diff man/lispref/commands.texi @ 2862:b95fe16005fd

[xemacs-hg @ 2005-07-17 20:08:40 by aidan] Restore the last argument to event-to-character, document that it's unused.
author aidan
date Sun, 17 Jul 2005 20:08:48 +0000
parents a25c824ed558
children 6772ce4d982b
line wrap: on
line diff
--- a/man/lispref/commands.texi	Sat Jul 16 21:51:13 2005 +0000
+++ b/man/lispref/commands.texi	Sun Jul 17 20:08:48 2005 +0000
@@ -1485,7 +1485,7 @@
 information than the XEmacs internal character encoding can store.
 @end defun
 
-@defun event-to-character event &optional allow-extra-modifiers allow-meta
+@defun event-to-character event &optional allow-extra-modifiers allow-meta allow-no-ascii
 This function returns the closest character approximation to
 @var{event}.  If the event isn't a keypress, this returns @code{nil}.
 
@@ -1504,6 +1504,20 @@
 Specifying @var{allow-meta} will give ambiguous results---@key{M-x} and
 @key{oslash} will return the same thing, for example---so you should
 probably not use it.
+
+@var{allow-non-ascii} is ignored; in previous versions of XEmacs, it
+controlled whether one particular type of mapping between X11 keysyms
+and characters would take place.  The intention was that this flag could
+be clear and you could be sure that if you got a Latin-1 character with
+the high bit set back, you could assume that the lower seven bits of the
+character were the ASCII code of the character in question, and that the
+Meta key was pressed at the same time.  This didn't work in the general
+case, however, because it left the other type of X11 keysym-to-character
+mapping in place, ready to give you a Latin-1 character for a Latin-1
+key.  If you feel the need to use such a flag, sit back and think about
+abstracting your code, and if you still feel the need, bear in mind that
+it will be buggy in earlier versions of XEmacs.
+
 @end defun
 
 @defun events-to-keys events &optional no-mice