X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FView.pm;h=60fc1df17f9e5e242d55a7f05a00ca2e7943c07e;hb=dfef0d2c34fbf21d665c40b5ad5069e9113c35c8;hp=d6711d5c257d06bb03869b7212c1a1b71d4655fa;hpb=b478f5be70f5e210347c7692333bb65ad3316f72;p=public-inbox.git diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm index d6711d5c..60fc1df1 100644 --- a/lib/PublicInbox/View.pm +++ b/lib/PublicInbox/View.pm @@ -1,19 +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 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 => ' '; @@ -21,15 +21,25 @@ use constant TCHILD => '` '; sub th_pfx ($) { $_[0] == 0 ? '' : TCHILD }; # public functions: (unstable) + sub msg_html { - my ($ctx, $mime) = @_; + my ($ctx, $mime, $more, $smsg) = @_; my $hdr = $mime->header_obj; - my $tip = _msg_html_prepare($hdr, $ctx); + 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 == $end) { # fake an EOF if generating the footer fails; # we want to at least show the message if something # here crashes: @@ -43,6 +53,64 @@ sub msg_html { }); } +sub msg_page { + my ($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) = @_; @@ -52,12 +120,25 @@ sub msg_reply { 'https://en.wikipedia.org/wiki/Posting_style#Interleaved_style'; my $info = ''; - if (my $url = $ctx->{-inbox}->{infourl}) { + 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); + my ($arg, $link, $reply_to_all) = + PublicInbox::Reply::mailto_arg_link($ibx, $hdr); + + # mailto: link only works if address obfuscation is disabled + if ($link) { + $link = <In-Reply-To header + via mailto: links, try the mailto: link +EOF + } + push @$arg, '/path/to/YOUR_REPLY'; $arg = ascii_html(join(" \\\n ", '', @$arg)); <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 $info -* Reply to all the recipients using the --to, --cc, - and --in-reply-to switches of git-send-email(1): +* Reply using the --to, --cc, and --in-reply-to + switches of git-send-email(1): git send-email$arg $se_url - -* If your mail client supports setting the In-Reply-To header - via mailto: links, try the mailto: link +$link EOF } sub in_reply_to { my ($hdr) = @_; - my $irt = $hdr->header_raw('In-Reply-To'); - - return mid_clean($irt) if defined $irt && $irt ne ''; - - 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))); @@ -117,47 +188,53 @@ sub nr_to_s ($$$) { # this is already inside a
 sub index_entry {
-	my ($mime, $ctx, $more) = @_;
+	my ($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 $subj = $smsg->subject;
+	my $mid_raw = $smsg->mid;
 	my $id = id_compress($mid_raw, 1);
 	my $id_m = 'm'.$id;
-	my $mid = PublicInbox::Hval->new_msgid($mid_raw);
 
 	my $root_anchor = $ctx->{root_anchor} || '';
-	my $irt = in_reply_to($hdr);
+	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);
+	$rv .= _th_index_lite($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";
+		}
 	}
-	$rv .= "From: "._hdr_names($hdr, 'From').' @ '._msg_date($hdr)." UTC";
+	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->as_href . '/';
+	my $mhref = $upfx . mid_escape($mid_raw) . '/';
 	$rv .= qq{ (permalink / };
 	$rv .= qq{raw)\n};
 	$rv .= '  '.join('; +', @tocc) . "\n" if @tocc;
 
 	my $mapping = $ctx->{mapping};
-	if (!$mapping && $irt) {
+	if (!$mapping && (defined($irt) || defined($irt = in_reply_to($hdr)))) {
 		my $mirt = PublicInbox::Hval->new_msgid($irt);
-		my $href = $upfx . $mirt->as_href . '/';
+		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 { $rv .= add_text_body($mhref, $_[0]) });
+	msg_iter($mime, sub { $rv .= add_text_body($mhref, $obfs_ibx, $_[0]) });
 
 	# add the footer
 	$rv .= "\n^ ".
@@ -170,18 +247,20 @@ sub index_entry {
 		$rv .= "\t[relevance $pct->{$mid_raw}%]";
 		$hr = 1;
 	} elsif ($mapping) {
-		my $threaded = 'threaded';
+		my $nested = 'nested';
 		my $flat = 'flat';
 		my $end = '';
 		if ($ctx->{flat}) {
 			$hr = 1;
 			$flat = "$flat";
 		} else {
-			$threaded = "$threaded";
+			$nested = "$nested";
 		}
 		$rv .= "\t[$flat";
-		$rv .= "|$threaded]";
+		$rv .= "|$nested]";
 		$rv .= " $ctx->{s_nr}";
+	} else {
+		$hr = $ctx->{-hr};
 	}
 
