X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FView.pm;h=60fc1df17f9e5e242d55a7f05a00ca2e7943c07e;hb=dfef0d2c34fbf21d665c40b5ad5069e9113c35c8;hp=38e35bf35c535ec38578ee98335c03caa5b804fb;hpb=073efc43a68cbd6852939ccaf644dbb380cdc143;p=public-inbox.git diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm index 38e35bf3..60fc1df1 100644 --- a/lib/PublicInbox/View.pm +++ b/lib/PublicInbox/View.pm @@ -1,21 +1,19 @@ -# Copyright (C) 2014-2015 all contributors -# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt) +# Copyright (C) 2014-2018 all contributors +# License: AGPL-3.0+ # # Used for displaying the HTML web interface. # See Documentation/design_www.txt for this. package PublicInbox::View; use strict; use warnings; -use URI::Escape qw/uri_escape_utf8/; -use Date::Parse qw/str2time/; -use Encode::MIME::Header; -use Plack::Util; -use PublicInbox::Hval qw/ascii_html/; +use PublicInbox::MsgTime qw(msg_datestamp); +use PublicInbox::Hval qw/ascii_html obfuscate_addrs/; use PublicInbox::Linkify; -use PublicInbox::MID qw/mid_clean id_compress mid_mime/; +use PublicInbox::MID qw/id_compress mid_escape mids references/; use PublicInbox::MsgIter; use PublicInbox::Address; use PublicInbox::WwwStream; +use PublicInbox::Reply; require POSIX; use constant INDENT => ' '; @@ -23,16 +21,25 @@ use constant TCHILD => '` '; sub th_pfx ($) { $_[0] == 0 ? '' : TCHILD }; # public functions: (unstable) + sub msg_html { - my ($ctx, $mime, $footer) = @_; + my ($ctx, $mime, $more, $smsg) = @_; my $hdr = $mime->header_obj; - my $tip = _msg_html_prepare($hdr, $ctx); - PublicInbox::WwwStream->new($ctx, sub { + 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) { - $tip . multipart_text_as_html($mime, '') . - '
' - } elsif ($nr == 2) { + # $more cannot be true w/o $smsg being defined: + my $upfx = $more ? '../'.mid_escape($smsg->mid).'/' : ''; + $tip . multipart_text_as_html($mime, $upfx, $obfs_ibx) . + '
' + } elsif ($more && @$more) { + ++$end; + msg_html_more($ctx, $more, $nr); + } elsif ($nr == $end) { # fake an EOF if generating the footer fails; # we want to at least show the message if something # here crashes: @@ -46,205 +53,419 @@ sub msg_html { }); } +sub msg_page { + my ($ctx) = @_; + my $mid = $ctx->{mid}; + my $ibx = $ctx->{-inbox}; + my ($first, $more, $head, $tail, $db); + my $smsg; + if (my $srch = $ibx->search) { + $srch->retry_reopen(sub { + ($head, $tail, $db) = $srch->each_smsg_by_mid($mid); + for (; !defined($first) && $head != $tail; $head++) { + my @args = ($head, $db, $mid); + $smsg = PublicInbox::SearchMsg->get(@args); + $first = $ibx->msg_by_smsg($smsg); + } + if ($head != $tail) { + $more = [ $head, $tail, $db ]; + } + }); + return unless $first; + } else { + $first = $ibx->msg_by_mid($mid) or return; + } + msg_html($ctx, PublicInbox::MIME->new($first), $more, $smsg); +} + +sub msg_html_more { + my ($ctx, $more, $nr) = @_; + my $str = eval { + my $smsg; + my ($head, $tail, $db) = @$more; + my $mid = $ctx->{mid}; + for (; !defined($smsg) && $head != $tail; $head++) { + my $m = PublicInbox::SearchMsg->get($head, $db, $mid); + $smsg = $ctx->{-inbox}->smsg_mime($m); + } + if ($head == $tail) { # done + @$more = (); + } else { + $more->[0] = $head; + } + if ($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, + $ctx->{-obfs_ibx}) . + '
' + } else { + ''; + } + }; + if ($@) { + warn "Error lookup up additional messages: $@\n"; + $str = '
Error looking up additional messages
'; + } + $str; +} + # /$INBOX/$MESSAGE_ID/#R sub msg_reply { my ($ctx, $hdr) = @_; - my $se_url = 'https://git-htmldocs.bogomips.org/git-send-email.html'; + my $se_url = + 'https://kernel.org/pub/software/scm/git/docs/git-send-email.html'; + my $p_url = + 'https://en.wikipedia.org/wiki/Posting_style#Interleaved_style'; + + my $info = ''; + my $ibx = $ctx->{-inbox}; + if (my $url = $ibx->{infourl}) { + $url = PublicInbox::Hval::prurl($ctx->{env}, $url); + $info = qq(\n List information: $url\n); + } - my ($arg, $link) = mailto_arg_link($hdr); - push @$arg, '/path/to/YOUR_REPLY'; + my ($arg, $link, $reply_to_all) = + PublicInbox::Reply::mailto_arg_link($ibx, $hdr); + + # mailto: link only works if address obfuscation is disabled + if ($link) { + $link = <In-Reply-To header + via mailto: links, try the mailto: link +EOF + } - "". - "You may reply publically to this message via\n". - "plain-text email using any one of the following methods:\n\n" . - "* Save the following mbox file, import it into your mail client,\n" . - " and reply-to-all from there: mbox\n\n" . - "* Reply to all the recipients using the --to, --cc,\n" . - " and --in-reply-to switches of git-send-email(1):\n\n" . - " git send-email \\\n " . - join(" \\\n ", @$arg ). "\n\n" . - qq( $se_url\n\n) . - "* If your mail client supports setting the In-Reply-To" . - " header\n via mailto: links, try the " . - qq(mailto: link\n) . - ''; + push @$arg, '/path/to/YOUR_REPLY'; + $arg = ascii_html(join(" \\\n ", '', @$arg)); + <
Reply instructions:
+
+You may reply publically to this message via plain-text email
+using any one of the following methods:
+
+* Save the following mbox file, import it into your mail client,
+  and $reply_to_all from there: mbox
+
+  Avoid top-posting and favor interleaved quoting:
+  $p_url
+$info
+* Reply using the --to, --cc, and --in-reply-to
+  switches of git-send-email(1):
+
+  git send-email$arg
+
+  $se_url
+$link
+EOF } sub in_reply_to { my ($hdr) = @_; - my $irt = $hdr->header_raw('In-Reply-To'); - - return mid_clean($irt) if (defined $irt); - - my $refs = $hdr->header_raw('References'); - if ($refs && $refs =~ /<([^>]+)>\s*\z/s) { - return $1; - } - undef; + my $refs = references($hdr); + $refs->[-1]; } -sub _hdr_names ($$) { +sub _hdr_names_html ($$) { my ($hdr, $field) = @_; my $val = $hdr->header($field) or return ''; ascii_html(join(', ', PublicInbox::Address::names($val))); } +sub nr_to_s ($$$) { + my ($nr, $singular, $plural) = @_; + return "0 $plural" if $nr == 0; + $nr == 1 ? "$nr $singular" : "$nr $plural"; +} + # this is already inside a
 sub index_entry {
-	my ($mime, $level, $state) = @_;
-	my $midx = $state->{anchor_idx}++;
-	my $ctx = $state->{ctx};
+	my ($smsg, $ctx, $more) = @_;
 	my $srch = $ctx->{srch};
-	my $hdr = $mime->header_obj;
-	my $subj = $hdr->header('Subject');
-
-	my $mid_raw = mid_clean(mid_mime($mime));
-	my $id = anchor_for($mid_raw);
-	my $seen = $state->{seen};
-	$seen->{$id} = "#$id"; # save the anchor for children, later
-
-	my $mid = PublicInbox::Hval->new_msgid($mid_raw);
-
-	my $root_anchor = $state->{root_anchor} || '';
-	my $path = $root_anchor ? '../../' : '';
-	my $href = $mid->as_href;
-	my $irt = in_reply_to($hdr);
-	my $parent_anchor = $seen->{anchor_for($irt)} if defined $irt;
-
-	$subj = ascii_html($subj);
-	$subj = "$subj";
-	$subj = "$subj" if $root_anchor eq $id;
-
-	my $ts = _msg_date($hdr);
-	my $rv = "";
-	$rv .= "$subj\n";
-	my $txt = "${path}$href/raw";
-	my $fh = $state->{fh};
-	my $from = _hdr_names($hdr, 'From');
-	$rv .= "- $from @ $ts UTC (raw)\n";
+	my $subj = $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};
+
+	my $rv = "* ";
+	$subj = ''.ascii_html($subj).'';
+	obfuscate_addrs($obfs_ibx, $subj) if $obfs_ibx;
+	$subj = "$subj" if $root_anchor eq $id_m;
+	$rv .= $subj . "\n";
+	$rv .= _th_index_lite($mid_raw, \$irt, $id, $ctx);
 	my @tocc;
+	my $mime = $smsg->{mime};
+	my $hdr = $mime->header_obj;
 	foreach my $f (qw(To Cc)) {
-		my $dst = _hdr_names($hdr, $f);
-		push @tocc, "$f: $dst" if $dst ne '';
+		my $dst = _hdr_names_html($hdr, $f);
+		if ($dst ne '') {
+			obfuscate_addrs($obfs_ibx, $dst) if $obfs_ibx;
+			push @tocc, "$f: $dst";
+		}
 	}
+	my $from = _hdr_names_html($hdr, 'From');
+	obfuscate_addrs($obfs_ibx, $from) if $obfs_ibx;
+	$rv .= "From: $from @ ".fmt_ts($smsg->ds)." UTC";
+	my $upfx = $ctx->{-upfx};
+	my $mhref = $upfx . mid_escape($mid_raw) . '/';
+	$rv .= qq{ (permalink / };
+	$rv .= qq{raw)\n};
 	$rv .= '  '.join('; +', @tocc) . "\n" if @tocc;
-	$fh->write($rv .= "\n");
 
