X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FView.pm;h=4058bee775507c70e0c1a73ca82235c4f9026f78;hb=476fc666c223f0fb;hp=61eb890f2ff46f3a076753c25e2ccf0343b707bb;hpb=f5eb6bb83558f3dabdb2fbbf7c1c926a726bfed9;p=public-inbox.git

diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm
index 61eb890f..4058bee7 100644
--- a/lib/PublicInbox/View.pm
+++ b/lib/PublicInbox/View.pm
@@ -11,10 +11,9 @@ use Date::Parse qw/str2time/;
 use Encode qw/find_encoding/;
 use Encode::MIME::Header;
 use Email::MIME::ContentType qw/parse_content_type/;
-use PublicInbox::Hval;
-use PublicInbox::MID qw/mid_clean id_compress mid2path/;
-use Digest::SHA qw/sha1_hex/;
-my $SALT = rand;
+use PublicInbox::Hval qw/ascii_html/;
+use PublicInbox::Linkify;
+use PublicInbox::MID qw/mid_clean id_compress mid2path mid_mime/;
 require POSIX;
 
 # TODO: make these constants tunable
@@ -23,8 +22,6 @@ use constant MAX_TRUNC_LEN => 72;
 use constant T_ANCHOR => '#u';
 use constant INDENT => '  ';
 
-*ascii_html = *PublicInbox::Hval::ascii_html;
-
 my $enc_utf8 = find_encoding('UTF-8');
 
 # public functions:
@@ -51,10 +48,9 @@ sub msg_reply {
 	$s = '(no subject)' if (!defined $s) || ($s eq '');
 	my $f = $hdr->header('From');
 	$f = '' unless defined $f;
-	$s = PublicInbox::Hval->new_oneline($s);
-	my $mid = $hdr->header('Message-ID');
+	my $mid = $hdr->header_raw('Message-ID');
 	$mid = PublicInbox::Hval->new_msgid($mid);
-	my $t = $s->as_html;
+	my $t = ascii_html($s);
 	my $se_url =
 	 'https://kernel.org/pub/software/scm/git/docs/git-send-email.html';
 
@@ -93,11 +89,11 @@ sub feed_entry {
 
 sub in_reply_to {
 	my ($hdr) = @_;
-	my $irt = $hdr->header('In-Reply-To');
+	my $irt = $hdr->header_raw('In-Reply-To');
 
 	return mid_clean($irt) if (defined $irt);
 
-	my $refs = $hdr->header('References');
+	my $refs = $hdr->header_raw('References');
 	if ($refs && $refs =~ /<([^>]+)>\s*\z/s) {
 		return $1;
 	}
@@ -116,24 +112,24 @@ sub index_entry {
 	my $enc = enc_for($hdr->header("Content-Type"));
 	my $subj = $hdr->header('Subject');
 
-	my $mid_raw = mid_clean($hdr->header('Message-ID'));
+	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 $from = PublicInbox::Hval->new_oneline($hdr->header('From'))->raw;
+	my $from = $hdr->header('From');
 	my @from = Email::Address->parse($from);
 	$from = $from[0]->name;
 
-	$from = PublicInbox::Hval->new_oneline($from)->as_html;
-	$subj = PublicInbox::Hval->new_oneline($subj)->as_html;
 	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;
 
+	$from = ascii_html($from);
+	$subj = ascii_html($subj);
 	if ($srch) {
 		my $t = $ctx->{flat} ? 'T' : 't';
 		$subj = "<a\nhref=\"${path}$href/$t/#u\">$subj</a>";
@@ -302,41 +298,6 @@ sub add_filename_line {
 	"$pad " . ascii_html($fn) . " $pad\n";
 }
 
