X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FView.pm;h=52d37a9ff690dbb32cc501ef0867137e0bc3b9fb;hb=5ca29b7f2789517f0a41f1482509f504b8b601e1;hp=530c878f70b1b3783ef7824a9b3ca8534c42402e;hpb=0761fa9470bf37bc2febd4f680b0fc4a304eae83;p=public-inbox.git

diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm
index 530c878f..52d37a9f 100644
--- a/lib/PublicInbox/View.pm
+++ b/lib/PublicInbox/View.pm
@@ -1,13 +1,13 @@
-# Copyright (C) 2014-2021 all contributors <meta@public-inbox.org>
+# Copyright (C) all contributors <meta@public-inbox.org>
 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 #
 # Used for displaying the HTML web interface.
 # See Documentation/design_www.txt for this.
 package PublicInbox::View;
 use strict;
-use warnings;
-use bytes (); # only for bytes::length
+use v5.10.1;
 use List::Util qw(max);
+use Text::Wrap qw(wrap); # stdlib, we need Perl 5.6+ for $huge
 use PublicInbox::MsgTime qw(msg_datestamp);
 use PublicInbox::Hval qw(ascii_html obfuscate_addrs prurl mid_href
 			ts2str fmt_ts);
@@ -20,8 +20,10 @@ use PublicInbox::WwwStream qw(html_oneshot);
 use PublicInbox::Reply;
 use PublicInbox::ViewDiff qw(flush_diff);
 use PublicInbox::Eml;
+use POSIX qw(strftime);
 use Time::Local qw(timegm);
 use PublicInbox::Smsg qw(subject_normalized);
+use PublicInbox::ContentHash qw(content_hash);
 use constant COLS => 72;
 use constant INDENT => '  ';
 use constant TCHILD => '` ';
@@ -31,15 +33,17 @@ sub msg_page_i {
 	my ($ctx, $eml) = @_;
 	if ($eml) { # called by WwwStream::async_eml or getline
 		my $smsg = $ctx->{smsg};
-		$ctx->{smsg} = $ctx->{over}->next_by_mid(@{$ctx->{next_arg}});
+		my $over = $ctx->{ibx}->over;
+		$ctx->{smsg} = $over ? $over->next_by_mid(@{$ctx->{next_arg}})
+				: $ctx->gone('over');
 		$ctx->{mhref} = ($ctx->{nr} || $ctx->{smsg}) ?
 				"../${\mid_href($smsg->{mid})}/" : '';
-		my $obuf = $ctx->{obuf} = _msg_page_prepare_obuf($eml, $ctx);
-		multipart_text_as_html($eml, $ctx);
-		delete $ctx->{obuf};
-		$$obuf .= '</pre><hr>';
-		$$obuf .= html_footer($ctx, $ctx->{first_hdr}) if !$ctx->{smsg};
-		$$obuf;
+		if (_msg_page_prepare($eml, $ctx)) {
+			$eml->each_part(\&add_text_body, $ctx, 1);
+			$ctx->zmore('</pre><hr>');
+		}
+		html_footer($ctx, $ctx->{first_hdr}) if !$ctx->{smsg};
+		delete($ctx->{obuf}) // \'';
 	} else { # called by WwwStream::async_next or getline
 		$ctx->{smsg}; # may be undef
 	}
@@ -52,12 +56,12 @@ sub no_over_html ($) {
 	my $eml = PublicInbox::Eml->new($bref);
 	$ctx->{mhref} = '';
 	PublicInbox::WwwStream::init($ctx);
-	my $obuf = $ctx->{obuf} = _msg_page_prepare_obuf($eml, $ctx);
-	multipart_text_as_html($eml, $ctx);
-	delete $ctx->{obuf};
-	$$obuf .= '</pre><hr>';
-	eval { $$obuf .= html_footer($ctx, $eml) };
-	html_oneshot($ctx, 200, $obuf);
+	if (_msg_page_prepare($eml, $ctx)) { # sets {-title_html}
+		$eml->each_part(\&add_text_body, $ctx, 1);
+		$ctx->zmore('</pre><hr>');
+	}
+	html_footer($ctx, $eml);
+	$ctx->html_done;
 }
 
 # public functions: (unstable)
@@ -66,7 +70,7 @@ sub msg_page {
 	my ($ctx) = @_;
 	my $ibx = $ctx->{ibx};
 	$ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
-	my $over = $ctx->{over} = $ibx->over or return no_over_html($ctx);
+	my $over = $ibx->over or return no_over_html($ctx);
 	my ($id, $prev);
 	my $next_arg = $ctx->{next_arg} = [ $ctx->{mid}, \$id, \$prev ];
 
