Hi Mike, hi all others,

\begin{CJK*}[dnp]{JIS}{min}

\tableofcontents % <- OK here because it's within \begin{CJK} ... \end{CJK}

\part{1n$bLZ$+$iMn$A$k}

\end{CJK*} Ok. That should work.

But I choose the "dirty" (i. e. ugly) method: I have to edit my toc file by hand, because I want to use bold math symbols in the chapter titles, so I add a CJK-environment as well. So my "solution" was: I wrote \part[some silly stuff -- my prof is an idiot ;-) Die Theorie $\TPA$]{Beweistheorie \`a la\\Gentzen,\\\begin{CJK*}[dnp]{JIS}{min}C]Fb\ (Takeuti) und\\?70f\ (Arai)\end{CJK*}\\[2ex]\thispagestyle{empty} Die Theorie $\TPA$} and edited the produced toc file by hand: Instead of \contentsline {part}{\numberline {II}some silly stuuff -- my prof is an idiot ;-) Die Theorie $\textnormal {\textbf {T}}_{\textnormal {\textbf {PA}}}$}{55} in the toc-file I wrote \contentsline {part}{\numberline {II}Beweistheorie \`a la Gentzen, \begin{CJK*}[dnp]{JIS}{min}C]Fb\ (Takeuti) und ?70f\ (Arai)\end{CJK*}\\Die Theorie $\textnormal {\textbf {T}}_{\textnormal {\textbf {PA}}}$}{55} The output is really beautiful.... Thank you, Mike. Best regards, Ludger

ugroh@t-online.de (Ulrich Groh) さんは書きました:

Am Freitag, 14. Februar 2003 17:22 schrieb Mike FABIAN:

...

Ludger Sicking さんは書きました:

...

When I wrote

\begin{CJK*}[dnp]{JIS}{min} \part{Beweistheorie \`a la Gentzen, ...} \end{CJK*}

the problem is that the kanjis don't arise in the table of contents. Have I to use an extra package? Does CJK provide such tricky movements of arguments or should I have to edit the <filename>.toc by myself??

See the attached example file.

I tried the file too. But if I have

% -*- coding: euc-jp -*-

in the file, latex will run into errors (see attachment) . If I delete the first line with the coding information, everything works fine.

Using xemacs on the original file I can see the japanese Hiragana/Kanji .

What goes wrong?

