X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FFeed.pm;h=cbdf5db99f8b8486e8a14caff84259c63e560fd9;hb=68f9da93b11412122fff6e6154a42ecabcdbada0;hp=07dce9eee83bf2458eb933548c02f638ca247f53;hpb=3b620d93cdc3bcf8eab02fd874a34db7ae013775;p=public-inbox.git

diff --git a/lib/PublicInbox/Feed.pm b/lib/PublicInbox/Feed.pm
index 07dce9ee..cbdf5db9 100644
--- a/lib/PublicInbox/Feed.pm
+++ b/lib/PublicInbox/Feed.pm
@@ -1,212 +1,94 @@
-# Copyright (C) 2013-2015 all contributors <meta@public-inbox.org>
-# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
+# Copyright (C) 2013-2020 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <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::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 PublicInbox::Address;
-use POSIX qw/strftime/;
-use constant {
-	DATEFMT => '%Y-%m-%dT%H:%M:%SZ', # Atom standard
-	MAX_PER_PAGE => 25, # this needs to be tunable
-};
+use PublicInbox::WwwAtomStream;
+use PublicInbox::Smsg; # this loads w/o Search::Xapian
+
+sub generate_i {
+	my ($ctx) = @_;
+	shift @{$ctx->{msgs}};
+}
 
 # main function
 sub generate {
 	my ($ctx) = @_;
-	sub { emit_atom($_[0], $ctx) };
+	my $msgs = $ctx->{msgs} = recent_msgs($ctx);
+	return _no_thread() unless @$msgs;
+	PublicInbox::WwwAtomStream->response($ctx, 200, \&generate_i);
 }
 
 sub generate_thread_atom {
 	my ($ctx) = @_;
-	sub { emit_atom_thread($_[0], $ctx) };
+	my $msgs = $ctx->{msgs} = $ctx->{-inbox}->over->get_thread($ctx->{mid});
+	return _no_thread() unless @$msgs;
+	PublicInbox::WwwAtomStream->response($ctx, 200, \&generate_i);
 }
 
 sub generate_html_index {
 	my ($ctx) = @_;
-	sub { emit_html_index($_[0], $ctx) };
-}
-
-# private subs
-
-sub title_tag {
-	my ($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>);
-}
+	# if the 'r' query parameter is given, it is a legacy permalink
+	# which we must continue supporting:
+	my $qp = $ctx->{qp};
+	my $ibx = $ctx->{-inbox};
+	if ($qp && !$qp->{r} && $ibx->over) {
+		return PublicInbox::View::index_topics($ctx);
+	}
 
-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 $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;
-		}
-		my $s = feed_entry($feed_opts, $path, $git) or return 0;
-		$fh->write($s);
-		1;
-	});
-	end_feed($fh);
+	my $env = $ctx->{env};
+	my $url = $ibx->base_url($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 _no_thread {
-	my ($cb) = @_;
-	$cb->([404, ['Content-Type', 'text/plain'],
-		["No feed found for thread\n"]]);
-}
+sub new_html_i {
+	my ($ctx, $eml) = @_;
+	$ctx->zmore($ctx->html_top) if exists $ctx->{-html_tip};
 
-sub end_feed {
-	my ($fh) = @_;
-	$fh->write('</feed>');
-	$fh->close;
+	$eml and return PublicInbox::View::eml_entry($ctx, $eml);
+	my $smsg = shift @{$ctx->{msgs}} or
+		$ctx->zmore(PublicInbox::View::pagination_footer(
+						$ctx, './new.html'));
+	$smsg;
 }
 
-sub emit_atom_thread {
-	my ($cb, $ctx) = @_;
-	my $res = $ctx->{srch}->get_thread($ctx->{mid});
-	return _no_thread($cb) unless $res->{total};
-	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 = $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir});
-	foreach my $msg (@{$res->{msgs}}) {
-		my $s = feed_entry($feed_opts, mid2path($msg->mid), $git);
-		$fh->write($s) if defined $s;
+sub new_html {
+	my ($ctx) = @_;
+	my $msgs = $ctx->{msgs} = recent_msgs($ctx);
+	if (!@$msgs) {
+		return [404, ['Content-Type', 'text/plain'],
+			["No messages, yet\n"] ];
 	}
-	end_feed($fh);
+	$ctx->{-html_tip} = '<pre>';
+	$ctx->{-upfx} = '';
+	$ctx->{-hr} = 1;
+	PublicInbox::WwwStream::aresponse($ctx, 200, \&new_html_i);
 }
 
-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) {
-		$top = qq{<form\naction=""><pre>$top} .
-			  qq{ <input\nname=q\ntype=text />} .
-			  qq{<input\ntype=submit\nvalue=search />} .
-			  q{</pre></form><pre>}
-	} else {
-		$top = '<pre>' . $top . "\n";
-	}
+# private subs
 
-	"<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 _no_thread () {
+	[404, ['Content-Type', 'text/plain'], ["No feed found for thread\n"]];
 }
 
-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, $srch));
+sub recent_msgs {
+	my ($ctx) = @_;
+	my $ibx = $ctx->{-inbox};
+	my $max = $ibx->{feedmax};
+	return PublicInbox::View::paginate_recent($ctx, $max) if $ibx->over;
 