 	$rv .= $more ? '

' : '
' if $hr; @@ -200,121 +279,122 @@ sub _th_index_lite { my $rv = ''; my $mapping = $ctx->{mapping} or return $rv; my $pad = ' '; - # map = [children, attr, node, idx, level] - my $map = $mapping->{$mid_raw}; - my $nr_c = scalar @{$map->[0]}; + 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 $level = $map->[4]; - my $idx = $map->[3]; - if (defined $irt) { - my $irt_map = $mapping->{$irt}; - my $siblings = $irt_map->[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->[1]; + $rv .= $pad . $irt_map->[0]; if ($idx > 0) { my $prev = $siblings->[$idx - 1]; - my $pmid = $prev->messageid; + 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]->messageid; - $rv .= $pad . $mapping->{$ppmid}->[1]; + my $ppmid = $siblings->[0]->{id}; + $rv .= $pad . $mapping->{$ppmid}->[0]; } - $rv .= $pad . $mapping->{$pmid}->[1]; + $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'); - my $this = $map->[1]; - $this =~ s!\n\z!\n!s; - $this =~ s! !!s; # no point in duplicating subject - $this =~ s!]+>([^<]+)!$1!s; # no point linking to self - $rv .= "@ $this"; - my $node = $map->[2]; - if (my $child = $node->child) { - my $cmid = $child->messageid; - $rv .= $pad . $mapping->{$cmid}->[1]; + $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 = $child->next) { - $rv .= $pad . $mapping->{$cn->messageid}->[1]; + } elsif (my $cn = $children->[1]) { + $rv .= $pad . $mapping->{$cn->{id}}->[0]; } } - if (my $next = $node->next) { - my $nmid = $next->messageid; - $rv .= $pad . $mapping->{$nmid}->[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 = $next->next) { - $rv .= $pad . $mapping->{$nn->messageid}->[1]; + } elsif (my $nn = $siblings->[$idx + 2]) { + $rv .= $pad . $mapping->{$nn->{id}}->[0]; } } $rv .= $pad ."$s_s, $s_c; $ctx->{s_nr}\n"; } sub walk_thread { - my ($th, $ctx, $cb) = @_; - my @q = map { (0, $_) } $th->rootset; + my ($rootset, $ctx, $cb) = @_; + my @q = map { (0, $_, -1) } @$rootset; while (@q) { - my $level = shift @q; - my $node = shift @q or next; - $cb->($ctx, $level, $node); - unshift @q, $level+1, $node->child, $level, $node->next; + 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}}; } } sub pre_thread { - my ($ctx, $level, $node) = @_; - my $mapping = $ctx->{mapping}; - my $idx = -1; - if (my $parent = $node->parent) { - my $m = $mapping->{$parent->messageid}->[0]; - $idx = scalar @$m; - push @$m, $node; - } - $mapping->{$node->messageid} = [ [], '', $node, $idx, $level ]; + my ($ctx, $level, $node, $idx) = @_; + $ctx->{mapping}->{$node->{id}} = [ '', $node, $idx, $level ]; skel_dump($ctx, $level, $node); } sub thread_index_entry { - my ($ctx, $level, $mime) = @_; + my ($ctx, $level, $smsg) = @_; my ($beg, $end) = thread_adj_level($ctx, $level); - $beg . '
' . index_entry($mime, $ctx, 0) . '
' . $end; + $beg . '
' . index_entry($smsg, $ctx, 0) . '
' . $end; } sub stream_thread ($$) { - my ($th, $ctx) = @_; + my ($rootset, $ctx) = @_; my $inbox = $ctx->{-inbox}; - my $mime; - my @q = map { (0, $_) } $th->rootset; + my @q = map { (0, $_) } @$rootset; my $level; + my $smsg; while (@q) { $level = shift @q; my $node = shift @q or next; - unshift @q, $level+1, $node->child, $level, $node->next; - $mime = $inbox->msg_by_mid($node->messageid) and last; + my $cl = $level + 1; + unshift @q, map { ($cl, $_) } @{$node->{children}}; + $smsg = $inbox->smsg_mime($node->{smsg}) and last; } - return missing_thread($ctx) unless $mime; + return missing_thread($ctx) unless $smsg; - $mime = Email::MIME->new($mime); - $ctx->{-title_html} = ascii_html($mime->header('Subject')); - $ctx->{-html_tip} = thread_index_entry($ctx, $level, $mime); + $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; - unshift @q, $level+1, $node->child, $level, $node->next; - my $mid = $node->messageid; - if ($mime = $inbox->msg_by_mid($mid)) { - $mime = Email::MIME->new($mime); - return thread_index_entry($ctx, $level, $mime); + 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, $mid); + return ghost_index_entry($ctx, $level, $node); } } my $ret = join('', thread_adj_level($ctx, 0)); @@ -327,8 +407,9 @@ sub stream_thread ($$) { sub thread_html { my ($ctx) = @_; my $mid = $ctx->{mid}; - my $sres = $ctx->{srch}->get_thread($mid, { asc => 1 }); - my $msgs = load_results($sres); + 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 = '
';
@@ -345,50 +426,46 @@ sub thread_html {
 	$ctx->{prev_attr} = '';
 	$ctx->{prev_level} = 0;
 	$ctx->{root_anchor} = anchor_for($mid);
-	$ctx->{seen} = {};
 	$ctx->{mapping} = {};
 	$ctx->{s_nr} = "$nr+ messages in thread";
 
