[KevinH]

Hi Nick,

Okay, I think xmlescape and HTMLparser both work better with full unicode strings. At that point, all metadata has already been encoded as utf-8, so I have modified mobi_opf.py to convert all required pieces from utf-8 to full unicode, pass through the xmlescape and escape methods, and then convert back to the needed utf-8 for the opf file.

So please give this mobi_opf.py version a try and let me know if it fixes your issues.

Thanks,

Kevin