From e022d3377fd2c50fd9931bf96394728958a90bf3 Mon Sep 17 00:00:00 2001
From: Eric Wong <e@80x24.org>
Date: Sat, 26 Apr 2014 01:01:10 +0000
Subject: huge refactor of encoding handling

Hopefully this simplifies and corrects our usage of Perl encoding
APIs.
---
 Documentation/design_www.txt |  9 +++++
 TODO                         |  1 -
 lib/PublicInbox/Feed.pm      | 78 ++++++++++++++++++--------------------------
 lib/PublicInbox/Filter.pm    | 57 ++++++++++++++++----------------
 lib/PublicInbox/Hval.pm      | 20 +++++++-----
 lib/PublicInbox/MDA.pm       |  9 +++--
 lib/PublicInbox/View.pm      | 35 +++++++++-----------
 public-inbox-learn           | 18 ++++------
 public-inbox-mda             | 23 +++++++------
 public-inbox.cgi             |  9 ++---
 t/filter.t                   | 66 +++++++++++++++----------------------
 t/mda.t                      | 21 ++++++++++++
 t/utf8.mbox                  | 12 +++++++
 13 files changed, 182 insertions(+), 176 deletions(-)
 create mode 100644 t/utf8.mbox

diff --git a/Documentation/design_www.txt b/Documentation/design_www.txt
index 50c97317..a265a890 100644
--- a/Documentation/design_www.txt
+++ b/Documentation/design_www.txt
@@ -24,6 +24,15 @@ Maybe TODO (these might be expensive)
 We use file name suffixes on all of these (except /) so URLs may easily
 cached/memoized using a static file server.
 
+Encoding notes
+--------------
+
+Raw HTML and XML should only contain us-ascii characters which render
+to UTF-8.
+
+Plain text (raw message) endpoints display in the original encoding(s)
+of the original email.
+
 Guidelines for using limited HTML
 ---------------------------------
 We mainly use HTML for linking pages together with <a>.
diff --git a/TODO b/TODO
index 76020a7f..e69de29b 100644
--- a/TODO
+++ b/TODO
@@ -1 +0,0 @@
-* header -> HTML/XML sanitization
diff --git a/lib/PublicInbox/Feed.pm b/lib/PublicInbox/Feed.pm
index abfc0a9a..3bee3a51 100644
--- a/lib/PublicInbox/Feed.pm
+++ b/lib/PublicInbox/Feed.pm
@@ -4,9 +4,7 @@ package PublicInbox::Feed;
 use strict;
 use warnings;
 use Email::Address;
-use Encode qw/find_encoding/;
-use Encode::MIME::Header;
-use CGI qw(escapeHTML);
+use Email::MIME;
 use Date::Parse qw(strptime str2time);
 use PublicInbox::Hval;
 eval { require Git }; # this is GPLv2+, so we are OK to use it
@@ -14,9 +12,6 @@ use constant {
 	DATEFMT => '%Y-%m-%dT%H:%M:%SZ',
 	MAX_PER_PAGE => 25,
 };
-my $enc_utf8 = find_encoding('utf8');
-my $enc_ascii = find_encoding('us-ascii');
-my $enc_mime = find_encoding('MIME-Header');
 
 # FIXME: workaround https://rt.cpan.org/Public/Bug/Display.html?id=22817
 
@@ -25,7 +20,6 @@ sub generate {
 	my ($class, $args) = @_;
 	require XML::Atom::SimpleFeed;
 	require PublicInbox::View;
-	require Email::MIME;
 	require POSIX;
 	my $max = $args->{max} || MAX_PER_PAGE;
 	my $top = $args->{top}; # bool
@@ -62,22 +56,25 @@ sub generate_html_index {
 	my $top = $args->{top}; # bool
 	local $ENV{GIT_DIR} = $args->{git_dir};
 	my $feed_opts = get_feedopts($args);
-	my $title = xs_html($feed_opts->{description} || "");
+
+	my $title = $feed_opts->{description} || '';
+	$title = PublicInbox::Hval->new_oneline($title)->as_html;
+
 	my @messages;
 	my $git = try_git_pm($args->{git_dir});
 	my ($first, $last) = each_recent_blob($args, sub {
-		my $simple = do_cat_mail($git, 'Email::Simple', $_[0])
+		my $mime = do_cat_mail($git, $_[0])
 			or return 0;
-		if ($top && ($simple->header("In-Reply-To") ||
-		             $simple->header("References"))) {
+		if ($top && ($mime->header('In-Reply-To') ||
+		             $mime->header('References'))) {
 			return 0;
 		}
-		$simple->body_set(""); # save some memory
+		$mime->body_set(''); # save some memory
 