-	my $th = thread_results($msgs);
-	walk_thread($th, $ctx, *pre_thread);
-	$skel .= '
'; - return stream_thread($th, $ctx) unless $ctx->{flat}; + my $rootset = thread_results($ctx, $msgs); - # flat display: lazy load the full message from mini_mime: + # reduce hash lookups in pre_thread->skel_dump my $inbox = $ctx->{-inbox}; - my $mime; - while ($mime = shift @$msgs) { - $mime = $inbox->msg_by_mid(mid_clean(mid_mime($mime))) and last; - } - return missing_thread($ctx) unless $mime; - $mime = Email::MIME->new($mime); - $ctx->{-title_html} = ascii_html($mime->header('Subject')); - $ctx->{-html_tip} = '
'.index_entry($mime, $ctx, scalar @$msgs);
-	$mime = undef;
+	$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 $msgs;
-		while ($mime = shift @$msgs) {
-			$mid = mid_clean(mid_mime($mime));
-			$mime = $inbox->msg_by_mid($mid) and last;
-		}
-		if ($mime) {
-			$mime = Email::MIME->new($mime);
-			return index_entry($mime, $ctx, scalar @$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; + $skel; }); } sub multipart_text_as_html { - my ($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; } @@ -406,14 +483,16 @@ sub flush_quote { $$s .= qq() . $rv . '' } -sub attach_link ($$$$) { - my ($upfx, $ct, $p, $fn) = @_; +sub attach_link ($$$$;$) { + my ($upfx, $ct, $p, $fn, $err) = @_; my ($part, $depth, @idx) = @$p; my $nl = $idx[-1] > 1 ? "\n" : ''; my $idx = join('.', @idx); my $size = bytes::length($part->body); - $ct ||= 'text/plain'; - $ct =~ s/;.*//; # no attributes + + # hide attributes normally, unless we want to aid users in + # spotting MUA problems: + $ct =~ s/;.*// unless $err; $ct = ascii_html($ct); my $desc = $part->header('Content-Description'); $desc = $fn unless defined $desc; @@ -426,72 +505,104 @@ sub attach_link ($$$$) { } 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;
@@ -500,6 +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) {
@@ -509,14 +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);
 	$rv .= "\n";
 }
@@ -524,37 +648,45 @@ sub _msg_html_prepare {
 sub thread_skel {
 	my ($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 = $srch->subject_path($hdr->header('Subject'));
-	$ctx->{seen} = { $subj => 1 };
+	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} = $dst;
-	walk_thread(thread_results(load_results($sres)), $ctx, *skel_dump);
+	$sres = load_results($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);
+
 	$ctx->{parent_msg} = $parent;
 }
 
@@ -566,7 +698,7 @@ sub _parent_headers {
 	if (defined $irt) {
 		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";
 	}
@@ -595,50 +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 = $hdr->header_raw('Message-ID');
-	push @arg, '--in-reply-to='.squote_maybe(mid_clean($mid));
-	my $irt = uri_escape_utf8($mid);
-	delete $cc{$to};
-	push @arg, "--to=$to";
-	$to = uri_escape_utf8($to);
-	$subj = uri_escape_utf8($subj);
-	my @cc = sort values %cc;
-	push(@arg, map { "--cc=$_" } @cc);
-	my $cc = uri_escape_utf8(join(',', @cc));
-	my $href = "mailto:$to?In-Reply-To=$irt&Cc=${cc}&Subject=$subj";
-	$href =~ s/%20/+/g;
-
-	(\@arg, ascii_html($href));
-}
-
 sub html_footer {
 	my ($hdr, $standalone, $ctx, $rhref) = @_;
 
