Hi, At Mon, 11 Dec 2000 01:37:00 +0900 (JST), [EMAIL PROTECTED] wrote: > > wen> When I upgrade potato to woody, everything seems ok except for > wen> xemacs21-mule-canna-wnn. > > After I changed LANG=zh_CN.GB2312 and LC_ALL=zh_CN.GB2312 to > LANG=ja_JP.ujis and LC_ALL=ja_JP.ujis, the installation of > xemacs21-mule-canna-wnn was passed. Probably we should consider it > as a bug.
Yes, this is the Bug#63805. This version of XEmacs define language information for chinese as "Chinese-GB" and "Chinese-BIG5" (in xemacs21-21.1.12/lisp/mule/chinese.el). However, it will try to set language information for chinese as "Chinese" (in xemacs21-21.1.12/lisp/mule/mule-init.el), which is not defined anywhere, so it errors. I think it could be fixed by changing auto-language-alist (in xemacs21-21.1.12/lisp/mule/mule-init.el) as follows: (defvar auto-language-alist '(("^ja" . "Japanese") ("^zh_.*.GB.*" . "Chinese-GB") ("^zh_.*.BIG5.*" . "Chinese-BIG5") ("^ko" . "Korean")) "Alist of LANG patterns vs. corresponding language environment. Each element looks like (REGEXP . LANGUAGE-ENVIRONMENT). It the value of the environment variable LANG matches the regexp REGEXP, then `set-language-environment' is called with LANGUAGE-ENVIRONMENT.") With this defvar, xemacs will try (set-language-environment "Chinese-GB") (if LANG=zh_.*.GB.*) or (set-language-environment "Chinese-BIG5") (if LANG=zh_.*.BIG5.*) in lisp/mule/mule-init.el Regards, Fumitoshi UKAI