| Subject: | [Fwd: IMAPUTF7 patch] |
| Date: | Tue, 14 Apr 2009 20:39:17 -0700 |
| To: | bug-encode-imaputf7 [...] rt.cpan.org |
| From: | Nick Czeczulin <nick [...] packetfrenzy.com> |
| Subject: | IMAPUTF7 patch |
| Date: | Fri, 10 Apr 2009 21:11:29 -0700 |
| To: | peter [...] makholm.net |
| From: | Nick Czeczulin <nick [...] packetfrenzy.com> |
Hi Peter,
My apologies for not using rt, but I do not have a registered account.
Two minor changes:
1. Changed to use ascii ranges to explicitly match the rfc. One of the
flaws with the current module is \r\n\t should be encoded.
2. The other thing I noticed is if you have an '&' following a
non-printable character it will get base64 encoded as well.
Thanks for maintaining this module.
Cheers!
-nick
--- IMAPUTF7.pm 2009-04-10 20:24:04.000000000 -0700
+++ IMAPUTF7-new.pm 2009-04-10 20:56:38.000000000 -0700
@@ -15,11 +15,8 @@
# Code directly borrowed from Encode::Unicode::UTF7 by Dan Kogai
#
-my $specials = quotemeta "!\"#$%'()*+,-./:;<=>?@[\\]^_`{|}~";
-# \s will not work because it matches U+3000 DEOGRAPHIC SPACE
-# We use qr/[\n\r\t\ ] instead
-my $re_asis = qr/(?:[\n\r\t\ A-Za-z0-9$specials])/;
-my $re_encoded = qr/(?:[^\n\r\t\ A-Za-z0-9$specials])/;
+my $re_asis = qr/(?:[\x20-\x25\x27-\x7e])/;
+my $re_encoded = qr/(?:[^\x20-\x7e])/;
my $e_utf16 = find_encoding("UTF-16BE");
sub needs_lines { 1 };
@@ -32,16 +29,15 @@
while (pos($str) < $len){
if ($str =~ /\G($re_asis+)/ogc){
$bytes .= $1;
+ }elsif($str =~ /\G&/ogc)
+ {
+ $bytes .= "&-";
}elsif($str =~ /\G($re_encoded+)/ogsc){
- if ($1 eq "&"){
- $bytes .= "&-";
- }else{
- my $s = $1;
+ my $s = $1;
my $base64 = encode_base64($e_utf16->encode($s), '');
$base64 =~ s/=+$//;
$base64 =~ s/\//,/g;
$bytes .= "&$base64-";
- }
}else{
die "This should not happen! (pos=" . pos($str) . ")";
}