CODEWITHSUNDEEP
clibxml2xmltextwriter
hank
hank

Reputation: 9853

Writing cyrillic characters into an xml file

I'm using xmlTextWriter from the libxml2 to write some xml files. And I need to write cyrillic characters into them. I do it this way:

xmlTextWriterStartDocument(writer, NULL, "utf-8", NULL);
...
snprintf(buf, sizeof(buf), "%s", "тест");
xmlTextWriterWriteAttribute(writer, 
                            (const xmlChar*)"test_attribute", 
                            (const xmlChar*)buf);

But when I open the resulting xml file I see html representation of my text, just like this: test_attribute="тест"

How can I fix this?

Upvotes: 1

Views: 3376

Answers (1)

Viktor Latypov
Viktor Latypov

Reputation: 14467

You need to use the separate utf-8 encoder.

In snprintf() your text is in CP-1251 (single-byte ASCII-era encoding), not in UTF-8 (variable-width encoding).

See this link for the reference implementation: http://7maze.ru/node/29

The comments are in russian, but all you need is a conversion table and the

string convertToUtf8(const char* chars, int len)

function at the end.

The "тест" string you used should look like "РўРчС_С'" (absolutely meaningless) while encoded.

An old C code from one old project. It uses the CP-866 encoding (another "popular" encoding from the MS-DOS), but the conversion from CP-1251 is straightforward.

/// CP866 to UTF-8
char *dosstrtou(char *buffer,const char *dosstr)
{
    char *buf1=buffer;
    while (*dosstr)
    {
        if ( (*dosstr>127)&&(*dosstr<176) )
        {
            *buf1=208;
            buf1++;
            *buf1 = (char)(*dosstr+16);
            dosstr++;
            buf1++;
            continue;
        }       
        if ( (*dosstr>223)&&(*dosstr<240) )
        {
            *buf1=209;
            buf1++;
            *buf1 = (char)(*dosstr-96);
            dosstr++;
            buf1++;
            continue;
        }       
        if (*dosstr==240)
        {
            *buf1=208;
            buf1++;
            *buf1=129;
            dosstr++;
            buf1++;
            continue;
        }       
        if (*dosstr==241)
        {
            *buf1=209;
            buf1++;
            *buf1=145;
            dosstr++;
            buf1++;
        }
        *buf1=*dosstr;
        buf1++;
        dosstr++;
    }
    *buf1='\0';
    return (buffer);
}

/// CP1251 to CP866
char *winstrtodos(char *buffer){
    char *ptr=buffer;
    while (*ptr!='\0')
    {
        if ( (*ptr>=0x80+0x40)&&(*ptr<=0xAF+0x40) )
            *ptr =(char)(*ptr-0x40);
        if ( (*ptr>=0xE0+0x10)&&(*ptr<=0xEF+0x10) )
            *ptr = (char)(*ptr-0x10);
        if (*ptr==0xA8) *ptr=0xF0;  
        if (*ptr==0xB8) *ptr=0xF1;
        ptr++;
    }
    return (buffer);
}

Just be careful with the memory.

Upvotes: 2

Related Questions