-	my $mhref = "${path}$href/";
+	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;
+		$rv .= qq(In-Reply-To: <$html>\n)
+	}
+	$rv .= "\n";
 
 	# scan through all parts, looking for displayable text
-	msg_iter($mime, sub { index_walk($fh, $mhref, $_[0]) });
-	$rv = "\n" . html_footer($hdr, 0, $ctx, "$path$href/#R");
-
-	if (defined $irt) {
-		unless (defined $parent_anchor) {
-			my $v = PublicInbox::Hval->new_msgid($irt, 1);
-			$v = $v->as_href;
-			$parent_anchor = "${path}$v/";
-		}
-		$rv .= " parent";
-	}
-	if (my $pct = $state->{pct}) { # used by SearchView.pm
-		$rv .= " [relevance $pct->{$mid_raw}%]";
-	} elsif ($srch) {
-		my $threaded = 'threaded';
+	msg_iter($mime, sub { $rv .= add_text_body($mhref, $obfs_ibx, $_[0]) });
+
+	# add the footer
+	$rv .= "\n^ ".
+		"permalink" .
+		" raw" .
+		" reply";
+
+	my $hr;
+	if (my $pct = $ctx->{pct}) { # used by SearchView.pm
+		$rv .= "\t[relevance $pct->{$mid_raw}%]";
+		$hr = 1;
+	} elsif ($mapping) {
+		my $nested = 'nested';
 		my $flat = 'flat';
 		my $end = '';
 		if ($ctx->{flat}) {
+			$hr = 1;
 			$flat = "$flat";
-			$end = "\n"; # for lynx
 		} else {
-			$threaded = "$threaded";
+			$nested = "$nested";
 		}
-		$rv .= " [$threaded";
-		$rv .= "|$flat]$end";
+		$rv .= "\t[$flat";
+		$rv .= "|$nested]";
+		$rv .= " $ctx->{s_nr}";
+	} else {
+		$hr = $ctx->{-hr};
 	}
-	$fh->write($rv .= '
'); -} -sub thread_html { - my ($ctx, $foot, $srch) = @_; - # $_[0] in sub is the Plack callback - sub { emit_thread_html($_[0], $ctx, $foot, $srch) } + $rv .= $more ? '
' : '
' if $hr; + $rv; } -sub walk_thread { - my ($th, $state, $cb) = @_; - my @q = map { (0, $_) } $th->rootset; - while (@q) { - my $level = shift @q; - my $node = shift @q or next; - $cb->($state, $level, $node); - unshift @q, $level+1, $node->child, $level, $node->next; - } +sub pad_link ($$;$) { + my ($mid, $level, $s) = @_; + $s ||= '...'; + my $id = id_compress($mid, 1); + (' 'x19).indent_for($level).th_pfx($level)."($s)\n"; } -# only private functions below. - -sub emit_thread_html { - my ($res, $ctx, $foot, $srch) = @_; - my $mid = $ctx->{mid}; - my $flat = $ctx->{flat}; - my $msgs = load_results($srch->get_thread($mid, { asc => $flat })); - my $nr = scalar @$msgs; - return missing_thread($res, $ctx) if $nr == 0; - my $seen = {}; - my $state = { - res => $res, - ctx => $ctx, - seen => $seen, - root_anchor => anchor_for($mid), - anchor_idx => 0, - cur_level => 0, - }; +sub _th_index_lite { + my ($mid_raw, $irt, $id, $ctx) = @_; + my $rv = ''; + my $mapping = $ctx->{mapping} or return $rv; + my $pad = ' '; + my $mid_map = $mapping->{$mid_raw}; + defined $mid_map or + return 'public-inbox BUG: '.ascii_html($mid_raw).' not mapped'; + my ($attr, $node, $idx, $level) = @$mid_map; + my $children = $node->{children}; + my $nr_c = scalar @$children; + my $nr_s = 0; + my $siblings; + if (my $smsg = $node->{smsg}) { + ($$irt) = (($smsg->{references} || '') =~ m/<([^>]+)>\z/); + } + my $irt_map = $mapping->{$$irt} if defined $$irt; + if (defined $irt_map) { + $siblings = $irt_map->[1]->{children}; + $nr_s = scalar(@$siblings) - 1; + $rv .= $pad . $irt_map->[0]; + if ($idx > 0) { + my $prev = $siblings->[$idx - 1]; + my $pmid = $prev->{id}; + if ($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 . $mapping->{$pmid}->[0]; + } + } + my $s_s = nr_to_s($nr_s, 'sibling', 'siblings'); + my $s_c = nr_to_s($nr_c, 'reply', 'replies'); + $attr =~ s!\n\z!\n!s; + $attr =~ s! !!s; # no point in duplicating subject + $attr =~ s!]+>([^<]+)!$1!s; # no point linking to self + $rv .= "@ $attr"; + if ($nr_c) { + my $cmid = $children->[0]->{id}; + $rv .= $pad . $mapping->{$cmid}->[0]; + 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]; + } + } - require PublicInbox::Git; - $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir}); - if ($flat) { - pre_anchor_entry($seen, $_) for (@$msgs); - __thread_entry($state, $_, 0) for (@$msgs); - } else { - walk_thread(thread_results($msgs), $state, *thread_entry); - if (my $max = $state->{cur_level}) { - $state->{fh}->write( - ('' x ($max - 1)) . ''); + my $next = $siblings->[$idx+1] if $siblings && $idx >= 0; + if ($next) { + my $nmid = $next->{id}; + $rv .= $pad . $mapping->{$nmid}->[0]; + 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 .= $pad ."$s_s, $s_c; $ctx->{s_nr}\n"; +} - # there could be a race due to a message being deleted in git - # but still being in the Xapian index: - my $fh = delete $state->{fh} or return missing_thread($res, $ctx); +sub walk_thread { + my ($rootset, $ctx, $cb) = @_; + my @q = map { (0, $_, -1) } @$rootset; + while (@q) { + my ($level, $node, $i) = splice(@q, 0, 3); + defined $node or next; + $cb->($ctx, $level, $node, $i); + ++$level; + $i = 0; + unshift @q, map { ($level, $_, $i++) } @{$node->{children}}; + } +} - my $final_anchor = $state->{anchor_idx}; - my $next = ""; - $next .= $final_anchor == 1 ? 'only message in' : 'end of'; - $next .= " thread, back to index"; - $next .= "\ndownload thread: "; - $next .= "mbox.gz"; - $next .= " / follow: Atom feed"; - $fh->write('
' . $next . "\n\n".
-			$foot .  '
'); - $fh->close; +sub pre_thread { + my ($ctx, $level, $node, $idx) = @_; + $ctx->{mapping}->{$node->{id}} = [ '', $node, $idx, $level ]; + skel_dump($ctx, $level, $node); } -sub index_walk { - my ($fh, $upfx, $p) = @_; - my $s = add_text_body($upfx, $p); +sub thread_index_entry { + my ($ctx, $level, $smsg) = @_; + my ($beg, $end) = thread_adj_level($ctx, $level); + $beg . '
' . index_entry($smsg, $ctx, 0) . '
' . $end; +} - return if $s eq ''; +sub stream_thread ($$) { + my ($rootset, $ctx) = @_; + my $inbox = $ctx->{-inbox}; + my @q = map { (0, $_) } @$rootset; + my $level; + my $smsg; + 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; + } + return missing_thread($ctx) unless $smsg; + + $ctx->{-obfs_ibx} = $inbox->{obfuscate} ? $inbox : 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; + }); +} - $fh->write($s); +sub thread_html { + my ($ctx) = @_; + my $mid = $ctx->{mid}; + my $srch = $ctx->{srch}; + my $sres = $srch->get_thread($mid); + my $msgs = load_results($srch, $sres); + my $nr = $sres->{total}; + return missing_thread($ctx) if $nr == 0; + my $skel = '
';
+	$skel .= $nr == 1 ? 'only message in thread' : 'end of thread';
+	$skel .= ", back to index\n\n";
+	$skel .= "Thread overview: ";
+	$skel .= $nr == 1 ? '(only message)' : "$nr+ messages";
+	$skel .= " (download: mbox.gz";
+	$skel .= " / follow: Atom feed)\n";
+	$skel .= "-- links below jump to the message on this page --\n";
+	$ctx->{-upfx} = '../../';
+	$ctx->{cur_level} = 0;
+	$ctx->{dst} = \$skel;
+	$ctx->{prev_attr} = '';
+	$ctx->{prev_level} = 0;
+	$ctx->{root_anchor} = anchor_for($mid);
+	$ctx->{mapping} = {};
+	$ctx->{s_nr} = "$nr+ messages 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);
+
+	$skel .= '
'; + return stream_thread($rootset, $ctx) unless $ctx->{flat}; + + # flat display: lazy load the full message from smsg + my $smsg; + while (my $m = shift @$msgs) { + $smsg = $inbox->smsg_mime($m) and last; + } + return missing_thread($ctx) unless $smsg; + $ctx->{-title_html} = ascii_html($smsg->subject); + $ctx->{-html_tip} = '
'.index_entry($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;
+	});
 }
 
 sub multipart_text_as_html {
-	my ($mime, $upfx) = @_;
+	my ($mime, $upfx, $obfs_ibx) = @_;
 	my $rv = "";
 
 	# scan through all parts, looking for displayable text
-	msg_iter($mime, sub {
-		my ($p) = @_;
-		$rv .= add_text_body($upfx, $p);
-	});
+	msg_iter($mime, sub { $rv .= add_text_body($upfx, $obfs_ibx, $_[0]) });
 	$rv;
 }
 