@@ -76,7 +80,7 @@ sub msg_page {
 	# allow user to easily browse the range around this message if
 	# they have ->over
 	$ctx->{-t_max} = $smsg->{ts};
-	PublicInbox::WwwStream::aresponse($ctx, 200, \&msg_page_i);
+	PublicInbox::WwwStream::aresponse($ctx, \&msg_page_i);
 }
 
 # /$INBOX/$MESSAGE_ID/#R
@@ -136,6 +140,9 @@ $info
   <a
 href="$se_url">$se_url</a>
 $link</pre>
+
+  Be sure your reply has a <b>Subject:</b> header at the top and a blank line
+  before the message body.
 EOF
 }
 
@@ -237,16 +244,20 @@ sub eml_entry {
 
 	# scan through all parts, looking for displayable text
 	$ctx->{mhref} = $mhref;
+	$ctx->{changed_href} = "#e$id"; # for diffstat "files? changed,"
 	$ctx->{obuf} = \$rv;
-	$eml->each_part(\&add_text_body, $ctx, 1);
-	delete $ctx->{obuf};
+	$eml->each_part(\&add_text_body, $ctx, 1); # expensive
+	$ctx->zmore; # TODO: remove once add_text_body is updated
 
 	# add the footer
-	$rv .= "\n<a\nhref=#$id_m\nid=e$id>^</a> ".
+	$rv = "\n<a\nhref=#$id_m\nid=e$id>^</a> ".
 		"<a\nhref=\"$mhref\">permalink</a>" .
 		" <a\nhref=\"${mhref}raw\">raw</a>" .
 		" <a\nhref=\"${mhref}#R\">reply</a>";
 
+	delete($ctx->{-qry}) and
+		$rv .= qq[ <a\nhref="${mhref}#related">related</a>];
+
 	my $hr;
 	if (defined(my $pct = $smsg->{pct})) { # used by SearchView.pm
 		$rv .= "\t[relevance $pct%]";
@@ -254,7 +265,6 @@ sub eml_entry {
 	} elsif ($mapping) {
 		my $nested = 'nested';
 		my $flat = 'flat';
-		my $end = '';
 		if ($ctx->{flat}) {
 			$hr = 1;
 			$flat = "<b>$flat</b>";
@@ -276,8 +286,15 @@ sub eml_entry {
 sub pad_link ($$;$) {
 	my ($mid, $level, $s) = @_;
 	$s ||= '...';
-	my $id = id_compress($mid, 1);
-	(' 'x19).indent_for($level).th_pfx($level)."<a\nhref=#r$id>($s)</a>\n";
+	my $href = defined($mid) ?
+		("<a\nhref=#r".id_compress($mid, 1).">($s)</a>\n") :
+		"($s)\n";
+	(' 'x19).indent_for($level).th_pfx($level).$href;
+}
+
+sub _skel_hdr {
+	# my ($mapping, $mid) = @_;
+	($_[0]->{$_[1] // \'bogus'} // [ "(?)\n" ])->[0];
 }
 
 sub _th_index_lite {
@@ -285,8 +302,7 @@ sub _th_index_lite {
 	my $rv = '';
 	my $mapping = $ctx->{mapping} or return $rv;
 	my $pad = '  ';
-	my $mid_map = $mapping->{$mid_raw};
-	defined $mid_map or
+	my $mid_map = $mapping->{$mid_raw} //
 		return 'public-inbox BUG: '.ascii_html($mid_raw).' not mapped';
 	my ($attr, $node, $idx, $level) = @$mid_map;
 	my $children = $node->{children};
@@ -309,39 +325,40 @@ sub _th_index_lite {
 				my $s = ($idx - 1). ' preceding siblings ...';
 				$rv .= pad_link($pmid, $level, $s);
 			} elsif ($idx == 2) {
-				my $ppmid = $siblings->[0]->{mid};
-				$rv .= $pad . $mapping->{$ppmid}->[0];
+				$rv .= $pad . _skel_hdr($mapping,
+						$siblings->[0] ?
+						$siblings->[0]->{mid} : undef);
 			}
-			$rv .= $pad . $mapping->{$pmid}->[0];
+			$rv .= $pad . _skel_hdr($mapping, $pmid);
 		}
 	}
 	my $s_s = nr_to_s($nr_s, 'sibling', 'siblings');
 	my $s_c = nr_to_s($nr_c, 'reply', 'replies');
-	$attr =~ s!\n\z!</b>\n!s;
+	chop $attr; # remove "\n"
 	$attr =~ s!<a\nhref.*</a> (?:&#34; )?!!s; # no point in dup subject
 	$attr =~ s!<a\nhref=[^>]+>([^<]+)</a>!$1!s; # no point linking to self
-	$rv .= "<b>@ $attr";
+	$rv .= "<b>@ $attr</b>\n";
 	if ($nr_c) {
-		my $cmid = $children->[0]->{mid};
-		$rv .= $pad . $mapping->{$cmid}->[0];
+		my $cmid = $children->[0] ? $children->[0]->{mid} : undef;
+		$rv .= $pad . _skel_hdr($mapping, $cmid);
 		if ($nr_c > 2) {
 			my $s = ($nr_c - 1). ' more replies';
 			$rv .= pad_link($cmid, $level + 1, $s);
 		} elsif (my $cn = $children->[1]) {
-			$rv .= $pad . $mapping->{$cn->{mid}}->[0];
+			$rv .= $pad . _skel_hdr($mapping, $cn->{mid});
 		}
 	}
 
