You may need to install additional fonts to display some international characters. Here are some sources for these fonts.
This package is best integrated with Emacs. Emacs has two FontSets by default; the intlfonts package has fonts for the
‘standard’ one. However, the latest free X distributions (XFree86 and X.org) now have more comprehensive sets.
You can also install these as packages from your favorite distribution. In Debian, for example, the fonts have been split into smaller packages. This is what it looks like in dselect, for example:
-- Opt x11 xfonts-intl- <none> 1.1-9 International fonts for X -- Asian. -- Opt x11 xfonts-intl- <none> 1.1-9 International fonts for X -- Chinese. -- Opt x11 xfonts-intl- <none> 1.1-9 International fonts for X -- Chinese big. -- Opt x11 xfonts-intl- <none> 1.1-9 International fonts for X -- European. *** Opt x11 xfonts-intl- 1.1-9 1.1-9 International fonts for X -- Japanese. *** Opt x11 xfonts-intl- 1.1-9 1.1-9 International fonts for X -- Japanese big. -- Opt x11 xfonts-intl- <none> 1.1-9 International fonts for X -- Phonetic Alphabet.
If you want to use the UnicodeEncoding, you will need appropriate fonts, too. The fonts must be for the ISO 10646-1 encoding. You can get the one from Markus Kuhn’s page. This page is a good introduction to the problems, too.
Note that many Unicode fonts are incomplete, so the language you are interested in may not yet be in the font. Check before you download.
In fact, you don’t necessarily need a 10646-encoded font to display unicoded text. Firstly, the text doesn’t have to be decoded into the mule-unicode charsets (e.g. the typical configuration of Mule-UCS). Second, you can map a character repertoire on to a font with a more-or-less arbitrary encoding which supports that repertoire—see
‘font-ccl-encoder-alist’. Here’s an example setup to use Latin-1 fonts for characters from other Emacs charsets which are unified with Latin-1:
(define-ccl-program encode-iso8859-1-font `(0 (translate-character ucs-8859-1-encode-table r0 r1) (if (r0 === ,(charset-id 'latin-iso8859-1)) (r1 = (r1 + 160))))) (map-char-table (lambda (k v) (if v (set-fontset-font t k '(nil . "iso8859-1")))) ucs-8859-1-encode-table) (add-to-list 'font-ccl-encoder-alist '("iso8859-1" . encode-iso8859-1-font))
Also, complex scripts encoded as utf-8 may require specialized fonts for the relevant glyphs such that you can’t display them with an iso10646-1 font. See the Indian support added after Emacs 21.3, for instance—the glyphs required to display the combining characters generally don’t correspond to Unicode characters.
Another attempt at creating a Unicode font. (This URL is probably obsolete. You can get an up-to-date version of unifont from Debian archives.)