Monks, maybe I am lame for calling perl from dos batch files. But I am used to dos, and use batches for testing and other stuff. (Winxp German with Active State Perl 5.8.)

Anyway, I have a situation where if I call a perl script from a dos batch file, I get one output (incorrect). If I call the same script directly from the command line, I get another output (correct). The problem is with the command line options, parsed with getopt::long, with german characters (äöüß). So, for example

perl germanCharProblems.pl -opt wordae="ärgerlich" -opt wordoe="börse"
+ -opt wordue="übersicht"
[download]

causes a problem when called from germanCharProblems.bat, which just contains the above line. But, it is okay straight from the command line.

I don't want to give up my ability to keep stuff in batch files, so I wrote a workaround function -- strange_dos_chars_to_germanchars -- to fix stuff passed in on the options when run from batch. Obviously, this is not a good solution. For one thing, it changes _ to ü if I have an underscore in an argument, obviously not a good thing.

I have a vague idea this *might* have something to do with the concept of perl "locale" but every time I attempt to look at the documentation on this my eyes glaze over. And it might *not* be the locale. Maybe it's activestate perl. Maybe if I tried this on cygwin it would work. Or maybe it's a windows setting?

Here's the script, to make the problem easy to isolate and reproduce:

#!perl
use strict;
use warnings; 

#cpan  modules.
use Getopt::Long; # Option handling

###################
# GLOBAL VARS
###################

# Our options
my %options;

# Grab the options from the command line
GetOptions ("opt=s" => \%options);

wipe_log();

for (keys %options) {
    logMessage('$options{' . "$_" . '}' . ": $options{$_} \n" );
}
#outputs: 
#$options{wordoe}: b÷rse 
#$options{wordue}: ³bersicht 
#$options{wordae}: õrgerlich 

for (keys %options) {
    $options{$_} = strange_dos_chars_to_germanchars($options{$_});
    logMessage('$options{' . "$_" . '}' . ": $options{$_} \n" );
}
#outputs:
#$options{wordoe}: börse 
#$options{wordue}: übersicht 
#$options{wordae}: ärgerlich 

# Takes a variable and spits it back out with the proper german charac
+ters
sub strange_dos_chars_to_germanchars {
    my $var = shift;
    my %table = (    'ß' => '¯' , 
                    'ä' => 'õ' , 
                    'ö' => '÷' , 
                    'ü' => '³' , 
                    'Ä' => '-' , 
                    'Ö' => 'Í' , 
                    'Ü' => '_');
    while (my ($k,$v) = each %table) {
        $var =~ s/$v/$k/g;
    }
    return $var;
}

sub wipe_log {
    my $logFileName = "log.txt";
    open F, "> $logFileName" or die "Cannot open log file: $logFileNam
+e";
    close F;
}

sub logMessage {
    my $message = shift;
    my $message = "$message\n";
    print $message; # maybe could turn this on and off with command li
+ne option.  do later?

    my $logFileName = "log.txt";
    open F, ">> $logFileName" or die "Cannot open log file: $logFileNa
+me";
    print F $message,"\n";
    close F;    
}
[download]

I would really, really appreciate anyone who could shed some light on this. Thanks!

thomas.