Re: Unicode support

[Pawel Kot]

On Thu, 26 Sep 2002, Hu Gang wrote:

> 00_fix_bug    :  fix current cvs bug

Good spotting. Thank you. Applied.

> 01_fix_unicode:  support for unicode. passed in my test.

Notes and questions:
-       if (mbtowc(&retval, &value, 1) == -1) return '?';
+       if (mbtowc(&retval, &value, 1) == -1) return  0;

Why that? It will truncate the string on the character that is not
supported by the current locale.

It doesn't look good:
+               if (tmp != 0)
+                       dest[i - 1] = tmp; /* is right ? */
and
+               if (tmp != 0) dest[i - 1] = tmp;
It will crash on i == 0;

I would propose something like this:
static unsigned int char_decode_uni_alphabet(wchar_t value, unsigned char *dest)
{
        unsigned int retval;

        switch (retval = wctomb(dest, value)) {
        case -1:
                *dest = '?';
                return 1;
        default:
                return retval;
        }
}
[...]
void char_decode_ucs2(unsigned char* dest, const unsigned char* src, int len)
{
[...]
        i += (char_decode_unicode(strtol(buf, NULL, 16), dest + i) - 1);
[...]
void char_decode_unicode(unsigned char* dest, const unsigned char* src, int len)
{
[...]
        i += (char_decode_unicode(wc, dest + i) - 1);
[...]

It is completly untested, it's rather an idea how to do it.

pkot
PS. I guess encoding neither works with chineses characters. Does it?
-- 
mailto:address@hidden :: mailto:address@hidden
http://kt.linuxnews.pl/ :: Kernel Traffic po polsku