Gnus/XEmacs again converted the encoding in my attachment when sending it. The attachement you saved isn't really EUC-JP encoded anymore, it is ISO-2022-JP encoded. To see that this is so, save the attachement again to disk (don't edit!) and execute the following in a UTF-8 capable terminal: mfabian@magellan:/tmp$ iconv -f euc-jp -t utf-8 < ludger-sicking.tex % -*- coding: euc-jp -*- \documentclass[12pt]{book} \usepackage{CJK} \begin{document} %\tableofcontents % <- *NOT* OK here because it's outside of \begin{CJK} ... \end{CJK} \begin{CJK*}[dnp]{JIS}{min} \tableofcontents % <- OK here because it's within \begin{CJK} ... \end{CJK} \part{猿も木から落ちる} \end{CJK*} \end{document} mfabian@magellan:/tmp$ You see garbage instead of the Japanese, because the file wasn't euc-jp encoded, therefore it didn't display in an UTF-8 capable terminal after converting to UTF-8. But if you convert from ISO-2022-JP to UTF-8, it works: mfabian@magellan:/tmp$ iconv -f iso-2022-jp -t utf-8 < ludger-sicking.tex % -*- coding: euc-jp -*- \documentclass[12pt]{book} \usepackage{CJK} \begin{document} %\tableofcontents % <- *NOT* OK here because it's outside of \begin{CJK} ... \end{CJK} \begin{CJK*}[dnp]{JIS}{min} \tableofcontents % <- OK here because it's within \begin{CJK} ... \end{CJK} \part{猿も木から落ちる} \end{CJK*} \end{document} mfabian@magellan:/tmp$ The Japanese is displayed correctly in the UTF-8 capable terminal, i.e. this was really ISO-2022-JP encoded and the conversion to UTF-8 therefore worked. CJK-LaTeX needs EUC-JP encoding when using \begin{CJK*}[dnp]{JIS}{min}. By removing the coding system cookie and saving the file, you have converted it to EUC-JP again. Emacs opened the file in EUC-JP mode, because of the coding system cookie. You can also see that in the modeline at the lower left. Nevertheless Emacs detected the ISO-2022-JP encoding automatically and displayed it correctly. When you removed the coding system cookie and saved the changed file to disk, it was written back to disk in EUC-JP encoding again. The file was originally EUC-JP encoded on my disk, but it was unfortunately automatically converted to ISO-2022-JP when I did sent the mail. I had the same problem recently with the UTF-8 encoded korean-Baekmuk-Gulim-UniKS-UTF8-H.ps file which I attached to a mail to ghugh Song , see also http://lists.suse.com/archive/m17n/2003-Feb/0009.html This one was converted from UTF-8 to EUC-KR by my mailer. In both cases I did sent the attachments as MIME type "text/plain" because type "application/octet-stream" (which means binary) is removed by the mailing list software running this list for security reasons (binary attachments might contain a virus). But when I send "text/plain" attachments, my mail program treats them the same way as normal text which I type into the body of the mail. I.e. Gnus/XEmacs automagically checks the content and try to convert to a "reasonable" encoding. For example, the mail I am currently typing contains a bit of text which was originally UTF-8 encoded (the Japanese in the output of the "iconv -f iso-2022-jp -t utf-8 < ludger-sicking.tex" line). When I pasted this in to XEmacs, it was converted to the XEmacs internal mule encoding, when sending the mail, it will be converted to ISO-2022-JP because the whole mail contains only ASCII and Japanese, therefore sending it as UTF-8 is not necessary. If I would type only a single German Umlaut into this mail, the whole mail would be converted to UTF-8. For a "normal" mail body, this behaviour is OK, it avoids UTF-8 unless really necessary, which ensures that Japanese and Korean remains readable even for users which old mail programs which don't handle UTF-8. I would like to send all my mails in UTF-8, but unfortunately many users are still using old mail software, especially most of the web based mail systems still can't handle UTF-8. But for attachments this automatic conversion is *very* annoying. I included these examples as attachments because I wanted to avoid that automatic conversion, but unfortunately this conversion is still done for text/plain attachments. And application/octet-stream attachments are not allowed. I don't know what to do about this, probably I can only save example files somewhere and give an URL where to download them. -- Mike Fabian http://www.suse.de/~mfabian 睡眠不足はいい仕事の敵だ。

Am Samstag, 15. Februar 2003 12:43 schrieben Sie:

ugroh@t-online.de (Ulrich Groh) さんは書きました:

To see that this is so, save the attachement again to disk (don't edit!) and execute the following in a UTF-8 capable terminal:

Maybe (sure) a stupid question: But how to start a UTF-8 capable terminal? -- Mit freundlichen Gruessen / Best regards Ulrich Groh

Am Samstag, 15. Februar 2003 12:43 schrieb Mike FABIAN:

ugroh@t-online.de (Ulrich Groh) さんは書きました:

mfabian@magellan:/tmp$ iconv -f euc-jp -t utf-8 < ludger-sicking.tex

With this I got garbage.

But if you convert from ISO-2022-JP to UTF-8, it works:

True . But still it don't work with latex. -- Mit freundlichen Gruessen / Best regards Ulrich Groh

Am Samstag, 15. Februar 2003 17:13 schrieben Sie:

ugroh@t-online.de (Ulrich Groh) さんは書きました:

...

Am Samstag, 15. Februar 2003 12:43 schrieb Mike FABIAN:

...

ugroh@t-online.de (Ulrich Groh) さんは書きました:

mfabian@magellan:/tmp$ iconv -f euc-jp -t utf-8 < ludger-sicking.tex

With this I got garbage.

...

But if you convert from ISO-2022-JP to UTF-8, it works:

True .

But still it don't work with latex.

Of course this file doesn't work with latex in UTF-8 encoding. If you want to use UTF-8 encoding, you have to change the \begin{CJK}... statement. The file was intended to be EUC-JP encoded, but it was converted to ISO-2022-JP by the mail software. If you convert it back to EUC-JP encoding, it works with latex.

No I got it. It is working!!! Thanks a lot for helping me to understand the system. -- Mit freundlichen Gruessen / Best regards Ulrich Groh