Unicode + Name Resolution

So I guess gethostbyname is deprecated, and I am trying to be all polite, and trying to make a unicode compatible application with the new name resolution technique of getaddrinfo + getnameinfo. I end up with a dotted IP address representing the host, as a Unicode string. I do not know how to convert this to an IN_ADDR structure, inet_addr only takes ANSI strings.

Code:

---

        ADDRINFOT * addr = { 0 };
        TCHAR *    host = _T("www.yahoo.com");
        TCHAR      hostBuffer[256];


        int ret = ::GetAddrInfo(host, 0, 0, &addr);
        if(ret)
        {
                ErrorMessage(_T("GetAddrInfo: %d"), ::WSAGetLastError());
        }
        if(addr)
        {
                ::GetNameInfo
                        (addr->ai_addr, sizeof sockaddr,
                        hostBuffer,    sizeof hostBuffer,
                        0, 0, NI_NUMERICHOST);

                // unsigned long ip = inet_addr(hostBuffer); //
                //      how how how how how how how how      //
        }
        ::FreeAddrInfo(addr);

---

check out wcstombs()

Kuphryn

Thanks. Quick question:

Quote:

---

Return Value

If wcstombs successfully converts the multibyte string, it returns the number of bytes written into the multibyte output string, excluding the terminating NULL (if any). If the mbstr argument is NULL, wcstombs returns the required size in bytes of the destination string. If wcstombs encounters a wide character it cannot convert to a multibyte character, it returns –1 cast to type size_t and sets errno to EILSEQ.

Remarks

The wcstombs function converts the wide-character string pointed to by wcstr to the corresponding multibyte characters and stores the results in the mbstr array. The count parameter indicates the maximum number of bytes that can be stored in the multibyte output string (that is, the size of mbstr). In general, it is not known how many bytes will be required when converting a wide-character string. Some wide characters will require only one byte in the output string; others require two. If there are two bytes in the multibyte output string for every wide character in the input string (including the wide character NULL), the result is guaranteed to fit.

---

I can't tell what they mean by that second statement. Is it okay for me to say

Code:

---

                int    len = wcstombs(NULL, hostBuffer, 0);
                char * mbs = new char[len];

---

And then call wcstombs again with the dest buffer.

size_t a = wcstombs(NULL, b, wcslen(b)) + 1

Kuphryn

Is there any point of me supporting unicode there if I'm just going to convert it back?

all api should be unicode and 32-bit api are legacy code

Kuphryn

GetAddrInfo returns an address. There should be no need to call GetNameInfo or inet_addr. If you just need a sockaddr it can be obtained with addr->ai_addr. This has the added advantage of being compatible with other address systems such as IPv6. If you really need an in_addr or even an unsigned long, you can obtain them too:

Code:

---

GetAddrInfo(host, NULL, NULL, &addr);

/* Make sure that we have an IPv4 address before treating it as such. */
if ( addr->ai_family  == PF_INET && 
    addr->ai_addrlen == sizeof(struct sockaddr_in) )
{
    /* You can get an in_addr: */
    struct in_addr ia = ((struct sockaddr_in*)(addr->ai_addr))->sin_addr;

    /* and an unsigned long: */
    unsigned long ip = ia.S_addr;
}

---

Quote:

---

Is there any point of me supporting unicode there if I'm just going to convert it back?

---

Generally, no. However, it can make your application easier to update down the track if its mostly unicode rather than zero unicode.