-		my $t = eval { str2time($simple->header('Date')) };
+		my $t = eval { str2time($mime->header('Date')) };
 		defined($t) or $t = 0;
-		$simple->header_set('X-PI-Date', $t);
-		push @messages, $simple;
+		$mime->header_set('X-PI-Date', $t);
+		push @messages, $mime;
 		1;
 	});
 
@@ -234,13 +231,9 @@ sub get_feedopts {
 	\%rv;
 }
 
-sub utf8_header {
-	my ($simple, $name) = @_;
-	my $val = $simple->header($name);
-	return "" unless defined $val;
-	$val =~ tr/\t\n / /s;
-	$val =~ tr/\r//d;
-	$enc_utf8->encode($enc_mime->decode($val));
+sub mime_header {
+	my ($mime, $name) = @_;
+	PublicInbox::Hval->new_oneline($mime->header($name))->raw;
 }
 
 sub feed_date {
@@ -254,27 +247,25 @@ sub feed_date {
 sub add_to_feed {
 	my ($feed_opts, $feed, $add, $top, $git) = @_;
 
-	my $mime = do_cat_mail($git, 'Email::MIME', $add) or return 0;
-	if ($top && $mime->header("In-Reply-To")) {
+	my $mime = do_cat_mail($git, $add) or return 0;
+	if ($top && $mime->header('In-Reply-To')) {
 		return 0;
 	}
 
 	my $midurl = $feed_opts->{midurl} || 'http://example.com/m/';
 	my $fullurl = $feed_opts->{fullurl} || 'http://example.com/f/';
 
-	my $mid = $mime->header('Message-ID');
+	my $mid = $mime->header_obj->header_raw('Message-ID');
+	defined $mid or return 0;
 	$mid = PublicInbox::Hval->new_msgid($mid);
 	my $href = $mid->as_href;
 	my $content = PublicInbox::View->as_feed_entry($mime,
 							"$fullurl$href.html");
 	defined($content) or return 0;
 
-	my $subject = utf8_header($mime, "Subject") || "";
-	length($subject) or return 0;
-
-	my $from = $mime->header('From') or return 0;
-
+	my $subject = mime_header($mime, 'Subject') or return 0;
 
+	my $from = mime_header($mime, 'From') or return 0;
 	my @from = Email::Address->parse($from);
 	my $name = $from[0]->name;
 	defined $name or $name = "";
@@ -282,9 +273,8 @@ sub add_to_feed {
 	defined $email or $email = "";
 
 	my $date = $mime->header('Date');
-	$date or return 0;
 	$date = PublicInbox::Hval->new_oneline($date);
-	$date = feed_date($date->as_utf8) or return 0;
+	$date = feed_date($date->raw) or return 0;
 	$feed->add_entry(
 		author => { name => $name, email => $email },
 		title => $subject,
@@ -300,36 +290,32 @@ sub dump_html_line {
 	my ($self, $level, $args) = @_; # args => [ $html, $midurl ]
 	if ($self->message) {
 		$args->[0] .= (' ' x $level);
-		my $simple = $self->message;
-		my $subj = $simple->header('Subject');
-		my $mid = $simple->header('Message-ID');
+		my $mime = $self->message;
+		my $subj = $mime->header('Subject');
+		my $mid = $mime->header_obj->header_raw('Message-ID');
 		$mid = PublicInbox::Hval->new_msgid($mid);
 		my $url = $args->[1] . $mid->as_href;
-		my $from = utf8_header($simple, "From");
+		my $from = mime_header($mime, 'From');
+
 		my @from = Email::Address->parse($from);
 		$from = $from[0]->name;
 		(defined($from) && length($from)) or $from = $from[0]->address;
-		$from = xs_html($from);
-		$subj = PublicInbox::Hval->new_oneline($subj);
-		$subj = $subj->as_html;
+
+		$from = PublicInbox::Hval->new_oneline($from)->as_html;
+		$subj = PublicInbox::Hval->new_oneline($subj)->as_html;
 		$args->[0] .= "<a href=\"$url.html\">$subj</a> $from\n";
 	}
 	dump_html_line($self->child, $level+1, $args) if $self->child;
 	dump_html_line($self->next, $level, $args) if $self->next;
 }
 
-sub xs_html {
-	$enc_ascii->encode(escapeHTML($enc_utf8->decode($_[0])),
-			Encode::HTMLCREF);
-}
-
 sub try_git_pm {
 	my ($dir) = @_;
 	eval { Git->repository(Directory => $dir) };
 };
 
 sub do_cat_mail {
-	my ($git, $class, $sha1) = @_;
+	my ($git, $sha1) = @_;
 	my $str;
 	if ($git) {
 		open my $fh, '>', \$str or
@@ -342,7 +328,7 @@ sub do_cat_mail {
 		$str = `git cat-file blob $sha1`;
 		return if $? != 0 || length($str) == 0;
 	}
-	$class->new($str);
+	Email::MIME->new($str);
 }
 
