uri.c

xml开源解析代码.版本为libxml2-2.6.29,可支持GB3212.网络消息发送XML时很有用.

        ret = xmlStrcat(ret, BAD_CAST ":");
        xmlFree(segment);
    }

    if (uri->authority) {
        segment =
            xmlURIEscapeStr(BAD_CAST uri->authority, BAD_CAST "/?;:@");
        NULLCHK(segment)
        ret = xmlStrcat(ret, BAD_CAST "//");
        ret = xmlStrcat(ret, segment);
        xmlFree(segment);
    }

    if (uri->user) {
        segment = xmlURIEscapeStr(BAD_CAST uri->user, BAD_CAST ";:&=+$,");
        NULLCHK(segment)
		ret = xmlStrcat(ret,BAD_CAST "//");	
        ret = xmlStrcat(ret, segment);
        ret = xmlStrcat(ret, BAD_CAST "@");
        xmlFree(segment);
    }

    if (uri->server) {
        segment = xmlURIEscapeStr(BAD_CAST uri->server, BAD_CAST "/?;:@");
        NULLCHK(segment)
		if (uri->user == NULL)
       		ret = xmlStrcat(ret, BAD_CAST "//");
        ret = xmlStrcat(ret, segment);
        xmlFree(segment);
    }

    if (uri->port) {
        xmlChar port[10];

        snprintf((char *) port, 10, "%d", uri->port);
        ret = xmlStrcat(ret, BAD_CAST ":");
        ret = xmlStrcat(ret, port);
    }

    if (uri->path) {
        segment =
            xmlURIEscapeStr(BAD_CAST uri->path, BAD_CAST ":@&=+$,/?;");
        NULLCHK(segment)
        ret = xmlStrcat(ret, segment);
        xmlFree(segment);
    }

    if (uri->query_raw) {
        ret = xmlStrcat(ret, BAD_CAST "?");
        ret = xmlStrcat(ret, BAD_CAST uri->query_raw);
    }
    else if (uri->query) {
        segment =
            xmlURIEscapeStr(BAD_CAST uri->query, BAD_CAST ";/?:@&=+,$");
        NULLCHK(segment)
        ret = xmlStrcat(ret, BAD_CAST "?");
        ret = xmlStrcat(ret, segment);
        xmlFree(segment);
    }

    if (uri->opaque) {
        segment = xmlURIEscapeStr(BAD_CAST uri->opaque, BAD_CAST "");
        NULLCHK(segment)
        ret = xmlStrcat(ret, segment);
        xmlFree(segment);
    }

    if (uri->fragment) {
        segment = xmlURIEscapeStr(BAD_CAST uri->fragment, BAD_CAST "#");
        NULLCHK(segment)
        ret = xmlStrcat(ret, BAD_CAST "#");
        ret = xmlStrcat(ret, segment);
        xmlFree(segment);
    }

    xmlFreeURI(uri);
#undef NULLCHK

    return (ret);
}

/************************************************************************
 *									*
 *			Escaped URI parsing				*
 *									*
 ************************************************************************/

/**
 * xmlParseURIFragment:
 * @uri:  pointer to an URI structure
 * @str:  pointer to the string to analyze
 *
 * Parse an URI fragment string and fills in the appropriate fields
 * of the @uri structure.
 * 
 * fragment = *uric
 *
 * Returns 0 or the error code
 */
static int
xmlParseURIFragment(xmlURIPtr uri, const char **str)
{
    const char *cur;
    
    if (str == NULL)
        return (-1);

    cur = *str;

    while (IS_URIC(cur) || IS_UNWISE(cur))
        NEXT(cur);
    if (uri != NULL) {
        if (uri->fragment != NULL)
            xmlFree(uri->fragment);
	if (uri->cleanup & 2)
	    uri->fragment = STRNDUP(*str, cur - *str);
	else
	    uri->fragment = xmlURIUnescapeString(*str, cur - *str, NULL);
    }
    *str = cur;
    return (0);
}

/**
 * xmlParseURIQuery:
 * @uri:  pointer to an URI structure
 * @str:  pointer to the string to analyze
 *
 * Parse the query part of an URI
 * 
 * query = *uric
 *
 * Returns 0 or the error code
 */
static int
xmlParseURIQuery(xmlURIPtr uri, const char **str)
{
    const char *cur;

    if (str == NULL)
        return (-1);

    cur = *str;

    while ((IS_URIC(cur)) ||
           ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))
        NEXT(cur);
    if (uri != NULL) {
        if (uri->query != NULL)
            xmlFree(uri->query);
	if (uri->cleanup & 2)
	    uri->query = STRNDUP(*str, cur - *str);
	else
	    uri->query = xmlURIUnescapeString(*str, cur - *str, NULL);

	/* Save the raw bytes of the query as well.
	 * See: http://mail.gnome.org/archives/xml/2007-April/thread.html#00114
	 */
	if (uri->query_raw != NULL)
	    xmlFree (uri->query_raw);
	uri->query_raw = STRNDUP (*str, cur - *str);
    }
    *str = cur;
    return (0);
}

