X-Git-Url: http://www.git.stargrave.org/?p=public-inbox.git;a=blobdiff_plain;f=lib%2FPublicInbox%2FView.pm;h=33b323dce248670dc1da248b795419470b4c0019;hp=eb002aebbb443d919502a6ac9b6e2ac0d0b77593;hb=78792fee4031aee0a8bd532bff98efc2b9409d68;hpb=0c22c3e3a99527ebf04a647ca287ea2fc9bdeddd

diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm
index eb002aeb..33b323dc 100644
--- a/lib/PublicInbox/View.pm
+++ b/lib/PublicInbox/View.pm
@@ -1,4 +1,4 @@
-# Copyright (C) 2014-2018 all contributors <meta@public-inbox.org>
+# Copyright (C) 2014-2020 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.
@@ -6,104 +6,83 @@
 package PublicInbox::View;
 use strict;
 use warnings;
+use bytes (); # only for bytes::length
 use PublicInbox::MsgTime qw(msg_datestamp);
-use PublicInbox::Hval qw/ascii_html obfuscate_addrs/;
+use PublicInbox::Hval qw(ascii_html obfuscate_addrs prurl mid_href);
 use PublicInbox::Linkify;
-use PublicInbox::MID qw/id_compress mid_escape mids references/;
+use PublicInbox::MID qw(id_compress mids mids_for_index references
+			$MID_EXTRACT);
 use PublicInbox::MsgIter;
 use PublicInbox::Address;
 use PublicInbox::WwwStream;
 use PublicInbox::Reply;
-require POSIX;
+use PublicInbox::ViewDiff qw(flush_diff);
+use POSIX qw(strftime);
 use Time::Local qw(timegm);
-
+use PublicInbox::Smsg qw(subject_normalized);
 use constant COLS => 72;
 use constant INDENT => '  ';
 use constant TCHILD => '` ';
 sub th_pfx ($) { $_[0] == 0 ? '' : TCHILD };
 
-# public functions: (unstable)
-
-sub msg_html {
-	my ($ctx, $mime, $more, $smsg) = @_;
-	my $hdr = $mime->header_obj;
-	my $ibx = $ctx->{-inbox};
-	my $obfs_ibx = $ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
-	my $tip = _msg_html_prepare($hdr, $ctx, $more, 0);
-	my $end = 2;
-	PublicInbox::WwwStream->response($ctx, 200, sub {
-		my ($nr, undef) = @_;
-		if ($nr == 1) {
-			# $more cannot be true w/o $smsg being defined:
-			my $upfx = $more ? '../'.mid_escape($smsg->mid).'/' : '';
-			$tip . multipart_text_as_html($mime, $upfx, $obfs_ibx) .
-				'</pre><hr>'
-		} elsif ($more && @$more) {
-			++$end;
-			msg_html_more($ctx, $more, $nr);
-		} elsif ($nr == $end) {
-			# fake an EOF if generating the footer fails;
-			# we want to at least show the message if something
-			# here crashes:
-			eval {
-				'<pre>' . html_footer($hdr, 1, $ctx) .
-				'</pre>' . msg_reply($ctx, $hdr)
-			};
-		} else {
-			undef
-		}
-	});
+sub msg_page_i {
+	my ($nr, $ctx) = @_;
+	if (my $more = delete $ctx->{more}) { # unlikely
+		# fake an EOF if $more retrieval fails;
+		eval { msg_page_more($ctx, $nr, @$more) };
+	} elsif (my $hdr = delete $ctx->{hdr}) {
+		# fake an EOF if generating the footer fails;
+		# we want to at least show the message if something
+		# here crashes:
+		eval { html_footer($ctx, $hdr) };
+	} else {
+		undef
+	}
 }
 
+# public functions: (unstable)
+
 sub msg_page {
 	my ($ctx) = @_;
 	my $mid = $ctx->{mid};
 	my $ibx = $ctx->{-inbox};
-	my ($first, $more);
-	my $smsg;
-	if (my $srch = $ibx->search) {
+	my ($smsg, $first, $next);
+	if (my $over = $ibx->over) {
 		my ($id, $prev);
-		$smsg = $srch->next_by_mid($mid, \$id, \$prev);
-		$first = $ibx->msg_by_smsg($smsg) if $smsg;
-		if ($first) {
-			my $next = $srch->next_by_mid($mid, \$id, \$prev);
-			$more = [ $id, $prev, $next ] if $next;
-		}
-		return unless $first;
+		$smsg = $over->next_by_mid($mid, \$id, \$prev) or return;
+		$first = $ibx->msg_by_smsg($smsg) or return;
+		$next = $over->next_by_mid($mid, \$id, \$prev);
+		$ctx->{more} = [ $id, $prev, $next ] if $next;
 	} else {
 		$first = $ibx->msg_by_mid($mid) or return;
 	}
-	msg_html($ctx, PublicInbox::MIME->new($first), $more, $smsg);
+	my $mime = PublicInbox::MIME->new($first);
+	$ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
+	my $hdr = $ctx->{hdr} = $mime->header_obj;
+	$ctx->{obuf} = _msg_page_prepare_obuf($hdr, $ctx, 0);
+	$ctx->{smsg} = $smsg;
+	# $next cannot be true w/o $smsg being defined:
+	$ctx->{mhref} = $next ? '../'.mid_href($smsg->{mid}).'/' : '';
+	multipart_text_as_html($mime, $ctx);
+	$ctx->{-html_tip} = (${delete $ctx->{obuf}} .= '</pre><hr>');
+	PublicInbox::WwwStream->response($ctx, 200, \&msg_page_i);
 }
 
-sub msg_html_more {
-	my ($ctx, $more, $nr) = @_;
-	my $str = eval {
-		my ($id, $prev, $smsg) = @$more;
-		my $mid = $ctx->{mid};
-		$smsg = $ctx->{-inbox}->smsg_mime($smsg);
-		my $next = $ctx->{srch}->next_by_mid($mid, \$id, \$prev);
-		@$more = $next ? ($id, $prev, $next) : ();
-		if ($smsg) {
-			my $mime = $smsg->{mime};
-			my $upfx = '../' . mid_escape($smsg->mid) . '/';
-			_msg_html_prepare($mime->header_obj, $ctx, $more, $nr) .
-				multipart_text_as_html($mime, $upfx,
-							$ctx->{-obfs_ibx}) .
-				'</pre><hr>'
-		} else {
-			'';
-		}
-	};
-	if ($@) {
-		warn "Error lookup up additional messages: $@\n";
-		$str = '<pre>Error looking up additional messages</pre>';
-	}
-	$str;
+sub msg_page_more { # cold
+	my ($ctx, $nr, $id, $prev, $smsg) = @_;
+	my $ibx = $ctx->{-inbox};
+	my $next = $ibx->over->next_by_mid($ctx->{mid}, \$id, \$prev);
+	$ctx->{more} = [ $id, $prev, $next ] if $next;
+	$smsg = $ibx->smsg_mime($smsg) or return '';
+	$ctx->{mhref} = '../' . mid_href($smsg->{mid}) . '/';
+	my $mime = delete $smsg->{mime};
+	$ctx->{obuf} = _msg_page_prepare_obuf($mime->header_obj, $ctx, $nr);
+	multipart_text_as_html($mime, $ctx);
+	${delete $ctx->{obuf}} .= '</pre><hr>';
 }
 
