accents in file names

[Chuck Swiger]

On Feb 12, 2009, at 2:50 PM, Wojciech Puchar wrote:
>>> accented letter to my freebsd box, the accented letter simply  
>>> disappear.
>>
>> UFS supports 8-bit characters except for "/" and "\0", but you also  
>> need to run a terminal with UTF8 support and use a correct font to  
>> view such things.
>
> why? i use ISO-8859-2

You've answered "why" when you state that you set up a locale which  
supports ISO Latin-X charset.  If you are running in the default C/ 
POSIX locale, using the US-ASCII character set and a font that only  
knows about 7-bit ASCII glyphs, then you won't get accented characters.

> UFS doesn't deal with encoding at all, just store what you give

That's right, which means you need to use filenames encoded in UTF8  
rather than in arbitrary Unicode.  People in Asia tend to want UTF-16  
or UTF-32 encoding (although historical encodings like Big5, Shift- 
JIS, and now GB18030 for China are still rather popular, and those are  
multibyte encodings), and things like gcc's implementation of  
widechars or Python are standardizing on UTF-32.

Regards,
-- 
-Chuck