/**
 * xmlParseURIScheme:
 * @uri:  pointer to an URI structure
 * @str:  pointer to the string to analyze
 *
 * Parse an URI scheme
 * 
 * scheme = alpha *( alpha | digit | "+" | "-" | "." )
 *
 * Returns 0 or the error code
 */
static int
xmlParseURIScheme(xmlURIPtr uri, const char **str) {
    const char *cur;

    if (str == NULL)
	return(-1);
    
    cur = *str;
    if (!IS_ALPHA(*cur))
	return(2);
    cur++;
    while (IS_SCHEME(*cur)) cur++;
    if (uri != NULL) {
	if (uri->scheme != NULL) xmlFree(uri->scheme);
	uri->scheme = STRNDUP(*str, cur - *str);
    }
    *str = cur;
    return(0);
}

/**
 * xmlParseURIOpaquePart:
 * @uri:  pointer to an URI structure
 * @str:  pointer to the string to analyze
 *
 * Parse an URI opaque part
 * 
 * opaque_part = uric_no_slash *uric
 *
 * Returns 0 or the error code
 */
static int
xmlParseURIOpaquePart(xmlURIPtr uri, const char **str)
{
    const char *cur;

    if (str == NULL)
        return (-1);

    cur = *str;
    if (!((IS_URIC_NO_SLASH(cur)) ||
          ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))) {
        return (3);
    }
    NEXT(cur);
    while ((IS_URIC(cur)) ||
           ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))
        NEXT(cur);
    if (uri != NULL) {
        if (uri->opaque != NULL)
            xmlFree(uri->opaque);
	if (uri->cleanup & 2)
	    uri->opaque = STRNDUP(*str, cur - *str);
	else
	    uri->opaque = xmlURIUnescapeString(*str, cur - *str, NULL);
    }
    *str = cur;
    return (0);
}

/**
 * xmlParseURIServer:
 * @uri:  pointer to an URI structure
 * @str:  pointer to the string to analyze
 *
 * Parse a server subpart of an URI, it's a finer grain analysis
 * of the authority part.
 * 
 * server        = [ [ userinfo "@" ] hostport ]
 * userinfo      = *( unreserved | escaped |
 *                       ";" | ":" | "&" | "=" | "+" | "$" | "," )
 * hostport      = host [ ":" port ]
 * host          = hostname | IPv4address | IPv6reference
 * hostname      = *( domainlabel "." ) toplabel [ "." ]
 * domainlabel   = alphanum | alphanum *( alphanum | "-" ) alphanum
 * toplabel      = alpha | alpha *( alphanum | "-" ) alphanum
 * IPv6reference = "[" IPv6address "]"
 * IPv6address   = hexpart [ ":" IPv4address ]
 * IPv4address   = 1*3digit "." 1*3digit "." 1*3digit "." 1*3digit
 * hexpart       = hexseq | hexseq "::" [ hexseq ]| "::" [ hexseq ]
 * hexseq        = hex4 *( ":" hex4)
 * hex4          = 1*4hexdig
 * port          = *digit
 *
 * Returns 0 or the error code
 */
static int
xmlParseURIServer(xmlURIPtr uri, const char **str) {
    const char *cur;
    const char *host, *tmp;
    const int IPV4max = 4;
    const int IPV6max = 8;
    int oct;

    if (str == NULL)
	return(-1);
    
    cur = *str;