 # /$INBOX/$MESSAGE_ID/#R
-sub msg_reply {
+sub msg_reply ($$) {
 	my ($ctx, $hdr) = @_;
 	my $se_url =
 	 'https://kernel.org/pub/software/scm/git/docs/git-send-email.html';
@@ -113,12 +92,15 @@ sub msg_reply {
 	my $info = '';
 	my $ibx = $ctx->{-inbox};
 	if (my $url = $ibx->{infourl}) {
-		$url = PublicInbox::Hval::prurl($ctx->{env}, $url);
+		$url = prurl($ctx->{env}, $url);
 		$info = qq(\n  List information: <a\nhref="$url">$url</a>\n);
 	}
 
 	my ($arg, $link, $reply_to_all) =
 			PublicInbox::Reply::mailto_arg_link($ibx, $hdr);
+	if (ref($arg) eq 'SCALAR') {
+		return '<pre id=R>'.ascii_html($$arg).'</pre>';
+	}
 
 	# mailto: link only works if address obfuscation is disabled
 	if ($link) {
@@ -136,7 +118,7 @@ EOF
 <hr><pre
 id=R><b>Reply instructions:</b>
 
-You may reply publically to <a
+You may reply publicly to <a
 href=#t>this message</a> via plain-text email
 using any one of the following methods:
 
@@ -185,8 +167,8 @@ sub fold_addresses ($) {
 
 sub _hdr_names_html ($$) {
 	my ($hdr, $field) = @_;
-	my $val = $hdr->header($field) or return '';
-	ascii_html(join(', ', PublicInbox::Address::names($val)));
+	my @vals = $hdr->header($field) or return '';
+	ascii_html(join(', ', PublicInbox::Address::names(join(',', @vals))));
 }
 
 sub nr_to_s ($$$) {
@@ -195,10 +177,12 @@ sub nr_to_s ($$$) {
 	$nr == 1 ? "$nr $singular" : "$nr $plural";
 }
 
+# human-friendly format
+sub fmt_ts ($) { strftime('%Y-%m-%d %k:%M', gmtime($_[0])) }
+
 # this is already inside a <pre>
 sub index_entry {
 	my ($smsg, $ctx, $more) = @_;
-	my $srch = $ctx->{srch};
 	my $subj = $smsg->subject;
 	my $mid_raw = $smsg->mid;
 	my $id = id_compress($mid_raw, 1);
@@ -208,6 +192,7 @@ sub index_entry {
 	my $irt;
 	my $obfs_ibx = $ctx->{-obfs_ibx};
 
+	$subj = '(no subject)' if $subj eq '';
 	my $rv = "<a\nhref=#e$id\nid=m$id>*</a> ";
 	$subj = '<b>'.ascii_html($subj).'</b>';
 	obfuscate_addrs($obfs_ibx, $subj) if $obfs_ibx;
@@ -215,13 +200,17 @@ sub index_entry {
 	$rv .= $subj . "\n";
 	$rv .= _th_index_lite($mid_raw, \$irt, $id, $ctx);
 	my @tocc;
-	my $mime = $smsg->{mime};
+	my $ds = $smsg->ds; # for v1 non-Xapian/SQLite users
+	# deleting {mime} is critical to memory use,
+	# the rest of the fields saves about 400K as we iterate across 1K msgs
+	my ($mime) = delete @$smsg{qw(mime ds ts blob subject)};
+
 	my $hdr = $mime->header_obj;
 	my $from = _hdr_names_html($hdr, 'From');
 	obfuscate_addrs($obfs_ibx, $from) if $obfs_ibx;
-	$rv .= "From: $from @ ".fmt_ts($smsg->ds)." UTC";
+	$rv .= "From: $from @ ".fmt_ts($ds)." UTC";
 	my $upfx = $ctx->{-upfx};
-	my $mhref = $upfx . mid_escape($mid_raw) . '/';
+	my $mhref = $upfx . mid_href($mid_raw) . '/';
 	$rv .= qq{ (<a\nhref="$mhref">permalink</a> / };
 	$rv .= qq{<a\nhref="${mhref}raw">raw</a>)\n};
 	my $to = fold_addresses(_hdr_names_html($hdr, 'To'));
@@ -245,15 +234,17 @@ sub index_entry {
 
 	my $mapping = $ctx->{mapping};
 	if (!$mapping && (defined($irt) || defined($irt = in_reply_to($hdr)))) {
-		my $mirt = PublicInbox::Hval->new_msgid($irt);
-		my $href = $upfx . $mirt->{href}. '/';
-		my $html = $mirt->as_html;
+		my $href = $upfx . mid_href($irt) . '/';
+		my $html = ascii_html($irt);
 		$rv .= qq(In-Reply-To: &lt;<a\nhref="$href">$html</a>&gt;\n)
 	}
 	$rv .= "\n";
 
 	# scan through all parts, looking for displayable text
-	msg_iter($mime, sub { $rv .= add_text_body($mhref, $obfs_ibx, $_[0]) });
+	$ctx->{mhref} = $mhref;
+	$ctx->{obuf} = \$rv;
+	msg_iter($mime, \&add_text_body, $ctx, 1);
+	delete $ctx->{obuf};
 
 	# add the footer
 	$rv .= "\n<a\nhref=#$id_m\nid=e$id>^</a> ".
@@ -262,8 +253,8 @@ sub index_entry {
 		" <a\nhref=\"${mhref}#R\">reply</a>";
 
 	my $hr;
-	if (my $pct = $ctx->{pct}) { # used by SearchView.pm
-		$rv .= "\t[relevance $pct->{$mid_raw}%]";
+	if (defined(my $pct = $smsg->{pct})) { # used by SearchView.pm
+		$rv .= "\t[relevance $pct%]";
 		$hr = 1;
 	} elsif ($mapping) {
 		my $nested = 'nested';
@@ -307,7 +298,10 @@ sub _th_index_lite {
 	my $nr_s = 0;
 	my $siblings;
 	if (my $smsg = $node->{smsg}) {
-		($$irt) = (($smsg->{references} || '') =~ m/<([^>]+)>\z/);
+		# delete saves about 200KB on a 1K message thread
+		if (my $refs = delete $smsg->{references}) {
+			($$irt) = ($refs =~ m/$MID_EXTRACT\z/o);
+		}
 	}
 	my $irt_map = $mapping->{$$irt} if defined $$irt;
 	if (defined $irt_map) {
@@ -359,7 +353,8 @@ sub _th_index_lite {
 	$rv .= $pad ."<a\nhref=#r$id>$s_s, $s_c; $ctx->{s_nr}</a>\n";
 }
 
