如何在C ++中使用libiconv将ISO-8859-1转换为UTF-8

Question

I'm using libcurl to fetch some HTML pages.

The HTML pages contain some character references like: סלקום

When I read this using libxml2 I'm getting: ׳₪׳¨׳˜׳ ׳¨

is it the ISO-8859-1 encoding?

If so, how do I convert it to UTF-8 to get the correct word.

Thanks

EDIT: I got the solution, MSalters was right, libxml2 does use UTF-8.

I added this to eclipse.ini

-Dfile.encoding=utf-8

and finally I got Hebrew characters on my Eclipse console. Thanks

Answer 1

Have you seen the libxml2 page on i18n ? It explains how libxml2 solves these problems.

You will get a ס from libxml2. However, you said that you get something like ׳₪׳¨׳˜׳ ׳¨ . Why do you think that you got that? You get an XMLchar* . How did you convert that pointer into the string above? Did you perhaps use a debugger? Does that debugger know how to render a XMLchar* ? My bet is that the XMLchar* is correct, but you used a debugger that cannot render the Unicode in a XMLchar*

To answer your last question, a XMLchar* is already UTF-8 and needs no further conversion.

Answer 2

No. Those entities correspond t the decimal value of the Unicode sequence number of your characters. See this page for example.

You can therefore store your Unicode values as int egers and use an algorithm to transform those integers to an UTF-8 multibyte character. See UTF-8 specification for this.

Answer 3

This answer was given in the assumpltion that the encoded text is returned as UTF-16, which as it turns out, isn't the case.

---

I would guess the encoding is UTF-16 or UCS2. Specify this as input for iconv. There might also be an endian issue, have a look here

---

The c-style way would be (no checking for clarity):

iconv_t ic = iconv_open("UCS-2", "UTF-8");
iconv(ic, myUCS2_Text, inputSize, myUTF8-Text, outputSize);
iconv_close(ic);