@@ -262,14 +483,16 @@ sub flush_quote {
 	$$s .= qq() . $rv . ''
 }
 
-sub attach_link ($$$$) {
-	my ($upfx, $ct, $p, $fn) = @_;
+sub attach_link ($$$$;$) {
+	my ($upfx, $ct, $p, $fn, $err) = @_;
 	my ($part, $depth, @idx) = @$p;
 	my $nl = $idx[-1] > 1 ? "\n" : '';
 	my $idx = join('.', @idx);
 	my $size = bytes::length($part->body);
-	$ct ||= 'text/plain';
-	$ct =~ s/;.*//; # no attributes
+
+	# hide attributes normally, unless we want to aid users in
+	# spotting MUA problems:
+	$ct =~ s/;.*// unless $err;
 	$ct = ascii_html($ct);
 	my $desc = $part->header('Content-Description');
 	$desc = $fn unless defined $desc;
@@ -282,72 +505,104 @@ sub attach_link ($$$$) {
 	} else {
 		$sfn = 'a.bin';
 	}
-	my @ret = qq($nl[-- Attachment #$idx: );
+	my $ret = qq($nl);
+	if ($err) {
+		$ret .=
+"[-- Warning: decoded text below may be mangled --]\n";
+	}
+	$ret .= "[-- Attachment #$idx: ";
 	my $ts = "Type: $ct, Size: $size bytes";
-	push(@ret, ($desc eq '') ? "$ts --]" : "$desc --]\n[-- $ts --]");
-	join('', @ret, "\n");
+	$desc = ascii_html($desc);
+	$ret .= ($desc eq '') ? "$ts --]" : "$desc --]\n[-- $ts --]";
+	$ret .= "\n";
 }
 
 sub add_text_body {
-	my ($upfx, $p) = @_; # from msg_iter: [ Email::MIME, depth, @idx ]
-	my ($part, $depth, @idx) = @$p;
-	my $ct = $part->content_type;
+	my ($upfx, $obfs_ibx, $p) = @_;
+	# $p - from msg_iter: [ Email::MIME, depth, @idx ]
+	my ($part, $depth) = @$p; # attachment @idx is unused
+	my $ct = $part->content_type || 'text/plain';
 	my $fn = $part->filename;
 
-	if (defined $ct && $ct =~ m!\btext/x?html\b!i) {
+	if ($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!
-	return attach_link($upfx, $ct, $p, $fn) if $@;
+	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 @lines = split(/^/m, $s);
 	$s = '';
-	if (defined($fn) || $depth > 0) {
-		$s .= attach_link($upfx, $ct, $p, $fn);
+	if (defined($fn) || $depth > 0 || $err) {
+		$s .= attach_link($upfx, $ct, $p, $fn, $err);
 		$s .= "\n";
 	}
 	my @quot;
 	my $l = PublicInbox::Linkify->new;
-	while (defined(my $cur = shift @lines)) {
+	foreach my $cur (@lines) {
 		if ($cur !~ /^>/) {
 			# show the previously buffered quote inline
 			flush_quote(\$s, $l, \@quot) if @quot;
 
 			# regular line, OK
-			$cur = $l->linkify_1($cur);
-			$cur = ascii_html($cur);
-			$s .= $l->linkify_2($cur);
+			$l->linkify_1($cur);
+			$s .= $l->linkify_2(ascii_html($cur));
 		} else {
 			push @quot, $cur;
 		}
 	}
 
-	my $end = "\n";
-	if (@quot) {
-		$end = '';
+	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";
+		}
 	}
-	$s =~ s/[ \t]+$//sgm; # kill per-line trailing whitespace
-	$s =~ s/\A\n+//s; # kill leading blank lines
-	$s =~ s/\s+\z//s; # kill all trailing spaces
-	$s .= $end;
 }
 
 sub _msg_html_prepare {
-	my ($hdr, $ctx) = @_;
+	my ($hdr, $ctx, $more, $nr) = @_;
 	my $srch = $ctx->{srch} if $ctx;
 	my $atom = '';
-	my $rv = ""; # anchor for body start
-
+	my $obfs_ibx = $ctx->{-obfs_ibx};
+	my $rv = '';
+	my $mids = mids($hdr);
+	my $multiple = scalar(@$mids) > 1; # zero, one, infinity
+	if ($nr == 0) {
+		if ($more) {
+			$rv .=
+"
WARNING: multiple messages refer to this Message-ID\n
"; + } + $rv .= ""; # anchor for body start + } else { + $rv .= '
';
+	}
 	if ($srch) {
 		$ctx->{-upfx} = '../';
 	}
 	my @title;
-	my $mid = $hdr->header_raw('Message-ID');
-	$mid = PublicInbox::Hval->new_msgid($mid);
 	foreach my $h (qw(From To Cc Subject Date)) {
 		my $v = $hdr->header($h);
 		defined($v) && ($v ne '') or next;
@@ -356,6 +611,7 @@ sub _msg_html_prepare {
 		if ($h eq 'From') {
 			my @n = PublicInbox::Address::names($v->raw);
 			$title[1] = ascii_html(join(', ', @n));
+			obfuscate_addrs($obfs_ibx, $title[1]) if $obfs_ibx;
 		} elsif ($h eq 'Subject') {
 			$title[0] = $v->as_html;
 			if ($srch) {
@@ -365,14 +621,26 @@ sub _msg_html_prepare {
 			}
 		}
 		$v = $v->as_html;
-		$v =~ s/(\@[^,]+,) /$1\n\t/g if ($h eq 'Cc' || $h eq 'To');
+		obfuscate_addrs($obfs_ibx, $v) if $obfs_ibx;
 		$rv .= "$h: $v\n";
 