-sub walk_thread {
+# non-recursive thread walker
+sub walk_thread ($$$) {
 	my ($rootset, $ctx, $cb) = @_;
 	my @q = map { (0, $_, -1) } @$rootset;
 	while (@q) {
@@ -372,7 +367,7 @@ sub walk_thread {
 	}
 }
 
-sub pre_thread  {
+sub pre_thread  { # walk_thread callback
 	my ($ctx, $level, $node, $idx) = @_;
 	$ctx->{mapping}->{$node->{id}} = [ '', $node, $idx, $level ];
 	skel_dump($ctx, $level, $node);
@@ -384,50 +379,51 @@ sub thread_index_entry {
 	$beg . '<pre>' . index_entry($smsg, $ctx, 0) . '</pre>' . $end;
 }
 
+sub stream_thread_i { # PublicInbox::WwwStream::getline callback
+	my ($nr, $ctx) = @_;
+	return unless exists($ctx->{skel});
+	my $q = $ctx->{-queue};
+	while (@$q) {
+		my $level = shift @$q;
+		my $node = shift @$q or next;
+		my $cl = $level + 1;
+		unshift @$q, map { ($cl, $_) } @{$node->{children}};
+		if (my $smsg = $ctx->{-inbox}->smsg_mime($node->{smsg})) {
+			return thread_index_entry($ctx, $level, $smsg);
+		} else {
+			return ghost_index_entry($ctx, $level, $node);
+		}
+	}
+	join('', thread_adj_level($ctx, 0)) . ${delete $ctx->{skel}};
+}
+
 sub stream_thread ($$) {
 	my ($rootset, $ctx) = @_;
-	my $inbox = $ctx->{-inbox};
+	my $ibx = $ctx->{-inbox};
 	my @q = map { (0, $_) } @$rootset;
-	my $level;
-	my $smsg;
+	my ($smsg, $level);
 	while (@q) {
 		$level = shift @q;
 		my $node = shift @q or next;
 		my $cl = $level + 1;
 		unshift @q, map { ($cl, $_) } @{$node->{children}};
-		$smsg = $inbox->smsg_mime($node->{smsg}) and last;
+		$smsg = $ibx->smsg_mime($node->{smsg}) and last;
 	}
 	return missing_thread($ctx) unless $smsg;
 
-	$ctx->{-obfs_ibx} = $inbox->{obfuscate} ? $inbox : undef;
-	$ctx->{-title_html} = ascii_html($smsg->subject);
+	$ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
+	$ctx->{-title_html} = ascii_html($smsg->{subject});
 	$ctx->{-html_tip} = thread_index_entry($ctx, $level, $smsg);
-	$smsg = undef;
-	PublicInbox::WwwStream->response($ctx, 200, sub {
-		return unless $ctx;
-		while (@q) {
-			$level = shift @q;
-			my $node = shift @q or next;
-			my $cl = $level + 1;
-			unshift @q, map { ($cl, $_) } @{$node->{children}};
-			if ($smsg = $inbox->smsg_mime($node->{smsg})) {
-				return thread_index_entry($ctx, $level, $smsg);
-			} else {
-				return ghost_index_entry($ctx, $level, $node);
-			}
-		}
-		my $ret = join('', thread_adj_level($ctx, 0));
-		$ret .= ${$ctx->{dst}}; # skel
-		$ctx = undef;
-		$ret;
-	});
+	$ctx->{-queue} = \@q;
+	PublicInbox::WwwStream->response($ctx, 200, \&stream_thread_i);
 }
 
+# /$INBOX/$MESSAGE_ID/t/
 sub thread_html {
 	my ($ctx) = @_;
 	my $mid = $ctx->{mid};
-	my $srch = $ctx->{srch};
-	my ($nr, $msgs) = $srch->get_thread($mid);
+	my $ibx = $ctx->{-inbox};
+	my ($nr, $msgs) = $ibx->over->get_thread($mid);
 	return missing_thread($ctx) if $nr == 0;
 	my $skel = '<hr><pre>';
 	$skel .= $nr == 1 ? 'only message in thread' : 'end of thread';
@@ -439,19 +435,19 @@ sub thread_html {
 	$skel .= "-- links below jump to the message on this page --\n";
 	$ctx->{-upfx} = '../../';
 	$ctx->{cur_level} = 0;
-	$ctx->{dst} = \$skel;
+	$ctx->{skel} = \$skel;
 	$ctx->{prev_attr} = '';
 	$ctx->{prev_level} = 0;
 	$ctx->{root_anchor} = anchor_for($mid);
 	$ctx->{mapping} = {};
-	$ctx->{s_nr} = "$nr+ messages in thread";
+	$ctx->{s_nr} = ($nr > 1 ? "$nr+ messages" : 'only message')
+	               .' in thread';
 
 	my $rootset = thread_results($ctx, $msgs);
 
 	# reduce hash lookups in pre_thread->skel_dump
-	my $inbox = $ctx->{-inbox};
-	$ctx->{-obfs_ibx} = $inbox->{obfuscate} ? $inbox : undef;
-	walk_thread($rootset, $ctx, *pre_thread);
+	$ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
+	walk_thread($rootset, $ctx, \&pre_thread);
 
 	$skel .= '</pre>';
 	return stream_thread($rootset, $ctx) unless $ctx->{flat};
@@ -459,49 +455,35 @@ sub thread_html {
 	# flat display: lazy load the full message from smsg
 	my $smsg;
 	while (my $m = shift @$msgs) {
-		$smsg = $inbox->smsg_mime($m) and last;
+		$smsg = $ibx->smsg_mime($m) and last;
 	}
 	return missing_thread($ctx) unless $smsg;
-	$ctx->{-title_html} = ascii_html($smsg->subject);
+	$ctx->{-title_html} = ascii_html($smsg->{subject});
 	$ctx->{-html_tip} = '<pre>'.index_entry($smsg, $ctx, scalar @$msgs);
-	$smsg = undef;
-	PublicInbox::WwwStream->response($ctx, 200, sub {
-		return unless $msgs;
-		$smsg = undef;
-		while (my $m = shift @$msgs) {
-			$smsg = $inbox->smsg_mime($m) and last;
-		}
-		return index_entry($smsg, $ctx, scalar @$msgs) if $smsg;
-		$msgs = undef;
-		$skel;
-	});
+	$ctx->{msgs} = $msgs;
+	PublicInbox::WwwStream->response($ctx, 200, \&thread_html_i);
 }
 
