X-Git-Url: http://www.git.stargrave.org/?p=public-inbox.git;a=blobdiff_plain;f=lib%2FPublicInbox%2FView.pm;h=fa96cca337d87f9113425a209c5b78aafa2b6dfa;hp=83b622fbde0cc43dd9f6b83ce26ef7ef5fe090ab;hb=4eee5af6011cc8cdefb66c9729952c7eff5c0b0b;hpb=021b862ac94f1c226f0778a336d4f3d566b86f9b

diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm
index 83b622fb..fa96cca3 100644
--- a/lib/PublicInbox/View.pm
+++ b/lib/PublicInbox/View.pm
@@ -1,119 +1,92 @@
-# Copyright (C) 2014-2019 all contributors <meta@public-inbox.org>
+# Copyright (C) 2014-2021 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 PublicInbox::MsgTime qw(msg_datestamp);
-use PublicInbox::Hval qw(ascii_html obfuscate_addrs prurl);
+use PublicInbox::Hval qw(ascii_html obfuscate_addrs prurl mid_href
+			ts2str fmt_ts);
 use PublicInbox::Linkify;
-use PublicInbox::MID qw/id_compress mid_escape mids mids_for_index references/;
+use PublicInbox::MID qw(id_compress mids mids_for_index references
+			$MID_EXTRACT);
 use PublicInbox::MsgIter;
 use PublicInbox::Address;
-use PublicInbox::WwwStream;
+use PublicInbox::WwwStream qw(html_oneshot);
 use PublicInbox::Reply;
 use PublicInbox::ViewDiff qw(flush_diff);
-use POSIX qw(strftime);
+use PublicInbox::Eml;
 use Time::Local qw(timegm);
-use PublicInbox::SearchMsg qw(subject_normalized);
+use PublicInbox::Smsg qw(subject_normalized);
+use PublicInbox::ContentHash qw(content_hash);
 use constant COLS => 72;
 use constant INDENT => '  ';
 use constant TCHILD => '` ';
 sub th_pfx ($) { $_[0] == 0 ? '' : TCHILD };
 
-sub msg_html_i {
-	my ($nr, $ctx) = @_;
-	my $more = $ctx->{more};
-	if ($nr == 1) {
-		# $more cannot be true w/o $smsg being defined:
-		my $upfx = $more ? '../'.mid_escape($ctx->{smsg}->mid).'/' : '';
-		$ctx->{tip} .
-			multipart_text_as_html(delete $ctx->{mime}, $upfx,
-						$ctx) . '</pre><hr>'
-	} elsif ($more && @$more) {
-		++$ctx->{end_nr};
-		msg_html_more($ctx, $more, $nr);
-	} elsif ($nr == $ctx->{end_nr}) {
-		# fake an EOF if generating the footer fails;
-		# we want to at least show the message if something
-		# here crashes:
-		eval {
-			my $hdr = delete($ctx->{hdr});
-			'<pre>' . html_footer($hdr, 1, $ctx) .
-			'</pre>' . msg_reply($ctx, $hdr)
-		};
-	} else {
-		undef
+sub msg_page_i {
+	my ($ctx, $eml) = @_;
+	if ($eml) { # called by WwwStream::async_eml or getline
+		my $smsg = $ctx->{smsg};
+		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);
+		if (length($$obuf)) {
+			multipart_text_as_html($eml, $ctx);
+			$$obuf .= '</pre><hr>';
+		}
+		delete $ctx->{obuf};
+		$$obuf .= html_footer($ctx, $ctx->{first_hdr}) if !$ctx->{smsg};
+		$$obuf;
+	} else { # called by WwwStream::async_next or getline
+		$ctx->{smsg}; # may be undef
 	}
 }
 
-# public functions: (unstable)
-
-sub msg_html {
-	my ($ctx, $mime, $more, $smsg) = @_;
-	my $ibx = $ctx->{-inbox};
-	$ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
-	my $hdr = $ctx->{hdr} = $mime->header_obj;
-	$ctx->{tip} = _msg_html_prepare($hdr, $ctx, $more, 0);
-	$ctx->{more} = $more;
-	$ctx->{end_nr} = 2;
-	$ctx->{smsg} = $smsg;
-	$ctx->{mime} = $mime;
-	PublicInbox::WwwStream->response($ctx, 200, \&msg_html_i);
+# /$INBOX/$MSGID/ for unindexed v1 inboxes
+sub no_over_html ($) {
+	my ($ctx) = @_;
+	my $bref = $ctx->{ibx}->msg_by_mid($ctx->{mid}) or return; # 404
+	my $eml = PublicInbox::Eml->new($bref);
+	$ctx->{mhref} = '';
+	PublicInbox::WwwStream::init($ctx);
+	my $obuf = $ctx->{obuf} = _msg_page_prepare_obuf($eml, $ctx);
+	if (length($$obuf)) {
+		multipart_text_as_html($eml, $ctx);
+		$$obuf .= '</pre><hr>';
+	}
+	delete $ctx->{obuf};
+	eval { $$obuf .= html_footer($ctx, $eml) };
+	html_oneshot($ctx, 200, $obuf);
 }
 