-my $LINK_RE = qr!\b((?:ftp|https?|nntp)://
-		 [\@:\w\.-]+/
-		 ?[\@\w\+\&\?\.\%\;/#=-]*)!x;
-
-sub linkify_1 {
-	my ($link_map, $s) = @_;
-	$s =~ s!$LINK_RE!
-		my $url = $1;
-		# salt this, as this could be exploited to show
-		# links in the HTML which don't show up in the raw mail.
-		my $key = sha1_hex($url . $SALT);
-		$link_map->{$key} = $url;
-		'PI-LINK-'. $key;
-	!ge;
-	$s;
-}
-
-sub linkify_2 {
-	my ($link_map, $s) = @_;
-
-	# Added "PI-LINK-" prefix to avoid false-positives on git commits
-	$s =~ s!\bPI-LINK-([a-f0-9]{40})\b!
-		my $key = $1;
-		my $url = $link_map->{$key};
-		if (defined $url) {
-			$url = ascii_html($url);
-			"<a\nhref=\"$url\">$url</a>";
-		} else {
-			# false positive or somebody tried to mess with us
-			$key;
-		}
-	!ge;
-	$s;
-}
-
 sub flush_quote {
 	my ($quot, $n, $part_nr, $full_pfx, $final, $do_anchor) = @_;
 
@@ -346,11 +307,11 @@ sub flush_quote {
 	if ($full_pfx) {
 		if (!$final && scalar(@$quot) <= MAX_INLINE_QUOTED) {
 			# show quote inline
-			my %l;
-			my $rv = join('', map { linkify_1(\%l, $_) } @$quot);
+			my $l = PublicInbox::Linkify->new;
+			my $rv = join('', map { $l->linkify_1($_) } @$quot);
 			@$quot = ();
 			$rv = ascii_html($rv);
-			return linkify_2(\%l, $rv);
+			return $l->linkify_2($rv);
 		}
 
 		# show a short snippet of quoted text and link to full version:
@@ -375,13 +336,13 @@ sub flush_quote {
 	} else {
 		# show everything in the full version with anchor from
 		# short version (see above)
-		my %l;
-		my $rv .= join('', map { linkify_1(\%l, $_) } @$quot);
+		my $l = PublicInbox::Linkify->new;
+		my $rv .= join('', map { $l->linkify_1($_) } @$quot);
 		@$quot = ();
 		$rv = ascii_html($rv);
-		return linkify_2(\%l, $rv) unless $do_anchor;
+		return $l->linkify_2($rv) unless $do_anchor;
 		my $nr = ++$$n;
-		"<a\nid=q${part_nr}_$nr></a>" . linkify_2(\%l, $rv);
+		"<a\nid=q${part_nr}_$nr></a>" . $l->linkify_2($rv);
 	}
 }
 
@@ -420,10 +381,10 @@ sub add_text_body {
 			}
 
 			# regular line, OK
-			my %l;
-			$cur = linkify_1(\%l, $cur);
+			my $l = PublicInbox::Linkify->new;
+			$cur = $l->linkify_1($cur);
 			$cur = ascii_html($cur);
-			$s .= linkify_2(\%l, $cur);
+			$s .= $l->linkify_2($cur);
 		} else {
 			push @quot, $cur;
 		}
@@ -445,12 +406,12 @@ sub headers_to_html_header {
 	my $srch = $ctx->{srch} if $ctx;
 	my $rv = "";
 	my @title;
-	my $mid = $hdr->header('Message-ID');
+	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;
-		$v = PublicInbox::Hval->new_oneline($v);
+		$v = PublicInbox::Hval->new($v);
 
 		if ($h eq 'From') {
 			my @from = Email::Address->parse($v->raw);
@@ -488,7 +449,7 @@ sub headers_to_html_header {
 sub thread_inline {
 	my ($dst, $ctx, $hdr, $upfx) = @_;
 	my $srch = $ctx->{srch};
-	my $mid = mid_clean($hdr->header('Message-ID'));
+	my $mid = mid_clean($hdr->header_raw('Message-ID'));
 	my $res = $srch->get_thread($mid);
 	my $nr = $res->{total};
 	my $expand = "<a\nhref=\"${upfx}t/#u\">expand</a> " .
@@ -545,7 +506,7 @@ sub _parent_headers_nosrch {
 		$rv .= "<a\nhref=\"../$href/\">$html</a>&gt;\n";
 	}
 