    /*
     * is there a userinfo ?
     */
    while (IS_USERINFO(cur)) NEXT(cur);
    if (*cur == '@') {
	if (uri != NULL) {
	    if (uri->user != NULL) xmlFree(uri->user);
	    if (uri->cleanup & 2)
		uri->user = STRNDUP(*str, cur - *str);
	    else
		uri->user = xmlURIUnescapeString(*str, cur - *str, NULL);
	}
	cur++;
    } else {
	if (uri != NULL) {
	    if (uri->user != NULL) xmlFree(uri->user);
	    uri->user = NULL;
	}
        cur = *str;
    }
    /*
     * This can be empty in the case where there is no server
     */
    host = cur;
    if (*cur == '/') {
	if (uri != NULL) {
	    if (uri->authority != NULL) xmlFree(uri->authority);
	    uri->authority = NULL;
	    if (uri->server != NULL) xmlFree(uri->server);
	    uri->server = NULL;
	    uri->port = 0;
	}
	return(0);
    }
    /*
     * host part of hostport can denote an IPV4 address, an IPV6 address
     * or an unresolved name. Check the IP first, its easier to detect
     * errors if wrong one.
     * An IPV6 address must start with a '[' and end with a ']'.
     */
    if (*cur == '[') {
	int compress=0;
	cur++;
	for (oct = 0; oct < IPV6max; ++oct) {
	    if (*cur == ':') {
		if (compress)
		    return(3);	/* multiple compression attempted */
		if (!oct) { 	/* initial char is compression */
		    if (*++cur != ':')
			return(3);
		}
		compress = 1;	/* set compression-encountered flag */
		cur++;		/* skip over the second ':' */
		continue;
	    }
	    while(IS_HEX(*cur)) cur++;
	    if (oct == (IPV6max-1))
		continue;
	    if (*cur != ':')
		break;
	    cur++;
	}
	if ((!compress) && (oct != IPV6max))
	    return(3);
	if (*cur != ']')
	    return(3);
	if (uri != NULL) {
	    if (uri->server != NULL) xmlFree(uri->server);
	    uri->server = (char *)xmlStrndup((xmlChar *)host+1,
			(cur-host)-1);
	}
	cur++;
    } else {
	/*
	 * Not IPV6, maybe IPV4
	 */
	for (oct = 0; oct < IPV4max; ++oct) {
            if (*cur == '.') 
                return(3); /* e.g. http://.xml/ or http://18.29..30/ */
            while(IS_DIGIT(*cur)) cur++;
            if (oct == (IPV4max-1))
                continue;
            if (*cur != '.')
	        break;
            cur++;
	}
    }
    if ((host[0] != '[') && (oct < IPV4max || (*cur == '.' && cur++) ||
			     IS_ALPHA(*cur))) {
        /* maybe host_name */
        if (!IS_ALPHANUM(*cur))
            return(4); /* e.g. http://xml.$oft */
        do {
            do ++cur; while (IS_ALPHANUM(*cur));
            if (*cur == '-') {
	        --cur;
                if (*cur == '.')
                    return(5); /* e.g. http://xml.-soft */
	        ++cur;
		continue;
            }
    	    if (*cur == '.') {
	        --cur;
                if (*cur == '-')
                    return(6); /* e.g. http://xml-.soft */
                if (*cur == '.')
                    return(7); /* e.g. http://xml..soft */
	        ++cur;
		continue;
            }
	    break;
        } while (1);
        tmp = cur;
        if (tmp[-1] == '.')
            --tmp; /* e.g. http://xml.$Oft/ */
        do --tmp; while (tmp >= host && IS_ALPHANUM(*tmp));
        if ((++tmp == host || tmp[-1] == '.') && !IS_ALPHA(*tmp))
            return(8); /* e.g. http://xmlsOft.0rg/ */
    }
    if (uri != NULL) {
	if (uri->authority != NULL) xmlFree(uri->authority);
	uri->authority = NULL;
	if (host[0] != '[') {	/* it's not an IPV6 addr */
	    if (uri->server != NULL) xmlFree(uri->server);
	    if (uri->cleanup & 2)
		uri->server = STRNDUP(host, cur - host);
	    else
		uri->server = xmlURIUnescapeString(host, cur - host, NULL);
	}
    }
    /*
     * finish by checking for a port presence.
     */
    if (*cur == ':') {
        cur++;
	if (IS_DIGIT(*cur)) {
	    if (uri != NULL)
	        uri->port = 0;
	    while (IS_DIGIT(*cur)) {
	        if (uri != NULL)
		    uri->port = uri->port * 10 + (*cur - '0');
		cur++;
	    }
	}
    }
    *str = cur;
    return(0);
}	

/**
 * xmlParseURIRelSegment:
 * @uri:  pointer to an URI structure
 * @str:  pointer to the string to analyze
 *
 * Parse an URI relative segment
 * 
 * rel_segment = 1*( unreserved | escaped | ";" | "@" | "&" | "=" |
 *                          "+" | "$" | "," )
 *
 * Returns 0 or the error code
 */
static int
xmlParseURIRelSegment(xmlURIPtr uri, const char **str)
{
    const char *cur;

    if (str == NULL)
        return (-1);

    cur = *str;
    if (!((IS_SEGMENT(cur)) ||
          ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))) {
        return (3);
    }
    NEXT(cur);
    while ((IS_SEGMENT(cur)) ||
           ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))
        NEXT(cur);
    if (uri != NULL) {
        if (uri->path != NULL)
            xmlFree(uri->path);
	if (uri->cleanup & 2)
	    uri->path = STRNDUP(*str, cur - *str);
	else
	    uri->path = xmlURIUnescapeString(*str, cur - *str, NULL);
    }
    *str = cur;
    return (0);
}

/**
 * xmlParseURIPathSegments:
 * @uri:  pointer to an URI structure
 * @str:  pointer to the string to analyze
 * @slash:  should we add a leading slash
 *
 * Parse an URI set of path segments
 * 
 * path_segments = segment *( "/" segment )
 * segment       = *pchar *( ";" param )
 * param         = *pchar
 *
 * Returns 0 or the error code
 */
static int
xmlParseURIPathSegments(xmlURIPtr uri, const char **str, int slash)
{
    const char *cur;

    if (str == NULL)
        return (-1);

    cur = *str;

    do {
        while ((IS_PCHAR(cur)) ||
	       ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))
            NEXT(cur);
        while (*cur == ';') {
            cur++;
            while ((IS_PCHAR(cur)) ||
	           ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))
                NEXT(cur);
        }
        if (*cur != '/')
            break;
        cur++;
    } while (1);
    if (uri != NULL) {
        int len, len2 = 0;
        char *path;