Mike Fabian writes:

Friedrich Dimmling writes:

...

I tried to use mule-ucs with GNU Emacs from SuSE 8.0 (21.1.1) to prepare a LaTeX file containing Chinese (simplified, mule input mode chinese-py) characters together with German umlaut characters, using the ucs.sty.

Everything seems to work fine. However when I load the .tex file (containing the -*- coding utf-8 -*- line) a second time into GNU Emcacs, part of the characters is taken from the jiantizi font (rather bold face), part of the characters is taken from the fantizi font (rather slim chars) and some characters are just empty boxes, as if there is no available font for these characters. I have the ifntchia.rpm installed. A translation by LaTeX still gives correct results. Did I miss something?

mule-ucs mixes many fonts trying to cover as much of possible of Unicode. This can lead to ugly combinations and missing characters.

Can you send me a small file where this problem occurs? Then I can try whether it is possible to tweak the font setup of Emacs/mule-ucs to display it better.

I uploaded a new Mule-UCS package: ftp://ftp.suse.com/pub/people/mfabian/Mule-UCS-0.84-28.noarch.rpm This new package creates a few additional fontsets which use Unicode fonts for all CJK characters. This makes a more uniform, nicer display of mixed Chinese/Japanese texts possible, i.e. it should solve the problem described above. If you also install the efont-unicode package ftp://ftp.suse.com/pub/people/mfabian/8.0-noarch/efont-unicode.rpm you will have the following additional fontsets available after starting Emacs: "12_efont_unicode" "13_misc_unicode" "14_efont_unicode" "16_efont_unicode" "16_gnu_unicode" "18_misc_unicode" "24_efont_unicode" These fontsets can be selected with shift+left mouse button -> fontset or selected in ~/.emacs using something like: (setq default-frame-alist `((font . "fontset-16_efont_unicode"))) -- Mike Fabian http://www.suse.de/~mfabian 睡眠不足はいい仕事の敵だ。