> I cannot get emacs 23 to make the same decision as emacs 22 about > what charset the current char should be - but strangely enough, I > have a result which seems to work (attached, and also the emacs 22 > result from my system). It seems to make different font choices.
Thanks for your tests, but your results with emacs 23 are not yet working well; for example, it selects a CJK font for the `ü' character in German. > [...] emacs 23 seems to be switching somewhat freely between > > latin-iso8859-1 <->chinese-gb2312, > chinese-big5-1 <->chinese-gb2312 > japanese-jisx0208<->chinese-gb2312 > chinese-gb2312<->korean-ksc5601 > japanese-jisx0208<->korean-ksc5601 > cyrillic-iso8859-5<->korean-ksc5601 Well, yes. For input in Unicode, without additional information, emacs 23 has no other chance than to follow the user setup (in `.emacs'). However, for files in emacs-mule encoding, even emacs 23 correctly selects the fonts. In other words, if I load CJKbabel.tex , the file looks just fine (at least for me on my GNU/Linux box), and the `charset' property for each character gives the used charset. However, if I save the file in UTF-8 encoding (removing the `coding' tag at the end of the file before doing that), a reloading of the file no longer selects the right fonts. For example, Japanese CJK characters are represented with the SimSun font on my computer. > I vaguely understand the decision is somewhat arbitrary - and this > also affect the font choice, hence the Russian section and the Greek > section are obviously using different fonts in the emacs 23 version > vs the emacs 22 version. For emacs-mule encoding, I don't see a different behaviour. > I also tried your suggestion with getting at the charset text > property, but it seems that whenever I tried that the code get stuck > in a loop. Since this is definitely the way to go (this is, handling the legacy emacs-mule encoding within emacs 23) I'll ask on the emacs-devel mailing list if I find some time, which is currently not the case, unfortunately. But maybe you are faster than me :-) > Also I have a small confusion - most of the examples have emacs-mule > encoding (including the thai section in the CJKbabel example); the > thai example has tis620. How does cjk-enc.el works on both? cjk-enc works with all encodings which are charsets at the same time, including tis620. However, for Unicode, this is not the case: Within emacs, this is a virtual encoding, and emacs use a large number of different charsets to represent it, if necessary. Werner _______________________________________________ Cjk maillist - Cjk@ffii.org https://lists.ffii.org/mailman/listinfo/cjk