X-Git-Url: http://www.git.stargrave.org/?p=public-inbox.git;a=blobdiff_plain;f=lib%2FPublicInbox%2FView.pm;h=687a0acb9367737410d5ff4aae641513fd49b629;hp=5d5808fad5b2176307e579d5108ad4e7284fd2db;hb=ef9ce6f2dfc9fcbb91e3cec0b7a2b7f7e359be5d;hpb=123031737a09edb3297e0fc6ea9e566fc2fecdf5 diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm index 5d5808fa..687a0acb 100644 --- a/lib/PublicInbox/View.pm +++ b/lib/PublicInbox/View.pm @@ -6,14 +6,14 @@ 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::Hval qw/ascii_html obfuscate_addrs/; use PublicInbox::Linkify; use PublicInbox::MID qw/mid_clean id_compress mid_mime mid_escape/; use PublicInbox::MsgIter; use PublicInbox::Address; use PublicInbox::WwwStream; +use PublicInbox::Reply; require POSIX; use constant INDENT => ' '; @@ -24,11 +24,13 @@ sub th_pfx ($) { $_[0] == 0 ? '' : TCHILD }; sub msg_html { my ($ctx, $mime) = @_; my $hdr = $mime->header_obj; - my $tip = _msg_html_prepare($hdr, $ctx); + my $obfs = $ctx->{-inbox}->{obfuscate}; + my $tip = _msg_html_prepare($hdr, $ctx, $obfs); PublicInbox::WwwStream->response($ctx, 200, sub { my ($nr, undef) = @_; if ($nr == 1) { - $tip . multipart_text_as_html($mime, '') . '
' + $tip . multipart_text_as_html($mime, '', $obfs) . + '
' } elsif ($nr == 2) { # fake an EOF if generating the footer fails; # we want to at least show the message if something @@ -52,12 +54,13 @@ 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, $link) = PublicInbox::Reply::mailto_arg_link($ibx, $hdr); push @$arg, '/path/to/YOUR_REPLY'; $arg = ascii_html(join(" \\\n ", '', @$arg)); <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; + my %mid = map { $_ => 1 } $hdr->header_raw('Message-ID'); + my @refs = ($hdr->header_raw('References'), + $hdr->header_raw('In-Reply-To')); + @refs = ((join(' ', @refs)) =~ /<([^>]+)>/g); + while (defined(my $irt = pop @refs)) { + next if $mid{"<$irt>"}; + return $irt; } undef; } -sub _hdr_names ($$) { +sub _hdr_names_html ($$) { my ($hdr, $field) = @_; my $val = $hdr->header($field) or return ''; ascii_html(join(', ', PublicInbox::Address::names($val))); @@ -127,19 +130,26 @@ sub index_entry { my $id_m = 'm'.$id; my $root_anchor = $ctx->{root_anchor} || ''; - my $irt = in_reply_to($hdr); + my $irt; + my $obfs = $ctx->{-obfuscate}; my $rv = "* "; $subj = ''.ascii_html($subj).''; + obfuscate_addrs($subj) if $obfs; $subj = "$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; 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($dst) if $obfs; + push @tocc, "$f: $dst"; + } } - $rv .= "From: "._hdr_names($hdr, 'From').' @ '._msg_date($hdr)." UTC"; + my $from = _hdr_names_html($hdr, 'From'); + obfuscate_addrs($from) if $obfs; + $rv .= "From: $from @ "._msg_date($hdr)." UTC"; my $upfx = $ctx->{-upfx}; my $mhref = $upfx . mid_escape($mid_raw) . '/'; $rv .= qq{ (permalink / }; @@ -147,7 +157,7 @@ sub index_entry { $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->{href}. '/'; my $html = $mirt->as_html; @@ -156,7 +166,7 @@ sub index_entry { $rv .= "\n"; # scan through all parts, looking for displayable text - msg_iter($mime, sub { $rv .= add_text_body($mhref, $_[0]) }); + msg_iter($mime, sub { $rv .= add_text_body($mhref, $obfs, $_[0]) }); # add the footer $rv .= "\n^ ". @@ -201,12 +211,18 @@ sub _th_index_lite { my $rv = ''; my $mapping = $ctx->{mapping} or return $rv; my $pad = ' '; - my ($attr, $node, $idx, $level) = @{$mapping->{$mid_raw}}; + 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; - my $irt_map = $mapping->{$irt} if defined $irt; + if (my $smsg = $node->{smsg}) { + ($$irt) = (($smsg->{references} || '') =~ m/<([^>]+)>\z/); + } + my $irt_map = $mapping->{$$irt} if defined $$irt; if (defined $irt_map) { $siblings = $irt_map->[1]->{children}; $nr_s = scalar(@$siblings) - 1; @@ -257,8 +273,8 @@ sub _th_index_lite { } sub walk_thread { - my ($th, $ctx, $cb) = @_; - my @q = map { (0, $_, -1) } @{$th->{rootset}}; + my ($rootset, $ctx, $cb) = @_; + my @q = map { (0, $_, -1) } @$rootset; while (@q) { my ($level, $node, $i) = splice(@q, 0, 3); defined $node or next; @@ -282,10 +298,10 @@ sub thread_index_entry { } sub stream_thread ($$) { - 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; while (@q) { $level = shift @q; @@ -296,7 +312,8 @@ sub stream_thread ($$) { } return missing_thread($ctx) unless $mime; - $mime = Email::MIME->new($mime); + $ctx->{-obfuscate} = $ctx->{-inbox}->{obfuscate}; + $mime = PublicInbox::MIME->new($mime); $ctx->{-title_html} = ascii_html($mime->header('Subject')); $ctx->{-html_tip} = thread_index_entry($ctx, $level, $mime); PublicInbox::WwwStream->response($ctx, 200, sub { @@ -308,7 +325,7 @@ sub stream_thread ($$) { unshift @q, map { ($cl, $_) } @{$node->{children}}; my $mid = $node->{id}; if ($mime = $inbox->msg_by_smsg($node->{smsg})) { - $mime = Email::MIME->new($mime); + $mime = PublicInbox::MIME->new($mime); return thread_index_entry($ctx, $level, $mime); } else { return ghost_index_entry($ctx, $level, $node); @@ -324,8 +341,9 @@ 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 = '
';
@@ -346,10 +364,14 @@ sub thread_html {
 	$ctx->{mapping} = {};
 	$ctx->{s_nr} = "$nr+ messages in thread";
 
