Re: encoding problem

[Bernt Hansen <bernt@norang.ca>]

Nick Dokos <nicholas.dokos@hp.com> writes:

> Bernt Hansen <bernt@norang.ca> wrote:
>
>> Eric S Fraga <e.fraga@ucl.ac.uk> writes:
>> 
>> > Bernt Hansen <bernt@norang.ca> writes:
>> >
>> >> Julien Cubizolles <j.cubizolles@free.fr> writes:
>> >>
>> >>> I'm having a very strange problem with character encoding. I write all
>> >>> my text files with emacs, with non-ascii characters (I'm french). I keep
>> >>> a copy of many files (latex/org/...) on separate machines using
>> >>> unison. Very often after a synchronization, the non-ascii charaters are
>> >>> completely displayed wrong (à for à, ç for ç) in the org files, but
>> >>> never in the latex files.
>> >>>
>> >>> I guess it's more an Emacs than org files but I can't see what's special
>> >>> in the org files that makes them more prone to such errors.
>> >>>
>> >>> Is there a way to *fix* easily these corruptions on a file, ie searching
>> >>> for all "weird" characters to replace ?
>> >>>
>> >>> How could I prevent this from happening again (checking/changing
>> >>> character encoding maybe ?)
>> >>>
>> >>> Thanks for your help,
>> >>>
>> >>> Julien.
>> >>
>> >> Hi Julien,
>> >>
>> >> I get prompts for encoding when saving/exporting (on Windows only) so I
>> >> put the following at the top of my org-files
>> >>
>> >> # -*- coding: utf-8 -*-
>> >>
>> >> which seems to fix the problem for me.  Maybe this will help?
>> >
>> > I used to have this problem and it was incredibly annoying.  I also
>> > started adding the line Bernt suggests but I kept forgetting for new
>> > files.  I finally solved this problem by adding the following lines to
>> > my emacs initialisation:
>> >
>> > #+begin_src emacs-lisp
>> > (prefer-coding-system 'utf-8)
>> > (set-charset-priority 'unicode)
>> > (setq default-process-coding-system '(utf-8-unix . utf-8-unix))
>> > #+end_src
>> >
>> > I couldn't tell you which of these matter or whether they are all
>> > necessary but I don't have these problems any longer so I haven't
>> > investigated any further!
>> 
>> Thanks Eric!
>> 
>> I'll try this and drop my mode line setting in each org file.  I still
>> encounter this when archiving for the first time to a new file -- since
>> I'm archive utf-8 content and the new target org file prompts for
>> encoding with my current setup.
>> 
>> Regards,
>> Bernt
>
> Isn't the setting of LANG used during initialization to set these things?
> I have LANG set to en_US.UTF-8  and new buffers are in utf-8-unix
> (except for mail composition buffers: they are in undecided-unix).
> I'm pretty sure I'm not mucking with coding systems anywhere in my emacs
> initialization otherwise.
>
> Nick


Maybe - I'm having this issue on Windows...

-Bernt