Hello,
Just installed mule-enabled 21.5.9 on Solaris 8 and noticed
that no longer can input japanese via xim. Instead of valid
characters the garbage is input, that seems to be raw euc-jp
characters as set my locale. The comment in event-Xt.c suggest
using composed-input-coding-system variable to determine
coding system for convertion. After initializing this variable
to coding system for current locale in startup file,
everything seems to work perfectly. No side efects noticed
yet. Not sure how conceptual this change is, anyway patch follows.
(setq composed-input-coding-system
(get-coding-system-from-locale (current-locale)))
Beta 21.5.9 Solaris 8、euc-jp ですが XIMで 日本語入力して文
字化けになりました。"composed-input-coding-system"のヴァリアブ
ルでXIMの出力を変えて使えます。
93a94,95
Lisp_Object Vcomposed_input_coding_system;
1065c1067
< (XLSTREAM (fb_instream), Qundecided, CODING_DECODE, 0);
---
(XLSTREAM (fb_instream), get_coding_system_for_text_file
(Vcomposed_input_coding_system, 0), CODING_DECODE, 0);
3555a3558,3563
DEFVAR_LISP("composed-input-coding-system",&Vcomposed_input_coding_system
/*
Use specified coding system for decoding XIM output
*/ );
Vcomposed_input_coding_system = Qundecided;