 	}
 	$title[0] ||= '(no subject)';
 	$ctx->{-title_html} = join(' - ', @title);
-	$rv .= 'Message-ID: <' . $mid->as_html . '> ';
-	$rv .= "(raw)\n";
+	foreach (@$mids) {
+		my $mid = PublicInbox::Hval->new_msgid($_) ;
+		my $mhtml = $mid->as_html;
+		if ($multiple) {
+			my $href = $mid->{href};
+			$rv .= "Message-ID: ";
+			$rv .= "";
+			$rv .= "<$mhtml> ";
+			$rv .= "(raw)\n";
+		} else {
+			$rv .= "Message-ID: <$mhtml> ";
+			$rv .= "(raw)\n";
+		}
+	}
 	$rv .= _parent_headers($hdr, $srch);
 	$rv .= "\n";
 }
@@ -380,41 +648,45 @@ sub _msg_html_prepare {
 sub thread_skel {
 	my ($dst, $ctx, $hdr, $tpfx) = @_;
 	my $srch = $ctx->{srch};
-	my $mid = mid_clean($hdr->header_raw('Message-ID'));
+	my $mid = mids($hdr)->[0];
 	my $sres = $srch->get_thread($mid);
 	my $nr = $sres->{total};
-	my $expand = qq(expand ) .
-			qq(/ mbox.gz ) .
-			qq(/ Atom feed);
+	my $expand = qq(expand[flat) .
+	                qq(|nested]  ) .
+			qq(mbox.gz  ) .
+			qq(Atom feed);
 
 	my $parent = in_reply_to($hdr);
+	$$dst .= "\nThread overview: ";
 	if ($nr <= 1) {
 		if (defined $parent) {
-			$$dst .= "($expand)\n ";
+			$$dst .= "$expand\n ";
 			$$dst .= ghost_parent("$tpfx../", $parent) . "\n";
 		} else {
-			$$dst .= "[no followups, yet] ($expand)\n";
+			$$dst .= "[no followups] $expand\n";
 		}
 		$ctx->{next_msg} = undef;
 		$ctx->{parent_msg} = $parent;
 		return;
 	}
 
-	$$dst .= "$nr+ messages in thread ($expand";
-	$$dst .= qq! / [top])\n!;
+	$$dst .= "$nr+ messages / $expand";
+	$$dst .= qq!  top\n!;
+
+	my $subj = $hdr->header('Subject');
+	defined $subj or $subj = '';
+	$ctx->{prev_subj} = [ split(/ /, $srch->subject_normalized($subj)) ];
+	$ctx->{cur} = $mid;
+	$ctx->{prev_attr} = '';
+	$ctx->{prev_level} = 0;
+	$ctx->{dst} = $dst;
+	$sres = load_results($srch, $sres);
+
+	# reduce hash lookups in skel_dump
+	my $ibx = $ctx->{-inbox};
+	$ctx->{-obfs_ibx} = $ibx->{obfuscate} ? $ibx : undef;
+	walk_thread(thread_results($ctx, $sres), $ctx, *skel_dump);
 
-	my $subj = $srch->subject_path($hdr->header('Subject'));
-	my $state = {
-		seen => { $subj => 1 },
-		srch => $srch,
-		cur => $mid,
-		prev_attr => '',
-		prev_level => 0,
-		upfx => "$tpfx../",
-		dst => $dst,
-	};
-	walk_thread(thread_results(load_results($sres)), $state, *skel_dump);
-	$ctx->{next_msg} = $state->{next_msg};
 	$ctx->{parent_msg} = $parent;
 }
 
@@ -424,9 +696,9 @@ sub _parent_headers {
 
 	my $irt = in_reply_to($hdr);
 	if (defined $irt) {
-		my $v = PublicInbox::Hval->new_msgid($irt, 1);
+		my $v = PublicInbox::Hval->new_msgid($irt);
 		my $html = $v->as_html;
-		my $href = $v->as_href;
+		my $href = $v->{href};
 		$rv .= "In-Reply-To: <";
 		$rv .= "$html>\n";
 	}
@@ -455,50 +727,6 @@ sub _parent_headers {
 	$rv;
 }
 
-sub squote_maybe ($) {
-	my ($val) = @_;
-	if ($val =~ m{([^\w@\./,\%\+\-])}) {
-		$val =~ s/(['!])/'\\$1'/g; # '!' for csh
-		return "'$val'";
-	}
-	$val;
-}
-
-sub mailto_arg_link {
-	my ($hdr) = @_;
-	my %cc; # everyone else
-	my $to; # this is the From address
-
-	foreach my $h (qw(From To Cc)) {
-		my $v = $hdr->header($h);
-		defined($v) && ($v ne '') or next;
-		my @addrs = PublicInbox::Address::emails($v);
-		foreach my $address (@addrs) {
-			my $dst = lc($address);
-			$cc{$dst} ||= $address;
-			$to ||= $dst;
-		}
-	}
-	my @arg;
-
-	my $subj = $hdr->header('Subject') || '';
-	$subj = "Re: $subj" unless $subj =~ /\bRe:/i;
-	my $mid = $hdr->header_raw('Message-ID');
-	push @arg, '--in-reply-to='.ascii_html(squote_maybe(mid_clean($mid)));
-	my $irt = uri_escape_utf8($mid);
-	delete $cc{$to};
-	push @arg, '--to=' . ascii_html($to);
-	$to = uri_escape_utf8($to);
-	$subj = uri_escape_utf8($subj);
-	my $cc = join(',', sort values %cc);
-	push @arg, '--cc=' . ascii_html($cc);
-	$cc = uri_escape_utf8($cc);
-	my $href = "mailto:$to?In-Reply-To=$irt&Cc=${cc}&Subject=$subj";
-	$href =~ s/%20/+/g;
-
-	(\@arg, $href);
-}
-
 sub html_footer {
 	my ($hdr, $standalone, $ctx, $rhref) = @_;
 
@@ -510,21 +738,35 @@ sub html_footer {
 	if ($idx && $srch) {
 		$idx .= "\n";
 		thread_skel(\$idx, $ctx, $hdr, $tpfx);
-		my $p = $ctx->{parent_msg};
-		my $next = $ctx->{next_msg};
-		if ($p) {
-			$p = PublicInbox::Hval->new_msgid($p);
-			$p = $p->as_href;
-			$irt = "parent ";
-		} else {
-			$irt = ' ' x length('parent ');
+		my ($next, $prev);
+		my $parent = '       ';
+		$next = $prev = '    ';
+
+		if (my $n = $ctx->{next_msg}) {
+			$n = PublicInbox::Hval->new_msgid($n)->{href};
+			$next = "next";
 		}
-		if ($next) {
-			my $n = PublicInbox::Hval->new_msgid($next)->as_href;
-			$irt .= "next ";
-		} else {
-			$irt .= ' ' x length('next ');
+		my $u;
+		my $par = $ctx->{parent_msg};
+		if ($par) {
+			$u = PublicInbox::Hval->new_msgid($par)->{href};
+			$u = "$upfx$u/";
 		}
+		if (my $p = $ctx->{prev_msg}) {
+			$prev = PublicInbox::Hval->new_msgid($p)->{href};
+			if ($p && $par && $p eq $par) {
+				$prev = "prev parent';
+				$parent = '';
+			} else {
+				$prev = "prev';
+				$parent = " parent" if $u;
+			}
+		} elsif ($u) { # unlikely
+			$parent = " parent";
+		}
+		$irt = "$next $prev$parent ";
 	} else {
 		$irt = '';
 	}
@@ -534,347 +776,357 @@ sub html_footer {
 }
 
