feed->as_string;
+ my ($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 Mail::Thread;
-
- my $max = $args->{max} || MAX_PER_PAGE;
- my $top = $args->{top}; # bool
- local $ENV{GIT_DIR} = $args->{git_dir};
- my $feed_opts = get_feedopts($args);
- my $title = xs_html($feed_opts->{description} || "");
- my @messages;
- my $git = try_git_pm($args->{git_dir});
- my ($first, $last) = each_recent_blob($args, sub {
- my $simple = do_cat_mail($git, 'Email::Simple', $_[0])
- or return 0;
- if ($top && ($simple->header("In-Reply-To") ||
- $simple->header("References"))) {
- return 0;
- }
- $simple->body_set(""); # save some memory
+ my ($ctx) = @_;
+ sub { emit_html_index($_[0], $ctx) };
+}
- my $t = eval { str2time($simple->header('Date')) };
- defined($t) or $t = 0;
- $simple->header_set('X-PI-Date', $t);
- push @messages, $simple;
- 1;
- });
+# private subs
+
+sub title_tag {
+ my ($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>";
+}
+
+sub atom_header {
+ my ($feed_opts, $title) = @_;
+
+ $title = title_tag($feed_opts->{description}) unless (defined $title);
+
+ qq(<?xml version="1.0" encoding="us-ascii"?>\n) .
+ qq{<feed\nxmlns="http://www.w3.org/2005/Atom">} .
+ qq{$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 $th = Mail::Thread->new(@messages);
- $th->thread;
- my @out = (
- "<html><head><title>$title</title>" .
- '<link rel=alternate title=Atom.feed href="' .
- $feed_opts->{atomurl} . '" type="application/atom+xml"/>' .
- '</head><body><pre>');
- push @out, $feed_opts->{midurl};
-
- # sort by date, most recent at top
- $th->order(sub {
- sort {
- $b->topmost->message->header('X-PI-Date') <=>
- $a->topmost->message->header('X-PI-Date')
- } @_;
+sub emit_atom {
+ my ($cb, $ctx) = @_;
+ my $fh = $cb->([ 200, ['Content-Type' => 'application/atom+xml']]);
+ my $max = $ctx->{max} || MAX_PER_PAGE;
+ my $feed_opts = get_feedopts($ctx);
+ my $x = atom_header($feed_opts);
+ my $git = $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir});
+ each_recent_blob($ctx, sub {
+ my ($path, undef, $ts) = @_;
+ if (defined $x) {
+ $fh->write($x . feed_updated(undef, $ts));
+ $x = undef;
+ }
+ add_to_feed($feed_opts, $fh, $path, $git);
});
- dump_html_line($_, 0, \@out) for $th->rootset;
+ end_feed($fh);
+}
- my $footer = nav_footer($args->{cgi}, $first, $last);
- $footer = "<hr /><pre>$footer</pre>" if $footer;
- $out[0] . "</pre>$footer</html>";
+sub _no_thread {
+ my ($cb) = @_;
+ my $fh = $cb->([404, ['Content-Type' => 'text/plain']]);
+ $fh->write("No feed found for thread\n");
+ $fh->close;
}
-# private subs
+sub end_feed {
+ my ($fh) = @_;
+ Email::Address->purge_cache;
+ $fh->write('</feed>');
+ $fh->close;
+}
+
+sub emit_atom_thread {
+ my ($cb, $ctx) = @_;
+ my $res = $ctx->{srch}->get_thread($ctx->{mid});
+ return _no_thread($cb) unless $res->{total};
+ my $fh = $cb->([200, ['Content-Type' => 'application/atom+xml']]);
+ my $feed_opts = get_feedopts($ctx);
+
+ my $html_url = $feed_opts->{atomurl} = $ctx->{self_url};
+ $html_url =~ s!/t\.atom\z!/!;
+ $feed_opts->{url} = $html_url;
+ $feed_opts->{emit_header} = 1;
+
+ my $git = $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir});
+ foreach my $msg (@{$res->{msgs}}) {
+ add_to_feed($feed_opts, $fh, mid2path($msg->mid), $git);
+ }
+ end_feed($fh);
+}
+
+sub emit_html_index {
+ my ($res, $ctx) = @_;
+ my $fh = $res->([200,['Content-Type'=>'text/html; charset=UTF-8']]);
+
+ my $max = $ctx->{max} || MAX_PER_PAGE;
+ my $feed_opts = get_feedopts($ctx);
+
+ my $title = ascii_html($feed_opts->{description} || '');
+ my ($footer, $param, $last);
+ my $state = { ctx => $ctx, seen => {}, anchor_idx => 0, fh => $fh };
+ my $srch = $ctx->{srch};
+
+ 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";
+ }
+
+ $fh->write("<html><head><title>$title</title>" .
