diff man/custom.texi @ 51274:78a8be5d896c

(Init Rebinding): Xref Non-ASCII Rebinding, for non-English letters. Explain how to set coding systems correctly and how to include the right coding cookie in the file.

--- a/man/custom.texi	Tue May 27 23:30:10 2003 +0000
+++ b/man/custom.texi	Wed May 28 08:56:29 2003 +0000
@@ -1712,6 +1712,42 @@
 by listing each of the characters within the square brackets that
 delimit the vector.
 
+  Language and encoding settings can be a reason for failing key
+binding.  For instance, say you have put the following in your init
+file:
+
+@example
+(global-set-key [?\M-ö] 'shell)
+@end example
+
+@noindent
+Now it can happen that the ö character read from the keyboard is
+actually different from the ö character read from your init file, even
+though they look the same.  In this case, Emacs will not do what you
+intended.  First you should check whether they are the same or
+different.  In the above example, you would position point (the cursor)
+on the ö character and hit @kbd{C-u C-x =} (the @kbd{C-u} part is
+important).  The second line of the output will contain the charset.
+Now hit the ö key and position point on the character thus produced.
+Again, hit @kbd{C-u C-x =}.  If the two charsets are different, then you
+have found the problem.  The solution involves putting a `coding cookie'
+in your init file, @pxref{File Variables}.  For example, suppose that
+the ö character from the init file has charset latin-iso8859-1 whereas
+the ö character from the keyboard has charset latin-iso8859-15.  The
+solution is to put a coding cookie into the first line of the
+@file{.emacs} file, as follows:
+
+@example
+;; -*- coding: iso8859-15; -*-
+@end example
+
+@noindent
+@xref{Non-ASCII Rebinding}, if the charset printed by @kbd{C-u C-x =} is
+`eight-bit-graphic'.
+
+  If the variable @code{keyboard-coding-system} is nil, it is probably
+best to set it to the right value, according to the locale you work in.
+
 @node Function Keys
 @subsection Rebinding Function Keys