 1;
diff --git a/lib/PublicInbox/Filter.pm b/lib/PublicInbox/Filter.pm
index bd33130c..a1587682 100644
--- a/lib/PublicInbox/Filter.pm
+++ b/lib/PublicInbox/Filter.pm
@@ -21,9 +21,9 @@ our $MIME_TEXT_ANY = qr!\btext/[a-z0-9\+\._-]+\b!i;
 # this is highly opinionated delivery
 # returns 0 only if there is nothing to deliver
 sub run {
-	my ($class, $simple) = @_;
+	my ($class, $mime) = @_;
 
-	my $content_type = $simple->header("Content-Type") || "text/plain";
+	my $content_type = $mime->header('Content-Type') || 'text/plain';
 
 	# kill potentially bad/confusing headers
 	# Note: ssoma already does this, but since we mangle the message,
@@ -32,35 +32,35 @@ sub run {
 	# the nature of public-inbox having no real subscribers.
 	foreach my $d (qw(status lines content-length
 			mail-followup-to mail-reply-to reply-to)) {
-		$simple->header_set($d);
+		$mime->header_set($d);
 	}
 
 	if ($content_type =~ m!\btext/plain\b!i) {
 		return 1; # yay, nothing to do
 	} elsif ($content_type =~ $MIME_HTML) {
 		# HTML-only, non-multipart
-		my $body = $simple->body;
+		my $body = $mime->body;
 		my $ct_parsed = parse_content_type($content_type);
 		dump_html(\$body, $ct_parsed->{attributes}->{charset});
-		replace_body($simple, $body);
+		replace_body($mime, $body);
 		return 1;
 	} elsif ($content_type =~ m!\bmultipart/!i) {
-		return strip_multipart($simple, $content_type);
+		return strip_multipart($mime, $content_type);
 	} else {
-		replace_body($simple, "$content_type message scrubbed");
+		replace_body($mime, "$content_type message scrubbed");
 		return 0;
 	}
 }
 
 sub replace_part {
-	my ($simple, $part, $type) = ($_[0], $_[1], $_[3]);
+	my ($mime, $part, $type) = ($_[0], $_[1], $_[3]);
 	# don't copy $_[2], that's the body (it may be huge)
 
 	# Email::MIME insists on setting Date:, so just set it consistently
 	# to avoid conflicts to avoid git merge conflicts in a split brain
 	# situation.
 	unless (defined $part->header('Date')) {
-		my $date = $simple->header('Date') ||
+		my $date = $mime->header('Date') ||
 		           'Thu, 01 Jan 1970 00:00:00 +0000';
 		$part->header_set('Date', $date);
 	}
@@ -77,11 +77,11 @@ sub replace_part {
 
 # converts one part of a multipart message to text
 sub html_part_to_text {
-	my ($simple, $part) = @_;
+	my ($mime, $part) = @_;
 	my $body = $part->body;
 	my $ct_parsed = parse_content_type($part->content_type);
 	dump_html(\$body, $ct_parsed->{attributes}->{charset});
-	replace_part($simple, $part, $body, 'text/plain');
+	replace_part($mime, $part, $body, 'text/plain');
 }
 
 # modifies $_[0] in place
@@ -110,8 +110,7 @@ sub dump_html {
 # unfortunately, too many people send HTML mail and we'll attempt to convert
 # it to something safer, smaller and harder-to-track.
 sub strip_multipart {
-	my ($simple, $content_type) = @_;
-	my $mime = Email::MIME->new($simple->as_string);
+	my ($mime, $content_type) = @_;
 
 	my (@html, @keep);
 	my $rejected = 0;
@@ -164,11 +163,11 @@ sub strip_multipart {
 
 	if ($content_type =~ m!\bmultipart/alternative\b!i) {
 		if (scalar @keep == 1) {
-			return collapse($simple, $keep[0]);
+			return collapse($mime, $keep[0]);
 		}
 	} else { # convert HTML parts to plain text
 		foreach my $part (@html) {
-			html_part_to_text($simple, $part);
+			html_part_to_text($mime, $part);
 			push @keep, $part;
 		}
 	}
@@ -186,34 +185,34 @@ sub strip_multipart {
 	}
 	if (scalar(@html) || $rejected) {
 		$mime->parts_set(\@keep);
-		$simple->body_set($mime->body_raw);
-		mark_changed($simple);
+		$mime->body_set($mime->body_raw);
+		mark_changed($mime);
 	} # else: no changes
 