+ "<link\nrel=alternate\ntitle=\"Atom feed\"\n".
+ "href=\"new.atom\"\ntype=\"application/atom+xml\"/>" .
+ PublicInbox::Hval::STYLE .
+ "</head><body>$top");
+
+ # if the 'r' query parameter is given, it is a legacy permalink
+ # which we must continue supporting:
+ my $cgi = $ctx->{cgi};
+ if ($cgi && !$cgi->param('r') && $srch) {
+ $state->{srch} = $srch;
+ $last = PublicInbox::View::emit_index_topics($state);
+ $param = 'o';
+ } else {
+ $last = emit_index_nosrch($ctx, $state);
+ $param = 'r';
+ }
+ $footer = nav_footer($cgi, $last, $feed_opts, $state, $param);
+ if ($footer) {
+ my $list_footer = $ctx->{footer};
+ $footer .= "\n\n" . $list_footer if $list_footer;
+ $footer = "<hr /><pre>$footer</pre>";
+ }
+ $fh->write("$footer</body></html>");
+ $fh->close;
+}
+
+sub emit_index_nosrch {
+ my ($ctx, $state) = @_;
+ my $git = $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir});
+ my (undef, $last) = each_recent_blob($ctx, sub {
+ my ($path, $commit, $ts, $u, $subj) = @_;
+ $state->{first} ||= $commit;
+
+ my $mime = do_cat_mail($git, $path) or return 0;
+ PublicInbox::View::index_entry($mime, 0, $state);
+ 1;
+ });
+ Email::Address->purge_cache;
+ $last;
+}
sub nav_footer {
- my ($cgi, $first, $last) = @_;
+ my ($cgi, $last, $feed_opts, $state, $param) = @_;
$cgi or return '';
- my $old_r = $cgi->param('r');
- my $prev = ' ';
+ my $old_r = $cgi->param($param);
+ my $head = ' ';
my $next = ' ';
- my %opts = (-path => 1, -query => 1, -relative => 1);
+ my $first = $state->{first};
+ my $anchor = $state->{anchor_idx};
if ($last) {
- $cgi->param('r', $last);
- $next = $cgi->url(%opts);
- $next = qq!<a href="$next">next</a>!;
+ $next = qq!<a\nhref="?$param=$last"\nrel=next>next</a>!;
}
- if ($first && $old_r) {
- $cgi->param('r', "$first..");
- $prev = $cgi->url(%opts);
- $prev = qq!<a href="$prev">prev</a>!;
+ if ($old_r) {
+ $head = $cgi->path_info;
+ $head = qq!<a\nhref="$head">head</a>!;
}
- "$prev $next";
+ my $atom = "<a\nhref=\"$feed_opts->{atomurl}\">Atom feed</a>";
+ "<a\nname=\"s$anchor\">page:</a> $next $head $atom";
}
sub each_recent_blob {
- my ($args, $cb) = @_;
- my $max = $args->{max} || MAX_PER_PAGE;
- my $refhex = qr/[a-f0-9]{4,40}(?:~\d+)?/;
- my $cgi = $args->{cgi};
+ my ($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 $refhex = qr/(?:HEAD|${hex}{4,40})(?:~\d+)?/;
+ my $cgi = $ctx->{cgi};
# revision ranges may be specified
- my $reverse;
my $range = 'HEAD';
my $r = $cgi->param('r') if $cgi;
- if ($r) {
- if ($r =~ /\A(?:$refhex\.\.)?$refhex\z/o) {
- $range = $r;
- } elsif ($r =~ /\A(?:$refhex\.\.)\z/o) {
- $reverse = 1;
- $range = $r;
- }
