Different results of Junit Test in different IDEs

Question

import org.jdom2.Document;
import org.jdom2.input.SAXBuilder;
import java.io.FileReader;

public class Test1 {

    @org.junit.Test
    public void main() throws Exception {
        SAXBuilder sax = new SAXBuilder();
        Document doc = sax.build(new FileReader("resources/file.xml"));
        System.out.println(doc.getRootElement().getText());
    }
}

file.xml contains this: <root>©</root> encoding is UTF-8.

Used libs jdom2-2.06, hamcrest-core-1.3, junit-4.11.

When I run this in IntelliJ output is such: ©.

When I run this in NetBeans output is such: ©.

If I put the code to public static void main and run it - everything is ok.

If I change FileReader to FileInputStream - everything is ok.

If I change FileReader to StringReader("<root>©</root>") - everything is ok.

What can it be?

Answer 1

You are not specifying a character set when reading the file, so it uses the JVM default, afaik running from IntelliJ usually defaults to UTF-8, while Eclipse (at least on Windows) defaults to the default non-unicode character set (eg Cp1252 in Western Europe).

You will need to be explicit, as described in the documentation of FileReader:

The constructors of this class assume that the default character encoding and the default byte-buffer size are appropriate. To specify these values yourself, construct an InputStreamReader on a FileInputStream.

In other words:

new InputStreamReader(new FileInputStream("resources/file.xml"), StandardCharsets.UTF_8)

Or alternatively, let SAXBuilder handle this for you and just give it an InputStream. I believe - but I am not 100 % sure - this will determine character set by the XML declaration:

sax.build(new FileInputStream("resources/file.xml"))