-sub multipart_text_as_html {
-	my ($mime, $upfx, $obfs_ibx) = @_;
-	my $rv = "";
-
-	# scan through all parts, looking for displayable text
-	msg_iter($mime, sub { $rv .= add_text_body($upfx, $obfs_ibx, $_[0]) });
-	$rv;
+sub thread_html_i { # PublicInbox::WwwStream::getline callback
+	my ($nr, $ctx) = @_;
+	my $msgs = $ctx->{msgs} or return;
+	while (my $smsg = shift @$msgs) {
+		$ctx->{-inbox}->smsg_mime($smsg) or next;
+		return index_entry($smsg, $ctx, scalar @$msgs);
+	}
+	my ($skel) = delete @$ctx{qw(skel msgs)};
+	$$skel;
 }
 
-sub flush_quote {
-	my ($s, $l, $quot) = @_;
-
-	# show everything in the full version with anchor from
-	# short version (see above)
-	my $rv = $l->linkify_1(join('', @$quot));
-	@$quot = ();
+sub multipart_text_as_html {
+	# ($mime, $ctx) = @_; # msg_iter will do "$_[0] = undef"
 
-	# we use a <span> here to allow users to specify their own
-	# color for quoted text
-	$rv = $l->linkify_2(ascii_html($rv));
-	$$s .= qq(<span\nclass="q">) . $rv . '</span>'
+	# scan through all parts, looking for displayable text
+	msg_iter($_[0], \&add_text_body, $_[1], 1);
 }
 
 sub attach_link ($$$$;$) {
-	my ($upfx, $ct, $p, $fn, $err) = @_;
+	my ($ctx, $ct, $p, $fn, $err) = @_;
 	my ($part, $depth, @idx) = @$p;
 	my $nl = $idx[-1] > 1 ? "\n" : '';
 	my $idx = join('.', @idx);
@@ -511,121 +493,130 @@ sub attach_link ($$$$;$) {
 	# spotting MUA problems:
 	$ct =~ s/;.*// unless $err;
 	$ct = ascii_html($ct);
-	my $desc = $part->header('Content-Description');
-	$desc = $fn unless defined $desc;
-	$desc = '' unless defined $desc;
 	my $sfn;
-	if (defined $fn && $fn =~ /\A[[:alnum:]][\w\.-]+[[:alnum:]]\z/) {
+	if (defined $fn && $fn =~ /\A$PublicInbox::Hval::FN\z/o) {
 		$sfn = $fn;
 	} elsif ($ct eq 'text/plain') {
 		$sfn = 'a.txt';
 	} else {
 		$sfn = 'a.bin';
 	}
-	my $ret = qq($nl<a\nhref="$upfx$idx-$sfn">);
+	my $rv = $ctx->{obuf};
+	$$rv .= qq($nl<a\nhref="$ctx->{mhref}$idx-$sfn">);
 	if ($err) {
-		$ret .=
-"[-- Warning: decoded text below may be mangled --]\n";
+		$$rv .= <<EOF;
+[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
+EOF
 	}
-	$ret .= "[-- Attachment #$idx: ";
+	$$rv .= "[-- Attachment #$idx: ";
 	my $ts = "Type: $ct, Size: $size bytes";
+	my $desc = $part->header('Content-Description') // $fn // '';
 	$desc = ascii_html($desc);
-	$ret .= ($desc eq '') ? "$ts --]" : "$desc --]\n[-- $ts --]";
-	$ret .= "</a>\n";
+	$$rv .= ($desc eq '') ? "$ts --]" : "$desc --]\n[-- $ts --]";
+	$$rv .= "</a>\n";
+	undef;
 }
 
-sub add_text_body {
-	my ($upfx, $obfs_ibx, $p) = @_;
+sub add_text_body { # callback for msg_iter
+	my ($p, $ctx) = @_;
+	my $upfx = $ctx->{mhref};
+	my $ibx = $ctx->{-inbox};
 	# $p - from msg_iter: [ Email::MIME, depth, @idx ]
-	my ($part, $depth) = @$p; # attachment @idx is unused
+	my ($part, $depth, @idx) = @$p;
 	my $ct = $part->content_type || 'text/plain';
 	my $fn = $part->filename;
-
-	if ($ct =~ m!\btext/x?html\b!i) {
-		return attach_link($upfx, $ct, $p, $fn);
-	}
-
-	my $s = eval { $part->body_str };
-
-	# badly-encoded message? tell the world about it!
-	my $err = $@;
-	if ($err) {
-		if ($ct =~ m!\btext/plain\b!i) {
-			# Try to assume UTF-8 because Alpine seems to
-			# do wacky things and set charset=X-UNKNOWN
-			$part->charset_set('UTF-8');
-			$s = eval { $part->body_str };
-
-			# If forcing charset=UTF-8 failed,
-			# attach_link will warn further down...
-			$s = $part->body if $@;
-		} else {
-			return attach_link($upfx, $ct, $p, $fn);
+	my ($s, $err) = msg_part_text($part, $ct);
+	return attach_link($ctx, $ct, $p, $fn) unless defined $s;
+
+	# makes no difference to browsers, and don't screw up filename
+	# link generation in diffs with the extra '%0D'
+	$s =~ s/\r\n/\n/sg;
+
+	# will be escaped to `&#8226;' in HTML
+	obfuscate_addrs($ibx, $s, "\x{2022}") if $ibx->{obfuscate};
+
+	# always support diff-highlighting, but we can't linkify hunk
+	# headers for solver unless some coderepo are configured:
+	my $diff;
+	if ($s =~ /^(?:diff|---|\+{3}) /ms) {
+		# diffstat anchors do not link across attachments or messages:
+		$idx[0] = $upfx . $idx[0] if $upfx ne '';
+		$ctx->{-apfx} = join('/', @idx);
+		$ctx->{-anchors} = {}; # attr => filename
+		$diff = 1;
+		delete $ctx->{-long_path};
+		my $spfx;
+		if ($ibx->{-repo_objs}) {
+			if (index($upfx, '//') >= 0) { # absolute URL (Atom feeds)
+				$spfx = $upfx;
+				$spfx =~ s!/([^/]*)/\z!/!;
+			} else {
+				my $n_slash = $upfx =~ tr!/!/!;
+				if ($n_slash == 0) {
+					$spfx = '../';
+				} elsif ($n_slash == 1) {
+					$spfx = '';
+				} else { # nslash == 2
+					$spfx = '../../';
+				}
+			}
 		}
-	}
+		$ctx->{-spfx} = $spfx;
+	};
+
+	# some editors don't put trailing newlines at the end:
+	$s .= "\n" unless $s =~ /\n\z/s;
 