-	my $refs = $hdr->header('References');
+	my $refs = $hdr->header_raw('References');
 	if ($refs) {
 		# avoid redundant URLs wasting bandwidth
 		my %seen;
@@ -586,7 +547,7 @@ sub mailto_arg_link {
 
 	my $subj = $hdr->header('Subject') || '';
 	$subj = "Re: $subj" unless $subj =~ /\bRe:/i;
-	my $mid = $hdr->header('Message-ID');
+	my $mid = $hdr->header_raw('Message-ID');
 	push @arg, "--in-reply-to='" . ascii_html($mid) . "'";
 	my $irt = uri_escape_utf8($mid);
 	delete $cc{$to};
@@ -618,7 +579,7 @@ sub html_footer {
 		my $p = $ctx->{parent_msg};
 		my $next = $ctx->{next_msg};
 		if ($p) {
-			$p = PublicInbox::Hval->new_oneline($p);
+			$p = PublicInbox::Hval->new_msgid($p);
 			$p = $p->as_href;
 			$irt = "<a\nhref=\"$upfx$p/\">parent</a> ";
 		} else {
@@ -662,8 +623,7 @@ sub thread_html_head {
 	my ($cb, $header, $state) = @_;
 	$$cb = $$cb->([200, ['Content-Type'=> 'text/html; charset=UTF-8']]);
 
-	my $s = PublicInbox::Hval->new_oneline($header->header('Subject'));
-	$s = $s->as_html;
+	my $s = ascii_html($header->header('Subject'));
 	$$cb->write("<html><head><title>$s</title>".
 		qq{<link\nrel=alternate\ntitle="Atom feed"\n} .
 		qq!href="../t.atom"\ntype="application/atom+xml"/>! .
@@ -673,7 +633,7 @@ sub thread_html_head {
 
 sub pre_anchor_entry {
 	my ($seen, $mime) = @_;
-	my $id = anchor_for($mime->header('Message-ID'));
+	my $id = anchor_for(mid_mime($mime));
 	$seen->{$id} = "#$id"; # save the anchor for children, later
 }
 
@@ -726,7 +686,7 @@ sub __thread_entry {
 
 	# lazy load the full message from mini_mime:
 	$mime = eval {
-		my $path = mid2path(mid_clean($mime->header('Message-ID')));
+		my $path = mid2path(mid_clean(mid_mime($mime)));
 		Email::MIME->new($git->cat_file('HEAD:'.$path));
 	} or return;
 
@@ -816,10 +776,9 @@ sub _inline_header {
 	my $dot = $level == 0 ? '' : '` ';
 
 	my $cur = $state->{cur};
-	my $mid = mid_clean($hdr->header('Message-ID'));
-	my $f = $hdr->header('X-PI-From');
+	my $mid = mid_clean($hdr->header_raw('Message-ID'));
+	my $f = ascii_html($hdr->header('X-PI-From'));
 	my $d = _msg_date($hdr);
-	$f = PublicInbox::Hval->new_oneline($f)->as_html;
 	my $pfx = ' ' . $d . ' ' . indent_for($level);
 	my $attr = $f;
 	$state->{first_level} ||= $level;
@@ -869,7 +828,7 @@ sub inline_dump {
 	return unless $node;
 	if (my $mime = $node->message) {
 		my $hdr = $mime->header_obj;
-		my $mid = mid_clean($hdr->header('Message-ID'));
+		my $mid = mid_clean($hdr->header_raw('Message-ID'));
 		if ($mid eq $state->{parent_cmp}) {
 			$state->{parent} = $mid;
 		}
@@ -917,7 +876,7 @@ sub add_topic {
 			push @{$state->{order}}, [ $level, $subj ];
 		}
 
-		my $mid = mid_clean($x->header('Message-ID'));
+		my $mid = mid_clean($x->header_raw('Message-ID'));
 
 		my $ts = $x->header('X-PI-TS');
 		my $exist = $state->{latest}->{$subj};