 	return $ok;
 }
 
 sub mark_changed {
-	my ($simple) = @_;
-	$simple->header_set("X-Content-Filtered-By", __PACKAGE__ ." $VERSION");
+	my ($mime) = @_;
+	$mime->header_set('X-Content-Filtered-By', __PACKAGE__ ." $VERSION");
 }
 
 sub collapse {
-	my ($simple, $part) = @_;
-	$simple->header_set("Content-Type", $part->content_type);
-	$simple->body_set($part->body_raw);
-	mark_changed($simple);
+	my ($mime, $part) = @_;
+	$mime->header_set('Content-Type', $part->content_type);
+	$mime->body_set($part->body_raw);
+	mark_changed($mime);
 	return 1;
 }
 
 sub replace_body {
-	my $simple = $_[0];
-	$simple->body_set($_[1]);
-	$simple->header_set("Content-Type", "text/plain");
-	if ($simple->header("Content-Transfer-Encoding")) {
-		$simple->header_set("Content-Transfer-Encoding", undef);
+	my $mime = $_[0];
+	$mime->body_set($_[1]);
+	$mime->header_set('Content-Type', 'text/plain');
+	if ($mime->header('Content-Transfer-Encoding')) {
+		$mime->header_set('Content-Transfer-Encoding', undef);
 	}
-	mark_changed($simple);
+	mark_changed($mime);
 }
 
 # Check for display-able text, no messed up binaries
diff --git a/lib/PublicInbox/Hval.pm b/lib/PublicInbox/Hval.pm
index 47b41a58..09ae2676 100644
--- a/lib/PublicInbox/Hval.pm
+++ b/lib/PublicInbox/Hval.pm
@@ -5,12 +5,11 @@
 package PublicInbox::Hval;
 use strict;
 use warnings;
-use fields qw(raw -as_utf8);
+use fields qw(raw);
 use Encode qw(find_encoding);
 use CGI qw(escapeHTML);
 use URI::Escape qw(uri_escape);
 
-my $enc_utf8 = find_encoding('utf8');
 my $enc_ascii = find_encoding('us-ascii');
 
 sub new {
@@ -39,14 +38,17 @@ sub new_oneline {
 	$class->new($raw);
 }
 
-sub as_utf8 {
-	my ($self) = @_;
-	$self->{-as_utf8} ||= $enc_utf8->encode($self->{raw});
-}
-
 sub ascii_html { $enc_ascii->encode(escapeHTML($_[0]), Encode::HTMLCREF) }
 
-sub as_html { ascii_html($_[0]->as_utf8) }
-sub as_href { ascii_html(uri_escape($_[0]->as_utf8)) }
+sub as_html { ascii_html($_[0]->{raw}) }
+sub as_href { ascii_html(uri_escape($_[0]->{raw})) }
+
+sub raw {
+	if (defined $_[1]) {
+		$_[0]->{raw} = $_[1];
+	} else {
+		$_[0]->{raw};
+	}
+}
 
 1;
diff --git a/lib/PublicInbox/MDA.pm b/lib/PublicInbox/MDA.pm
index bb14ae50..ed8d74da 100644
--- a/lib/PublicInbox/MDA.pm
+++ b/lib/PublicInbox/MDA.pm
@@ -4,7 +4,7 @@ package PublicInbox::MDA;
 use strict;
 use warnings;
 use Email::Address;
-use Encode qw/decode encode/;
+use Encode qw/decode/;
 use Date::Parse qw(strptime);
 use constant MAX_SIZE => 1024 * 500; # same as spamc default
 use constant cmd => qw/ssoma-mda -1/;
@@ -71,16 +71,15 @@ sub set_list_headers {
 
 # returns a 3-element array: name, email, date
 sub author_info {
-	my ($class, $simple) = @_;
+	my ($class, $mime) = @_;
 
-	my $from = decode('MIME-Header', $simple->header('From'));
-	$from = encode('utf8', $from);
+	my $from = $mime->header('From');
 	my @from = Email::Address->parse($from);
 	my $name = $from[0]->name;
 	defined $name or $name = '';
 	my $email = $from[0]->address;
 	defined $email or $email = '';
-	($name, $email, $simple->header('Date'));
+	($name, $email, $mime->header('Date'));
 }
 
 1;
diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm
index 811e4703..0a8ab549 100644
--- a/lib/PublicInbox/View.pm
+++ b/lib/PublicInbox/View.pm
@@ -5,7 +5,6 @@ use strict;
 use warnings;
 use PublicInbox::Hval;
 use URI::Escape qw/uri_escape/;
-use CGI qw/escapeHTML/;
 use Encode qw/find_encoding/;
 use Encode::MIME::Header;
 use Email::MIME::ContentType qw/parse_content_type/;
@@ -13,8 +12,7 @@ use constant MAX_INLINE_QUOTED => 5;
 use constant MAX_TRUNC_LEN => 72;
 *ascii_html = *PublicInbox::Hval::ascii_html;
 