-	my @lines = split(/^/m, $s);
-	$s = '';
+	# split off quoted and unquoted blocks:
+	my @sections = PublicInbox::MsgIter::split_quotes($s);
+	undef $s; # free memory
+	my $rv = $ctx->{obuf};
 	if (defined($fn) || $depth > 0 || $err) {
-		$s .= attach_link($upfx, $ct, $p, $fn, $err);
-		$s .= "\n";
+		# badly-encoded message with $err? tell the world about it!
+		attach_link($ctx, $ct, $p, $fn, $err);
+		$$rv .= "\n";
 	}
-	my @quot;
-	my $l = PublicInbox::Linkify->new;
-	foreach my $cur (@lines) {
-		if ($cur !~ /^>/) {
-			# show the previously buffered quote inline
-			flush_quote(\$s, $l, \@quot) if @quot;
-
-			# regular line, OK
-			$l->linkify_1($cur);
-			$s .= $l->linkify_2(ascii_html($cur));
+	my $l = $ctx->{-linkify} //= PublicInbox::Linkify->new;
+	foreach my $cur (@sections) {
+		if ($cur =~ /\A>/) {
+			# we use a <span> here to allow users to specify
+			# their own color for quoted text
+			$$rv .= qq(<span\nclass="q">);
+			$$rv .= $l->to_html($cur);
+			$$rv .= '</span>';
+		} elsif ($diff) {
+			flush_diff($ctx, \$cur);
 		} else {
-			push @quot, $cur;
-		}
-	}
-
-	if (@quot) { # ugh, top posted
-		flush_quote(\$s, $l, \@quot);
-		obfuscate_addrs($obfs_ibx, $s) if $obfs_ibx;
-		$s;
-	} else {
-		obfuscate_addrs($obfs_ibx, $s) if $obfs_ibx;
-		if ($s =~ /\n\z/s) { # common, last line ends with a newline
-			$s;
-		} else { # some editors don't do newlines...
-			$s .= "\n";
+			# regular lines, OK
+			$$rv .= $l->to_html($cur);
 		}
+		undef $cur; # free memory
 	}
 }
 
-sub _msg_html_prepare {
-	my ($hdr, $ctx, $more, $nr) = @_;
-	my $srch = $ctx->{srch} if $ctx;
-	my $atom = '';
+sub _msg_page_prepare_obuf {
+	my ($hdr, $ctx, $nr) = @_;
+	my $over = $ctx->{-inbox}->over;
 	my $obfs_ibx = $ctx->{-obfs_ibx};
 	my $rv = '';
-	my $mids = mids($hdr);
-	my $multiple = scalar(@$mids) > 1; # zero, one, infinity
+	my $mids = mids_for_index($hdr);
 	if ($nr == 0) {
-		if ($more) {
+		if ($ctx->{more}) {
 			$rv .=
-"<pre>WARNING: multiple messages refer to this Message-ID\n</pre>";
+"<pre>WARNING: multiple messages have this Message-ID\n</pre>";
 		}
 		$rv .= "<pre\nid=b>"; # anchor for body start
 	} else {
 		$rv .= '<pre>';
 	}
-	if ($srch) {
+	if ($over) {
 		$ctx->{-upfx} = '../';
 	}
-	my @title;
-	my $v;
-	if (defined($v = $hdr->header('From'))) {
-		$v = PublicInbox::Hval->new($v);
-		my @n = PublicInbox::Address::names($v->raw);
-		$title[1] = ascii_html(join(', ', @n));
-		$v = $v->as_html;
+	my @title; # (Subject[0], From[0])
+	for my $v ($hdr->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]);
@@ -633,144 +624,159 @@ sub _msg_html_prepare {
 		$rv .= "From: $v\n" if $v ne '';
 	}
 	foreach my $h (qw(To Cc)) {
-		defined($v = $hdr->header($h)) or next;
-		fold_addresses($v);
-		$v = ascii_html($v);
-		obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx;
-		$rv .= "$h: $v\n" if $v ne '';
+		for my $v ($hdr->header($h)) {
+			fold_addresses($v);
+			$v = ascii_html($v);
+			obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx;
+			$rv .= "$h: $v\n" if $v ne '';
+		}
 	}
-	if (defined($v = $hdr->header('Subject')) && ($v ne '')) {
-		$v = ascii_html($v);
+	my @subj = $hdr->header('Subject');
+	if (@subj) {
+		my $v = ascii_html(shift @subj);
 		obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx;
-		if ($srch) {
-			$rv .= qq(Subject: <a\nhref="#r"\nid=t>$v</a>\n);
-		} else {
+		$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";
 		}
-		$title[0] = $v;
 	} else { # dummy anchor for thread skeleton at bottom of page
-		$rv .= qq(<a\nhref="#r"\nid=t></a>) if $srch;
+		$rv .= qq(<a\nhref="#r"\nid=t></a>) if $over;
 		$title[0] = '(no subject)';
 	}
-	if (defined($v = $hdr->header('Date'))) {
+	for my $v ($hdr->header('Date')) {
 		$v = ascii_html($v);
 		obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx; # possible :P
 		$rv .= "Date: $v\n";
 	}
 	$ctx->{-title_html} = join(' - ', @title);
-	foreach (@$mids) {
-		my $mid = PublicInbox::Hval->new_msgid($_) ;
-		my $mhtml = $mid->as_html;
-		if ($multiple) {
-			my $href = $mid->{href};
-			$rv .= "Message-ID: ";
-			$rv .= "<a\nhref=\"../$href/\">";
-			$rv .= "&lt;$mhtml&gt;</a> ";
-			$rv .= "(<a\nhref=\"../$href/raw\">raw</a>)\n";
-		} else {
-			$rv .= "Message-ID: &lt;$mhtml&gt; ";
-			$rv .= "(<a\nhref=\"raw\">raw</a>)\n";
+	if (scalar(@$mids) == 1) { # common case
+		my $mhtml = ascii_html($mids->[0]);
+		$rv .= "Message-ID: &lt;$mhtml&gt; ";
+		$rv .= "(<a\nhref=\"raw\">raw</a>)\n";
+	} else {
+		# 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: $_\n" for ($hdr->header_raw($h));
 		}
+		$lnk->linkify_mids('..', \$s, 1);
+		$rv .= $s;
 	}
-	$rv .= _parent_headers($hdr, $srch);
+	$rv .= _parent_headers($hdr, $over);
 	$rv .= "\n";
+	\$rv;
 }
 
-sub thread_skel {
-	my ($dst, $ctx, $hdr, $tpfx) = @_;
-	my $srch = $ctx->{srch};
-	my $mid = mids($hdr)->[0];
-	my ($nr, $msgs) = $srch->get_thread($mid);
-	my $expand = qq(expand[<a\nhref="${tpfx}T/#u">flat</a>) .
-	                qq(|<a\nhref="${tpfx}t/#u">nested</a>]  ) .
-			qq(<a\nhref="${tpfx}t.mbox.gz">mbox.gz</a>  ) .
-			qq(<a\nhref="${tpfx}t.atom">Atom feed</a>);
+sub SKEL_EXPAND () {
+	qq(expand[<a\nhref="T/#u">flat</a>) .
+		qq(|<a\nhref="t/#u">nested</a>]  ) .
+		qq(<a\nhref="t.mbox.gz">mbox.gz</a>  ) .
+		qq(<a\nhref="t.atom">Atom feed</a>);
+}
 