 	my $next = $siblings->[$idx+1] if $siblings && $idx >= 0;
 	if ($next) {
 		my $nmid = $next->{mid};
-		$rv .= $pad . $mapping->{$nmid}->[0];
+		$rv .= $pad . _skel_hdr($mapping, $nmid);
 		my $nnext = $nr_s - $idx;
 		if ($nnext > 2) {
 			my $s = ($nnext - 1).' subsequent siblings';
 			$rv .= pad_link($nmid, $level, $s);
 		} elsif (my $nn = $siblings->[$idx + 2]) {
-			$rv .= $pad . $mapping->{$nn->{mid}}->[0];
+			$rv .= $pad . _skel_hdr($mapping, $nn->{mid});
 		}
 	}
 	$rv .= $pad ."<a\nhref=#r$id>$s_s, $s_c; $ctx->{s_nr}</a>\n";
@@ -370,7 +387,8 @@ sub pre_thread  { # walk_thread callback
 sub thread_eml_entry {
 	my ($ctx, $eml) = @_;
 	my ($beg, $end) = thread_adj_level($ctx, $ctx->{level});
-	$beg . '<pre>' . eml_entry($ctx, $eml) . '</pre>' . $end;
+	$ctx->zmore($beg.'<pre>');
+	eml_entry($ctx, $eml) . '</pre>' . $end;
 }
 
 sub next_in_queue ($$) {
@@ -413,13 +431,14 @@ sub stream_thread_i { # PublicInbox::WwwStream::getline callback
 
 sub stream_thread ($$) {
 	my ($rootset, $ctx) = @_;
-	$ctx->{-queue} = [ map { (0, $_) } @$rootset ];
-	PublicInbox::WwwStream::aresponse($ctx, 200, \&stream_thread_i);
+	@{$ctx->{-queue}} = map { (0, $_) } @$rootset;
+	PublicInbox::WwwStream::aresponse($ctx, \&stream_thread_i);
 }
 
 # /$INBOX/$MSGID/t/ and /$INBOX/$MSGID/T/
 sub thread_html {
 	my ($ctx) = @_;
+	$ctx->{-upfx} = '../../';
 	my $mid = $ctx->{mid};
 	my $ibx = $ctx->{ibx};
 	my ($nr, $msgs) = $ibx->over->get_thread($mid);
@@ -443,13 +462,12 @@ EOF
 	$skel .= " (download: <a\nhref=\"../t.mbox.gz\">mbox.gz</a>";
 	$skel .= " / follow: <a\nhref=\"../t.atom\">Atom feed</a>)\n";
 	$skel .= "-- links below jump to the message on this page --\n";
-	$ctx->{-upfx} = '../../';
 	$ctx->{cur_level} = 0;
 	$ctx->{skel} = \$skel;
 	$ctx->{prev_attr} = '';
 	$ctx->{prev_level} = 0;
-	$ctx->{root_anchor} = anchor_for($mid);
-	$ctx->{mapping} = {};
+	$ctx->{root_anchor} = 'm' . id_compress($mid, 1);
+	$ctx->{mapping} = {}; # mid -> [ header_summary, node, idx, level ]
 	$ctx->{s_nr} = ($nr > 1 ? "$nr+ messages" : 'only message')
 	               .' in thread';
 
@@ -465,7 +483,7 @@ EOF
 	# flat display: lazy load the full message from smsg
 	$ctx->{msgs} = $msgs;
 	$ctx->{-html_tip} = '<pre>';
-	PublicInbox::WwwStream::aresponse($ctx, 200, \&thread_html_i);
+	PublicInbox::WwwStream::aresponse($ctx, \&thread_html_i);
 }
 
 sub thread_html_i { # PublicInbox::WwwStream::getline callback
@@ -487,26 +505,14 @@ sub thread_html_i { # PublicInbox::WwwStream::getline callback
 	}
 }
 
