Who would have thought that Unicode support would be so difficult? I had no idea the battle that was to be fought with double encodings, utf flags, and finicky XML parsers.
I'm about to completely replace one of the major Grazr API modules with a new one that should be much better in terms of proper character encoding without sacrificing any other functionality (fingers crossed). The server shouldn't be down more than a minute.
There are still some feeds that don't properly report their encoding and in those cases there's very little Grazr can do right now. If Grazr encounters an encoding it's not familiar with, it will report an error (and an error will be logged on the server). As more 'unknown' encodings are found I'll add them to the repertoire of available encodings.
For now though, I hope my international users will begin to see the proper characters and that the random display glitches will be gone.
Update: Alas I spoke too soon. As soon as the feed is cached, the output get garbled again. Still working.
Update2: Ok I think it's fixed now, even with the cache. There are still some bad cached feeds working through the system, but in about an hour it should all be served from the new module.