-my $enc_utf8 = find_encoding('utf8');
-my $enc_ascii = find_encoding('us-ascii');
+my $enc_utf8 = find_encoding('UTF-8');
 my $enc_mime = find_encoding('MIME-Header');
 
 # public functions:
@@ -138,33 +136,30 @@ sub add_text_body_full {
 }
 
 sub headers_to_html_header {
-	my ($simple, $full_pfx) = @_;
+	my ($mime, $full_pfx) = @_;
 
 	my $rv = "";
 	my @title;
 	foreach my $h (qw(From To Cc Subject Date)) {
-		my $v = $simple->header($h);
-		defined $v or next;
-		$v =~ tr/\n/ /s;
-		$v =~ tr/\r//d;
-		my $raw = $enc_mime->decode($v);
-		$v = ascii_html($raw);
-		$rv .= "$h: $v\n";
+		my $v = $mime->header($h);
+		defined($v) && length($v) or next;
+		$v = PublicInbox::Hval->new_oneline($v);
+		$rv .= "$h: " . $v->as_html . "\n";
 
 		if ($h eq 'From') {
-			my @from = Email::Address->parse($raw);
-			$raw = $from[0]->name;
-			unless (defined($raw) && length($raw)) {
-				$raw = '<' . $from[0]->address . '>';
+			my @from = Email::Address->parse($v->raw);
+			$v = $from[0]->name;
+			unless (defined($v) && length($v)) {
+				$v = '<' . $from[0]->address . '>';
 			}
-			$title[1] = ascii_html($raw);
-
+			$title[1] = ascii_html($v);
 		} elsif ($h eq 'Subject') {
-			$title[0] = $v;
+			$title[0] = $v->as_html;
 		}
 	}
 
-	my $mid = $simple->header('Message-ID');
+	my $header_obj = $mime->header_obj;
+	my $mid = $header_obj->header_raw('Message-ID');
 	if (defined $mid) {
 		$mid = PublicInbox::Hval->new_msgid($mid);
 		$rv .= 'Message-ID: &lt;' . $mid->as_html . '&gt; ';
@@ -173,7 +168,7 @@ sub headers_to_html_header {
 		$rv .= "(<a href=\"$href.txt\">original</a>)\n";
 	}
 
-	my $irp = $simple->header('In-Reply-To');
+	my $irp = $header_obj->header_raw('In-Reply-To');
 	if (defined $irp) {
 		$irp = PublicInbox::Hval->new_msgid($irp);
 		my $html = $irp->as_html;
diff --git a/public-inbox-learn b/public-inbox-learn
index 2c2bbfb5..bb130b21 100755
--- a/public-inbox-learn
+++ b/public-inbox-learn
@@ -5,7 +5,7 @@ my $usage = "$0 (spam|ham) < /path/to/message";
 use strict;
 use warnings;
 use PublicInbox::Config;
-use Email::Simple;
+use Email::MIME;
 use Email::Address;
 use IPC::Run qw/run/;
 my $train = shift or die "usage: $usage\n";
@@ -14,22 +14,18 @@ if ($train !~ /\A(?:ham|spam)\z/) {
 }
 
 my $pi_config = PublicInbox::Config->new;
-my $simple;
-{
-	local $/;
-	$simple = Email::Simple->new(<>);
-}
+my $mime = Email::MIME->new(eval { local $/; <> });
 
 # get all recipients
 my %dests;
 foreach my $h (qw(Cc To)) {
-	foreach my $recipient (Email::Address->parse($simple->header($h))) {
+	foreach my $recipient (Email::Address->parse($mime->header($h))) {
 		$dests{lc($recipient->address)} = 1;
 	}
 }
 
-my $in = $simple->as_string;
-$simple->body_set("");
+my $in = $mime->as_string;
+$mime->body_set('');
 
 my $err = 0;
 my @output = qw(> /dev/null > /dev/null);
@@ -59,10 +55,10 @@ foreach my $recipient (keys %dests) {
 