-sub multipart_text_as_html {
-	# ($mime, $ctx) = @_; # each_part may do "$_[0] = undef"
-
-	# scan through all parts, looking for displayable text
-	$_[0]->each_part(\&add_text_body, $_[1], 1);
-}
-
 sub submsg_hdr ($$) {
 	my ($ctx, $eml) = @_;
-	my $obfs_ibx = $ctx->{-obfs_ibx};
-	my $rv = $ctx->{obuf};
-	$$rv .= "\n";
+	my $s = "\n";
 	for my $h (qw(From To Cc Subject Date Message-ID X-Alt-Message-ID)) {
-		my @v = $eml->header($h);
-		for my $v (@v) {
-			obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx;
-			$v = ascii_html($v);
-			$$rv .= "$h: $v\n";
-		}
+		$s .= "$h: $_\n" for $eml->header($h);
 	}
+	obfuscate_addrs($ctx->{-obfs_ibx}, $s) if $ctx->{-obfs_ibx};
+	ascii_html($s);
 }
 
 sub attach_link ($$$$;$) {
@@ -517,8 +523,8 @@ sub attach_link ($$$$;$) {
 	# downloads for 0-byte multipart attachments
 	return unless $part->{bdy};
 
-	my $nl = $idx eq '1' ? '' : "\n"; # like join("\n", ...)
-	my $size = bytes::length($part->body);
+	my $size = length($part->body);
+	delete $part->{bdy}; # save memory
 
 	# hide attributes normally, unless we want to aid users in
 	# spotting MUA problems:
@@ -532,23 +538,17 @@ sub attach_link ($$$$;$) {
 	} else {
 		$sfn = 'a.bin';
 	}
-	my $rv = $ctx->{obuf};
-	$$rv .= qq($nl<a\nhref="$ctx->{mhref}$idx-$sfn">);
-	if ($err) {
-		$$rv .= <<EOF;
+	my $rv = $idx eq '1' ? '' : "\n"; # like join("\n", ...)
+	$rv .= qq(<a\nhref="$ctx->{mhref}$idx-$sfn">);
+	$rv .= <<EOF if $err;
 [-- Warning: decoded text below may be mangled, UTF-8 assumed --]
 EOF
-	}
-	$$rv .= "[-- Attachment #$idx: ";
-	my $ts = "Type: $ct, Size: $size bytes";
+	$rv .= "[-- Attachment #$idx: ";
 	my $desc = $part->header('Content-Description') // $fn // '';
-	$desc = ascii_html($desc);
-	$$rv .= ($desc eq '') ? "$ts --]" : "$desc --]\n[-- $ts --]";
-	$$rv .= "</a>\n";
-
-	submsg_hdr($ctx, $part) if $part->{is_submsg};
-
-	undef;
+	$rv .= ascii_html($desc)." --]\n[-- " if $desc ne '';
+	$rv .= "Type: $ct, Size: $size bytes --]</a>\n";
+	$rv .= submsg_hdr($ctx, $part) if $part->{is_submsg};
+	$rv;
 }
 
 sub add_text_body { # callback for each_part
@@ -560,18 +560,17 @@ sub add_text_body { # callback for each_part
 	my ($part, $depth, $idx) = @$p;
 	my $ct = $part->content_type || 'text/plain';
 	my $fn = $part->filename;
-	my ($s, $err) = msg_part_text($part, $ct);
-	return attach_link($ctx, $ct, $p, $fn) unless defined $s;
-
 	my $rv = $ctx->{obuf};
+	my ($s, $err) = msg_part_text($part, $ct);
+	$s // return $$rv .= (attach_link($ctx, $ct, $p, $fn) // '');
 	if ($part->{is_submsg}) {
-		submsg_hdr($ctx, $part);
+		$$rv .= submsg_hdr($ctx, $part);
 		$$rv .= "\n";
 	}
 
 	# makes no difference to browsers, and don't screw up filename
 	# link generation in diffs with the extra '%0D'
-	$s =~ s/\r\n/\n/sg;
+	$s =~ s/\r+\n/\n/sg;
 
 	# will be escaped to `&#8226;' in HTML
 	obfuscate_addrs($ibx, $s, "\x{2022}") if $ibx->{obfuscate};
@@ -610,17 +609,15 @@ sub add_text_body { # callback for each_part
 		$ctx->{-spfx} = $spfx;
 	};
 
-	# some editors don't put trailing newlines at the end:
-	$s .= "\n" unless $s =~ /\n\z/s;
-
 	# split off quoted and unquoted blocks:
 	my @sections = PublicInbox::MsgIter::split_quotes($s);
 	undef $s; # free memory
 	if (defined($fn) || ($depth > 0 && !$part->{is_submsg}) || $err) {
 		# badly-encoded message with $err? tell the world about it!
-		attach_link($ctx, $ct, $p, $fn, $err);
+		$$rv .= attach_link($ctx, $ct, $p, $fn, $err);
 		$$rv .= "\n";
 	}
+	delete $part->{bdy}; # save memory
 	foreach my $cur (@sections) {
 		if ($cur =~ /\A>/) {
 			# we use a <span> here to allow users to specify
@@ -638,86 +635,97 @@ sub add_text_body { # callback for each_part
 	}
 }
 
