orrence has asked for the wisdom of the Perl Monks concerning the following question:

Hello all,

I am using a module which parses a text file with german umlaut characters (äöü etc.). In this module the 'use locale'-pragma is used and everything works fine executing my script inside an xterm. Problems arise when I want to do the same thing in a CGI-Script (apache-1.3-server).

The script fails because the geman umlaut-characters (äöü etc.) are not part of \w in the CGI-environment, causing some regexes to fail, although I set

$ENV{LANG}='de_DE@euro'
$ENV{LC_CTYPE}='de_DE@euro'
$ENV{LC_COLLATE}='POSIX'

just like it is set in bash-environment. According to documentation of 'use locale' in the camel-book it uses the values in LC_CTYPE for regex-comparisons and in LC_COLLATE for string comparisons.

What can I do to tell apache respective CGI that I am in a german environment and want special german characters to be interpreted as alphanumeric characters?

Thanks,
Daniel.

P.S.: using perl 5.8.0 on a Linux box (SuSe 8.2, kernel 2.4.20)

Replies are listed 'Best First'.
Re: locale problems with cgi and apache
by Roger (Parson) on Nov 05, 2003 at 13:38 UTC
    Perhaps you should consider using unicode strings. See utf8 on www.cpan.org for more info, or simply google with 'perl, utf, locale, german' and you will get hundreds of hits.

[reply]
Re: locale problems with cgi and apache
by chromatic (Archbishop) on Nov 06, 2003 at 07:01 UTC

    When and where do you set these variables? I assume they must be set before locale loads. If you're not setting them in the server's parent process or in httpd.conf before running your program, you might try:

    BEGIN
{
    $ENV{LANG}       = 'de_DE@euro';
    $ENV{LC_CTYPE}   = 'de_DE@euro';
    $ENV{LC_COLLATE} = 'POSIX';
}

use locale;

    [download]
[reply]
[d/l]
[select]
      Thank you for thinking about my problem. I found a solution using the module POSIX and its function setlocale like this:

      &setlocale( LC_CTYPE, 'de_DE@euro');

      Now it works the way I expect it to ...

      Daniel.
[reply]

Back to Seekers of Perl Wisdom