@@ -655,17 +743,17 @@ sub html_footer {
 		$next = $prev = '    ';
 
 		if (my $n = $ctx->{next_msg}) {
-			$n = PublicInbox::Hval->new_msgid($n)->as_href;
+			$n = PublicInbox::Hval->new_msgid($n)->{href};
 			$next = "next";
 		}
 		my $u;
 		my $par = $ctx->{parent_msg};
 		if ($par) {
-			$u = PublicInbox::Hval->new_msgid($par)->as_href;
+			$u = PublicInbox::Hval->new_msgid($par)->{href};
 			$u = "$upfx$u/";
 		}
 		if (my $p = $ctx->{prev_msg}) {
-			$prev = PublicInbox::Hval->new_msgid($p)->as_href;
+			$prev = PublicInbox::Hval->new_msgid($p)->{href};
 			if ($p && $par && $p eq $par) {
 				$prev = "prev parent';
@@ -690,7 +778,7 @@ sub html_footer {
 sub linkify_ref_nosrch {
 	my $v = PublicInbox::Hval->new_msgid($_[0]);
 	my $html = $v->as_html;
-	my $href = $v->as_href;
+	my $href = $v->{href};
 	"<$html>";
 }
 
@@ -701,39 +789,28 @@ sub anchor_for {
 
 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 indent_for {
 	my ($level) = @_;
-	INDENT x ($level - 1);
+	$level ? INDENT x ($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 {
@@ -744,20 +821,43 @@ sub missing_thread {
 
 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 ($ctx, $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 $mid = mid_clean($hdr->header_raw('Message-ID'));
-	my $f = ascii_html($hdr->header('X-PI-From'));
-	my $d = _msg_date($hdr) . ' ' . indent_for($level) . th_pfx($level);
+	my $mid = $smsg->{mid};
+
+	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;
 	$ctx->{first_level} ||= $level;
 
@@ -782,76 +882,67 @@ sub _skel_header {
 	# 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 = $ctx->{srch}->subject_path($s);
-	if ($ctx->{seen}->{$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 {
-		$ctx->{seen}->{$h} = 1;
-		$s = PublicInbox::Hval->new($s);
-		$s = $s->as_html;
+		$end = "$f\n";
 	}
-	my $m = PublicInbox::Hval->new_msgid($mid);
+	my $m;
 	my $id = '';
 	my $mapping = $ctx->{mapping};
-	my $end = defined($s) ? "$s $f\n" : "$f\n";
 	if ($mapping) {
 		my $map = $mapping->{$mid};
 		$id = id_compress($mid, 1);
 		$m = '#m'.$id;
-		$map->[1] = "$d$end";
+		$map->[0] = "$d$end";
 		$id = "\nid=r".$id;
 	} else {
-		$m = $ctx->{-upfx}.$m->as_href.'/';
+		$m = $ctx->{-upfx}.mid_escape($mid).'/';
 	}
 	$$dst .=  $d . "" . $end;
 }
 
-sub skel_dump {
+sub _skel_ghost {
 	my ($ctx, $level, $node) = @_;
-	if (my $mime = $node->message) {
-		_skel_header($ctx, $mime->header_obj, $level);
+
+	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 {
-		my $mid = $node->messageid;
-		my $dst = $ctx->{dst};
-		my $mapping = $ctx->{mapping};
-		my $map = $mapping->{$mid} if $mapping;
-		if ($mid eq 'subject dummy') {
-			my $ncp = "\t[no common parent]\n";
-			$map->[1] = $ncp if $map;
-			$$dst .= $ncp;
-			return;
-		}
-		my $d = $ctx->{pct} ? '    [irrelevant] ' # search result
-				    : '     [not found] ';
-		$d .= indent_for($level) . th_pfx($level);
-		my $upfx = $ctx->{-upfx};
-		my $m = PublicInbox::Hval->new_msgid($mid);
-		my $href = $upfx . $m->as_href . '/';
-		my $html = $m->as_html;
-
-		if ($map) {
-			my $id = id_compress($mid, 1);
-			$map->[1] = $d . qq{<$html>\n};
-			$d .= qq{<$html>\n};
-		} else {
-			$d .= qq{<$html>\n};
-		}
-		$$dst .= $d;
+		$d .= qq{<$html>\n};
 	}
+	my $dst = $ctx->{dst};
+	$$dst .= $d;
 }
 
