X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FFeed.pm;h=36802fa1fc79def41fd380907e274185bc587226;hb=5422a844b7384c32b3532d128e15e0b50d24435b;hp=dbb80cad0073ca8e4fc5621ed123d59f622928fa;hpb=1e7fe3de7021427f8326a5f7710347035f7a63c5;p=public-inbox.git

diff --git a/lib/PublicInbox/Feed.pm b/lib/PublicInbox/Feed.pm
index dbb80cad..36802fa1 100644
--- a/lib/PublicInbox/Feed.pm
+++ b/lib/PublicInbox/Feed.pm
@@ -1,24 +1,23 @@
-# Copyright (C) 2013-2015, all contributors <meta@public-inbox.org>
+# 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);
-use PublicInbox::Hval;
-use PublicInbox::GitCatFile;
+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', # atom standard
+	DATEFMT => '%Y-%m-%dT%H:%M:%SZ', # Atom standard
 	MAX_PER_PAGE => 25, # this needs to be tunable
 };
 
-use Encode qw/find_encoding/;
-my $enc_utf8 = find_encoding('UTF-8');
-
 # main function
 sub generate {
 	my ($ctx) = @_;
@@ -39,8 +38,9 @@ sub generate_html_index {
 
 sub title_tag {
 	my ($title) = @_;
+	$title =~ tr/\t\n / /s; # squeeze spaces
 	# try to avoid the type attribute in title:
-	$title = PublicInbox::Hval->new_oneline($title)->as_html;
+	$title = ascii_html($title);
 	my $type = index($title, '&') >= 0 ? "\ntype=\"html\"" : '';
 	"<title$type>$title</title>";
 }
@@ -53,42 +53,40 @@ sub atom_header {
 	qq(<?xml version="1.0" encoding="us-ascii"?>\n) .
 	qq{<feed\nxmlns="http://www.w3.org/2005/Atom">} .
 	qq{$title} .
-	qq(<link\nhref="$feed_opts->{url}"/>) .
+	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>);
 }
 
 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 $feed_opts = get_feedopts($ctx);
 	my $x = atom_header($feed_opts);
-	my $git = PublicInbox::GitCatFile->new($ctx->{git_dir});
+	my $ibx = $ctx->{-inbox};
 	each_recent_blob($ctx, sub {
 		my ($path, undef, $ts) = @_;
 		if (defined $x) {
-			$fh->write($x . '<updated>' .
-				   strftime(DATEFMT, gmtime($ts)) .
-				   '</updated>');
+			$fh->write($x . feed_updated(undef, $ts));
 			$x = undef;
 		}
-		add_to_feed($feed_opts, $fh, $path, $git);
+		my $s = feed_entry($feed_opts, $path, $ibx) or return 0;
+		$fh->write($s);
+		1;
 	});
-	$git = undef; # destroy pipes
-	_end_feed($fh);
+	end_feed($fh);
 }
 
 sub _no_thread {
 	my ($cb) = @_;
-	my $fh = $cb->([404, ['Content-Type' => 'text/plain']]);
-	$fh->write("No feed found for thread\n");
-	$fh->close;
+	$cb->([404, ['Content-Type', 'text/plain'],
+		["No feed found for thread\n"]]);
 }
 
-sub _end_feed {
+sub end_feed {
 	my ($fh) = @_;
-	Email::Address->purge_cache;
 	$fh->write('</feed>');
 	$fh->close;
 }
@@ -97,65 +95,67 @@ sub emit_atom_thread {
 	my ($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 $fh = $cb->([200, ['Content-Type' => 'application/atom+xml']]);
 
 	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 $git = PublicInbox::GitCatFile->new($ctx->{git_dir});
+	my $ibx = $ctx->{-inbox};
 	foreach my $msg (@{$res->{msgs}}) {
-		add_to_feed($feed_opts, $fh, mid2path($msg->mid), $git);
+		my $s = feed_entry($feed_opts, mid2path($msg->mid), $ibx);
+		$fh->write($s) if defined $s;
 	}
-	$git = undef; # destroy pipes
-	_end_feed($fh);
+	end_feed($fh);
 }
 
-sub emit_html_index {
-	my ($cb, $ctx) = @_;
-	my $fh = $cb->([200,['Content-Type'=>'text/html; charset=UTF-8']]);
-
-	my $max = $ctx->{max} || MAX_PER_PAGE;
-	my $feed_opts = get_feedopts($ctx);
-
-	my $title = $feed_opts->{description} || '';
-	$title = PublicInbox::Hval->new_oneline($title)->as_html;
-	my $atom_url = $feed_opts->{atomurl};
-	my ($footer, $param, $last);
-	my $state = { ctx => $ctx, seen => {}, anchor_idx => 0 };
-	my $srch = $ctx->{srch};
-
-	my $top = "<b>$title</b> (<a\nhref=\"$atom_url\">Atom feed</a>)";
+sub _html_index_top {
+	my ($feed_opts, $srch) = @_;
 
