Re: Wide character support enabling

Hi,

On Friday 17 February 2006 23:15, Leon Bottou wrote:
> On Friday 17 February 2006 03:09 pm, Raymond Martin wrote:
> > I have read that more recent implementations of wchar_t can easily map to unicode
> > and that the regular C string functions work with multi-byte characters now. Will
> > have to do some more experimenting on this.
> 
> Not so simple.
> 
> On Linux, the wchar_t type uses unicode encoding on 32 bits
> and does not depend on the locale. But this is not the 
> case under Solaris. On Windows, the wchar_t has only 16 bits. 
> Therefore unicode characters beyond 64K are coded on two words.

So a nice bunch of glue code is needed for portable string implementation.

I'm not familiar with issues on Solaris or Windows so I will have to concern 
myself with making things work on Linux before anything else. Lush's current 
string support is fine for me right now (Latin-1 chars for English & French is 
what I need for starters).

Maybe an interface to an external implementation would be better for specialized
text purposes, that does not interfere with or use the base string functions or
types.

One implementation is ICU (http://icu.sf.net), by IBM, that provides a full set of 
unicode functionality (with C/C++ & Java libraries). This is close to the definitive 
unicode offering and it basically forms the internationalization support in Java, 
with additional functionality.


Raymond



__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642