-	my $th = thread_results($msgs);
-	walk_thread($th, $ctx, *pre_thread);
+	my $rootset = thread_results($msgs);
+
+	# reduce hash lookups in pre_thread->skel_dump
+	$ctx->{-obfuscate} = $ctx->{-inbox}->{obfuscate};
+	walk_thread($rootset, $ctx, *pre_thread);
+
 	$skel .= '
'; - return stream_thread($th, $ctx) unless $ctx->{flat}; + return stream_thread($rootset, $ctx) unless $ctx->{flat}; # flat display: lazy load the full message from smsg my $inbox = $ctx->{-inbox}; @@ -358,7 +380,7 @@ sub thread_html { $mime = $inbox->msg_by_smsg($mime) and last; } return missing_thread($ctx) unless $mime; - $mime = Email::MIME->new($mime); + $mime = PublicInbox::MIME->new($mime); $ctx->{-title_html} = ascii_html($mime->header('Subject')); $ctx->{-html_tip} = '
'.index_entry($mime, $ctx, scalar @$msgs);
 	$mime = undef;
@@ -368,7 +390,7 @@ sub thread_html {
 			$mime = $inbox->msg_by_smsg($mime) and last;
 		}
 		if ($mime) {
-			$mime = Email::MIME->new($mime);
+			$mime = PublicInbox::MIME->new($mime);
 			return index_entry($mime, $ctx, scalar @$msgs);
 		}
 		$msgs = undef;
@@ -377,14 +399,11 @@ sub thread_html {
 }
 
 sub multipart_text_as_html {
-	my ($mime, $upfx) = @_;
+	my ($mime, $upfx, $obfs) = @_;
 	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, $_[0]) });
 	$rv;
 }
 
@@ -431,13 +450,15 @@ sub attach_link ($$$$;$) {
 	}
 	$ret .= "[-- Attachment #$idx: ";
 	my $ts = "Type: $ct, Size: $size bytes";
+	$desc = ascii_html($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 ($upfx, $obfs, $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;
 
@@ -472,33 +493,35 @@ sub add_text_body {
 	}
 	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, $l, \@quot);
+		obfuscate_addrs($s) if $obfs;
+		$s;
+	} else {
+		obfuscate_addrs($s) if $obfs;
+		if ($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, $obfs) = @_;
 	my $srch = $ctx->{srch} if $ctx;
 	my $atom = '';
 	my $rv = ""; # anchor for body start
@@ -517,6 +540,7 @@ sub _msg_html_prepare {
 		if ($h eq 'From') {
 			my @n = PublicInbox::Address::names($v->raw);
 			$title[1] = ascii_html(join(', ', @n));
+			obfuscate_addrs($title[1]) if $obfs;
 		} elsif ($h eq 'Subject') {
 			$title[0] = $v->as_html;
 			if ($srch) {
@@ -526,7 +550,7 @@ sub _msg_html_prepare {
 			}
 		}
 		$v = $v->as_html;
-		$v =~ s/(\@[^,]+,) /$1\n\t/g if ($h eq 'Cc' || $h eq 'To');
+		obfuscate_addrs($v) if $obfs;
 		$rv .= "$h: $v\n";
 
