Input and Output Disciplines
There is currently no easy way to mark data read from a file or other external source as being utf8. This will be one of the major areas of focus in the near future.

So part of the problem may be this: You expect your query parameter is encoded in UTF-8 (I'm assuming), but your script just sees a sequence of extended-ASCII characters. You might be able to get around this by explicitly using pack "U",... to reconstruct UTF-8 characters from the input one at a time, but I don't recall if I ever got that technique to work reliably.

If you're just trying to ensure that an input string doesn't exceed a particular character length, you should be able to use length($string) to get its length in characters rather than bytes. That assumes that you already have it stored internally as UTF-8, of course, and that you haven't done a use bytes.

Unicode support in 5.8 is supposed to be much improved, but I haven't yet had a chance to try it for myself yet.

        $perlmonks{seattlejohn} = 'John Clyman';

Sorry, I guess I wasn't clear.

First, I'm using Perl 5.8

As for the input from the CGI, it's originally in EUC-JP, and then I change it to when I receive it from the browser to UTF-8. This is because we eventually shove it in XML format. I want to wrap THAT utf8 string at a certain column

<META NAME="Content-Type" CONTENT="text/HTML;CHARSET=utf-8">

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

There's a well-defined method for doing this in Unicode: UAX#14: Line Breaking Properties.

I don't know what can be wrong...

What about following code?

...
use utf8;
my $max_text_len = 40;
...

sub split_text {
  my $intext = shift || return '';
  my @result = ();
  my $index  = 0;

  foreach my $word (split(' ', $intext)) {
    next if !length($word);

    if (length("$result[$index] $word") > $max_text_len) {
      $index++;
      $result[$index] = $word;
    } else {
      $result[$index] .= ($result[$index] ? ' ' : '').$word;
    }
  }
  return join("\n", @result);
}
[download]

Dingus