codes_convert.c

用C语言与网页通迅的时候,难免会有进行编码的,此函数可以对GB2312转换成Unicode,然后再进行UTF8的页面编码,生成网页的％EB等格式

#include "codes_convert.h"

extern int errno;

char urls[] = "0123456789ABCDEF";

int geturlspos(char ch)
{
	int i;

	for(i=0; i<strlen(urls); i++)
	{
		if (urls[i] == ch)
			return i;
	}

	return 0;
}

/* normal to url */
/* ascii no deal, ' ' to '+', others to %XY, X is high, Y is lower  */
int urlencode(char *ib, int il, char *ob)
{
	int i=0;
	int j=0;
	
	for(i=0; i<il; i++, j++)
	{
		if (!isascii(ib[i]))
		{
			ob[j++] = '%';
			ob[j++] = urls[(ib[i] >> 4) & 0x0f]; 
			ob[j] = urls[ib[i] & 0x0f];
		}
		else
		{
			ob[j] = ib[i];
		}
	}

	ob[j] = '\0';

	return j;
}

/* url to normal */
/* ascii no deal, '+' to ' ', others to normal, %XY->X is high, Y is lower */
int urldecode(char *ib, int il, char *ob)
{
	int i=0;
	int j=0;

	for(i=0; i<il; i++)
	{
		if (ib[i] == '%')
		{
			ob[j] =	((geturlspos(ib[i+1]) << 4) & 0xf0) | (geturlspos(ib[i+2]) & 0x0f); 
			i += 2;
		}
		else
		{
			ob[j] = ib[i];
		}

		j++;
	}

	return j;
}

/* code convert */
/* fc: from chatset, tc: to chatset */
/* ib: inbuf, il: inlen, ob: outbuf, ol: outlen */
int code_convert(char *fc, char *tc, char *ib, int il, char *ob, int ol)
{
	iconv_t cd;

	/*
		iconv_t iconv_open(const char *tocode, const char *fromcode);
	*/

	cd = iconv_open(tc, fc);
	if (cd == (iconv_t)-1)
	{
		printf("sys alm: call function iconv_open error, errno[%d]\n", errno);
		return -1;
	}

	/*
	size_t iconv(iconv_t cd,
                     char **inbuf, size_t *inbytesleft,
                     char **outbuf, size_t *outbytesleft);
	*/

	if (iconv(cd, (char**)&ib, (size_t*)&il, (char**)&ob,(size_t*)&ol) == (size_t)-1)
	{
		printf("sys alm: call function iconv error, errno[%d]\n", errno);
		return -1;
	}
	
	/*
		int iconv_close(iconv_t cd);
	*/
	iconv_close(cd);

	return ol;
}

/* gb2312 to utf8 */
int gb2312_to_utf8(char *ib, int il, char *ob, int ol)
{
	return code_convert("GB2312", "UTF-8", ib, il, ob, ol);
}

/* utf8 to gb2312 */
int utf8_to_gb2312(char *ib, int il, char *ob, int ol)
{
	return code_convert("UTF-8", "GB2312", ib, il, ob, ol);
}