-# Copyright (C) 2014, Eric Wong <normalperson@yhbt.net> and all contributors
+# Copyright (C) 2014-2015 all contributors <meta@public-inbox.org>
# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
#
# represents a header value in various forms
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);
+use PublicInbox::MID qw/mid_clean/;
-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, $no_compress) = @_;
+ $msgid = mid_clean($msgid);
+ $class->new($msgid, $msgid);
}
sub new_oneline {
$class->new($raw);
}
-sub as_utf8 {
- my ($self) = @_;
- $self->{-as_utf8} ||= $enc_utf8->encode($self->{raw});
+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 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;