 		# no checking for errors here, we assume the message has
 		# been reviewed by a human at this point:
-		PublicInbox::Filter->run($simple);
+		PublicInbox::Filter->run($mime);
 
 		my ($name, $email, $date) =
-				PublicInbox::MDA->author_info($simple);
+				PublicInbox::MDA->author_info($mime);
 		local $ENV{GIT_AUTHOR_NAME} = $name;
 		local $ENV{GIT_AUTHOR_EMAIL} = $email;
 		local $ENV{GIT_AUTHOR_DATE} = $date;
diff --git a/public-inbox-mda b/public-inbox-mda
index 6c446749..50805da2 100755
--- a/public-inbox-mda
+++ b/public-inbox-mda
@@ -6,9 +6,8 @@ use warnings;
 my $usage = 'public-inbox-mda < rfc2822_message';
 
 use Email::Filter;
+use Email::MIME;
 use Email::Address;
-use Encode qw/decode encode/;
-use Encode::MIME::Header;
 use File::Path::Expand qw/expand_filename/;
 use IPC::Run qw(run);
 use PublicInbox::MDA;
@@ -35,19 +34,19 @@ my $filtered; # string dest
 if (PublicInbox::MDA->precheck($filter, $recipient) &&
     do_spamc($filter->simple, \$filtered)) {
 	# update our message with SA headers (in case our filter rejects it)
-	my $simple = Email::Simple->new($filtered);
+	my $msg = Email::Simple->new($filtered);
 	$filtered = undef;
-	$filter->simple($simple);
+	$filter->simple($msg);
 
-	if (PublicInbox::Filter->run($simple)) {
+	if (PublicInbox::Filter->run($msg)) {
 		# run spamc again on the HTML-free message
-		if (do_spamc($simple, \$filtered)) {
-			$simple = Email::Simple->new($filtered);
-			PublicInbox::MDA->set_list_headers($simple, $dst);
-			$filter->simple($simple);
+		if (do_spamc($msg, \$filtered)) {
+			$msg = Email::MIME->new($filtered);
+			PublicInbox::MDA->set_list_headers($msg, $dst);
+			$filter->simple($msg);
 
 			my ($name, $email, $date) =
-					PublicInbox::MDA->author_info($simple);
+					PublicInbox::MDA->author_info($msg);
 			local $ENV{GIT_AUTHOR_NAME} = $name;
 			local $ENV{GIT_AUTHOR_EMAIL} = $email;
 			local $ENV{GIT_AUTHOR_DATE} = $date;
@@ -64,9 +63,9 @@ exit 0; # goes to emergency
 # not using Email::Filter->pipe here since we want the stdout of
 # the command even on failure (spamc will set $? on error).
 sub do_spamc {
-	my ($simple, $out) = @_;
+	my ($msg, $out) = @_;
 	eval {
-		my $orig = $simple->as_string;
+		my $orig = $msg->as_string;
 		run([qw/spamc -E --headers/], \$orig, $out);
 	};
 
diff --git a/public-inbox.cgi b/public-inbox.cgi
index 9bc3dc2b..3c51c314 100755
--- a/public-inbox.cgi
+++ b/public-inbox.cgi
@@ -12,10 +12,11 @@
 use 5.008;
 use strict;
 use warnings;
-use CGI qw(:cgi :escapeHTML -nosticky); # PSGI/FastCGI/mod_perl compat
-use Encode qw(decode_utf8);
+use CGI qw(:cgi -nosticky); # PSGI/FastCGI/mod_perl compat
+use Encode qw(find_encoding);
 use PublicInbox::Config;
 use URI::Escape qw(uri_escape uri_unescape);
+my $enc_utf8 = find_encoding('UTF-8');
 our $LISTNAME_RE = qr!\A/([\w\.\-]+)!;
 our $pi_config;
 BEGIN {
@@ -56,7 +57,7 @@ sub main {
 	if ($cgi->request_method !~ /\AGET|HEAD\z/) {
 		return r("405 Method Not Allowed");
 	}
-	my $path_info = decode_utf8($cgi->path_info);
+	my $path_info = $enc_utf8->decode($cgi->path_info);
 
 	# top-level indices and feeds
 	if ($path_info eq "/") {
@@ -213,6 +214,6 @@ sub set_binmode {
 		# no way to validate raw messages, mixed encoding is possible.
 		binmode STDOUT;
 	} else { # strict encoding for HTML and XML
-		binmode STDOUT, ':encoding(UTF-8)';
+		binmode STDOUT, ':encoding(us-ascii)';
 	}
 }
diff --git a/t/filter.t b/t/filter.t
index ac9f1f6f..c3cd39f4 100644
--- a/t/filter.t
+++ b/t/filter.t
@@ -4,7 +4,6 @@ use strict;
 use warnings;
 use Test::More;
 use Email::MIME;
-use Email::Filter;
 use PublicInbox::Filter;
 
 sub count_body_parts {
@@ -18,7 +17,7 @@ sub count_body_parts {
 
 # plain-text email is passed through unchanged
 {
-	my $s = Email::Simple->create(
+	my $s = Email::MIME->create(
 		header => [
 			From => 'a@example.com',
 			To => 'b@example.com',
@@ -27,14 +26,12 @@ sub count_body_parts {
 		],
 		body => "hello world\n",
 	);
-	my $f = Email::Filter->new(data => $s->as_string);
-	is(1, PublicInbox::Filter->run($f->simple), "run was a success");
-	is($s->as_string, $f->simple->as_string, "plain email unchanged");
+	is(1, PublicInbox::Filter->run($s), "run was a success");
 }
 
