hval: to_attr: support wide characters

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

diff --git a/lib/PublicInbox/Hval.pm b/lib/PublicInbox/Hval.pm
index 7e007027c8ba9d27fc58cf6c6d942b83b1d18c4f..39256ee01e0380bdfafd8c655265a82ceb796100 100644 (file)
--- a/lib/PublicInbox/Hval.pm
+++ b/lib/PublicInbox/Hval.pm
@@ -139,10 +139,12 @@ sub to_attr ($) {
        return if index($str, '//') >= 0;
 
        my $first = '';
+       utf8::encode($str); # to octets
        if ($str =~ s/\A([^A-Ya-z])//ms) { # start with a letter
                  $first = sprintf('Z%02x', ord($1));
        }
        $str =~ s/([^A-Za-z0-9_\.\-])/$ESCAPES{$1}/egms;
+       utf8::decode($str); # allow wide chars
        $first . $str;
 }
 
@@ -155,6 +157,7 @@ sub from_attr ($) {
        }
        $str =~ s!::([a-f0-9]{2})!chr(hex($1))!egms;
        $str =~ tr!:!/!;
+       utf8::decode($str);
        $first . $str;
 }