🔒 Locked

Unicode String

Started by blanky Sep 19, 2005 at 9:12 PM 10 replies 800+ views

Original Post

Sep 19, 2005 at 9:12 PM

I need to know how to enable Unicode support in the String STL class. I read it in a book that it supports Unicode, or WCHAR, but I haven't found anything on the topic, thanks.

SiCrane

Mod

473

SiCrane

Sep 19, 2005 at 9:16 PM

You may want to try using std::wstring, which is like std::string, except that it uses wchar_ts intead of chars. However, the standard doesn't guarantee that std::wstring will use a unicode encoding. Typical C++ library implementations use UCS-2 encodings which can be considered a Unicode subset.

Code-R

Sep 20, 2005 at 3:20 AM

std::basic_string is a template. you can support the latest 31bit unicode by doing:

typedef ustring std::basic_string<int32>;

:D:D

Fruny

203

Fruny

Sep 20, 2005 at 3:33 AM

Assuming you have a type named int32. There's int32_t in stdint.h, but that's C99, not C++ (yet).

zoggo

Sep 20, 2005 at 6:51 AM

Firstly, anyone who declares typedefs like that and gets away with it needs their compiler looking at...

typedef ustring std::basic_string<int32>;

should have the order of names reversed.

Secondly, the standard is a little sketchy about what wchar_t actually is:

Quote:
values can represent distince codes for all memebers of the largest extended character set specified among supported locales.

This has the unfortunate consequence of meaning your wide string use 2 bytes per character on Windows and 4 bytes per character on most *nix (including Mac OSX) systems.

SiCrane

Mod

473

SiCrane

Sep 20, 2005 at 7:33 AM

Quote:
Original post by Code-R
std::basic_string is a template. you can support the latest 31bit unicode by doing:
typedef ustring std::basic_string<int32>;
:D:D

That typedef does not guarantee a unicode encoding. Furthermore, typedefs for basic_string other than std::string or std::wstring are inherently difficult to use since you then have to define nearly all the functions necessary to use them. Ex: specialize std::ctype, create your own versions of tolower() and toupper(), etc. This means you'll need to essentially write your own Unicode support from scratch. The typedef by itself essentially gives you nothing.

njpaul

Sep 20, 2005 at 8:39 AM

The only time I had problems with unicode strings was when I was trying to write them to a file and I would get a space after every letter. I switched to the Windows way of writing files and it worked.