KY> Hi, KY> I'm using latest 21.2 CVS. I found a problem that the function KY> `looking-at' does not work for Japanese characters as follows. KY> (let ((oops (decode-coding-string "\e$B$&$C$W$9\e(B" 'iso-2022-jp))) KY> (with-temp-buffer KY> (insert oops) KY> (goto-char (point-min)) KY> (looking-at (regexp-quote oops)))) KY> => nil KY> What's happened? I can't write Japanese kanji text using egg-sj3 KY> because of this. I can confirm that this is broken in current CVS, but works in CVS as of Nov.9. Whoever introduced this, please fix. Likely suspects: 2000-11-14 Yoshiki Hayashi <yoshiki(a)xemacs.org&gt; * buffer.c: Remove if 0'ed entry. * buffer.h: Rewrite TRT to use char-table. * bufslots.h: Remove mirror tables. * casetab.c: Remove mirror tables. (CASE_TABLE_P): New macro. (Fcase_table_p): Element of a case table is string or char-table. (make_mirror_trt_table): Removed. (set_case_table): Setup char-table from strings for backward compatibility. * dired.c (Fdirectory_files): * dired-msw.c: (mswindows_get_files): * lisp.h: Change prototype of re_pattern_buffer. * regex.c: (RE_TRANSLATE): New macro. (TRANSLATE_P): Ditto. Change translate to type RE_TRANSLATE_TYPE. * regex.h: Define RE_TRANSLATE_TYPE * search.c (TRANSLATE): New macro. (TRANSLATE_ASCII): New macro. Translate table is changed to Lisp_Object. (signal_failure): Inhibit return. 2000-07-21 Jan Vroonhof <jan(a)xemacs.org&gt; * dumped-lisp.el (preloaded-file-list): Load code-cmds.el * code-cmds.el: New file * mule/mule-cmds.el: * mule/mule-misc.el (coding-keymap): New keymap. Define coding system keyboard commands on file-coding builds too. (coding-system-change-eol-conversion): (universal-coding-system-argument): (set-default-coding-systems): (prefer-coding-system): Moved from mule-cmds.el (set-buffer-process-coding-system): Moved from mule-misc.el

yoshiki> That's my fault. Two occurence of RE_TRANSLATE without yoshiki> range checking just slipped my eyes. This patch is a bit yoshiki> ugly, but it does the job until properly MULEized. Oh, my most favorite tool has come back. Thanks a lot! -- Katsumi Yamaoka <yamaoka(a)jpl.org&gt;