I'm converting a big whopping xml file to lots of little html files. One of the things I use are Zope's page templates. But one way or another I keep getting unicode errors when running the templates... Which is irritating as I only use unicode as input, not some mix of 8-bit and unicode (which is the standard recipe for disaster).

One tool that's really handy is a unicode character lookup database. This way you can at least check whether the unicode characters themselves are OK.