Java: String UTF-8 encoding difference between local app and Google App Engine

I am trying to get a Google App Engine web app to send back UTF-8 encoded text to the browser. I do this, by writing this:

byte[] utf8Bytes = "æøå".getBytes("UTF-8");

When I do this locally, I get a byte array with 6 bytes back. When I do this on Google App Engine, I get an array with 12 bytes back. Weird, eh?

Does anyone know why?

I have succeeded in writing UTF-8 encoded text from GAE, by encoding the bytes myself, and write the raw bytes back. Like this:

output.write(new byte[]{(byte)0xc3, (byte)0xa5, (byte) 0xc3, (byte)0xa6, (byte)0xc3, (byte)0xb8 });

And this actually works. But, does anyone have an answer to why the String's are encoded differently on GAE, than they are locally?

Note: Encoding the characters via unicode escapes worked - like this:

byte[] utf8Bytes = "\u开发者_高级运维00E5\u00F8\u00E6".getBytes("UTF-8");

The bytes you are getting from GAE makes me assume that the source code file with the "æøå" literal is saved as UTF-8, but compiled with a compiler which is expecting the source files to be encoded as ISO-8859-1, ISO-8859-15 or Cp1252.

If you are building your source code with Ant or Maven, you have to specify the source file encoding in your build.xml or pom.xml.

Are you sure you have set the content-encoding in your HttpServletResponse before getting a Writer?