I've seen similar in the wild with XML, except the charset was wrong and escaping was incomplete. So you get the XML header with UTF8 encoding, a root tag, an

I've seen similar in the wild with XML, except the charset was wrong and escaping was incomplete. So you get the XML header with UTF8 encoding, a root tag, an indented child tag. 3 lines. Then ye olde CSV, with the 5 required entities for ascii escaped, but if you manage to put accented chars in it, they end up in ISO 8859-1. Then 2 lines footer. If you wrote out the file, the header and footer had to be byte for byte identical. Entities encoding non ascii just crashed the parser. Of course, the CSV lines were sensitive to indentation, they just dropped the first 8 chars of each line. Still, XML was officially supported.