 sub linkify_ref_nosrch {
-	my $v = PublicInbox::Hval->new_msgid($_[0], 1);
+	my $v = PublicInbox::Hval->new_msgid($_[0]);
 	my $html = $v->as_html;
-	my $href = $v->as_href;
+	my $href = $v->{href};
 	"<$html>";
 }
 
 sub anchor_for {
 	my ($msgid) = @_;
-	my $id = $msgid;
-	if ($id !~ /\A[a-f0-9]{40}\z/) {
-		$id = id_compress(mid_clean($id), 1);
-	}
-	'm' . $id;
-}
-
-sub thread_html_head {
-	my ($hdr, $state) = @_;
-	my $res = delete $state->{res} or die "BUG: no Plack callback in {res}";
-	my $fh = $res->([200, ['Content-Type'=> 'text/html; charset=UTF-8']]);
-	$state->{fh} = $fh;
-
-	my $s = ascii_html($hdr->header('Subject'));
-	$fh->write("$s".
-		qq{! .
-		PublicInbox::Hval::STYLE .
-		"");
-}
-
-sub pre_anchor_entry {
-	my ($seen, $mime) = @_;
-	my $id = anchor_for(mid_mime($mime));
-	$seen->{$id} = "#$id"; # save the anchor for children, later
+	'm' . id_compress($msgid, 1);
 }
 
 sub ghost_parent {
 	my ($upfx, $mid) = @_;
-	# 'subject dummy' is used internally by Mail::Thread
-	return '[no common parent]' if ($mid eq 'subject dummy');
 
 	$mid = PublicInbox::Hval->new_msgid($mid);
-	my $href = $mid->as_href;
+	my $href = $mid->{href};
 	my $html = $mid->as_html;
 	qq{[parent not found: <$html>]};
 }
 