+sub thread_skel ($$$) {
+	my ($skel, $ctx, $hdr) = @_;
+	my $mid = mids($hdr)->[0];
+	my $ibx = $ctx->{-inbox};
+	my ($nr, $msgs) = $ibx->over->get_thread($mid);
 	my $parent = in_reply_to($hdr);
-	$$dst .= "\n<b>Thread overview: </b>";
+	$$skel .= "\n<b>Thread overview: </b>";
 	if ($nr <= 1) {
 		if (defined $parent) {
-			$$dst .= "$expand\n ";
-			$$dst .= ghost_parent("$tpfx../", $parent) . "\n";
+			$$skel .= SKEL_EXPAND."\n ";
+			$$skel .= ghost_parent('../', $parent) . "\n";
 		} else {
-			$$dst .= "[no followups] $expand\n";
+			$$skel .= '[no followups] '.SKEL_EXPAND."\n";
 		}
 		$ctx->{next_msg} = undef;
 		$ctx->{parent_msg} = $parent;
 		return;
 	}
 
-	$$dst .= "$nr+ messages / $expand";
-	$$dst .= qq!  <a\nhref="#b">top</a>\n!;
+	$$skel .= $nr;
+	$$skel .= '+ messages / '.SKEL_EXPAND.qq!  <a\nhref="#b">top</a>\n!;
 
-	my $subj = $hdr->header('Subject');
-	defined $subj or $subj = '';
-	$ctx->{prev_subj} = [ split(/ /, $srch->subject_normalized($subj)) ];
+	# nb: mutt only shows the first Subject in the index pane
+	# 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;
-	$ctx->{dst} = $dst;
+	$ctx->{skel} = $skel;
 
 	# reduce hash lookups in skel_dump
-	my $ibx = $ctx->{-inbox};
 	$ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
-	walk_thread(thread_results($ctx, $msgs), $ctx, *skel_dump);
+	walk_thread(thread_results($ctx, $msgs), $ctx, \&skel_dump);
 
 	$ctx->{parent_msg} = $parent;
 }
 
 sub _parent_headers {
-	my ($hdr, $srch) = @_;
+	my ($hdr, $over) = @_;
 	my $rv = '';
-
-	my $refs = references($hdr);
-	my $irt = pop @$refs;
-	if (defined $irt) {
-		my $v = PublicInbox::Hval->new_msgid($irt);
-		my $html = $v->as_html;
-		my $href = $v->{href};
-		$rv .= "In-Reply-To: &lt;";
-		$rv .= "<a\nhref=\"../$href/\">$html</a>&gt;\n";
+	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 $srch;
+	return $rv if $over;
 
+	$refs //= references($hdr);
 	if (@$refs) {
-		@$refs = map { linkify_ref_nosrch($_) } @$refs;
+		@$refs = map { linkify_ref_no_over($_) } @$refs;
 		$rv .= 'References: '. join("\n\t", @$refs) . "\n";
 	}
 	$rv;
 }
 
+# returns a string buffer via ->getline
 sub html_footer {
-	my ($hdr, $standalone, $ctx, $rhref) = @_;
-
-	my $srch = $ctx->{srch} if $ctx;
+	my ($ctx, $hdr) = @_;
+	my $ibx = $ctx->{-inbox};
 	my $upfx = '../';
-	my $tpfx = '';
-	my $idx = $standalone ? " <a\nhref=\"$upfx\">index</a>" : '';
-	my $irt = '';
-	if ($idx && $srch) {
-		$idx .= "\n";
-		thread_skel(\$idx, $ctx, $hdr, $tpfx);
+	my $skel = " <a\nhref=\"$upfx\">index</a>";
+	my $rv = '<pre>';
+	if ($ibx->over) {
+		$skel .= "\n";
+		thread_skel(\$skel, $ctx, $hdr);
 		my ($next, $prev);
 		my $parent = '       ';
 		$next = $prev = '    ';
 
 		if (my $n = $ctx->{next_msg}) {
-			$n = PublicInbox::Hval->new_msgid($n)->{href};
+			$n = mid_href($n);
 			$next = "<a\nhref=\"$upfx$n/\"\nrel=next>next</a>";
 		}
 		my $u;
 		my $par = $ctx->{parent_msg};
 		if ($par) {
-			$u = PublicInbox::Hval->new_msgid($par)->{href};
+			$u = mid_href($par);
 			$u = "$upfx$u/";
 		}
 		if (my $p = $ctx->{prev_msg}) {
-			$prev = PublicInbox::Hval->new_msgid($p)->{href};
+			$prev = mid_href($p);
 			if ($p && $par && $p eq $par) {
 				$prev = "<a\nhref=\"$upfx$prev/\"\n" .
 					'rel=prev>prev parent</a>';
@@ -783,19 +789,18 @@ sub html_footer {
 		} elsif ($u) { # unlikely
 			$parent = " <a\nhref=\"$u\"\nrel=prev>parent</a>";
 		}
-		$irt = "$next $prev$parent ";
-	} else {
-		$irt = '';
+		$rv .= "$next $prev$parent ";
 	}
-	$rhref ||= '#R';
-	$irt .= qq(<a\nhref="$rhref">reply</a>);
-	$irt .= $idx;
+	$rv .= qq(<a\nhref="#R">reply</a>);
+	$rv .= $skel;
+	$rv .= '</pre>';
+	$rv .= msg_reply($ctx, $hdr);
 }
 
-sub linkify_ref_nosrch {
-	my $v = PublicInbox::Hval->new_msgid($_[0]);
-	my $html = $v->as_html;
-	my $href = $v->{href};
+sub linkify_ref_no_over {
+	my ($mid) = @_;
+	my $href = mid_href($mid);
+	my $html = ascii_html($mid);
 	"&lt;<a\nhref=\"../$href/\">$html</a>&gt;";
 }
 
@@ -807,9 +812,8 @@ sub anchor_for {
 sub ghost_parent {
 	my ($upfx, $mid) = @_;
 
-	$mid = PublicInbox::Hval->new_msgid($mid);
-	my $href = $mid->{href};
-	my $html = $mid->as_html;
+	my $href = mid_href($mid);
+	my $html = ascii_html($mid);
 	qq{[parent not found: &lt;<a\nhref="$upfx$href/">$html</a>&gt;]};
 }
 
@@ -843,8 +847,7 @@ sub strict_loose_note ($) {
 sub thread_results {
 	my ($ctx, $msgs) = @_;
 	require PublicInbox::SearchThread;
-	my $ibx = $ctx->{-inbox};
-	my $rootset = PublicInbox::SearchThread::thread($msgs, *sort_ds, $ibx);
+	my $rootset = PublicInbox::SearchThread::thread($msgs, \&sort_ds, $ctx);
 
