Re: Problem with national characters in XHTML

[Lennart Borgman]

Piet van Oostrum wrote:

> > > > > > address@hidden (Tomas Zerolo) (TZ) wrote:
> > > > > >            
>
>  
>
> > TZ> Ah. You have to distinguish between Emacs's internal representation
> > TZ> (that's possibly the 2276 you mention), which doesn't change (al least
> > TZ> unless you try hard ;) and what is in the file (how Emacs writes or
> > TZ> interprets what it reads). You can change those things changing the
> > TZ> coding system (look for something like `multilingual environment').
> >    
>
> By default Emacs uses different internal representations for the "same"
> character in different coding systems. So a iso-8859-1 "ä" is a different
> thing than a utf-8 "ä". This difference will disappear when Emacs switches
> to Unicode internally. For the time being the OP could use Unicode
> unification, if his Emacs version is young enough. I have used this for
> some years now without any problems. Maybe it solves the original problem.
>
> (require 'ucs-tables)
> (unify-8859-on-encoding-mode 1)
> (unify-8859-on-decoding-mode 1)
>  
The values I have I have in CVS emacs.exe -Q is

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

Though I do not understand what it means right now ;-)

Evaling (unify-8859-on-decoding-mode 1) does not change the behaviour of 
C-q 3 4 4 RET. It still enters a character that (following-char) reports as

 2276 (04344, 0x8e4)

I did not notice before that there only seem to be on bit that differs 
(see the second figure) - if that in some way matters.