+# public functions: (unstable)
+
 sub msg_page {
 	my ($ctx) = @_;
-	my $mid = $ctx->{mid};
-	my $ibx = $ctx->{-inbox};
-	my ($first, $more);
-	my $smsg;
-	if (my $over = $ibx->over) {
-		my ($id, $prev);
-		$smsg = $over->next_by_mid($mid, \$id, \$prev);
-		$first = $ibx->msg_by_smsg($smsg) if $smsg;
-		if ($first) {
-			my $next = $over->next_by_mid($mid, \$id, \$prev);
-			$more = [ $id, $prev, $next ] if $next;
-		}
-		return unless $first;
-	} else {
-		$first = $ibx->msg_by_mid($mid) or return;
-	}
-	msg_html($ctx, PublicInbox::MIME->new($first), $more, $smsg);
-}
+	my $ibx = $ctx->{ibx};
+	$ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
+	my $over = $ibx->over or return no_over_html($ctx);
+	my ($id, $prev);
+	my $next_arg = $ctx->{next_arg} = [ $ctx->{mid}, \$id, \$prev ];
 
-sub msg_html_more {
-	my ($ctx, $more, $nr) = @_;
-	my $str = eval {
-		my ($id, $prev, $smsg) = @$more;
-		my $mid = $ctx->{mid};
-		my $ibx = $ctx->{-inbox};
-		$smsg = $ibx->smsg_mime($smsg);
-		my $next = $ibx->over->next_by_mid($mid, \$id, \$prev);
-		@$more = $next ? ($id, $prev, $next) : ();
-		if ($smsg) {
-			my $upfx = '../' . mid_escape($smsg->mid) . '/';
-			my $mime = delete $smsg->{mime};
-			_msg_html_prepare($mime->header_obj, $ctx, $more, $nr) .
-				multipart_text_as_html($mime, $upfx, $ctx) .
-				'</pre><hr>'
-		} else {
-			'';
-		}
-	};
-	if ($@) {
-		warn "Error lookup up additional messages: $@\n";
-		$str = '<pre>Error looking up additional messages</pre>';
-	}
-	$str;
+	my $smsg = $ctx->{smsg} = $over->next_by_mid(@$next_arg) or
+		return; # undef == 404
+
+	# 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);
 }
 
 # /$INBOX/$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';
@@ -121,7 +94,7 @@ sub msg_reply {
 	 'https://en.wikipedia.org/wiki/Posting_style#Interleaved_style';
 
 	my $info = '';
-	my $ibx = $ctx->{-inbox};
+	my $ibx = $ctx->{ibx};
 	if (my $url = $ibx->{infourl}) {
 		$url = prurl($ctx->{env}, $url);
 		$info = qq(\n  List information: <a\nhref="$url">$url</a>\n);
@@ -169,6 +142,9 @@ $info
   <a
 href="$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
 }
 
@@ -181,7 +157,7 @@ sub in_reply_to {
 sub fold_addresses ($) {
 	return $_[0] if length($_[0]) <= COLS;
 	# try to fold on commas after non-word chars before $lim chars,
-	# Try to get the "," preceeded by ">" or ")", but avoid folding
+	# Try to get the "," preceded by ">" or ")", but avoid folding
 	# on the comma where somebody uses "Lastname, Firstname".
 	# We also try to keep the last and penultimate addresses in
 	# the list on the same line if possible, hence the extra \z
@@ -208,17 +184,15 @@ sub nr_to_s ($$$) {
 	$nr == 1 ? "$nr $singular" : "$nr $plural";
 }
 
-# human-friendly format
-sub fmt_ts ($) { strftime('%Y-%m-%d %k:%M', gmtime($_[0])) }
-
+# Displays the text of of the message for /$INBOX/$MSGID/[Tt]/ endpoint
 # this is already inside a <pre>
-sub index_entry {
-	my ($smsg, $ctx, $more) = @_;
-	my $subj = $smsg->subject;
-	my $mid_raw = $smsg->mid;
+sub eml_entry {
+	my ($ctx, $eml) = @_;
+	my $smsg = delete $ctx->{smsg};
+	my $subj = delete $smsg->{subject};
+	my $mid_raw = $smsg->{mid};
 	my $id = id_compress($mid_raw, 1);
 	my $id_m = 'm'.$id;
-
 	my $root_anchor = $ctx->{root_anchor} || '';
 	my $irt;
 	my $obfs_ibx = $ctx->{-obfs_ibx};
@@ -231,21 +205,20 @@ sub index_entry {
 	$rv .= $subj . "\n";
 	$rv .= _th_index_lite($mid_raw, \$irt, $id, $ctx);
 	my @tocc;
-	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 $ds = delete $smsg->{ds}; # for v1 non-Xapian/SQLite users
 
-	my $hdr = $mime->header_obj;
-	my $from = _hdr_names_html($hdr, 'From');
+	# Deleting these fields saves about 400K as we iterate across 1K msgs
+	delete @$smsg{qw(ts blob)};
+
+	my $from = _hdr_names_html($eml, 'From');
 	obfuscate_addrs($obfs_ibx, $from) if $obfs_ibx;
 	$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'));
-	my $cc = fold_addresses(_hdr_names_html($hdr, 'Cc'));
+	my $to = fold_addresses(_hdr_names_html($eml, 'To'));
+	my $cc = fold_addresses(_hdr_names_html($eml, 'Cc'));
 	my ($tlen, $clen) = (length($to), length($cc));
 	my $to_cc = '';
 	if (($tlen + $clen) > COLS) {
@@ -264,19 +237,18 @@ sub index_entry {
 	$rv .= $to_cc;
 