-sub thread_adj_level {
-	my ($state, $level) = @_;
-
-	my $max = $state->{cur_level};
-	if ($level <= 0) {
-		return '' if $max == 0; # flat output
-
-		# reset existing lists
-		my $x = $max > 1 ? ('' x ($max - 1)) : '';
-		$state->{fh}->write($x . '');
-		$state->{cur_level} = 0;
-		return '';
-	}
-	if ($level == $max) { # continue existing list
-		$state->{fh}->write('
  • '); - } elsif ($level < $max) { - my $x = $max > 1 ? ('
  • ' x ($max - $level)) : ''; - $state->{fh}->write($x .= '
  • '); - $state->{cur_level} = $level; - } else { # ($level > $max) # start a new level - $state->{cur_level} = $level; - $state->{fh}->write(($max ? '
  • ' : '') . '
    • '); - } - '
    • '; -} - -sub ghost_flush { - my ($state, $upfx, $mid, $level) = @_; - my $end = '
      '. ghost_parent($upfx, $mid) . '
      '; - $state->{fh}->write($end .= thread_adj_level($state, $level)); -} - -sub __thread_entry { - my ($state, $mime, $level) = @_; - - # lazy load the full message from mini_mime: - $mime = eval { - my $mid = mid_clean(mid_mime($mime)); - $state->{ctx}->{-inbox}->msg_by_mid($mid); - } or return; - $mime = Email::MIME->new($mime); - - thread_html_head($mime, $state) if $state->{anchor_idx} == 0; - if (my $ghost = delete $state->{ghost}) { - # n.b. ghost messages may only be parents, not children - foreach my $g (@$ghost) { - ghost_flush($state, '../../', @$g); - } - } - my $end = thread_adj_level($state, $level); - index_entry($mime, $level, $state); - $state->{fh}->write($end) if $end; - - 1; -} - sub indent_for { my ($level) = @_; - INDENT x ($level - 1); -} - -sub __ghost_prepare { - my ($state, $node, $level) = @_; - my $ghost = $state->{ghost} ||= []; - push @$ghost, [ $node->messageid, $level ]; -} - -sub thread_entry { - my ($state, $level, $node) = @_; - if (my $mime = $node->message) { - unless (__thread_entry($state, $mime, $level)) { - __ghost_prepare($state, $node, $level); - } - } else { - __ghost_prepare($state, $node, $level); - } + $level ? INDENT x ($level - 1) : ''; } sub load_results { - my ($sres) = @_; - - [ map { $_->mini_mime } @{delete $sres->{msgs}} ]; -} - -sub msg_timestamp { - my ($hdr) = @_; - my $ts = eval { str2time($hdr->header('Date')) }; - defined($ts) ? $ts : 0; + my ($srch, $sres) = @_; + my $msgs = delete $sres->{msgs}; + $srch->retry_reopen(sub { [ map { $_->mid; $_ } @$msgs ] }); } sub thread_results { - my ($msgs) = @_; - require PublicInbox::Thread; - my $th = PublicInbox::Thread->new(@$msgs); - $th->thread; - $th->order(*sort_ts); - $th + my ($ctx, $msgs) = @_; + require PublicInbox::SearchThread; + PublicInbox::SearchThread::thread($msgs, *sort_ds, $ctx->{-inbox}); } sub missing_thread { - my ($res, $ctx) = @_; + my ($ctx) = @_; require PublicInbox::ExtMsg; - - $res->(PublicInbox::ExtMsg::ext_msg($ctx)) + PublicInbox::ExtMsg::ext_msg($ctx); } sub _msg_date { my ($hdr) = @_; - my $ts = $hdr->header('X-PI-TS') || msg_timestamp($hdr); - fmt_ts($ts); + fmt_ts(msg_datestamp($hdr)); } sub fmt_ts { POSIX::strftime('%Y-%m-%d %k:%M', gmtime($_[0])) } -sub _skel_header { - my ($state, $hdr, $level) = @_; +sub dedupe_subject { + my ($prev_subj, $subj, $val) = @_; + + 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 ||= $val; + } + pop @$subj if @$subj && $subj->[-1] =~ /^re:\s*/i; + if (scalar(@curr_pop) == 1) { + $omit = ''; + push @$prev_subj, @prev_pop; + push @$subj, @curr_pop; + } + $omit; +} + +sub skel_dump { + my ($ctx, $level, $node) = @_; + my $smsg = $node->{smsg} or return _skel_ghost($ctx, $level, $node); + + my $dst = $ctx->{dst}; + my $cur = $ctx->{cur}; + my $mid = $smsg->{mid}; - my $dst = $state->{dst}; - my $cur = $state->{cur}; - my $mid = mid_clean($hdr->header_raw('Message-ID')); - my $f = ascii_html($hdr->header('X-PI-From')); - my $d = _msg_date($hdr); - my $pfx = "$d " . indent_for($level) . th_pfx($level); + 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 $attr = $f; - $state->{first_level} ||= $level; + $ctx->{first_level} ||= $level; - if ($attr ne $state->{prev_attr} || $state->{prev_level} > $level) { - $state->{prev_attr} = $attr; - } else { - $attr = ''; + if ($attr ne $ctx->{prev_attr} || $ctx->{prev_level} > $level) { + $ctx->{prev_attr} = $attr; } - $state->{prev_level} = $level; + $ctx->{prev_level} = $level; if ($cur) { if ($cur eq $mid) { - delete $state->{cur}; - $$dst .= "$pfx". + delete $ctx->{cur}; + $$dst .