[prev in list] [next in list] [prev in thread] [next in thread] 

List:       aspell-user
Subject:    [Aspell-user] Aspell, Emacs and UTF-8
From:       Marco De Vitis <starless () spin ! it>
Date:       2005-10-20 10:21:39
Message-ID: dj7r3i$9es$1 () sea ! gmane ! org
[Download RAW message or body]

Hello,
I have a problem trying to understand how utf-8 support works when using 
aspell from within emacs.

I've got two "fresh" Linux installations: Mandriva 2006 and Fedora Core 
4. Both are at their default settings. "locale" shows it_IT.UTF-8 for 
both of them.

I have an UTF-8 XML file, containing Italian words with accented chars, 
e.g. "società".

In MDV2006, when I do a "aspell -H -c file.xml" from console it works 
fine, accented words are correctly accepted. Instead, if I open file.xml 
with emacs (which correctly enters UTF-8 mode and correctly shows 
accented chars), and then do a M-x ispell-buffer, the spell process 
chokes on accented words; e.g. when it encounters "società" it only 
highlights "societ" and amongst the suggestions shows "societÃ", which 
is how that UTF-8 word appears when decoded using iso-8859-1.
I tried playing around with spell command arguments in emacs, to no avail.
What is happening?

A curious thing is that in Fedora Core 4, instead, the opposite happens.
When I open the file with emacs, it correctly enters UTF-8 mode and 
correctly shows accented chars, again, and when doing M-x ispell buffer 
everything works as it should. Instead, when doing "aspell -H -c 
file.xml" from console, aspell chokes on accented words; but this time 
it can easily be fixed by using the --encoding=utf-8 argument.

Here are the software versions shipped with the distributions:

Mandriva Linux 2006:
aspell    0.60.3
aspell-it 0.53.0
emacs     21.4.2

Fedora Core 4:
aspell    0.50.5
aspell-it 0.53
emacs     21.4.1

I also tried downloading the latest Italian aspell dictionary 
(aspell6-it-2.2_20050523-0.tar.bz2) and installing it on MDV2006, but 
nothing changed.

Any useful info is appreciated, thanks.

-- 
Ciao,
   Marco.




[prev in list] [next in list] [prev in thread] [next in thread]