 	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 = $mirt->as_html;
+	if (!$mapping && (defined($irt) || defined($irt = in_reply_to($eml)))) {
+		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
 	$ctx->{mhref} = $mhref;
-	$ctx->{rv} = \$rv;
-	msg_iter($mime, \&add_text_body, $ctx, 1);
-	delete $ctx->{rv};
+	$ctx->{obuf} = \$rv;
+	$eml->each_part(\&add_text_body, $ctx, 1);
+	delete $ctx->{obuf};
 
 	# add the footer
 	$rv .= "\n<a\nhref=#$id_m\nid=e$id>^</a> ".
@@ -291,7 +263,6 @@ sub index_entry {
 	} elsif ($mapping) {
 		my $nested = 'nested';
 		my $flat = 'flat';
-		my $end = '';
 		if ($ctx->{flat}) {
 			$hr = 1;
 			$flat = "<b>$flat</b>";
@@ -305,15 +276,23 @@ sub index_entry {
 		$hr = $ctx->{-hr};
 	}
 
-	$rv .= $more ? '</pre><hr><pre>' : '</pre>' if $hr;
+	# do we have more messages? start a new <pre> if so
+	$rv .= scalar(@{$ctx->{msgs}}) ? '</pre><hr><pre>' : '</pre>' if $hr;
 	$rv;
 }
 
 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 {
@@ -329,11 +308,9 @@ sub _th_index_lite {
 	my $nr_c = scalar @$children;
 	my $nr_s = 0;
 	my $siblings;
-	if (my $smsg = $node->{smsg}) {
-		# delete saves about 200KB on a 1K message thread
-		if (my $refs = delete $smsg->{references}) {
-			($$irt) = ($refs =~ m/<([^>]+)>\z/);
-		}
+	# delete saves about 200KB on a 1K message thread
+	if (my $refs = delete $node->{references}) {
+		($$irt) = ($refs =~ m/$MID_EXTRACT\z/o);
 	}
 	my $irt_map = $mapping->{$$irt} if defined $$irt;
 	if (defined $irt_map) {
@@ -342,50 +319,52 @@ sub _th_index_lite {
 		$rv .= $pad . $irt_map->[0];
 		if ($idx > 0) {
 			my $prev = $siblings->[$idx - 1];
-			my $pmid = $prev->{id};
+			my $pmid = $prev->{mid};
 			if ($idx > 2) {
 				my $s = ($idx - 1). ' preceding siblings ...';
 				$rv .= pad_link($pmid, $level, $s);
 			} elsif ($idx == 2) {
-				my $ppmid = $siblings->[0]->{id};
-				$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;
-	$attr =~ s!<a\nhref.*</a> !!s; # no point in duplicating subject
+	$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>@ $attr";
 	if ($nr_c) {
-		my $cmid = $children->[0]->{id};
-		$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->{id}}->[0];
+			$rv .= $pad . _skel_hdr($mapping, $cn->{mid});
 		}
 	}
 
 	my $next = $siblings->[$idx+1] if $siblings && $idx >= 0;
 	if ($next) {
-		my $nmid = $next->{id};
-		$rv .= $pad . $mapping->{$nmid}->[0];
+		my $nmid = $next->{mid};
+		$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->{id}}->[0];
+			$rv .= $pad . _skel_hdr($mapping, $nn->{mid});
 		}
 	}
 	$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) {
@@ -398,78 +377,95 @@ sub walk_thread {
 	}
 }
 
-sub pre_thread  {
+sub pre_thread  { # walk_thread callback
 	my ($ctx, $level, $node, $idx) = @_;
-	$ctx->{mapping}->{$node->{id}} = [ '', $node, $idx, $level ];
+	$ctx->{mapping}->{$node->{mid}} = [ '', $node, $idx, $level ];
 	skel_dump($ctx, $level, $node);
 }
 
-sub thread_index_entry {
-	my ($ctx, $level, $smsg) = @_;
-	my ($beg, $end) = thread_adj_level($ctx, $level);
-	$beg . '<pre>' . index_entry($smsg, $ctx, 0) . '</pre>' . $end;
+sub thread_eml_entry {
+	my ($ctx, $eml) = @_;
+	my ($beg, $end) = thread_adj_level($ctx, $ctx->{level});
+	$beg . '<pre>' . eml_entry($ctx, $eml) . '</pre>' . $end;
 }
 