 # convert single-part HTML to plain-text
 {
-	my $s = Email::Simple->create(
+	my $s = Email::MIME->create(
 		header => [
 			From => 'a@example.com',
 			To => 'b@example.com',
@@ -43,13 +40,12 @@ sub count_body_parts {
 		],
 		body => "<html><body>bad body</body></html>\n",
 	);
-	my $f = Email::Filter->new(data => $s->as_string);
-	is(1, PublicInbox::Filter->run($f->simple), "run was a success");
-	unlike($f->simple->as_string, qr/<html>/, "HTML removed");
-	is("text/plain", $f->simple->header("Content-Type"),
+	is(1, PublicInbox::Filter->run($s), "run was a success");
+	unlike($s->as_string, qr/<html>/, "HTML removed");
+	is("text/plain", $s->header("Content-Type"),
 		"content-type changed");
-	like($f->simple->body, qr/\A\s*bad body\s*\z/, "body");
-	like($f->simple->header("X-Content-Filtered-By"),
+	like($s->body, qr/\A\s*bad body\s*\z/, "body");
+	like($s->header("X-Content-Filtered-By"),
 		qr/PublicInbox::Filter/, "XCFB header added");
 }
 
@@ -79,9 +75,8 @@ sub count_body_parts {
 		],
 		parts => $parts,
 	);
-	my $f = Email::Filter->new(data => $email->as_string);
-	is(1, PublicInbox::Filter->run($f->simple), "run was a success");
-	my $parsed = Email::MIME->new($f->simple->as_string);
+	is(1, PublicInbox::Filter->run($email), "run was a success");
+	my $parsed = Email::MIME->new($email->as_string);
 	is("text/plain", $parsed->header("Content-Type"));
 	is(scalar $parsed->parts, 1, "HTML part removed");
 	my %bodies;
@@ -110,9 +105,8 @@ sub count_body_parts {
 		header_str => [ From => 'a@example.com', Subject => 'blah' ],
 		parts => $parts,
 	);
-	my $f = Email::Filter->new(data => $email->as_string);
-	is(1, PublicInbox::Filter->run($f->simple), "run was a success");
-	my $parsed = Email::MIME->new($f->simple->as_string);
+	is(1, PublicInbox::Filter->run($email), "run was a success");
+	my $parsed = Email::MIME->new($email->as_string);
 	is(scalar $parsed->parts, 2, "still 2 parts");
 	my %bodies;
 	$parsed->walk_parts(sub {
@@ -149,9 +143,8 @@ sub count_body_parts {
 		header_str => [ From => 'a@example.com', Subject => 'blah' ],
 		parts => $parts,
 	);
-	my $f = Email::Filter->new(data => $email->as_string);
-	is(1, PublicInbox::Filter->run($f->simple), "run was a success");
-	my $parsed = Email::MIME->new($f->simple->as_string);
+	is(1, PublicInbox::Filter->run($email), "run was a success");
+	my $parsed = Email::MIME->new($email->as_string);
 	is(scalar $parsed->parts, 2, "still 2 parts");
 	my %bodies;
 	$parsed->walk_parts(sub {
@@ -186,9 +179,8 @@ sub count_body_parts {
 		header_str => [ From => 'a@example.com', Subject => 'blah' ],
 		parts => $parts,
 	);
-	my $f = Email::Filter->new(data => $email->as_string);
-	is(1, PublicInbox::Filter->run($f->simple), "run was a success");
-	my $parsed = Email::MIME->new($f->simple->as_string);
+	is(1, PublicInbox::Filter->run($email), "run was a success");
+	my $parsed = Email::MIME->new($email->as_string);
 	is(scalar $parsed->parts, 1, "image part removed");
 	my %bodies;
 	$parsed->walk_parts(sub {
@@ -226,10 +218,9 @@ sub count_body_parts {
 		header_str => [ From => 'a@example.com', Subject => 'blah' ],
 		parts => $parts,
 	);
-	my $f = Email::Filter->new(data => $email->as_string);
-	is(0, PublicInbox::Filter->run($f->simple),
+	is(0, PublicInbox::Filter->run($email),
 		"run signaled to stop delivery");
-	my $parsed = Email::MIME->new($f->simple->as_string);
+	my $parsed = Email::MIME->new($email->as_string);
 	is(scalar $parsed->parts, 1, "bad parts removed");
 	my %bodies;
 	$parsed->walk_parts(sub {
@@ -245,7 +236,7 @@ sub count_body_parts {
 }
 
