The code below behaves correctly in Windows XP with ORO version 2.0.6. The substitute method returns 0 and the string does not change (the e with the accent in the input string does not match the e with the backward accent in the substitution pattern.)
The same code in Solaris with 2.0.6 does perform one substitution and the resulting string is "abcdf". Is there a system library or something I need to upgrade that prevents ORO from handling the special characters? Or is there a gap somewhere in my understanding? If I load the java source code in vim, I do see the two e's correctly, with their respective accents. Thanks in advance for any help. I've browsed the archive and haven't seen this discussed. Todd import org.apache.oro.text.perl.Perl5Util; public class perltest { public static void main(String[] args) { Perl5Util util = new Perl5Util(); String input = "abcdéf"; StringBuffer result = new StringBuffer(); int numSubs = util.substitute(result, "s#è##g", input); System.out.println("num: " + numSubs); // returns 0 in windows, 1 in solaris System.out.println(input); System.out.println(result); // returns the input string in windows, "abcdf" in solaris } } -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>