When I issue
cursortoxml(alias(), lcXmlFileName, 1, 48+512)
the xml I get has the proper utf-8 in the header line, but the file itself is not encoded in utf-8 at all. The codepage of the cursor (1250 at the moment, but there may be others in the future) seems to be irrelevant, it is exported as 1252.
So what, do I have to do a filetostr(), strconv() twice and add the BOM to get a proper utf-8 encoded xml file? I can do that, no problem, I just want to know if there's something in the cursortoxml() above that I forgot.