+	my $title = ascii_html($feed_opts->{description} || '');
+	my $top = "<b>$title</b> (<a\nhref=\"new.atom\">Atom feed</a>)";
 	if ($srch) {
-		$top = qq{<form\naction=""><tt>$top} .
+		$top = qq{<form\naction=""><pre>$top} .
 			  qq{ <input\nname=q\ntype=text />} .
 			  qq{<input\ntype=submit\nvalue=search />} .
-			  qq{</tt></form>} .
-			  PublicInbox::View::PRE_WRAP;
+			  q{</pre></form><pre>}
 	} else {
-		$top = PublicInbox::View::PRE_WRAP . $top . "\n";
+		$top = '<pre>' . $top . "\n";
 	}
 
-	$fh->write("<html><head><title>$title</title>" .
-		   "<link\nrel=alternate\ntitle=\"Atom feed\"\n".
-		   "href=\"$atom_url\"\ntype=\"application/atom+xml\"/>" .
-		   "</head><body>$top");
+	"<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 emit_html_index {
+	my ($res, $ctx) = @_;
+	my $feed_opts = get_feedopts($ctx);
+	my $fh = $res->([200,['Content-Type'=>'text/html; charset=UTF-8']]);
+
+	my $max = $ctx->{max} || MAX_PER_PAGE;
+
+	my ($footer, $param, $last);
+	my $state = { ctx => $ctx, seen => {}, anchor_idx => 0, fh => $fh };
+	my $srch = $ctx->{srch};
+	$fh->write(_html_index_top($feed_opts, $srch));
 
 	# 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) {
+	my $qp = $ctx->{qp};
+	if ($qp && !$qp->{r} && $srch) {
 		$state->{srch} = $srch;
-		$last = PublicInbox::View::emit_index_topics($state, $fh);
+		$last = PublicInbox::View::emit_index_topics($state);
 		$param = 'o';
 	} else {
-		$last = emit_index_nosrch($ctx, $state, $fh);
+		$last = emit_index_nosrch($ctx, $state);
 		$param = 'r';
 	}
-	$footer = nav_footer($cgi, $last, $feed_opts, $state, $param);
+	$footer = nav_footer($ctx, $last, $feed_opts, $state, $param);
 	if ($footer) {
 		my $list_footer = $ctx->{footer};
 		$footer .= "\n\n" . $list_footer if $list_footer;
@@ -166,37 +166,37 @@ sub emit_html_index {
 }
 
 sub emit_index_nosrch {
-	my ($ctx, $state, $fh) = @_;
-	my $git = PublicInbox::GitCatFile->new($ctx->{git_dir});
+	my ($ctx, $state) = @_;
+	my $ibx = $ctx->{-inbox};
+	my $fh = $state->{fh};
 	my (undef, $last) = each_recent_blob($ctx, sub {
 		my ($path, $commit, $ts, $u, $subj) = @_;
 		$state->{first} ||= $commit;
 
-		my $mime = do_cat_mail($git, $path) or return 0;
-		PublicInbox::View::index_entry($fh, $mime, 0, $state);
+		my $mime = do_cat_mail($ibx, $path) or return 0;
+		$fh->write(PublicInbox::View::index_entry($mime, $state));
 		1;
 	});
-	Email::Address->purge_cache;
 	$last;
 }
 
 sub nav_footer {
-	my ($cgi, $last, $feed_opts, $state, $param) = @_;
-	$cgi or return '';
-	my $old_r = $cgi->param($param);
+	my ($ctx, $last, $feed_opts, $state, $param) = @_;
+	my $qp = $ctx->{qp} or return '';
+	my $old_r = $qp->{$param};
 	my $head = '    ';
 	my $next = '    ';
 	my $first = $state->{first};
 	my $anchor = $state->{anchor_idx};
 
 	if ($last) {
-		$next = qq!<a\nhref="?$param=$last">next</a>!;
+		$next = qq!<a\nhref="?$param=$last"\nrel=next>next</a>!;
 	}
 	if ($old_r) {
-		$head = $cgi->path_info;
+		$head = $ctx->{env}->{PATH_INFO};
 		$head = qq!<a\nhref="$head">head</a>!;
 	}
-	my $atom = "<a\nhref=\"$feed_opts->{atomurl}\">atom</a>";
+	my $atom = "<a\nhref=\"$feed_opts->{atomurl}\">Atom feed</a>";
 	"<a\nname=\"s$anchor\">page:</a> $next $head $atom";
 }
 
