Am Mittwoch, 30. Januar 2013 schrieb Ludovic Courtès:
> The attached patch removes that UTF-8 assumption. Can you test whether
> it fixes the problem?
Yes, please commit!
Andreas