Write multibyte character string under Unicode mode

[kshea]

Hi, All:

I'm wondering if it is possible to write multibyte character strings to a file in project with build configuration of Unicode.

I have tried to write a few lines into a text file, but there exists extra bytes in the end of the string.

Any advice will be greatly appreciated.

Kay

 

[zakharin]

I'm assuming you are using the fstream classes in the STL here.

You need to be using a locale that supports multibyte strings. Alberto Barbati has submitted libraries to the Boost project to do this. In his words, it "can read and write UTF-8, UTF-16 and UTF-32 with all endianness variants." You can get it meanwhile at

http://groups.yahoo.com/group/boost/files/utf/

I use his older library which "has only UTF-8 and UCS-2 (which is similar to UTF-16 but lacks proper handling of surrogates and non-characters)." This is available at

http://groups.yahoo.com/group/boost/files/codecvt.zip.

You'll find that there are a few modifications that need to be made to get it to work with VC++, but it is short and can be included with any project. To imbue your fstream class with a UCS2 locale, you can include the following before opening the file:

filename.rdbuf()->pubimbue(_ADDFAC(locale(), new ucs2_conversion()));

where filename is an fstream derived class.

 

[Zyrenthian]

First you need functionality to write the string as a char*. The old school method of ifstream and ofstream use this so this is a possiblity. Then you need to convert the string from wide char to single byte. For this you can

#inlcude <atlconv.h>


void write(LPCTSTR str)
{
USES_CONVERSION;
std:fstream outputFile(&quot;output.txt&quot;
outputFile<<T2A(str)<<endl;

}


Matt