-sub _msg_page_prepare_obuf {
+sub _msg_page_prepare {
 	my ($eml, $ctx) = @_;
-	my $over = $ctx->{ibx}->over;
-	my $obfs_ibx = $ctx->{-obfs_ibx};
-	my $rv = '';
+	my $have_over = !!$ctx->{ibx}->over;
 	my $mids = mids_for_index($eml);
 	my $nr = $ctx->{nr}++;
 	if ($nr) { # unlikely
-		$rv .= '<pre>';
+		if ($ctx->{chash} eq content_hash($eml)) {
+			warn "W: BUG? @$mids not deduplicated properly\n";
+			return;
+		}
+		$ctx->{-html_tip} =
+"<pre>WARNING: multiple messages have this Message-ID\n</pre><pre>";
 	} else {
 		$ctx->{first_hdr} = $eml->header_obj;
-		if ($ctx->{smsg}) {
-			$rv .=
-"<pre>WARNING: multiple messages have this Message-ID\n</pre>";
-		}
-		$rv .= "<pre\nid=b>"; # anchor for body start
+		$ctx->{chash} = content_hash($eml) if $ctx->{smsg}; # reused MID
+		$ctx->{-html_tip} = "<pre\nid=b>"; # anchor for body start
 	}
-	$ctx->{-upfx} = '../' if $over;
+	$ctx->{-upfx} = '../';
 	my @title; # (Subject[0], From[0])
+	my $hbuf = '';
 	for my $v ($eml->header('From')) {
 		my @n = PublicInbox::Address::names($v);
-		$v = ascii_html($v);
-		$title[1] //= ascii_html(join(', ', @n));
-		if ($obfs_ibx) {
-			obfuscate_addrs($obfs_ibx, $v);
-			obfuscate_addrs($obfs_ibx, $title[1]);
-		}
-		$rv .= "From: $v\n" if $v ne '';
+		$title[1] //= join(', ', @n);
+		$hbuf .= "From: $v\n" if $v ne '';
 	}
-	foreach my $h (qw(To Cc)) {
+	for my $h (qw(To Cc)) {
 		for my $v ($eml->header($h)) {
 			fold_addresses($v);
-			$v = ascii_html($v);
-			obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx;
-			$rv .= "$h: $v\n" if $v ne '';
+			$hbuf .= "$h: $v\n" if $v ne '';
 		}
 	}
 	my @subj = $eml->header('Subject');
-	if (@subj) {
-		my $v = ascii_html(shift @subj);
-		obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx;
-		$rv .= 'Subject: ';
-		$rv .= $over ? qq(<a\nhref="#r"\nid=t>$v</a>\n) : "$v\n";
-		$title[0] = $v;
-		for $v (@subj) { # multi-Subject message :<
-			$v = ascii_html($v);
-			obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx;
-			$rv .= "Subject: $v\n";
-		}
-	} else { # dummy anchor for thread skeleton at bottom of page
-		$rv .= qq(<a\nhref="#r"\nid=t></a>) if $over;
-		$title[0] = '(no subject)';
+	$hbuf .= "Subject: $_\n" for @subj;
+	$title[0] = $subj[0] // '(no subject)';
+	$hbuf .= "Date: $_\n" for $eml->header('Date');
+	$hbuf = ascii_html($hbuf);
+	$ctx->{-title_html} = ascii_html(join(' - ', @title));
+	if (my $obfs_ibx = $ctx->{-obfs_ibx}) {
+		obfuscate_addrs($obfs_ibx, $hbuf);
