Re: fallbacks for nonexisting combining characters

[Markus Kuhn <Markus.Kuhn@xxxxxxxxxxxx>]

Roozbeh Pournader wrote on 2001-01-11 17:20 UTC:
> > Please change xterm such that a combining character missing from the
> > font is simply not displayed. Then this will also save me from having to
> > add to the charcell fonts space characters for all the zero-width and
> > control characters that we want to handle in xterm just like combining
> > characters (as we discussed a few weeks ago).
> 
> But how can a user find that a character is missing?

The stupid up to 786 kbyte large XCharStruct arrays in the per_char
array of the corresponding XFontStruct (even for CharCell fonts!) must
be good for something after all ...

> What about handling the missing combining characters as non-combining?

This would make the cursor position (and thus wcwidth()) depend on the
font coverage. No, I don't like that. An essential aspect of my UTF-8
vision is that terminal applications can use wcswidth() very reliably
and conveniently to predict how the cursor will move when the terminal
receives some known string in UTF-8 mode. That's why we have a normal
and a wide default char and that is also why combining characters should
never move the cursor, not even when they appear not after a graphic
base character or are not available in a font.

Markus

-- 
Markus G. Kuhn, Computer Laboratory, University of Cambridge, UK
Email: mkuhn at acm.org,  WWW: <http://www.cl.cam.ac.uk/~mgk25/>

-
Linux-UTF8:   i18n of Linux on all levels
Archive:      http://mail.nl.linux.org/lists/