-	# if the 'r' query parameter is given, it is a legacy permalink
-	# which we must continue supporting:
+	# only for rare v1 inboxes which aren't indexed at all
 	my $qp = $ctx->{qp};
-	if ($qp && !$qp->{r} && $srch) {
-		$state->{srch} = $srch;
-		$last = PublicInbox::View::emit_index_topics($state);
-		$param = 'o';
-	} else {
-		$last = emit_index_nosrch($ctx, $state);
-		$param = 'r';
-	}
-	$footer = nav_footer($ctx, $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;
-	});
-	$last;
-}
-
-sub nav_footer {
-	my ($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=$last"\nrel=next>next</a>!;
-	}
-	if ($old_r) {
-		$head = $ctx->{env}->{PATH_INFO};
-		$head = qq!<a\nhref="$head">head</a>!;
-	}
-	my $atom = "<a\nhref=\"$feed_opts->{atomurl}\">Atom feed</a>";
-	"<a\nname=\"s$anchor\">page:</a> $next $head $atom";
-}
-
-sub each_recent_blob {
-	my ($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 $qp = $ctx->{qp};
+	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})(?:~[0-9]+)?/;
 
 	# revision ranges may be specified
 	my $range = 'HEAD';
@@ -218,151 +100,44 @@ 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 $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 $log = $ibx->git->popen(qw/log
+				--no-notes --no-color --raw -r
+				--no-abbrev --abbrev-commit/,
+				"--format=%H", $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);
+	my $last_commit;
 	local $/ = "\n";
+	my @oids;
 	while (defined(my $line = <$log>)) {
 		if ($line =~ /$addmsg/o) {
 			my $add = $1;
 			next if $deleted{$add}; # optimization-only
-			$nr += $cb->($add, $cur_commit, $ts, $u, $subj);
-			if ($nr >= $max) {
+			push @oids, $add;
+			if (scalar(@oids) >= $max) {
 				$last = 1;
 				last;
 			}
 		} elsif ($line =~ /$delmsg/o) {
 			$deleted{$1} = 1;
-		} elsif ($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 ($line =~ /^(${hex}{7,40})/o) {
+			if ($line =~ /^(${hex}{7,40})/) {
 				$last_commit = $1;
 				last;
 			}
 		}
 	}
 
-	# for pagination
-	($first_commit, $last_commit);
-}
-
-# private functions below
-sub get_feedopts {
-	my ($ctx) = @_;
-	my $pi_config = $ctx->{pi_config};
-	my $inbox = $ctx->{inbox};
-	my $obj = $ctx->{-inbox};
-	my $cgi = $ctx->{cgi};
-	my %rv = ( description => $obj ? $obj->description : 'FIXME' );
-
-	if ($obj) {
-		$rv{address} = $obj->{address};
-		$rv{id_addr} = $obj->{-primary_address};
-	} elsif ($pi_config && defined $inbox && $inbox ne '') {
-		# TODO: remove
-		my $addr = $pi_config->get($inbox, 'address') || "";
-		$rv{address} = $addr;
-		$addr = $addr->[0] if ref($addr);
-		$rv{id_addr} = $addr;
-	}
-	$rv{id_addr} ||= 'public-inbox@example.com';
-
-	my $url_base;
-	if ($obj) {
-		$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 {
-			$rv{atomurl} = $url_base."new.atom";
-		}
-	} else {
-		$url_base = 'http://example.com/';
-		$rv{atomurl} = $url_base.'new.atom';
-	}
-	$rv{url} ||= $url_base;
-	$rv{midurl} = $url_base;
-
-	\%rv;
-}
-
-sub feed_updated {
-	my ($date, $ts) = @_;
-	my @t = eval { strptime($date) } if defined $date;
-	@t = gmtime($ts || time) unless scalar @t;
-
-	'<updated>' . strftime(DATEFMT, @t) . '</updated>';
-}
-
-# returns undef or string
-sub feed_entry {
-	my ($feed_opts, $add, $git) = @_;
-
-	my $mime = do_cat_mail($git, $add) or return;
-	my $url = $feed_opts->{url};
-	my $midurl = $feed_opts->{midurl};
-
-	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 = $midurl.$mid->as_href;
-
-	my $date = $header_obj->header('Date');
-	my $updated = feed_updated($date);
-
-	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 $name = PublicInbox::Address::from_name($from);
-	$name = ascii_html($name);
-	$email = ascii_html($email);
-
-	my $s = '';
-	if (delete $feed_opts->{emit_header}) {
-		$s .= atom_header($feed_opts, $title) . $updated;
-	}
-	$s .= "<entry><author><name>$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>";
-}
+	$last_commit and
+		$ctx->{next_page} = qq[<a\nhref="?r=$last_commit"\nrel=next>] .
+					'next (older)</a>';
 
-sub do_cat_mail {
-	my ($git, $path) = @_;
-	my $mime = eval {
-		my $str = $git->cat_file("HEAD:$path");
-		Email::MIME->new($str);
-	};
-	$@ ? undef : $mime;
+	[ map { bless {blob => $_ }, 'PublicInbox::Smsg' } @oids ];
 }
 
 1;