Re: Can't isearch 'ö'

[Peter Dyballa]

Am 10.04.2005 um 16:50 schrieb Stefan Monnier:

> > In a text buffer starting with
>
> >         ;;; -*- mode: Emacs-Lisp; coding: iso-8859-15; -*-
>
> > I have a line
>
> >             (message "konventionell-römisch")
>
> > which I can't find with C-s ö. It's always "Failing I-search: ö."

I think I got the reason for this behaviour:

 '(unify-8859-on-decoding-mode t nil (ucs-tables))


	unify-8859-on-decoding-mode is an interactive compiled Lisp function 
in `international/ucs-tables'.
        (unify-8859-on-decoding-mode &optional ARG)
        
        Set up translation-tables for unifying ISO 8859 characters on decoding.
        On decoding, i.e. input operations, non-ASCII characters from the
        built-in ISO 8859 charsets are unified by mapping them into the
        `iso-latin-1' and `mule-unicode-0100-24ff' charsets.
        
	Also sets `translation-table-for-input' globally, so that keyboard 
input
        produces unified characters.
        
        See also command `unify-8859-on-encoding-mode' and the user option
        `utf-fragment-on-decoding'.


Seems this customization makes keyboard input become iso-latin-1 and 
mule-unicode-0100-24ff, which are distinct from iso-latin-9 or 
iso-latin-2 or ... of the text buffers so that 'ö' and 'ö' aren't the 
same any more.

Changing the value in *Customization* buffer for the current session 
does not change the mode isearch works in, only relaunching GNU Emacs 
changes the behaviour and brings success.

--
Greetings

  Pete
                      ~  o
                       ~_\\_/\
                      ~  O   O