mid: mid_compressed => mid_compress

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

diff --git a/lib/PublicInbox/Hval.pm b/lib/PublicInbox/Hval.pm
index 1a83cf10a54561f698f45a6671d6ed1f99734548..21efe40ef87771668b204836076e438886f1fb0f 100644 (file)
--- a/lib/PublicInbox/Hval.pm
+++ b/lib/PublicInbox/Hval.pm
@@ -5,28 +5,29 @@
 package PublicInbox::Hval;
 use strict;
 use warnings;
-use fields qw(raw);
+use fields qw(raw href);
 use Encode qw(find_encoding);
-use URI::Escape qw(uri_escape);
+use URI::Escape qw(uri_escape_utf8);
+use PublicInbox::MID qw/mid_clean mid_compress/;
 
 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) = @_;
+       $msgid = mid_clean($msgid);
+       $class->new($msgid, mid_compress($msgid));
 }
 
 sub new_oneline {
@@ -47,12 +48,13 @@ my %xhtml_map = (
 
 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 as_html { ascii_html($_[0]->{raw}) }
-sub as_href { ascii_html(uri_escape($_[0]->{raw})) }
+sub as_href { ascii_html(uri_escape_utf8($_[0]->{href})) }
 
 sub raw {
        if (defined $_[1]) {