+ if ($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 = qw/git log --no-notes --no-color --raw -r --no-abbrev/;
- push @cmd, '--reverse' if $reverse;
- push @cmd, $range;
- my $first;
-
- my $pid = open(my $log, '-|', @cmd) or
- die('open `'.join(' ', @cmd) . " pipe failed: $!\n");
- my %deleted;
+ my $git = $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir});
+ my $log = $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>) {
- if ($line =~ /^:000000 100644 0{40} ([a-f0-9]{40})/) {
+ my ($cur_commit, $first_commit, $last_commit);
+ my ($ts, $subj, $u);
+ 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 =~ /^:100644 000000 ([a-f0-9]{40}) 0{40}/) {
+ } elsif ($line =~ /$delmsg/o) {
$deleted{$1} = 1;
- } elsif ($line =~ /^commit ([a-f0-9]{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) {
while (my $line = <$log>) {
- if ($line =~ /^commit ([a-f0-9]{40})/) {
- push @commits, $1;
+ if ($line =~ /^(${hex}{7,40})/o) {
+ $last_commit = $1;
last;
}
}
- } elsif ($reverse) {
- unshift @commits, undef
- } else {
- push @commits, undef;
}
- close $log; # we may EPIPE here
# for pagination
- $reverse ? ($commits[-1],$commits[0]) : ($commits[0],$commits[-1]);
+ ($first_commit, $last_commit);
}
# private functions below
sub get_feedopts {
- my ($args) = @_;
- my $pi_config = $args->{pi_config};
- my $listname = $args->{listname};
- my $cgi = $args->{cgi};
+ my ($ctx) = @_;
+ my $pi_config = $ctx->{pi_config};
+ my $listname = $ctx->{listname};
+ my $cgi = $ctx->{cgi};
my %rv;
- if (open my $fh, '<', "$args->{git_dir}/description") {
+ if (open my $fh, '<', "$ctx->{git_dir}/description") {
chomp($rv{description} = <$fh>);
- close $fh;
+ } else {
+ $rv{description} = '($GIT_DIR/description missing)';
}
- if ($pi_config && defined $listname && length $listname) {
- foreach my $key (qw(address)) {
- $rv{$key} = $pi_config->get($listname, $key) || "";
- }
+ if ($pi_config && defined $listname && $listname ne '') {
+ my $addr = $pi_config->get($listname, 'address') || "";
+ $rv{address} = $addr;
+ $addr = $addr->[0] if ref($addr);
+ $rv{id_addr} = $addr;
}
+ $rv{id_addr} ||= 'public-inbox@example.com';
+
my $url_base;
if ($cgi) {
- my $cgi_url = $cgi->url(-path=>1, -relative=>1);
- my $base = $cgi->url(-base);
- $url_base = $cgi_url;
- if ($url_base =~ s!/(?:|index\.html)?\z!!) {
- $rv{atomurl} = "$base$url_base/atom.xml";
+ $url_base = $cgi->base->as_string . $listname;
+ if (my $mid = $ctx->{mid}) { # per-thread feed:
+ $rv{atomurl} = "$url_base/$mid/t.atom";
} else {
- $url_base =~ s!/atom\.xml\z!!;
- $rv{atomurl} = $base . $cgi_url;
- $url_base = $base . $url_base; # XXX is this needed?
