Re: iconv message suppression

Il dom, 2003-11-30 alle 18:03, James K. Lowden ha scritto:
> On Sun, 30 Nov 2003, Alex Kiesel <kiesel@xxxxxxxxxx> wrote:
> > > It's not so good to use ucs2 for utf16 but it can work...
> > 
> > Is there any other way to convert the data?
> 
> I think Frediano is referring to the fact that UCS-2 and UTF-16 overlap
> but differ.  
> 
> http://www.cl.cam.ac.uk/~mgk25/unicode.html#ucsutf
> 
> If unichar data are actually encoded in UTF-16, they may theoretically
> include characters off the Basic Multilingual Plane.  Such characters are
> encoded in UTF-16 as two 16-bit "surrogate" characters, which obviously
> shouldn't be interpreted as two UCS-2 characters.  
> 
> The question is, can the server actually deliver/accept non-BMP
> characters?  If not, we can pretend we're dealing with UCS-2 data.  Else
> we have to do it right, to do it right.  ;-)
> 
> Regards, 
> 

Working on FreeTDS on Sunday too ??
I'm continually getting mail from samba-tng developing... well, I'm not
the only mad in the world :)

Another issue using ucs2le is... le! If client is big-endian it get
utf16be and trying to convert using ucs2le is not as good as client was
little-endian.

I finished my iconv test program and now my computer is crunching ...

$ ps awx | grep iconv_test
 3620 pts/0    R    356:40 ./iconv_test
15706 pts/1    S      0:00 grep iconv_test

freddy77