 	# FIXME: `tid' is broken on --reindex, so that needs to be fixed
 	# and preserved in the future.  This bug is hidden by `sid' matches
@@ -858,7 +861,7 @@ sub thread_results {
 	if (defined($mid) && scalar(@$rootset) > 1) {
 		$ctx->{root_idx} = -1;
 		my $nr = scalar @$msgs;
-		walk_thread($rootset, $ctx, *find_mid_root);
+		walk_thread($rootset, $ctx, \&find_mid_root);
 		my $idx = $ctx->{found_mid_at};
 		if (defined($idx) && $idx != 0) {
 			my $tip = splice(@$rootset, $idx, 1);
@@ -876,49 +879,52 @@ sub missing_thread {
 	PublicInbox::ExtMsg::ext_msg($ctx);
 }
 
-sub _msg_date {
-	my ($hdr) = @_;
-	fmt_ts(msg_datestamp($hdr));
-}
-
-sub fmt_ts { POSIX::strftime('%Y-%m-%d %k:%M', gmtime($_[0])) }
-
 sub dedupe_subject {
 	my ($prev_subj, $subj, $val) = @_;
 
-	my $omit = ''; # '"' denotes identical text omitted
+	my $omit; # '"' denotes identical text omitted
 	my (@prev_pop, @curr_pop);
 	while (@$prev_subj && @$subj && $subj->[-1] eq $prev_subj->[-1]) {
 		push(@prev_pop, pop(@$prev_subj));
 		push(@curr_pop, pop(@$subj));
-		$omit ||= $val;
+		$omit //= $val;
 	}
 	pop @$subj if @$subj && $subj->[-1] =~ /^re:\s*/i;
 	if (scalar(@curr_pop) == 1) {
-		$omit = '';
+		$omit = undef;
 		push @$prev_subj, @prev_pop;
 		push @$subj, @curr_pop;
 	}
-	$omit;
+	$omit // '';
 }
 
-sub skel_dump {
+sub skel_dump { # walk_thread callback
 	my ($ctx, $level, $node) = @_;
 	my $smsg = $node->{smsg} or return _skel_ghost($ctx, $level, $node);
 
-	my $dst = $ctx->{dst};
+	my $skel = $ctx->{skel};
 	my $cur = $ctx->{cur};
 	my $mid = $smsg->{mid};
 
 	if ($level == 0 && $ctx->{skel_dump_roots}++) {
-		$$dst .= delete $ctx->{sl_note} || '';
+		$$skel .= delete($ctx->{sl_note}) || '';
 	}
 
-	my $f = ascii_html($smsg->from_name);
+	my $f = ascii_html($smsg->{from_name});
 	my $obfs_ibx = $ctx->{-obfs_ibx};
 	obfuscate_addrs($obfs_ibx, $f) if $obfs_ibx;
 
-	my $d = fmt_ts($smsg->{ds}) . ' ' . indent_for($level) . th_pfx($level);
+	my $d = fmt_ts($smsg->{ds});
+	my $unmatched; # if lazy-loaded by SearchThread::Msg::visible()
+	if (exists $ctx->{searchview}) {
+		if (defined(my $pct = $smsg->{pct})) {
+			$d .= (sprintf(' % 2u', $pct) . '%');
+		} else {
+			$unmatched = 1;
+			$d .= '    ';
+		}
+	}
+	$d .= ' ' . indent_for($level) . th_pfx($level);
 	my $attr = $f;
 	$ctx->{first_level} ||= $level;
 
@@ -930,7 +936,7 @@ sub skel_dump {
 	if ($cur) {
 		if ($cur eq $mid) {
 			delete $ctx->{cur};
-			$$dst .= "<b>$d<a\nid=r\nhref=\"#t\">".
+			$$skel .= "<b>$d<a\nid=r\nhref=\"#t\">".
 				 "$attr [this message]</a></b>\n";
 			return 1;
 		} else {
@@ -943,8 +949,7 @@ sub skel_dump {
 	# Subject is never undef, this mail was loaded from
 	# our Xapian which would've resulted in '' if it were
 	# really missing (and Filter rejects empty subjects)
-	my @subj = split(/ /, $ctx->{srch}->subject_normalized($smsg->subject));
-
+	my @subj = split(/ /, subject_normalized($smsg->{subject}));
 	# remove common suffixes from the subject if it matches the previous,
 	# so we do not show redundant text at the end.
 	my $prev_subj = $ctx->{prev_subj} || [];
@@ -961,7 +966,7 @@ sub skel_dump {
 	}
 	my $m;
 	my $id = '';
-	my $mapping = $ctx->{mapping};
+	my $mapping = $unmatched ? undef : $ctx->{mapping};
 	if ($mapping) {
 		my $map = $mapping->{$mid};
 		$id = id_compress($mid, 1);
@@ -969,9 +974,9 @@ sub skel_dump {
 		$map->[0] = "$d<a\nhref=\"$m\">$end";
 		$id = "\nid=r".$id;
 	} else {
-		$m = $ctx->{-upfx}.mid_escape($mid).'/';
+		$m = $ctx->{-upfx}.mid_href($mid).'/';
 	}
-	$$dst .=  $d . "<a\nhref=\"$m\"$id>" . $end;
+	$$skel .=  $d . "<a\nhref=\"$m\"$id>" . $end;
 	1;
 }
 
@@ -979,13 +984,12 @@ sub _skel_ghost {
 	my ($ctx, $level, $node) = @_;
 
 	my $mid = $node->{id};
-	my $d = $ctx->{pct} ? '    [irrelevant] ' # search result
-			    : '     [not found] ';
+	my $d = '     [not found] ';
+	$d .= '    '  if exists $ctx->{searchview};
 	$d .= indent_for($level) . th_pfx($level);
 	my $upfx = $ctx->{-upfx};
-	my $m = PublicInbox::Hval->new_msgid($mid);
-	my $href = $upfx . $m->{href} . '/';
-	my $html = $m->as_html;
+	my $href = $upfx . mid_href($mid) . '/';
+	my $html = ascii_html($mid);
 
 	my $mapping = $ctx->{mapping};
 	my $map = $mapping->{$mid} if $mapping;
@@ -996,52 +1000,48 @@ sub _skel_ghost {
 	} else {
 		$d .= qq{&lt;<a\nhref="$href">$html</a>&gt;\n};
 	}
-	my $dst = $ctx->{dst};
-	$$dst .= $d;
+	${$ctx->{skel}} .= $d;
 	1;
 }
 
 sub sort_ds {
 	[ sort {
-		(eval { $a->topmost->{smsg}->ds } || 0) <=>
-		(eval { $b->topmost->{smsg}->ds } || 0)
+		(eval { $a->topmost->{smsg}->{ds} } || 0) <=>
+		(eval { $b->topmost->{smsg}->{ds} } || 0)
 	} @{$_[0]} ];
 }
 
