Import and extend GNU's descr-text.el, supporting prefix argument for C-x = 2008-05-25 Aidan Kehoe <kehoea@parhasard.net> * descr-text.el: New. Taken from GNU's GPLV2 version of 2007-02-14, with modifications for XEmacs support and extensions for Unihan.txt support and db/dbm caches. * simple.el (what-cursor-position): Support an optional prefix argument, as does GNU, calling #'describe-char to giving more detail on the character at point, notably from UnicodeData and (in our case, optionally) Unihan.txt. * syntax.el (syntax-after): Make this available for the sake of #'describe-char. * mule/mule-cmds.el (iso-2022-control-alist): Make this available, for the sake of #'encoded-string-description and #'describe-char. * mule/mule-cmds.el (encoded-string-description): Make this available, for the sake of #'describe-char. * unicode.el (unicode-error-default-translation-table): Make this a char table of type generic, not of type char. Makes it possible to have the relevant logic in #'describe-char reasonably clear; also, and this is undocumented, makes it much easier to implement #'frob-unicode-errors-region. I should document this, and revise #'frob-unicode-errors-region.

--- a/lisp/unicode.el	Wed May 21 21:09:20 2008 +0200
+++ b/lisp/unicode.el	Sun May 25 21:11:35 2008 +0200
@@ -506,7 +506,7 @@
 ;; point). Make them available to user code. 
 (defvar unicode-error-default-translation-table
   (loop 
-    with char-table = (make-char-table 'char)
+    with char-table = (make-char-table 'generic)
     for i from ?\x00 to ?\xFF
     initially (unless (featurep 'mule) (return))
     do