>>>>> "John" == John Spray <[EMAIL PROTECTED]> writes:
John> So it turns out the big character counts were actually -1 error John> return values being cast to a size_t. I've added an error check John> to the mbstowcs call that was generating them. It would be nice John> to handle all this encoding properly, but my plan is to cover my John> ears and hum while I wait for Unicode to happen. These "invalid" John> strings don't actually appear on the screen either, but I'm not John> sure why that is. OK. John> Okay to apply the attached? I'd rather not assert that the John> conversion should be valid, since that in fact isn't the case: John> the strings are fine, the gtk/xft code we've got is handling John> them wrongly I think. OK, as far as I am concerned. And I think the deal with gtk is that you are allowed to do what you want, since it is is just extra stuff (not really ready) for 1.4.0. But posting the patches is nice, of course. JMarc
