![](https://secure.gravatar.com/avatar/a0f6b595b0e3b7f3abba19eeb8261fc2.jpg?s=120&d=mm&r=g)
Hello Steffen, Yes, it seems we have forgotten your original problem in the discussion ;-) I'm not sure I can really help, though, but:
can you tell me then what is wrong in my minimal example here (it work like a charm in April!)?
Nothing is wrong with your input file, really, it's an (admittedly annoying) side-effect of changes that have been made recently in XeTeX-specific font selection options. I have the same problem here and, as frightening as XeTeX's messages may look, the output looks right. One thing: if you wish to typeset Arabic, you really have to say \font\Arab="Baghdad:script=arab", and not only \font\Arab="Baghdad", otherwise character shaping is not correctly done and the output is just messy. This behaviour needs to be fixed in ConTeXt but it's been around for a while now so it's really not that easy (fixing some things will break others, etc.). It's quite unfortunate that you had to experience all XeTeX+ConTeXt font problems at the same time, though, but judging from the other recent mails to the list I have the impression that you're not the only one today with weird problems and that everything seems to have gone astray ;-) I don't know, maybe it's just Weird Encoding Day; just this morning I was experimenting with yet another variation on ispell called hunspell, and I had the following extremely interesting crash which I would like to call the Locale-of-Death: [Current locale is fr_FR.UTF-8] $ hunspell -d morph Hunspell 1.2.1 [blah blah blah spell-checking ...] ^C $ unset LC_ALL $ hunspell -d morph zsh: bus error (core dumped) hunspell -d morph :-) Arthur