 {
-	my $s = Email::Simple->create(
+	my $s = Email::MIME->create(
 		header => [
 			From => 'a@example.com',
 			To => 'b@example.com',
@@ -254,13 +245,12 @@ sub count_body_parts {
 		],
 		body => "hello world\n",
 	);
-	my $f = Email::Filter->new(data => $s->as_string);
-	is(0, PublicInbox::Filter->run($f->simple), "run was a failure");
-	like($f->simple->as_string, qr/scrubbed/, "scrubbed message");
+	is(0, PublicInbox::Filter->run($s), "run was a failure");
+	like($s->as_string, qr/scrubbed/, "scrubbed message");
 }
 
 {
-	my $s = Email::Simple->create(
+	my $s = Email::MIME->create(
 		header => [
 			From => 'a@example.com',
 			To => 'b@example.com',
@@ -273,9 +263,8 @@ sub count_body_parts {
 
 	is('c@example.com', $s->header("Mail-Followup-To"),
 		"mft set correctly");
-	my $f = Email::Filter->new(data => $s->as_string);
-	is(1, PublicInbox::Filter->run($f->simple), "run succeeded for mft");
-	is(undef, $f->simple->header("Mail-Followup-To"), "mft stripped");
+	is(1, PublicInbox::Filter->run($s), "run succeeded for mft");
+	is(undef, $s->header("Mail-Followup-To"), "mft stripped");
 }
 
 # multi-part with application/octet-stream
@@ -308,11 +297,10 @@ EOF
 		header_str => [ From => 'a@example.com', Subject => 'blah' ],
 		parts => $parts,
 	);
-	my $f = Email::Filter->new(data => $email->as_string);
-	is(1, PublicInbox::Filter->run($f->simple), "run was a success");
-	my $parsed = Email::MIME->new($f->simple->as_string);
+	is(1, PublicInbox::Filter->run($email), "run was a success");
+	my $parsed = Email::MIME->new($email->as_string);
 	is(scalar $parsed->parts, 1, "only one remaining part");
-	like($f->simple->header("X-Content-Filtered-By"),
+	like($parsed->header("X-Content-Filtered-By"),
 		qr/PublicInbox::Filter/, "XCFB header added");
 }
 
diff --git a/t/mda.t b/t/mda.t
index 3357718c..fad96e5b 100644
--- a/t/mda.t
+++ b/t/mda.t
@@ -4,6 +4,7 @@ use strict;
 use warnings;
 use Test::More;
 use Email::MIME;
+use Email::Filter;
 use File::Temp qw/tempdir/;
 use Cwd;
 use IPC::Run qw(run);
@@ -43,6 +44,26 @@ my $failbox = "$home/fail.mbox";
 	}
 }
 
+local $ENV{GIT_COMMITTER_NAME} = eval {
+	use PublicInbox::MDA;
+	use Encode qw/encode/;
+	my $mbox = 't/utf8.mbox';
+	open(my $fh, '<', $mbox) or die "failed to open mbox: $mbox\n";
+	my $str = eval { local $/; <$fh> };
+	close $fh;
+	my $msg = Email::Filter->new(data => $str);
+	$msg = Email::MIME->new($msg->simple->as_string);
+	my ($author, $email, $date) = PublicInbox::MDA->author_info($msg);
+	is('El&#233;anor',
+		encode('us-ascii', my $tmp = $author, Encode::HTMLCREF),
+		'HTML conversion is correct');
+	is($email, 'e@example.com', 'email parsed correctly');
+	is($date, 'Thu, 01 Jan 1970 00:00:00 +0000',
+		'message date parsed correctly');
+	$author;
+};
+die $@ if $@;
+
 {
 	my $good_rev;
 	local $ENV{PI_EMERGENCY} = $failbox;
diff --git a/t/utf8.mbox b/t/utf8.mbox
new file mode 100644
index 00000000..cebaf9b0
--- /dev/null
+++ b/t/utf8.mbox
@@ -0,0 +1,12 @@
+From e@yhbt.net Thu Jan 01 00:00:00 1970
+Date: Thu, 01 Jan 1970 00:00:00 +0000
+To: =?utf-8?Q?El=C3=A9anor?= <e@example.com>
+From: =?utf-8?Q?El=C3=A9anor?= <e@example.com>
+Subject: Testing for =?utf-8?Q?El=C3=A9anor?=
+Message-ID: <testmessage@example.com>
+MIME-Version: 1.0
+Content-Type: text/plain; charset=utf-8
+Content-Disposition: inline
+Content-Transfer-Encoding: 8bit
+
+This is a test message for Eléanor
-- 
cgit v1.2.3-24-ge0c7