@@ -207,11 +207,11 @@ sub each_recent_blob {
 	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};
+	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 = $r;
 	}
@@ -219,19 +219,17 @@ sub each_recent_blob {
 	# get recent messages
 	# we could use git log -z, but, we already know ssoma will not
 	# leave us with filenames with spaces in them..
-	my @cmd = ('git', "--git-dir=$ctx->{git_dir}",
-			qw/log --no-notes --no-color --raw -r
-			   --abbrev=16 --abbrev-commit/,
-			"--format=%h%x00%ct%x00%an%x00%s%x00");
-	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 $nr = 0;
 	my ($cur_commit, $first_commit, $last_commit);
 	my ($ts, $subj, $u);
+	local $/ = "\n";
 	while (defined(my $line = <$log>)) {
 		if ($line =~ /$addmsg/o) {
 			my $add = $1;
@@ -252,6 +250,7 @@ sub each_recent_blob {
 	}
 
 	if ($last) {
+		local $/ = "\n";
 		while (my $line = <$log>) {
 			if ($line =~ /^(${hex}{7,40})/o) {
 				$last_commit = $1;
@@ -260,7 +259,6 @@ sub each_recent_blob {
 		}
 	}
 
-	close $log; # we may EPIPE here
 	# for pagination
 	($first_commit, $last_commit);
 }
@@ -269,117 +267,85 @@ sub each_recent_blob {
 sub get_feedopts {
 	my ($ctx) = @_;
 	my $pi_config = $ctx->{pi_config};
-	my $listname = $ctx->{listname};
+	my $inbox = $ctx->{inbox};
+	my $obj = $ctx->{-inbox};
 	my $cgi = $ctx->{cgi};
-	my %rv;
-	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 && $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 %rv = ( description => $obj->description );
 
+	$rv{address} = $obj->{address};
+	$rv{id_addr} = $obj->{-primary_address};
 	my $url_base;
-	if ($cgi) {
-		my $base;
-		if (ref($cgi) eq 'CGI') {
-			$base = $cgi->url(-base);
-		} else {
-			$base = $cgi->base->as_string;
-			$base =~ s!/\z!!;
-		}
-		$url_base = "$base/$listname";
-		if (my $mid = $ctx->{mid}) { # per-thread feed:
-			$rv{atomurl} = "$url_base/$mid/t.atom";
-		} else {
-			$rv{atomurl} = "$url_base/new.atom";
-		}
+	$url_base = $obj->base_url($cgi); # CGI may be undef
+	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/new.atom";
+		$rv{atomurl} = $url_base."new.atom";
 	}
-	$rv{url} ||= "$url_base/";
-	$rv{midurl} = "$url_base/";
+	$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) ? strftime(DATEFMT, @t) : 0;
+	'<updated>' . strftime(DATEFMT, @t) . '</updated>';
 }
 
-# returns 0 (skipped) or 1 (added)
-sub add_to_feed {
-	my ($feed_opts, $fh, $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 $mime = do_cat_mail($ibx, $add) or return;
 	my $url = $feed_opts->{url};
 	my $midurl = $feed_opts->{midurl};
 
 	my $header_obj = $mime->header_obj;
-	my $mid = $header_obj->header('Message-ID');
-	defined $mid or return 0;
+	my $mid = $header_obj->header_raw('Message-ID');
+	defined $mid or return;
 	$mid = PublicInbox::Hval->new_msgid($mid);
-	my $href = $mid->as_href;
-	my $content = PublicInbox::View->feed_entry($mime, "$midurl$href/f/");
-	defined($content) or return 0;
-	$mime = undef;
+	my $href = $midurl.$mid->as_href;
 
 	my $date = $header_obj->header('Date');
-	$date = PublicInbox::Hval->new_oneline($date);
-	$date = feed_date($date->raw) or return 0;
-	$date = "<updated>$date</updated>";
+	my $updated = feed_updated($date);
 
-	my $title = mime_header($header_obj, 'Subject') or return 0;
+	my $title = $header_obj->header('Subject');
+	defined $title or return;
 	$title = title_tag($title);
 
-	my $from = mime_header($header_obj, 'From') or return 0;
-	my @from = Email::Address->parse($from) or return 0;
-	my $name = PublicInbox::Hval->new_oneline($from[0]->name)->as_html;
-	my $email = $from[0]->address;
-	$email = PublicInbox::Hval->new_oneline($email)->as_html;
+	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}) {
-		$fh->write(atom_header($feed_opts, $title) . $date);
+		$s .= atom_header($feed_opts, $title) . $updated;
 	}
-	$fh->write("<entry><author><name>$name</name><email>$email</email>" .
-		   "</author>$title$date" .
-		   qq{<content\ntype="xhtml">} .
-		   qq{<div\nxmlns="http://www.w3.org/1999/xhtml">});
-	$fh->write($content);
+	$s .= "<entry><author><name>$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);
-	$fh->write(qq!</div></content><link\nhref="$midurl$href/"/>!.
-		   "<id>$id</id></entry>");
-	1;
+	$s .= qq!</div></content><link\nhref="$href/"/>!.
+		"<id>$id</id></entry>";
 }
 
 sub do_cat_mail {
-	my ($git, $path) = @_;
-	my $mime = eval {
-		my $str = $git->cat_file("HEAD:$path");
-		Email::MIME->new($str);
-	};
-	$@ ? undef : $mime;
+	my ($ibx, $path) = @_;
+	my $mime = eval { $ibx->msg_by_path($path) } or return;
+	Email::MIME->new($mime);
 }
 
 1;