+		obfuscate_addrs($obfs_ibx, $ctx->{-title_html});
 	}
-	for my $v ($eml->header('Date')) {
-		$v = ascii_html($v);
-		obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx; # possible :P
-		$rv .= qq{Date: $v\t<a\nhref="#r">[thread overview]</a>\n};
-	}
-	if (!$nr) { # first (and only) message, common case
-		$ctx->{-title_html} = join(' - ', @title);
-		$rv = $ctx->html_top . $rv;
+
+	# [thread overview] link is typically added after Date,
+	# but added after Subject, or even nothing.
+	if ($have_over) {
+		chop $hbuf; # drop "\n", or noop if $rv eq ''
+		$hbuf .= qq{\t<a\nhref="#r">[thread overview]</a>\n};
+		$hbuf =~ s!^Subject:\x20(.*?)(\n[A-Z]|\z)
+				!Subject: <a\nhref="#r"\nid=t>$1</a>$2!msx or
+			$hbuf .= qq(<a\nhref="#r\nid=t></a>);
 	}
 	if (scalar(@$mids) == 1) { # common case
-		my $mhtml = ascii_html($mids->[0]);
-		$rv .= "Message-ID: &lt;$mhtml&gt; ";
-		$rv .= "(<a\nhref=\"raw\">raw</a>)\n";
+		my $x = ascii_html($mids->[0]);
+		$hbuf .= qq[Message-ID: &lt;$x&gt; (<a href="raw">raw</a>)\n];
+	}
+	if (!$nr) { # first (and only) message, common case
+		$ctx->zmore($ctx->html_top, $hbuf);
 	} else {
+		delete $ctx->{-title_html};
+		$ctx->zmore($ctx->{-html_tip}, $hbuf);
+	}
+	$ctx->{-linkify} //= PublicInbox::Linkify->new;
+	$hbuf = '';
+	if (scalar(@$mids) != 1) { # unlikely, but it happens :<
 		# X-Alt-Message-ID can happen if a message is injected from
 		# public-inbox-nntpd because of multiple Message-ID headers.
-		my $lnk = PublicInbox::Linkify->new;
-		my $s = '';
 		for my $h (qw(Message-ID X-Alt-Message-ID)) {
-			$s .= "$h: $_\n" for ($eml->header_raw($h));
+			$hbuf .= "$h: $_\n" for ($eml->header_raw($h));
 		}
-		$lnk->linkify_mids('..', \$s, 1);
-		$rv .= $s;
+		$ctx->{-linkify}->linkify_mids('..', \$hbuf, 1); # escapes HTML
+		$ctx->zmore($hbuf);
+		$hbuf = '';
 	}
-	$rv .= _parent_headers($eml, $over);
-	$rv .= "\n";
-	\$rv;
+	my @irt = $eml->header_raw('In-Reply-To');
+	my $refs;
+	if (!@irt) {
+		$refs = references($eml);
+		$irt[0] = pop(@$refs) if scalar @$refs;
+	}
+	$hbuf .= "In-Reply-To: $_\n" for @irt;
+
+	# do not display References: if search is present,
+	# we show the thread skeleton at the bottom, instead.
+	if (!$have_over) {
+		$refs //= references($eml);
+		$hbuf .= 'References: <'.join(">\n\t<", @$refs).">\n" if @$refs;
+	}
+	$ctx->{-linkify}->linkify_mids('..', \$hbuf); # escapes HTML
+	$ctx->zmore($hbuf .= "\n");
+	${$ctx->{obuf}} = ''; # TODO remove
+	1;
 }
 
