This blog uses Unicode (UTF-8). Mostly this occurs when I post in Swedish.
After a lot of work getting the email interface to support UTF-8 I was pretty happy. My blog setup was more or less Linux from end to end — I used Putty to shell into the Linux server hosting the blog and used Emacs to write the posts. All the characters were correct, both in Emacs and published on the web.
However, I moved my shell account to a FreeBSD box, and suddenly
everything’s strange. I’m running screen with -U option, but when I
use Emacs now I either get the infamous question-mark-in-a-diamond or
a real Swedish character which unfortunately consists of multi-byte
chars which mess up the display. The only recourse is to use Emacs
under X on the Linux box at home to manually change all the weird
chars to HTML-entities, which sucks.
I don’t know which subsystem is to blame — FreeBSD, the Emacs running on it, or the screen that Emacs is running under. It seems to work OK in Putty, but I need to confirm that.
Damn, I thought I had this crap figured out…