-sub stream_thread_i { # PublicInbox::WwwStream::getline callback
-	my ($nr, $ctx) = @_;
-	return unless exists($ctx->{dst});
-	my $q = $ctx->{-queue};
+sub next_in_queue ($$) {
+	my ($q, $ghost_ok) = @_;
 	while (@$q) {
-		my $level = shift @$q;
-		my $node = shift @$q or next;
+		my ($level, $smsg) = splice(@$q, 0, 2);
 		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, $node);
+		unshift @$q, map { ($cl, $_) } @{$smsg->{children}};
+		return ($level, $smsg) if $ghost_ok || exists($smsg->{blob});
+	}
+	undef;
+}
+
+sub stream_thread_i { # PublicInbox::WwwStream::getline callback
+	my ($ctx, $eml) = @_;
+	return thread_eml_entry($ctx, $eml) if $eml;
+	return unless exists($ctx->{skel});
+	my $ghost_ok = $ctx->{nr}++;
+	while (1) {
+		my ($lvl, $smsg) = next_in_queue($ctx->{-queue}, $ghost_ok);
+		if ($smsg) {
+			if (exists $smsg->{blob}) { # next message for cat-file
+				$ctx->{level} = $lvl;
+				if (!$ghost_ok) { # first non-ghost
+					$ctx->{-title_html} =
+						ascii_html($smsg->{subject});
+					$ctx->zmore($ctx->html_top);
+				}
+				return $smsg;
+			}
+			# buffer the ghost entry and loop
+			$ctx->zmore(ghost_index_entry($ctx, $lvl, $smsg));
+		} else { # all done
+			$ctx->zmore(join('', thread_adj_level($ctx, 0)));
+			$ctx->zmore(${delete($ctx->{skel})});
+			return;
 		}
 	}
-	join('', thread_adj_level($ctx, 0)) . ${delete $ctx->{dst}}; # skel
 }
 
 sub stream_thread ($$) {
 	my ($rootset, $ctx) = @_;
-	my $ibx = $ctx->{-inbox};
-	my @q = map { (0, $_) } @$rootset;
-	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 = $ibx->smsg_mime($node->{smsg}) and last;
-	}
-	return missing_thread($ctx) unless $smsg;
-
-	$ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
-	$ctx->{-title_html} = ascii_html($smsg->subject);
-	$ctx->{-html_tip} = thread_index_entry($ctx, $level, $smsg);
-	$ctx->{-queue} = \@q;
-	PublicInbox::WwwStream->response($ctx, 200, \&stream_thread_i);
+	$ctx->{-queue} = [ map { (0, $_) } @$rootset ];
+	PublicInbox::WwwStream::aresponse($ctx, 200, \&stream_thread_i);
 }
 
+# /$INBOX/$MSGID/t/ and /$INBOX/$MSGID/T/
 sub thread_html {
 	my ($ctx) = @_;
+	$ctx->{-upfx} = '../../';
 	my $mid = $ctx->{mid};
-	my $ibx = $ctx->{-inbox};
+	my $ibx = $ctx->{ibx};
 	my ($nr, $msgs) = $ibx->over->get_thread($mid);
 	return missing_thread($ctx) if $nr == 0;
+
+	# link $INBOX_DIR/description text to "index_topics" view around
+	# the newest message in this thread
+	my $t = ts2str($ctx->{-t_max} = max(map { delete $_->{ts} } @$msgs));
+	my $t_fmt = fmt_ts($ctx->{-t_max});
+
 	my $skel = '<hr><pre>';
 	$skel .= $nr == 1 ? 'only message in thread' : 'end of thread';
-	$skel .= ", back to <a\nhref=\"../../\">index</a>\n\n";
+	$skel .= <<EOF;
+, other threads:[<a
+href="../../?t=$t">~$t_fmt UTC</a> | <a
+href="../../">newest</a>]
+
+EOF
 	$skel .= "<b\nid=t>Thread overview:</b> ";
 	$skel .= $nr == 1 ? '(only message)' : "$nr+ messages";
 	$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->{dst} = \$skel;
+	$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';
 
@@ -477,72 +473,74 @@ sub thread_html {
 
 	# reduce hash lookups in pre_thread->skel_dump
 	$ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
-	walk_thread($rootset, $ctx, *pre_thread);
+	walk_thread($rootset, $ctx, \&pre_thread);
 
 	$skel .= '</pre>';
 	return stream_thread($rootset, $ctx) unless $ctx->{flat};
 
 	# flat display: lazy load the full message from smsg
-	my $smsg;
-	while (my $m = shift @$msgs) {
-		$smsg = $ibx->smsg_mime($m) and last;
-	}
-	return missing_thread($ctx) unless $smsg;
-	$ctx->{-title_html} = ascii_html($smsg->subject);
-	$ctx->{-html_tip} = '<pre>'.index_entry($smsg, $ctx, scalar @$msgs);
 	$ctx->{msgs} = $msgs;
-	PublicInbox::WwwStream->response($ctx, 200, \&thread_html_i);
+	$ctx->{-html_tip} = '<pre>';
+	PublicInbox::WwwStream::aresponse($ctx, 200, \&thread_html_i);
 }
 