 sub SKEL_EXPAND () {
@@ -754,7 +762,6 @@ sub thread_skel ($$$) {
 	# when multiple Subject: headers are present, so we follow suit:
 	my $subj = $hdr->header('Subject') // '';
 	$subj = '(no subject)' if $subj eq '';
-	$ctx->{prev_subj} = [ split(/ /, subject_normalized($subj)) ];
 	$ctx->{cur} = $mid;
 	$ctx->{prev_attr} = '';
 	$ctx->{prev_level} = 0;
@@ -767,50 +774,40 @@ sub thread_skel ($$$) {
 	$ctx->{parent_msg} = $parent;
 }
 
-sub _parent_headers {
-	my ($hdr, $over) = @_;
-	my $rv = '';
-	my @irt = $hdr->header_raw('In-Reply-To');
-	my $refs;
-	if (@irt) {
-		my $lnk = PublicInbox::Linkify->new;
-		$rv .= "In-Reply-To: $_\n" for @irt;
-		$lnk->linkify_mids('..', \$rv);
-	} else {
-		$refs = references($hdr);
-		my $irt = pop @$refs;
-		if (defined $irt) {
-			my $html = ascii_html($irt);
-			my $href = mid_href($irt);
-			$rv .= "In-Reply-To: &lt;";
-			$rv .= "<a\nhref=\"../$href/\">$html</a>&gt;\n";
-		}
-	}
-
-	# do not display References: if search is present,
-	# we show the thread skeleton at the bottom, instead.
-	return $rv if $over;
-
-	$refs //= references($hdr);
-	if (@$refs) {
-		@$refs = map { linkify_ref_no_over($_) } @$refs;
-		$rv .= 'References: '. join("\n\t", @$refs) . "\n";
-	}
-	$rv;
-}
-
-# returns a string buffer
+# appends to obuf
 sub html_footer {
 	my ($ctx, $hdr) = @_;
-	my $ibx = $ctx->{ibx};
 	my $upfx = '../';
-	my $skel;
-	my $rv = '<pre>';
-	if ($ibx->over) {
+	my ($related, $skel);
+	my $qry = delete $ctx->{-qry};
+	if ($qry && $ctx->{ibx}->isrch) {
+		my $q = ''; # search for either ancestor or descendent patches
+		for (@{$qry->{dfpre}}, @{$qry->{dfpost}}) {
+			chop if length > 7; # include 1 abbrev "older" patches
+			$q .= "dfblob:$_ ";
+		}
+		chop $q; # omit trailing SP
+		local $Text::Wrap::columns = COLS;
+		local $Text::Wrap::huge = 'overflow';
+		$q = wrap('', '', $q);
+		my $rows = ($q =~ tr/\n/\n/) + 1;
+		$q = ascii_html($q);
+		$related = <<EOM;
+<form id=related
+action=$upfx
+><pre>find likely ancestor, descendant, or conflicting patches for <a
+href=#t>this message</a>:
+<textarea name=q cols=${\COLS} rows=$rows>$q</textarea>
+<input type=submit value=search
+/>\t(<a href=${upfx}_/text/help/#search>help</a>)</pre></form>
+EOM
+	}
+	if ($ctx->{ibx}->over) {
 		my $t = ts2str($ctx->{-t_max});
 		my $t_fmt = fmt_ts($ctx->{-t_max});
-		$skel .= <<EOF;
-	other threads:[<a
+		my $fallback = $related ? "\t" : "<a id=related>\t</a>";
+		$skel = <<EOF;
+${fallback}other threads:[<a
 href="$upfx?t=$t">~$t_fmt UTC</a>|<a
 href="$upfx">newest</a>]
 EOF
@@ -844,26 +841,15 @@ EOF
 		} elsif ($u) { # unlikely
 			$parent = " <a\nhref=\"$u\"\nrel=prev>parent</a>";
 		}
-		$rv .= "$next $prev$parent ";
+		${$ctx->{obuf}} .= "<pre>$next $prev$parent ";
 	} else { # unindexed inboxes w/o over
+		${$ctx->{obuf}} .= '<pre>';
 		$skel = qq( <a\nhref="$upfx">latest</a>);
 	}
-	$rv .= qq(<a\nhref="#R">reply</a>);
-	$rv .= $skel;
-	$rv .= '</pre>';
-	$rv .= msg_reply($ctx, $hdr);
-}
-
-sub linkify_ref_no_over {
-	my ($mid) = @_;
-	my $href = mid_href($mid);
-	my $html = ascii_html($mid);
-	"&lt;<a\nhref=\"../$href/\">$html</a>&gt;";
-}
-
-sub anchor_for {
-	my ($msgid) = @_;
-	'm' . id_compress($msgid, 1);
+	${$ctx->{obuf}} .= qq(<a\nhref="#R">reply</a>);
+	# $skel may be big for big threads, don't append it to obuf
+	$skel .= '</pre>' . ($related // '');
+	$ctx->zmore($skel .= msg_reply($ctx, $hdr)); # flushes obuf
 }
 
 sub ghost_parent {
@@ -967,7 +953,7 @@ sub skel_dump { # walk_thread callback
 		$$skel .= delete($ctx->{sl_note}) || '';
 	}
 
-	my $f = ascii_html($smsg->{from_name});
+	my $f = ascii_html(delete $smsg->{from_name});
 	my $obfs_ibx = $ctx->{-obfs_ibx};
 	obfuscate_addrs($obfs_ibx, $f) if $obfs_ibx;
 
@@ -1062,10 +1048,10 @@ sub _skel_ghost {
 }
 
 sub sort_ds {
-	[ sort {
+	@{$_[0]} = sort {
 		(eval { $a->topmost->{ds} } || 0) <=>
 		(eval { $b->topmost->{ds} } || 0)
-	} @{$_[0]} ];
+	} @{$_[0]};
 }
 
 # accumulate recent topics if search is supported
@@ -1118,9 +1104,10 @@ sub dump_topics {
 	}
 
 	my @out;
-	my $ibx = $ctx->{ibx};
-	my $obfs_ibx = $ibx->{obfuscate} ? $ibx : undef;
-
+	my $obfs_ibx = $ctx->{ibx}->{obfuscate} ? $ctx->{ibx} : undef;
+	if (my $note = delete $ctx->{t_note}) {
+		push @out, $note; # "messages from ... to ..."
+	}
 	# sort by recency, this allows new posts to "bump" old topics...
 	foreach my $topic (sort { $b->[0] <=> $a->[0] } @$order) {
 		my ($ds, $n, $seen, $top_subj, @extra) = @$topic;
@@ -1143,13 +1130,11 @@ sub dump_topics {
 			$anchor = '#t'; # thread skeleton
 		}
 
