From: Markus Kuhn (
Date: Sun Apr 11 1999 - 07:08:10 EDT

Erik van der Poel <> wrote on 1999-04-10 21:09 UTC:
> I can see their point. If you use 0xFFFD for DEFAULT_CHAR, then you
> can't tell whether that glyph is being displayed because *you* didn't
> have a *glyph* or because *Unicode* didn't have a *character* for
> whatever the user is trying to display. 0x0000 seems more like a valid
> control char to me. How about using 0xFFFF ("guaranteed not to be a
> Unicode character at all") for X11's DEFAULT_CHAR?

I have tried to use 0xFFFF, but curiously this has caused a space
character to be displayed for unavailable glyphs instead of
DEFAULT_CHAR. I have therefore used 0xFFFE as the DEFAULT_CHAR value,
which is equally guaranteed not to be a valid Unicode character. The
glyph shape that I have decided on is a white rectangle of the size of
an "H" with a dotted or dashed border. This glyph is easily
distinguishable from any existing Unicode character and it symbolizes
(at least for me) very well that there is something missing here.

You can download the updated 6x13.bdf font now from

This font will in the next few days probably be added to the forthcoming
XFree86 4.0 release, so now is your last chance of testing and reviewing
it. There will hopefully also be some UTF-8 support in xterm in XFree86
4.0, such that you can immediately start using it.


Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at,  WWW: <>

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:45 EDT