 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 ($ctx, $eml) = @_;
+	if ($eml) {
+		my $smsg = $ctx->{smsg};
+		if (exists $ctx->{-html_tip}) {
+			$ctx->{-title_html} = ascii_html($smsg->{subject});
+			$ctx->zmore($ctx->html_top);
+		}
+		return eml_entry($ctx, $eml);
+	} else {
+		while (my $smsg = shift @{$ctx->{msgs}}) {
+			return $smsg if exists($smsg->{blob});
+		}
+		my $skel = delete($ctx->{skel}) or return; # all done
+		$ctx->zmore($$skel);
+		undef;
 	}
-	my ($skel) = delete @$ctx{qw(dst msgs)};
-	$$skel;
 }
 
 sub multipart_text_as_html {
-	my (undef, $mhref, $ctx) = @_; # $mime = $_[0]
-	$ctx->{mhref} = $mhref;
-	$ctx->{rv} = \(my $rv = '');
+	# ($mime, $ctx) = @_; # each_part may do "$_[0] = undef"
 
 	# scan through all parts, looking for displayable text
-	msg_iter($_[0], \&add_text_body, $ctx, 1);
-	${delete $ctx->{rv}};
+	$_[0]->each_part(\&add_text_body, $_[1], 1);
 }
 
-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($$quot);
-
-	# we use a <span> here to allow users to specify their own
-	# color for quoted text
-	$rv = $l->linkify_2(ascii_html($rv));
-	$$quot = undef;
-	$$s .= qq(<span\nclass="q">) . $rv . '</span>'
+sub submsg_hdr ($$) {
+	my ($ctx, $eml) = @_;
+	my $obfs_ibx = $ctx->{-obfs_ibx};
+	my $rv = $ctx->{obuf};
+	$$rv .= "\n";
+	for my $h (qw(From To Cc Subject Date Message-ID X-Alt-Message-ID)) {
+		my @v = $eml->header($h);
+		for my $v (@v) {
+			obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx;
+			$v = ascii_html($v);
+			$$rv .= "$h: $v\n";
+		}
+	}
 }
 
 sub attach_link ($$$$;$) {
 	my ($ctx, $ct, $p, $fn, $err) = @_;
-	my ($part, $depth, @idx) = @$p;
-	my $nl = $idx[-1] > 1 ? "\n" : '';
-	my $idx = join('.', @idx);
-	my $size = bytes::length($part->body);
+	my ($part, $depth, $idx) = @$p;
+
+	# Eml iteration clobbers multipart ->{bdy}, so do not offer
+	# downloads for 0-byte multipart attachments
+	return unless $part->{bdy};
+
+	my $nl = $idx eq '1' ? '' : "\n"; # like join("\n", ...)
+	my $size = length($part->body);
+	delete $part->{bdy}; # save memory
 
 	# hide attributes normally, unless we want to aid users in
 	# 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 && $fn =~ /\A$PublicInbox::Hval::FN\z/o) {
 		$sfn = $fn;
@@ -551,47 +549,68 @@ sub attach_link ($$$$;$) {
 	} else {
 		$sfn = 'a.bin';
 	}
-	my $rv = $ctx->{rv};
+	my $rv = $ctx->{obuf};
 	$$rv .= qq($nl<a\nhref="$ctx->{mhref}$idx-$sfn">);
 	if ($err) {
-		$$rv .= "[-- Warning: decoded text below may be mangled --]\n";
+		$$rv .= <<EOF;
+[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
+EOF
 	}
 	$$rv .= "[-- Attachment #$idx: ";
 	my $ts = "Type: $ct, Size: $size bytes";
+	my $desc = $part->header('Content-Description') // $fn // '';
 	$desc = ascii_html($desc);
 	$$rv .= ($desc eq '') ? "$ts --]" : "$desc --]\n[-- $ts --]";
 	$$rv .= "</a>\n";
+
+	submsg_hdr($ctx, $part) if $part->{is_submsg};
+
 	undef;
 }
 
-sub add_text_body { # callback for msg_iter
+sub add_text_body { # callback for each_part
 	my ($p, $ctx) = @_;
 	my $upfx = $ctx->{mhref};
-	my $ibx = $ctx->{-inbox};
-	# $p - from msg_iter: [ Email::MIME, depth, @idx ]
-	my ($part, $depth, @idx) = @$p;
+	my $ibx = $ctx->{ibx};
+	my $l = $ctx->{-linkify} //= PublicInbox::Linkify->new;
+	# $p - from each_part: [ Email::MIME-like, depth, $idx ]
+	my ($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 = $ctx->{obuf};
+	if ($part->{is_submsg}) {
+		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};
 
