Weird character set problem

Robin Smidsrød

Okay, I'm trying to get my webapplication, which uses Locale::TextDomain to
display Norwegian characters, and it's turning out to be a real headache.

Now everything works, but all my national characters turn up on the webage as
question marks.

I use CGI.pm and mod_perl 2.0.1 (apache2), and initially in the application I
set $q->charset('UTF-8').

I generate this header:

Content-Type: text/html; charset=UTF-8

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">;
<html xmlns="http://www.w3.org/1999/xhtml"; xml:lang="nb_NO" lang="nb_NO">
<head>
<title>Grieg Multimedia - Startmappe</title>
<!-- Styles added in priority order -->
<link rel="stylesheet" type="text/css" href="/theme/style.css" />
<link rel="stylesheet" type="text/css" href="/theme/crystal/style.css" />
<link rel="shortcut icon" type="image/ico" href="/favicon.ico" />
<script language="javascript" type="text/javascript"
src="js/index.js"></script>
</head>
<body>

I use this code to initialize the locale-environment:

use POSIX qw(strftime :locale_h);
use Locale::TextDomain ('portfolio');
our $lang_default='nb_NO';
setlocale(LC_MESSAGES,$lang_default);

I've also tried with 'nb_NO.utf8', but it doesn't really make a difference.

The header in lib/LocaleData/nb/LC_MESSAGES/portfolio.po (ISO-8859-1 encoded
strings) is like this:

msgid ""
msgstr ""
"Plural-Forms: nplurals=2; plural=n != 1;\n"
"Project-Id-Version: portfolio\n"
"POT-Creation-Date: 2005-08-17 17:13+0200\n"
"PO-Revision-Date: 2005-08-31 16:24+0100\n"
"Last-Translator: Robin Smidsrød <>\n"
"Language-Team: Norwegian Bokmål <>\n"
"MIME-Version: 1.0\n"
"Content-Type: text/plain; charset=iso-8859-1\n"
"Content-Transfer-Encoding: 8bit\n"
"X-Poedit-Language: Norwegian Bokmal\n"
"X-Poedit-Country: NORWAY\n"
"X-Poedit-SourceCharset: iso-8859-1\n"

I've also added a UTF-8-encoded version (converted with iconv -f ISO-8859-1 t
UTF-8) of the same file in lib/LocaleData/nb_NO.utf8/LC_MESSAGES/portfolio.po:

msgid ""
msgstr ""
"Plural-Forms: nplurals=2; plural=n != 1;\n"
"Project-Id-Version: portfolio\n"
"POT-Creation-Date: 2005-08-17 17:13+0200\n"
"PO-Revision-Date: 2005-08-25 15:21+0100\n"
"Last-Translator: Robin SmidsrÃ¸d <>\n"
"Language-Team: Norwegian BokmÃ¥l <>\n"
"MIME-Version: 1.0\n"
"Content-Type: text/plain; charset=utf-8\n"
"Content-Transfer-Encoding: 8bit\n"
"X-Poedit-Language: Norwegian Bokmal\n"
"X-Poedit-Country: NORWAY\n"
"X-Poedit-SourceCharset: iso-8859-1\n"

Both of these files have been compiled with 'msgfmt -c -v -o portfolio.mo
portfolio.po' which succeds without errors.

I've tried changing the $q->charset() command to ISO-8859-1, but that doesn't
make difference to the output.

This my apache2 config:

       PerlSwitches -w -Mlib=/home/xxx/lib
       AddHandler perl-script .pl
       PerlOptions +ParseHeaders
       PerlOptions +SetupEnv
       PerlResponseHandler ModPerl::Registry
       PerlSetVar PORTFOLIO_BINDIR "/home/xxx/bin"
       # Automatic reload check of changed modules at each request, useful in
development
       PerlInitHandler Apache2::Reload

If you can pinpoint what I'm doing wrong I'd be a very happy man! I'm quite
certain it is something related to character conversion, and as far as I can
tell, it looks like my translations get converted to base US-ASCII for some
reason, because my æ (native Norwegian character) gets converted to ae in
output. This is just a hunch, though.

Please advice!

Regards,
Robin Smidsrød

libintl-perl