hval: fixup bad line endings in HTML output

[public-inbox.git] / lib / PublicInbox / Hval.pm

diff --git a/lib/PublicInbox/Hval.pm b/lib/PublicInbox/Hval.pm
index 47b41a5844c875ddff9eeabf271485025a157f68..68f895460821569e9ba67a01902af6248e9692c5 100644 (file)
--- a/lib/PublicInbox/Hval.pm
+++ b/lib/PublicInbox/Hval.pm
@@ -5,30 +5,36 @@
 package PublicInbox::Hval;
 use strict;
 use warnings;
-use fields qw(raw -as_utf8);
+use fields qw(raw href);
 use Encode qw(find_encoding);
-use CGI qw(escapeHTML);
-use URI::Escape qw(uri_escape);
+use URI::Escape qw(uri_escape_utf8);
 
-my $enc_utf8 = find_encoding('utf8');
 my $enc_ascii = find_encoding('us-ascii');
 
 sub new {
-       my ($class, $raw) = @_;
+       my ($class, $raw, $href) = @_;
        my $self = fields::new($class);
 
        # we never care about leading/trailing whitespace
        $raw =~ s/\A\s*//;
        $raw =~ s/\s*\z//;
        $self->{raw} = $raw;
+       $self->{href} = defined $href ? $href : $raw;
        $self;
 }
 
 sub new_msgid {
-       my ($class, $raw) = @_;
-       $raw =~ s/\A<//;
-       $raw =~ s/>\z//;
-       $class->new($raw);
+       my ($class, $msgid) = @_;
+       $msgid =~ s/\A\s*<?//;
+       $msgid =~ s/>?\s*\z//;
+
+       if (length($msgid) <= 40) {
+               $class->new($msgid);
+       } else {
+               require Digest::SHA;
+               my $hex = Digest::SHA::sha1_hex($msgid);
+               $class->new($msgid, $hex);
+       }
 }
 
 sub new_oneline {
@@ -39,14 +45,30 @@ sub new_oneline {
        $class->new($raw);
 }
 
-sub as_utf8 {
-       my ($self) = @_;
-       $self->{-as_utf8} ||= $enc_utf8->encode($self->{raw});
+my %xhtml_map = (
+       '"' => '&#34;',
+       '&' => '&#38;',
+       "'" => '&#39;',
+       '<' => '&lt;',
+       '>' => '&gt;',
+);
+
+sub ascii_html {
+       my ($s) = @_;
+       $s =~ s/\r\n/\n/sg; # fixup bad line endings
+       $s =~ s/([<>&'"])/$xhtml_map{$1}/ge;
+       $enc_ascii->encode($s, Encode::HTMLCREF);
 }
 
-sub ascii_html { $enc_ascii->encode(escapeHTML($_[0]), Encode::HTMLCREF) }
+sub as_html { ascii_html($_[0]->{raw}) }
+sub as_href { ascii_html(uri_escape_utf8($_[0]->{href})) }
 
-sub as_html { ascii_html($_[0]->as_utf8) }
-sub as_href { ascii_html(uri_escape($_[0]->as_utf8)) }
+sub raw {
+       if (defined $_[1]) {
+               $_[0]->{raw} = $_[1];
+       } else {
+               $_[0]->{raw};
+       }
+}
 
 1;