View Single Post
  #2   (View Single Post)  
Old 6th May 2008
lvlamb's Avatar
lvlamb lvlamb is offline
Real Name: Louis V. Lambrecht
Spam Deminer
 
Join Date: May 2008
Location: .be
Posts: 221
Default

There are several items to check.
First, ISO8859-15 should be your locale for accented characters (and the Euro sign).

Second, UTF-8, although somewhat compatible will have to be told you use fr_FR* locale.

Third, your system locale can be superseeded by you applications setups: for instance, I can use a system en_US.8859-1 locale, but have my xterm set to UTF-8 (uxterm) or emacs use whatever locale, or gnome set to us UTF-8 by default.
I can be on en_US on my terminal, but see Nautilus save my files as UTF-8.

Forth: Perl is going UTF-8 default. Currently there is an utf8 pragma (man utf8) allowing/disabling UTF-8.

Fifth: when you lauch an UTF-8 capable application, that application might call a subshell which will default run on your systems locale, which can be -as in my case- en_US. This sub-shell could eventually bork out with an error message as "malformed UTF-8 character string".

Most of the errors will be spitted out by bad written applications when those applications GNUishly presume your only possible setup is UTF-8 (if not UNICODE which is yet another non-standard standard).

My 2cents: check both your applications and DM (gnome, KDE) for proper encoding a common locale output.
Check the proper encoding choice when saving files.

Portable mails always should use 8859-1 text format.
You can read 8859-1 mails or webpages correctly, including diacritics from your most visited links by settin UTF-8 on your local browser setting. Local means that people using other locales will not read those pages the same way you do.

Fwiw, when switching from one code page to the other, include a setenv or export ENV or just change LC_* in your routine.
Remember that when you re-read the output, the output screen should also be set on the proper environment/locale. For instance, re-read your UTF-8 formatted mail with Thunderbird set to use UTF-8 also.
Note that UTF-8 will not be read correctly by 8859-15 set applications.
sécurité, is what you get.
__________________
da more I know I know I know nuttin'
Reply With Quote