 	# 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
-		$ctx->{-diff} = $diff = [];
+	if ($s =~ /^--- [^\n]+\n\+{3} [^\n]+\n@@ /ms) {
+		# diffstat anchors do not link across attachments or messages,
+		# -apfx is just a stable prefix for making diffstat anchors
+		# linkable to the first diff hunk w/o crossing attachments
+		$idx =~ tr!.!/!; # compatibility with previous versions
+		$ctx->{-apfx} = $upfx . $idx;
+
+		# do attr => filename mappings for diffstats in git diffs:
+		$ctx->{-anchors} = {} if $s =~ /^diff --git /sm;
+		$diff = 1;
 		delete $ctx->{-long_path};
 		my $spfx;
-		if ($ibx->{-repo_objs}) {
-			if (index($upfx, '//') >= 0) { # absolute URL (Atom feeds)
+		# absolute URL (Atom feeds)
+		if ($ibx->{coderepo}) {
+			if (index($upfx, '//') >= 0) {
 				$spfx = $upfx;
 				$spfx =~ s!/([^/]*)/\z!/!;
 			} else {
@@ -608,62 +627,58 @@ sub add_text_body { # callback for msg_iter
 		$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 = split(/((?:^>[^\n]*\n)+)/sm, $s);
-	$s = '';
-	my $rv = $ctx->{rv};
-	if (defined($fn) || $depth > 0 || $err) {
+	my @sections = PublicInbox::MsgIter::split_quotes($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);
 		$$rv .= "\n";
 	}
-	my $l = PublicInbox::Linkify->new;
+	delete $part->{bdy}; # save memory
 	foreach my $cur (@sections) {
 		if ($cur =~ /\A>/) {
-			flush_quote($rv, $l, \$cur);
+			# we use a <span> here to allow users to specify
+			# their own color for quoted text
+			$$rv .= qq(<span\nclass="q">);
+			$$rv .= $l->to_html($cur);
+			$$rv .= '</span>';
 		} elsif ($diff) {
-			@$diff = split(/^/m, $cur);
-			$cur = undef;
-			flush_diff($rv, $ctx, $l);
+			flush_diff($ctx, \$cur);
 		} else {
 			# regular lines, OK
-			$l->linkify_1($cur);
-			$$rv .= $l->linkify_2(ascii_html($cur));
-			$cur = undef;
+			$$rv .= $l->to_html($cur);
 		}
+		undef $cur; # free memory
 	}
-
-	obfuscate_addrs($ibx, $$rv) if $ibx->{obfuscate};
 }
 
-sub _msg_html_prepare {
-	my ($hdr, $ctx, $more, $nr) = @_;
-	my $atom = '';
-	my $over = $ctx->{-inbox}->over;
+sub _msg_page_prepare_obuf {
+	my ($eml, $ctx) = @_;
+	my $over = $ctx->{ibx}->over;
 	my $obfs_ibx = $ctx->{-obfs_ibx};
 	my $rv = '';
-	my $mids = mids_for_index($hdr);
-	if ($nr == 0) {
-		if ($more) {
-			$rv .=
-"<pre>WARNING: multiple messages have this Message-ID\n</pre>";
+	my $mids = mids_for_index($eml);
+	my $nr = $ctx->{nr}++;
+	if ($nr) { # unlikely
+		if ($ctx->{chash} eq content_hash($eml)) {
+			warn "W: BUG? @$mids not deduplicated properly\n";
+			return \$rv;
 		}
-		$rv .= "<pre\nid=b>"; # anchor for body start
-	} else {
+		$rv .=
+"<pre>WARNING: multiple messages have this Message-ID\n</pre>";
 		$rv .= '<pre>';
+	} else {
+		$ctx->{first_hdr} = $eml->header_obj;
+		$ctx->{chash} = content_hash($eml) if $ctx->{smsg}; # reused MID
+		$rv .= "<pre\nid=b>"; # anchor for body start
 	}
-	if ($over) {
-		$ctx->{-upfx} = '../';
-	}
+	$ctx->{-upfx} = '../' if $over;
 	my @title; # (Subject[0], From[0])
-	for my $v ($hdr->header('From')) {
-		$v = PublicInbox::Hval->new($v);
-		my @n = PublicInbox::Address::names($v->raw);
+	for my $v ($eml->header('From')) {
+		my @n = PublicInbox::Address::names($v);
+		$v = ascii_html($v);
 		$title[1] //= ascii_html(join(', ', @n));
-		$v = $v->as_html;
 		if ($obfs_ibx) {
 			obfuscate_addrs($obfs_ibx, $v);
 			obfuscate_addrs($obfs_ibx, $title[1]);
@@ -671,39 +686,40 @@ sub _msg_html_prepare {
 		$rv .= "From: $v\n" if $v ne '';
 	}
 	foreach my $h (qw(To Cc)) {
-		for my $v ($hdr->header($h)) {
+		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 '';
 		}
 	}
-	my @subj = $hdr->header('Subject');
+	my @subj = $eml->header('Subject');
 	if (@subj) {
-		for my $v (@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;
-			$rv .= 'Subject: ';
-			if ($over) {
-				$rv .= qq(<a\nhref="#r"\nid=t>$v</a>\n);
-			} else {
-				$rv .= "$v\n";
-			}
-			$title[0] //= $v;
+			$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)';
 	}
-	for my $v ($hdr->header('Date')) {
+	for my $v ($eml->header('Date')) {
 		$v = ascii_html($v);
 		obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx; # possible :P
-		$rv .= "Date: $v\n";
+		$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;
 	}
-	$ctx->{-title_html} = join(' - ', @title);
 	if (scalar(@$mids) == 1) { # common case
-		my $mid = PublicInbox::Hval->new_msgid($mids->[0]);
-		my $mhtml = $mid->as_html;
+		my $mhtml = ascii_html($mids->[0]);
 		$rv .= "Message-ID: &lt;$mhtml&gt; ";
 		$rv .= "(<a\nhref=\"raw\">raw</a>)\n";
 	} else {
@@ -712,41 +728,45 @@ sub _msg_html_prepare {
 		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));
+			$s .= "$h: $_\n" for ($eml->header_raw($h));
 		}
 		$lnk->linkify_mids('..', \$s, 1);
 		$rv .= $s;
 	}
-	$rv .= _parent_headers($hdr, $over);
+	$rv .= _parent_headers($eml, $over);
 	$rv .= "\n";
+	\$rv;
+}
+
+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 ($dst, $ctx, $hdr, $tpfx) = @_;
+sub thread_skel ($$$) {
+	my ($skel, $ctx, $hdr) = @_;
 	my $mid = mids($hdr)->[0];
-	my $ibx = $ctx->{-inbox};
+	my $ibx = $ctx->{ibx};
 	my ($nr, $msgs) = $ibx->over->get_thread($mid);
-	my $expand = qq(expand[<a\nhref="${tpfx}T/#u">flat</a>) .
-	                qq(|<a\nhref="${tpfx}t/#u">nested</a>]  ) .
-			qq(<a\nhref="${tpfx}t.mbox.gz">mbox.gz</a>  ) .
-			qq(<a\nhref="${tpfx}t.atom">Atom feed</a>);
-
 	my $parent = in_reply_to($hdr);
-	$$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 .= "<a\nid=r>[no followups]</a> ".
+					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!;
 