= "$d". "$attr [this message]\n"; - return; + } else { + $ctx->{prev_msg} = $mid; } } else { - $state->{next_msg} ||= $mid; + $ctx->{next_msg} ||= $mid; } # 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 $s = $hdr->header('Subject'); - my $h = $state->{srch}->subject_path($s); - if ($state->{seen}->{$h}) { - $s = undef; + my @subj = split(/ /, $ctx->{srch}->subject_normalized($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} || []; + $ctx->{prev_subj} = [ @subj ]; + my $omit = dedupe_subject($prev_subj, \@subj, '" '); + my $end; + if (@subj) { + my $subj = join(' ', @subj); + $subj = ascii_html($subj); + obfuscate_addrs($obfs_ibx, $subj) if $obfs_ibx; + $end = "$subj $omit$f\n" } else { - $state->{seen}->{$h} = 1; - $s = PublicInbox::Hval->new($s); - $s = $s->as_html; + $end = "$f\n"; } - my $m = PublicInbox::Hval->new_msgid($mid); - $m = $state->{upfx} . $m->as_href . '/'; - $$dst .= "$pfx"; - $$dst .= defined($s) ? "$s $f\n" : "$f\n"; -} - -sub skel_dump { - my ($state, $level, $node) = @_; - if (my $mime = $node->message) { - my $hdr = $mime->header_obj; - my $mid = mid_clean($hdr->header_raw('Message-ID')); - _skel_header($state, $hdr, $level); + my $m; + my $id = ''; + my $mapping = $ctx->{mapping}; + if ($mapping) { + my $map = $mapping->{$mid}; + $id = id_compress($mid, 1); + $m = '#m'.$id; + $map->[0] = "$d$end"; + $id = "\nid=r".$id; } else { - my $mid = $node->messageid; - my $dst = $state->{dst}; - if ($mid eq 'subject dummy') { - $$dst .= "\t[no common parent]\n"; - } else { - $$dst .= ' [not found] '; - $$dst .= indent_for($level) . th_pfx($level); - $mid = PublicInbox::Hval->new_msgid($mid); - my $href = $state->{upfx} . $mid->as_href . '/'; - my $html = $mid->as_html; - $$dst .= qq{<$html>\n}; - } + $m = $ctx->{-upfx}.mid_escape($mid).'/'; } + $$dst .= $d . "" . $end; } -sub sort_ts { - sort { - (eval { $a->topmost->message->header('X-PI-TS') } || 0) <=> - (eval { $b->topmost->message->header('X-PI-TS') } || 0) - } @_; +sub _skel_ghost { + my ($ctx, $level, $node) = @_; + + my $mid = $node->{id}; + my $d = $ctx->{pct} ? ' [irrelevant] ' # search result + : ' [not found] '; + $d .= indent_for($level) . th_pfx($level); + my $upfx = $ctx->{-upfx}; + my $m = PublicInbox::Hval->new_msgid($mid); + my $href = $upfx . $m->{href} . '/'; + my $html = $m->as_html; + + my $mapping = $ctx->{mapping}; + my $map = $mapping->{$mid} if $mapping; + if ($map) { + my $id = id_compress($mid, 1); + $map->[0] = $d . qq{<$html>\n}; + $d .= qq{<$html>\n}; + } else { + $d .= qq{<$html>\n}; + } + my $dst = $ctx->{dst}; + $$dst .= $d; } -sub _tryload_ghost ($$) { - my ($srch, $mid) = @_; - my $smsg = $srch->lookup_mail($mid) or return; - $smsg->mini_mime; +sub sort_ds { + [ sort { + (eval { $a->topmost->{smsg}->ds } || 0) <=> + (eval { $b->topmost->{smsg}->ds } || 0) + } @{$_[0]} ]; } # accumulate recent topics if search is supported -# returns 1 if done, undef if not -sub add_topic { - my ($state, $level, $node) = @_; - my $srch = $state->{srch}; - my $mid = $node->messageid; - my $x = $node->message || _tryload_ghost($srch, $mid); - my ($subj, $ts); +# returns 200 if done, 404 if not +sub acc_topic { + my ($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) { - $x = $x->header_obj; - $subj = $x->header('Subject'); + $subj = $x->subject; $subj = $srch->subject_normalized($subj); - $ts = $x->header('X-PI-TS'); - } else { # ghost message, do not bump level - $ts = -666; - $subj = "<$mid>"; - } - if (++$state->{subjs}->{$subj} == 1) { - push @{$state->{order}}, [ $level, $subj ]; - } - my $exist = $state->{latest}->{$subj}; - if (!$exist || $exist->[1] < $ts) { - $state->{latest}->{$subj} = [ $mid, $ts ]; - } -} - -sub emit_topics { - my ($state) = @_; - my $order = $state->{order}; - my $subjs = $state->{subjs}; - my $latest = $state->{latest}; - my $fh = $state->{fh}; - return $fh->write("\n[No topics in range]
  • ") unless scalar @$order; - my $pfx; - my $prev = 0; - my $prev_attr = ''; - my $cur; - my @recent; - while (defined(my $info = shift @$order)) { - my ($level, $subj) = @$info; - my $n = delete $subjs->{$subj}; - my ($mid, $ts) = @{delete $latest->{$subj}}; - $mid = PublicInbox::Hval->new_msgid($mid)->as_href; - $pfx = indent_for($level); - my $nl = $level == $prev ? "\n" : ''; - if ($nl && $cur) { - push @recent, $cur; - $cur = undef; + $ds = $x->ds; + if ($level == 0) { + $topic = [ $ds, 1, { $subj => $mid }, $subj ]; + $ctx->{-cur_topic} = $topic; + push @{$ctx->{order}}, $topic; + return; } - $cur ||= [ $ts, '' ]; - $cur->[0] = $ts if $ts > $cur->[0]; - $cur->[1] .= $nl . $pfx . th_pfx($level); - if ($ts == -666) { # ghost - $cur->[1] .= ghost_parent('', $mid) . "\n"; - next; # child will have mbox / atom link + + $topic = $ctx->{-cur_topic}; # should never be undef + $topic->[0] = $ds if $ds > $topic->[0]; + $topic->[1]++; + my $seen = $topic->[2]; + if (scalar(@$topic) == 3) { # parent was a ghost + push @$topic, $subj; + } elsif (!$seen->{$subj}) { + push @$topic, $level, $subj; } + $seen->{$subj} = $mid; # latest for subject + } else { # ghost message + return if $level != 0; # ignore child ghosts + $topic = [ -666, 0, {} ]; + $ctx->{-cur_topic} = $topic; + push @{$ctx->{order}}, $topic; + } +} - $subj = PublicInbox::Hval->new($subj)->as_html; - $cur->[1] .= "$subj\n"; - $ts = fmt_ts($ts); - my $attr = " $ts UTC"; +sub dump_topics { + my ($ctx) = @_; + my $order = delete $ctx->{order}; # [ ds, subj1, subj2, subj3, ... ] + if (!@$order) { + $ctx->{-html_tip} = '
