Re: [Q] NSCharacterSet difference between Cocoa and GNUstep?

[Frederic Stark]

Sungjin Chun wrote:

> Hi,
>
> During testing of XMLRPC framework, I've encountered character set 
> related problem. The test data is simple string of Korean and this works 
> well in Cocoa environment. But in GNUstep, this data, ie., korean string 
> is considered as illegalCharacterSet in NSCharacterSet. From Apple 
> documentation, Cocoa uses Unicode 3.2, but I do not know on GNUstep's case.
>
> Is this because GNUstep uses older data which does not have Korean 
> characterset?

We use korean characters with GNUstep/win32. I don't know any korean, 
but we use UTF-16 files with property lists containing korean 
characters, and everything is fine. This is on a US Windows and a French 
Windows, with korean installed. We had report of this also working on 
korean windows installation.

Maybe GNUstep does not include the whole korean character set, but it 
does include some.

We had some problem, thought:

* Some endianess problem (I think), because some part of windows did not 
support both FF FE and FE FF text files.
* Some UTF-8 problems. Our GNUstep was not built with proper UTF-8 support.
* Some end-line problems (maybe some files were not checked into cvs as 
binary, hence some extra 0d/0a were inserted, which is a /very/ bad idea 
in a unicode file).
* Some internal code problem. Some UTF8 strings were stored in normal 
(non-UTF strings). Under Mac OS X, everything looked fine, because 
Terminal.app is a UTF8 app (we did not realize that if you NSLog a 
string with the wrong encoding, Terminal.app would make it look correct)

Cheers,

--fred