 	# nb: mutt only shows the first Subject in the index pane
 	# when multiple Subject: headers are present, so we follow suit:
@@ -756,11 +776,11 @@ sub thread_skel {
 	$ctx->{cur} = $mid;
 	$ctx->{prev_attr} = '';
 	$ctx->{prev_level} = 0;
-	$ctx->{dst} = $dst;
+	$ctx->{skel} = $skel;
 
 	# reduce hash lookups in skel_dump
 	$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;
 }
@@ -778,9 +798,8 @@ sub _parent_headers {
 		$refs = references($hdr);
 		my $irt = pop @$refs;
 		if (defined $irt) {
-			my $v = PublicInbox::Hval->new_msgid($irt);
-			my $html = $v->as_html;
-			my $href = $v->{href};
+			my $html = ascii_html($irt);
+			my $href = mid_href($irt);
 			$rv .= "In-Reply-To: &lt;";
 			$rv .= "<a\nhref=\"../$href/\">$html</a>&gt;\n";
 		}
@@ -798,33 +817,39 @@ sub _parent_headers {
 	$rv;
 }
 
+# returns a string buffer
 sub html_footer {
-	my ($hdr, $standalone, $ctx, $rhref) = @_;
-
-	my $ibx = $ctx->{-inbox} if $ctx;
+	my ($ctx, $hdr) = @_;
+	my $ibx = $ctx->{ibx};
 	my $upfx = '../';
-	my $tpfx = '';
-	my $idx = $standalone ? " <a\nhref=\"$upfx\">index</a>" : '';
-	my $irt = '';
-	if ($idx && $ibx->over) {
-		$idx .= "\n";
-		thread_skel(\$idx, $ctx, $hdr, $tpfx);
+	my $skel;
+	my $rv = '<pre>';
+	if ($ibx->over) {
+		my $t = ts2str($ctx->{-t_max});
+		my $t_fmt = fmt_ts($ctx->{-t_max});
+		$skel .= <<EOF;
+	other threads:[<a
+href="$upfx?t=$t">~$t_fmt UTC</a>|<a
+href="$upfx">newest</a>]
+EOF
+
+		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$prev/\"\n" .
 					'rel=prev>prev parent</a>';
@@ -837,33 +862,28 @@ sub html_footer {
 		} elsif ($u) { # unlikely
 			$parent = " <a\nhref=\"$u\"\nrel=prev>parent</a>";
 		}
-		$irt = "$next $prev$parent ";
-	} else {
-		$irt = '';
+		$rv .= "$next $prev$parent ";
+	} else { # unindexed inboxes w/o over
+		$skel = qq( <a\nhref="$upfx">latest</a>);
 	}
-	$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_no_over {
-	my $v = PublicInbox::Hval->new_msgid($_[0]);
-	my $html = $v->as_html;
-	my $href = $v->{href};
+	my ($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);
-}
-
 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;]};
 }
 
@@ -875,9 +895,9 @@ sub indent_for {
 sub find_mid_root {
 	my ($ctx, $level, $node, $idx) = @_;
 	++$ctx->{root_idx} if $level == 0;
-	if ($node->{id} eq $ctx->{mid}) {
+	if ($node->{mid} eq $ctx->{mid}) {
 		$ctx->{found_mid_at} = $ctx->{root_idx};
-		return 0;
+		return 0; # stop iterating
 	}
 	1;
 }
@@ -911,7 +931,7 @@ sub thread_results {
 	if (defined($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, $idx, 1);
@@ -932,35 +952,35 @@ sub missing_thread {
 sub dedupe_subject {
 	my ($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 {
-	my ($ctx, $level, $node) = @_;
-	my $smsg = $node->{smsg} or return _skel_ghost($ctx, $level, $node);
+sub skel_dump { # walk_thread callback
+	my ($ctx, $level, $smsg) = @_;
+	$smsg->{blob} or return _skel_ghost($ctx, $level, $smsg);
 
-	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(delete $smsg->{from_name});
 	my $obfs_ibx = $ctx->{-obfs_ibx};
 	obfuscate_addrs($obfs_ibx, $f) if $obfs_ibx;
 
