Comment 1 for bug 855418

When reading the file stream it uses Encoding.Default. Ideally the code would be smart enough to work out that it was a unicode file that it was reading and change the character encoding used but I think a more practical way of doing this would be to have an extra attribute in the database.xml file e.g. encoding="Encoding.Unicode" that would indicate which encoding to use when reading. This would default to Encoding.Default if no attribute was present.