 	}
@@ -571,7 +595,12 @@ sub thread_skel {
 	$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
+	$ctx->{-obfuscate} = $ctx->{-inbox}->{obfuscate};
+	walk_thread(thread_results($sres), $ctx, *skel_dump);
+
 	$ctx->{parent_msg} = $parent;
 }
 
@@ -612,50 +641,6 @@ 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 $irt = mid_escape($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) = @_;
 
@@ -731,9 +716,9 @@ sub indent_for {
 }
 
 sub load_results {
-	my ($sres) = @_;
-
-	[ map { $_->ensure_metadata; $_ } @{delete $sres->{msgs}} ];
+	my ($srch, $sres) = @_;
+	my $msgs = delete $sres->{msgs};
+	$srch->retry_reopen(sub { [ map { $_->mid; $_ } @$msgs ] });
 }
 
 sub msg_timestamp {
@@ -745,10 +730,7 @@ sub msg_timestamp {
 sub thread_results {
 	my ($msgs) = @_;
 	require PublicInbox::SearchThread;
-	my $th = PublicInbox::SearchThread->new($msgs);
-	$th->thread;
-	$th->order(*sort_ts);
-	$th
+	PublicInbox::SearchThread::thread($msgs, *sort_ts);
 }
 
 sub missing_thread {
@@ -765,13 +747,18 @@ sub _msg_date {
 
 sub fmt_ts { POSIX::strftime('%Y-%m-%d %k:%M', gmtime($_[0])) }
 
-sub _skel_header {
-	my ($ctx, $smsg, $level) = @_;
+sub skel_dump {
+	my ($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 $f = ascii_html($smsg->from_name);
+	my $obfs = $ctx->{-obfuscate};
+	obfuscate_addrs($f) if $obfs;
+
 	my $d = fmt_ts($smsg->{ts}) . ' ' . indent_for($level) . th_pfx($level);
 	my $attr = $f;
 	$ctx->{first_level} ||= $level;
@@ -797,19 +784,20 @@ 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 = $smsg->subject;
-	my $h = $ctx->{srch}->subject_path($s);
+	my $subj = $smsg->subject;
+	my $h = $ctx->{srch}->subject_path($subj);
 	if ($ctx->{seen}->{$h}) {
-		$s = undef;
+		$subj = undef;
 	} else {
 		$ctx->{seen}->{$h} = 1;
-		$s = PublicInbox::Hval->new($s);
-		$s = $s->as_html;
+		$subj = PublicInbox::Hval->new($subj);
+		$subj = $subj->as_html;
+		obfuscate_addrs($subj) if $obfs;
 	}
 	my $m;
 	my $id = '';
 	my $mapping = $ctx->{mapping};
-	my $end = defined($s) ? "$s $f\n" : "$f\n";
+	my $end = defined($subj) ? "$subj $f\n" : "$f\n";
 	if ($mapping) {
 		my $map = $mapping->{$mid};
 		$id = id_compress($mid, 1);
@@ -822,32 +810,29 @@ sub _skel_header {
 	$$dst .=  $d . "" . $end;
 }
 
-sub skel_dump {
+sub _skel_ghost {
 	my ($ctx, $level, $node) = @_;
-	if (my $smsg = $node->{smsg}) {
-		_skel_header($ctx, $smsg, $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->{id};
-		my $dst = $ctx->{dst};
-		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};
-		}
-		$$dst .= $d;
+		$d .= qq{<$html>\n};
 	}
+	my $dst = $ctx->{dst};
+	$$dst .= $d;
 }
 
 sub sort_ts {
@@ -904,6 +889,7 @@ sub dump_topics {
 	}
 
 	my @out;
+	my $obfs = $ctx->{-inbox}->{obfuscate};
 
 	# sort by recency, this allows new posts to "bump" old topics...
 	foreach my $topic (sort { $b->[0] <=> $a->[0] } @$order) {
@@ -932,12 +918,13 @@ sub dump_topics {
 			" $ts 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;
+			my $subj = $ex[$i + 1];
+			$mid = delete $seen->{$subj};
+			$subj = ascii_html($subj);
+			obfuscate_addrs($subj) if $obfs;
 			$href = mid_escape($mid);
 			$s .= indent_for($level) . TCHILD;
-			$s .= "$sub\n";
+			$s .= "$subj\n";
 		}
 		push @out, $s;
 	}
@@ -975,10 +962,11 @@ sub index_topics {
 	my $opts = { offset => $off, limit => 200 };
 
 	$ctx->{order} = [];
-	my $sres = $ctx->{srch}->query('', $opts);
+	my $srch = $ctx->{srch};
+	my $sres = $srch->query('', $opts);
 	my $nr = scalar @{$sres->{msgs}};
 	if ($nr) {
-		$sres = load_results($sres);
+		$sres = load_results($srch, $sres);
 		walk_thread(thread_results($sres), $ctx, *acc_topic);
 	}
 	$ctx->{-next_o} = $off+ $nr;