-		my $mbox = qq(<a\nhref="$href/t.mbox.gz">mbox.gz</a>);
-		my $atom = qq(<a\nhref="$href/t.atom">Atom</a>);
 		my $s = "<a\nhref=\"$href/T/$anchor\">$top_subj</a>\n" .
-			" $ds UTC $n - $mbox / $atom\n";
-		for (my $i = 0; $i < scalar(@extra); $i += 2) {
-			my $level = $extra[$i];
-			my $subj = $extra[$i + 1]; # already normalized
+			" $ds UTC $n\n";
+		while (@extra) {
+			my $level = shift @extra;
+			my $subj = shift @extra; # already normalized
 			$mid = delete $seen->{$subj};
 			my @subj = split(/ /, $subj);
 			my @next_prev = @subj; # full copy
@@ -1181,7 +1166,7 @@ sub pagination_footer ($$) {
 		$next = $next ? "$next | " : '             | ';
 		$prev .= qq[ | <a\nhref="$latest">latest</a>];
 	}
-	"<hr><pre>page: $next$prev</pre>";
+	($next || $prev) ? "<hr><pre id=nav>page: $next$prev</pre>" : '';
 }
 
 sub paginate_recent ($$) {
@@ -1196,23 +1181,30 @@ sub paginate_recent ($$) {
 	$t =~ s/\A([0-9]{8,14})-// and $after = str2ts($1);
 	$t =~ /\A([0-9]{8,14})\z/ and $before = str2ts($1);
 
-	my $ibx = $ctx->{ibx};
-	my $msgs = $ibx->recent($opts, $after, $before);
-	my $nr = scalar @$msgs;
-	if ($nr < $lim && defined($after)) {
+	my $msgs = $ctx->{ibx}->over->recent($opts, $after, $before);
+	if (defined($after) && scalar(@$msgs) < $lim) {
 		$after = $before = undef;
-		$msgs = $ibx->recent($opts);
-		$nr = scalar @$msgs;
+		$msgs = $ctx->{ibx}->over->recent($opts);
 	}
-	my $more = $nr == $lim;
+	my $more = scalar(@$msgs) == $lim;
 	my ($newest, $oldest);
-	if ($nr) {
+	if (@$msgs) {
 		$newest = $msgs->[0]->{ts};
 		$oldest = $msgs->[-1]->{ts};
 		# if we only had $after, our SQL query in ->recent ordered
 		if ($newest < $oldest) {
 			($oldest, $newest) = ($newest, $oldest);
-			$more = 0 if defined($after) && $after < $oldest;
+			$more = undef if defined($after) && $after < $oldest;
+		}
+		if (defined($after // $before)) {
+			my $n = strftime('%Y-%m-%d %H:%M:%S', gmtime($newest));
+			my $o = strftime('%Y-%m-%d %H:%M:%S', gmtime($oldest));
+			$ctx->{t_note} = <<EOM;
+ messages from $o to $n UTC [<a href="#nav">more...</a>]
+EOM
+			my $s = ts2str($newest);
+			$ctx->{prev_page} = qq[<a\nhref="?t=$s-"\nrel=prev>] .
+						'prev (newer)</a>';
 		}
 	}
 	if (defined($oldest) && $more) {
@@ -1220,11 +1212,6 @@ sub paginate_recent ($$) {
 		$ctx->{next_page} = qq[<a\nhref="?t=$s"\nrel=next>] .
 					'next (older)</a>';
 	}
-	if (defined($newest) && (defined($before) || defined($after))) {
-		my $s = ts2str($newest);
-		$ctx->{prev_page} = qq[<a\nhref="?t=$s-"\nrel=prev>] .
-					'prev (newer)</a>';
-	}
 	$msgs;
 }
 
@@ -1232,11 +1219,8 @@ sub paginate_recent ($$) {
 sub index_topics {
 	my ($ctx) = @_;
 	my $msgs = paginate_recent($ctx, 200); # 200 is our window
-	if (@$msgs) {
-		walk_thread(thread_results($ctx, $msgs), $ctx, \&acc_topic);
-	}
-	html_oneshot($ctx, dump_topics($ctx), \pagination_footer($ctx, '.'));
-
+	walk_thread(thread_results($ctx, $msgs), $ctx, \&acc_topic) if @$msgs;
+	html_oneshot($ctx, dump_topics($ctx), pagination_footer($ctx, '.'));
 }
 
 sub thread_adj_level {
@@ -1266,7 +1250,7 @@ sub thread_adj_level {
 sub ghost_index_entry {
 	my ($ctx, $level, $node) = @_;
 	my ($beg, $end) = thread_adj_level($ctx,  $level);
-	$beg . '<pre>'. ghost_parent($ctx->{-upfx}, $node->{mid})
+	$beg . '<pre>'. ghost_parent($ctx->{-upfx}, $node->{mid} // '?')
 		. '</pre>' . $end;
 }