@@ -986,7 +1006,7 @@ 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 {
@@ -999,7 +1019,7 @@ 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(/ /, 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} || [];
@@ -1024,23 +1044,22 @@ 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;
 }
 
 sub _skel_ghost {
 	my ($ctx, $level, $node) = @_;
 
-	my $mid = $node->{id};
+	my $mid = $node->{mid};
 	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;
@@ -1051,52 +1070,53 @@ 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 { $b->topmost->{smsg}->ds } || 0)
-	} @{$_[0]} ];
+	@{$_[0]} = sort {
+		(eval { $a->topmost->{ds} } || 0) <=>
+		(eval { $b->topmost->{ds} } || 0)
+	} @{$_[0]};
 }
 
 # accumulate recent topics if search is supported
 # returns 200 if done, 404 if not
-sub acc_topic {
-	my ($ctx, $level, $node) = @_;
-	my $mid = $node->{id};
-	my $x = $node->{smsg} || $ctx->{-inbox}->smsg_by_mid($mid);
-	my ($subj, $ds);
-	my $topic;
-	if ($x) {
-		$subj = $x->subject;
-		$subj = subject_normalized($subj);
+sub acc_topic { # walk_thread callback
+	my ($ctx, $level, $smsg) = @_;
+	my $mid = $smsg->{mid};
+	my $has_blob = $smsg->{blob} // do {
+		if (my $by_mid = $ctx->{ibx}->smsg_by_mid($mid)) {
+			%$smsg = (%$smsg, %$by_mid);
+			1;
+		}
+	};
+	if ($has_blob) {
+		my $subj = subject_normalized($smsg->{subject});
 		$subj = '(no subject)' if $subj eq '';
-		$ds = $x->ds;
-		if ($level == 0) {
-			$topic = [ $ds, 1, { $subj => $mid }, $subj ];
+		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}}, $topic;
 	}
 	1;
@@ -1105,24 +1125,24 @@ sub acc_topic {
 sub dump_topics {
 	my ($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;
 	}
 
 	my @out;
-	my $ibx = $ctx->{-inbox};
+	my $ibx = $ctx->{ibx};
 	my $obfs_ibx = $ibx->{obfuscate} ? $ibx : undef;
 
 	# 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,
@@ -1136,23 +1156,22 @@ 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</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];
+		my $s = "<a\nhref=\"$href/T/$anchor\">$top_subj</a>\n" .
+			" $ds UTC $n\n";
+		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(/ /, 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 = 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;
 	}
@@ -1160,31 +1179,22 @@ sub dump_topics {
 	200;
 }
 
-# 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', $_[0]);
-	timegm($ss, $mm, $hh, $dd, $mon - 1, $yyyy);
+	timegm($ss || 0, $mm || 0, $hh || 0, $dd, $mon - 1, $yyyy);
 }
 
 sub pagination_footer ($$) {
 	my ($ctx, $latest) = @_;
-	delete $ctx->{qp} or return;
 	my $next = $ctx->{next_page} || '';
 	my $prev = $ctx->{prev_page} || '';
-	if ($prev) {
-		$next = $next ? "$next " : '     ';
-		$prev .= qq! <a\nhref='$latest'>latest</a>!;
+	if ($prev) {  # aligned padding for: 'next (older) | '
+		$next = $next ? "$next | " : '             | ';
+		$prev .= qq[ | <a\nhref="$latest">latest</a>];
 	}
 	"<hr><pre>page: $next$prev</pre>";
 }
 
-sub index_nav { # callback for WwwStream
-	my (undef, $ctx) = @_;
-	pagination_footer($ctx, '.')
-}
-
 sub paginate_recent ($$) {
 	my ($ctx, $lim) = @_;
 	my $t = $ctx->{qp}->{t} || '';
@@ -1197,7 +1207,7 @@ sub paginate_recent ($$) {
 	$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->{-inbox};
+	my $ibx = $ctx->{ibx};
 	my $msgs = $ibx->recent($opts, $after, $before);
 	my $nr = scalar @$msgs;
 	if ($nr < $lim && defined($after)) {
@@ -1218,22 +1228,26 @@ sub paginate_recent ($$) {
 	}
 	if (defined($oldest) && $more) {
 		my $s = ts2str($oldest);
-		$ctx->{next_page} = qq!<a\nhref="?t=$s"\nrel=next>next</a>!;
+		$ctx->{next_page} = qq[<a\nhref="?t=$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</a>!;
+		$ctx->{prev_page} = qq[<a\nhref="?t=$s-"\nrel=prev>] .
+					'prev (newer)</a>';
 	}
 	$msgs;
 }
 
+# GET /$INBOX - top-level inbox view for indexed inboxes
 sub index_topics {
 	my ($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($ctx), *index_nav);
+	html_oneshot($ctx, dump_topics($ctx), \pagination_footer($ctx, '.'));
+
 }
 
 sub thread_adj_level {
@@ -1263,7 +1277,7 @@ 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->{id})
+	$beg . '<pre>'. ghost_parent($ctx->{-upfx}, $node->{mid} // '?')
 		. '</pre>' . $end;
 }