-sub sort_ts {
-	sort {
-		(eval { $a->topmost->message->header('X-PI-TS') } || 0) <=>
-		(eval { $b->topmost->message->header('X-PI-TS') } || 0)
-	} @_;
-}
-
-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
@@ -859,24 +950,23 @@ sub _tryload_ghost ($$) {
 sub acc_topic {
 	my ($ctx, $level, $node) = @_;
 	my $srch = $ctx->{srch};
-	my $mid = $node->messageid;
-	my $x = $node->message || _tryload_ghost($srch, $mid);
-	my ($subj, $ts);
+	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');
+		$ds = $x->ds;
 		if ($level == 0) {
-			$topic = [ $ts, 1, { $subj => $mid }, $subj ];
+			$topic = [ $ds, 1, { $subj => $mid }, $subj ];
 			$ctx->{-cur_topic} = $topic;
 			push @{$ctx->{order}}, $topic;
 			return;
 		}
 
 		$topic = $ctx->{-cur_topic}; # should never be undef
-		$topic->[0] = $ts if $ts > $topic->[0];
+		$topic->[0] = $ds if $ds > $topic->[0];
 		$topic->[1]++;
 		my $seen = $topic->[2];
 		if (scalar(@$topic) == 3) { # parent was a ghost
@@ -895,23 +985,27 @@ sub acc_topic {
 
 sub dump_topics {
 	my ($ctx) = @_;
-	my $order = delete $ctx->{order}; # [ ts, subj1, subj2, subj3, ... ]
+	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 ($ts, $n, $seen, $top, @ex) = @$topic; + my ($ds, $n, $seen, $top, @ex) = @$topic; @$topic = (); next unless defined $top; # ghost topic my $mid = delete $seen->{$top}; - my $href = PublicInbox::Hval->new_msgid($mid)->as_href; + my $href = mid_escape($mid); + my $prev_subj = [ split(/ /, $top) ]; $top = PublicInbox::Hval->new($top)->as_html; - $ts = fmt_ts($ts); + $ds = fmt_ts($ds); # $n isn't the total number of posts on the topic, # just the number of posts in the current results window @@ -927,15 +1021,20 @@ sub dump_topics { my $mbox = qq(mbox.gz); my $atom = qq(Atom); my $s = "$top\n" . - " $ts UTC $n - $mbox / $atom\n"; + " $ds UTC $n - $mbox / $atom\n"; for (my $i = 0; $i < scalar(@ex); $i += 2) { my $level = $ex[$i]; - my $sub = $ex[$i + 1]; - $mid = delete $seen->{$sub}; - $sub = PublicInbox::Hval->new($sub)->as_html; - $href = PublicInbox::Hval->new_msgid($mid)->as_href; + 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; - $s .= "$sub\n"; + $s .= "$subj$omit\n"; } push @out, $s; } @@ -970,16 +1069,31 @@ sub index_nav { # callback for WwwStream sub index_topics { my ($ctx) = @_; my ($off) = (($ctx->{qp}->{o} || '0') =~ /(\d+)/); - my $opts = { offset => $off, limit => 200 }; + my $lim = 200; + my $opts = { offset => $off, limit => $lim }; $ctx->{order} = []; - my $sres = $ctx->{srch}->query('', $opts); + 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) { - $sres = load_results($sres); - walk_thread(thread_results($sres), $ctx, *acc_topic); + $sres = load_results($srch, $sres); + walk_thread(thread_results($ctx, $sres), $ctx, *acc_topic); } - $ctx->{-next_o} = $off+ $nr; + $ctx->{-next_o} = $off + $nr; $ctx->{-cur_o} = $off; PublicInbox::WwwStream->response($ctx, dump_topics($ctx), *index_nav); } @@ -1009,9 +1123,10 @@ sub thread_adj_level { } sub ghost_index_entry { - my ($ctx, $level, $mid) = @_; + my ($ctx, $level, $node) = @_; my ($beg, $end) = thread_adj_level($ctx, $level); - $beg . '
'. ghost_parent($ctx->{-upfx}, $mid) . '
' . $end; + $beg . '
'. ghost_parent($ctx->{-upfx}, $node->{id})
+		. '
' . $end; } 1;