+ $rv{atomurl} = "$url_base/new.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/";
+ $rv{midurl} = "$url_base/";
\%rv;
}
-sub utf8_header {
- my ($simple, $name) = @_;
- my $val = $simple->header($name);
- return "" unless defined $val;
- $val =~ tr/\t\n / /s;
- $val =~ tr/\r//d;
- $enc_utf8->encode($enc_mime->decode($val));
-}
-
-sub feed_date {
- my ($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, $top, $git) = @_;
-
- my $mime = do_cat_mail($git, 'Email::MIME', $add) or return 0;
- if ($top && $mime->header("In-Reply-To")) {
- return 0;
- }
+ my ($feed_opts, $fh, $add, $git) = @_;
- my $midurl = $feed_opts->{midurl} || 'http://example.com/m/';
- my $fullurl = $feed_opts->{fullurl} || 'http://example.com/f/';
+ my $mime = do_cat_mail($git, $add) or return 0;
+ my $url = $feed_opts->{url};
+ my $midurl = $feed_opts->{midurl};
- my $mid = $mime->header('Message-ID');
+ my $header_obj = $mime->header_obj;
+ my $mid = $header_obj->header_raw('Message-ID');
+ defined $mid or return 0;
$mid = PublicInbox::Hval->new_msgid($mid);
my $href = $mid->as_href;
- my $content = PublicInbox::View->as_feed_entry($mime,
- "$fullurl$href.html");
+ my $content = PublicInbox::View->feed_entry($mime);
defined($content) or return 0;
+ $mime = undef;
- my $subject = utf8_header($mime, "Subject") || "";
- length($subject) or return 0;
+ my $date = $header_obj->header('Date');
+ my $updated = feed_updated($date);
- my $from = $mime->header('From') or return 0;
+ my $title = $header_obj->header('Subject');
+ defined $title or return 0;
+ $title = title_tag($title);
-
- my @from = Email::Address->parse($from);
- my $name = $from[0]->name;
- defined $name or $name = "";
+ my $from = $header_obj->header('From') or return 0;
+ my @from = Email::Address->parse($from) or return 0;
+ my $name = ascii_html($from[0]->name);
my $email = $from[0]->address;
- defined $email or $email = "";
-
- my $date = $mime->header('Date');
- $date or return 0;
- $date = PublicInbox::Hval->new_oneline($date);
- $date = feed_date($date->as_utf8) or return 0;
- $feed->add_entry(
- author => { name => $name, email => $email },
- title => $subject,
- updated => $date,
- content => { type => "html", content => $content },
- link => $midurl . $href,
- id => $add,
- );
- 1;
-}
+ $email = ascii_html($email);
-sub dump_html_line {
- my ($self, $level, $args) = @_; # args => [ $html, $midurl ]
- if ($self->message) {
- $args->[0] .= (' ' x $level);
- my $simple = $self->message;
- my $subj = $simple->header('Subject');
- my $mid = $simple->header('Message-ID');
- $mid = PublicInbox::Hval->new_msgid($mid);
- my $url = $args->[1] . $mid->as_href;
- my $from = utf8_header($simple, "From");
- my @from = Email::Address->parse($from);
- $from = $from[0]->name;
- (defined($from) && length($from)) or $from = $from[0]->address;
- $from = xs_html($from);
- $subj = PublicInbox::Hval->new_oneline($subj);
- $subj = $subj->as_html;
- $args->[0] .= "<a href=\"$url.html\">$subj</a> $from\n";
+ if (delete $feed_opts->{emit_header}) {
+ $fh->write(atom_header($feed_opts, $title) . $updated);
}
- dump_html_line($self->child, $level+1, $args) if $self->child;
- dump_html_line($self->next, $level, $args) if $self->next;
-}
-
-sub xs_html {
- $enc_ascii->encode(escapeHTML($enc_utf8->decode($_[0])),
- Encode::HTMLCREF);
+ $fh->write("<entry><author><name>$name</name><email>$email</email>" .
+ "</author>$title$updated" .
+ qq{<content\ntype="xhtml">} .
+ qq{<div\nxmlns="http://www.w3.org/1999/xhtml">});
+ $fh->write($content);
+
+ $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);
+ $fh->write(qq!</div></content><link\nhref="$midurl$href/"/>!.
+ "<id>$id</id></entry>");
+ 1;
}
-sub try_git_pm {
- my ($dir) = @_;
- eval { Git->repository(Directory => $dir) };
-};
-
sub do_cat_mail {
- my ($git, $class, $sha1) = @_;
- my $str;
- if ($git) {
- open my $fh, '>', \$str or
- die "failed to setup string handle: $!\n";
- binmode $fh;
- my $bytes = $git->cat_blob($sha1, $fh);
- close $fh or die "failed to close string handle: $!\n";
- return if $bytes <= 0;
- } else {
- $str = `git cat-file blob $sha1`;
- return if $? != 0 || length($str) == 0;
- }
- $class->new($str);
+ my ($git, $path) = @_;
+ my $mime = eval {
+ my $str = $git->cat_file("HEAD:$path");
+ Email::MIME->new($str);
+ };
+ $@ ? undef : $mime;
}
1;