    [No topics in range]
    '; + return 404; + } + + 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; + @$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; + $ds = fmt_ts($ds); # $n isn't the total number of posts on the topic, # just the number of posts in the current results window - $n = $n == 1 ? '' : " ($n+ messages)"; - - if ($level == 0 || $attr ne $prev_attr) { - my $mbox = qq(mbox.gz); - my $atom = qq(Atom); - $pfx .= INDENT if $level > 0; - $cur->[1] .= $pfx . $attr . $n . " - $mbox / $atom\n"; - $prev_attr = $attr; + my $anchor; + if ($n == 1) { + $n = ''; + $anchor = '#u'; # top of only message + } else { + $n = " ($n+ messages)"; + $anchor = '#t'; # thread skeleton + } + + my $mbox = qq(mbox.gz); + my $atom = qq(Atom); + my $s = "$top\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]; + $mid = delete $seen->{$subj}; + my @subj = split(/ /, $srch->subject_normalized($subj)); + my @next_prev = @subj; # full copy + my $omit = dedupe_subject($prev_subj, \@subj, ' "'); + $prev_subj = \@next_prev; + $subj = ascii_html(join(' ', @subj)); + obfuscate_addrs($obfs_ibx, $subj) if $obfs_ibx; + $href = mid_escape($mid); + $s .= indent_for($level) . TCHILD; + $s .= "$subj$omit\n"; } + push @out, $s; } - push @recent, $cur if $cur; - @recent = map { $_->[1] } sort { $b->[0] <=> $a->[0] } @recent; - $fh->write(join('', @recent) . '
    '); -} - -sub emit_index_topics { - my ($state) = @_; - my ($off) = (($state->{ctx}->{cgi}->param('o') || '0') =~ /(\d+)/); - $state->{order} = []; - $state->{subjs} = {}; - $state->{latest} = {}; - my $max = 25; - my %opts = ( offset => $off, limit => $max * 4 ); - while (scalar @{$state->{order}} < $max) { - my $sres = $state->{srch}->query('', \%opts); - my $nr = scalar @{$sres->{msgs}} or last; - $sres = load_results($sres); - walk_thread(thread_results($sres), $state, *add_topic); - $opts{offset} += $nr; - } - - emit_topics($state); - $opts{offset}; + $ctx->{-html_tip} = '
    ' . join("\n", @out) . '
    '; + 200; +} + +sub index_nav { # callback for WwwStream + my (undef, $ctx) = @_; + delete $ctx->{qp} or return; + my ($next, $prev); + $next = $prev = ' '; + my $latest = ''; + + my $next_o = $ctx->{-next_o}; + if ($next_o) { + $next = qq!next!; + } + if (my $cur_o = $ctx->{-cur_o}) { + $latest = qq! latest!; + + my $o = $cur_o - ($next_o - $cur_o); + if ($o > 0) { + $prev = qq!prev!; + } elsif ($o == 0) { + $prev = qq!prev!; + } + } + "
    page: $next $prev$latest
    "; +} + +sub index_topics { + my ($ctx) = @_; + my ($off) = (($ctx->{qp}->{o} || '0') =~ /(\d+)/); + my $lim = 200; + my $opts = { offset => $off, limit => $lim }; + + $ctx->{order} = []; + my $srch = $ctx->{srch}; + + my $qs = ''; + # this complicated bit cuts loading time by over 400ms on my system: + if ($off == 0) { + my ($min, $max) = $ctx->{-inbox}->mm->minmax; + my $n = $max - $lim; + $n = $min if $n < $min; + for (; $qs eq '' && $n >= $min; --$n) { + my $smsg = $srch->lookup_article($n) or next; + $qs = POSIX::strftime('d:%Y%m%d..', gmtime($smsg->ts)); + } + } + + my $sres = $srch->query($qs, $opts); + my $nr = scalar @{$sres->{msgs}}; + if ($nr) { + $sres = load_results($srch, $sres); + walk_thread(thread_results($ctx, $sres), $ctx, *acc_topic); + } + $ctx->{-next_o} = $off + $nr; + $ctx->{-cur_o} = $off; + PublicInbox::WwwStream->response($ctx, dump_topics($ctx), *index_nav); +} + +sub thread_adj_level { + my ($ctx, $level) = @_; + + my $max = $ctx->{cur_level}; + if ($level <= 0) { + return ('', '') if $max == 0; # flat output + + # reset existing lists + my $beg = $max > 1 ? ('' x ($max - 1)) : ''; + $ctx->{cur_level} = 0; + ("$beg", ''); + } elsif ($level == $max) { # continue existing list + qw(
  • ); + } elsif ($level < $max) { + my $beg = $max > 1 ? ('' x ($max - $level)) : ''; + $ctx->{cur_level} = $level; + ("$beg
  • ", '
  • '); + } else { # ($level > $max) # start a new level + $ctx->{cur_level} = $level; + my $beg = ($max ? '
  • ' : '') . '
    • '; + ($beg, '
    • '); + } +} + +sub ghost_index_entry { + my ($ctx, $level, $node) = @_; + my ($beg, $end) = thread_adj_level($ctx, $level); + $beg . '
      '. ghost_parent($ctx->{-upfx}, $node->{id})
      +		. '
      ' . $end; } 1;