 # accumulate recent topics if search is supported
 # returns 200 if done, 404 if not
-sub acc_topic {
+sub acc_topic { # walk_thread callback
 	my ($ctx, $level, $node) = @_;
-	my $srch = $ctx->{srch};
 	my $mid = $node->{id};
-	my $x = $node->{smsg} || $ctx->{-inbox}->smsg_by_mid($mid);
-	my ($subj, $ds);
-	my $topic;
-	if ($x) {
-		$subj = $x->subject;
-		$subj = $srch->subject_normalized($subj);
-		$ds = $x->ds;
-		if ($level == 0) {
-			$topic = [ $ds, 1, { $subj => $mid }, $subj ];
+	my $smsg = $node->{smsg} // $ctx->{-inbox}->smsg_by_mid($mid);
+	if ($smsg) {
+		my $subj = subject_normalized($smsg->{subject});
+		$subj = '(no subject)' if $subj eq '';
+		my $ds = $smsg->{ds};
+		if ($level == 0) { # new, top-level topic
+			my $topic = [ $ds, 1, { $subj => $mid }, $subj ];
 			$ctx->{-cur_topic} = $topic;
 			push @{$ctx->{order}}, $topic;
 			return 1;
 		}
 
-		$topic = $ctx->{-cur_topic}; # should never be undef
+		# continue existing topic
+		my $topic = $ctx->{-cur_topic}; # should never be undef
 		$topic->[0] = $ds if $ds > $topic->[0];
-		$topic->[1]++;
+		$topic->[1]++; # bump N+ message counter
 		my $seen = $topic->[2];
 		if (scalar(@$topic) == 3) { # parent was a ghost
 			push @$topic, $subj;
-		} elsif (!$seen->{$subj}) {
-			push @$topic, $level, $subj;
+		} elsif (!defined($seen->{$subj})) {
+			push @$topic, $level, $subj; # @extra messages
 		}
 		$seen->{$subj} = $mid; # latest for subject
 	} else { # ghost message
 		return 1 if $level != 0; # ignore child ghosts
-		$topic = [ -666, 0, {} ];
-		$ctx->{-cur_topic} = $topic;
+		my $topic = $ctx->{-cur_topic} = [ -666, 0, {} ];
 		push @{$ctx->{order}}, $topic;
 	}
 	1;
@@ -1050,7 +1050,7 @@ sub acc_topic {
 sub dump_topics {
 	my ($ctx) = @_;
 	my $order = delete $ctx->{order}; # [ ds, subj1, subj2, subj3, ... ]
-	if (!@$order) {
+	unless ($order) {
 		$ctx->{-html_tip} = '<pre>[No topics in range]</pre>';
 		return 404;
 	}
@@ -1058,17 +1058,16 @@ sub dump_topics {
 	my @out;
 	my $ibx = $ctx->{-inbox};
 	my $obfs_ibx = $ibx->{obfuscate} ? $ibx : undef;
-	my $srch = $ctx->{srch};
 
 	# sort by recency, this allows new posts to "bump" old topics...
 	foreach my $topic (sort { $b->[0] <=> $a->[0] } @$order) {
-		my ($ds, $n, $seen, $top, @ex) = @$topic;
+		my ($ds, $n, $seen, $top_subj, @extra) = @$topic;
 		@$topic = ();
-		next unless defined $top;  # ghost topic
-		my $mid = delete $seen->{$top};
-		my $href = mid_escape($mid);
-		my $prev_subj = [ split(/ /, $top) ];
-		$top = PublicInbox::Hval->new($top)->as_html;
+		next unless defined $top_subj;  # ghost topic
+		my $mid = delete $seen->{$top_subj};
+		my $href = mid_href($mid);
+		my $prev_subj = [ split(/ /, $top_subj) ];
+		$top_subj = ascii_html($top_subj);
 		$ds = fmt_ts($ds);
 
 		# $n isn't the total number of posts on the topic,
@@ -1084,21 +1083,21 @@ sub dump_topics {
 
 		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/$anchor\"><b>$top</b></a>\n" .
+		my $s = "<a\nhref=\"$href/T/$anchor\">$top_subj</a>\n" .
 			" $ds UTC $n - $mbox / $atom\n";
-		for (my $i = 0; $i < scalar(@ex); $i += 2) {
-			my $level = $ex[$i];
-			my $subj = $ex[$i + 1];
+		for (my $i = 0; $i < scalar(@extra); $i += 2) {
+			my $level = $extra[$i];
+			my $subj = $extra[$i + 1]; # already normalized
 			$mid = delete $seen->{$subj};
-			my @subj = split(/ /, $srch->subject_normalized($subj));
+			my @subj = split(/ /, $subj);
 			my @next_prev = @subj; # full copy
 			my $omit = dedupe_subject($prev_subj, \@subj, ' &#34;');
 			$prev_subj = \@next_prev;
-			$subj = ascii_html(join(' ', @subj));
+			$subj = ascii_html($subj);
 			obfuscate_addrs($obfs_ibx, $subj) if $obfs_ibx;
-			$href = mid_escape($mid);
+			$href = mid_href($mid);
 			$s .= indent_for($level) . TCHILD;
-			$s .= "<a\nhref=\"$href/T/#u\">$subj</a>$omit\n";
+			$s .= qq(<a\nhref="$href/T/#u">$subj</a>$omit\n);
 		}
 		push @out, $s;
 	}
@@ -1106,10 +1105,8 @@ sub dump_topics {
 	200;
 }
 
-sub ts2str ($) {
-	my ($ts) = @_;
-	POSIX::strftime('%Y%m%d%H%M%S', gmtime($ts));
-}
+# only for the t= query parameter passed to overview DB
+sub ts2str ($) { strftime('%Y%m%d%H%M%S', gmtime($_[0])) };
 
 sub str2ts ($) {
 	my ($yyyy, $mon, $dd, $hh, $mm, $ss) = unpack('A4A2A2A2A2A2', $_[0]);
@@ -1142,8 +1139,8 @@ sub paginate_recent ($$) {
 	# Xapian uses '..' but '-' is perhaps friendier to URL linkifiers
 	# if only $after exists "YYYYMMDD.." because "." could be skipped
 	# if interpreted as an end-of-sentence
-	$t =~ s/\A(\d{8,14})-// and $after = str2ts($1);
-	$t =~ /\A(\d{8,14})\z/ and $before = str2ts($1);
+	$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->{-inbox};
 	my $msgs = $ibx->recent($opts, $after, $before);
@@ -1179,9 +1176,9 @@ sub index_topics {
 	my ($ctx) = @_;
 	my $msgs = paginate_recent($ctx, 200); # 200 is our window
 	if (@$msgs) {
-		walk_thread(thread_results($ctx, $msgs), $ctx, *acc_topic);
+		walk_thread(thread_results($ctx, $msgs), $ctx, \&acc_topic);
 	}
-	PublicInbox::WwwStream->response($ctx, dump_topics($ctx), *index_nav);
+	PublicInbox::WwwStream->response($ctx, dump_topics($ctx), \&index_nav);
 }
 
 sub thread_adj_level {