xmluri.cpp

来自「IBM的解析xml的工具Xerces的源代码」· C++ 代码 · 共 2,222 行 · 第 1/5 页

    if (index == trimmedUriSpecLen || (foundScheme && (trimmedUriSpec[index] == chPound)))
    {
        ThrowXMLwithMemMgr1(MalformedURLException
                , XMLExcepts::XMLNUM_URI_Component_Empty
                , errMsg_PATH
                , fMemoryManager);
    }

	// two slashes means generic URI syntax, so we get the authority
    XMLCh* authUriSpec = (XMLCh*) fMemoryManager->allocate
    (
        (trimmedUriSpecLen+1) * sizeof(XMLCh)
    );//new XMLCh[trimmedUriSpecLen+1];
    ArrayJanitor<XMLCh> authName(authUriSpec, fMemoryManager);
    XMLString::subString(authUriSpec, trimmedUriSpec, index, trimmedUriSpecLen, fMemoryManager);

    if (((index+1) < trimmedUriSpecLen) &&
        XMLString::startsWith(authUriSpec, DOUBLE_SLASH))
    {
        index += 2;
        int startPos = index;

        // get authority - everything up to path, query or fragment
        XMLCh testChar;
        while (index < trimmedUriSpecLen)
        {
            testChar = trimmedUriSpec[index];
            if (testChar == chForwardSlash ||
                testChar == chQuestion     ||
                testChar == chPound         )
            {
                break;
            }

            index++;
        }

        // if we found authority, parse it out, otherwise we set the
        // host to empty string
        if (index > startPos)
        {
            XMLString::subString(authUriSpec, trimmedUriSpec, startPos, index, fMemoryManager);
            initializeAuthority(authUriSpec);
        }
        else
        {
            //fHost = 0;
            setHost(XMLUni::fgZeroLenString);
        }
    }

    // we need to check if index has exceed the lenght or not
    if (index >= trimmedUriSpecLen)
        return;

    XMLCh* pathUriSpec = (XMLCh*) fMemoryManager->allocate
    (
        (trimmedUriSpecLen+1) * sizeof(XMLCh)
    );//new XMLCh[trimmedUriSpecLen+1];
    ArrayJanitor<XMLCh> pathUriSpecName(pathUriSpec, fMemoryManager);
    XMLString::subString(pathUriSpec, trimmedUriSpec, index, trimmedUriSpecLen, fMemoryManager);

	initializePath(pathUriSpec);

	// Resolve relative URI to base URI - see RFC 2396 Section 5.2
	// In some cases, it might make more sense to throw an exception
	// (when scheme is specified is the string spec and the base URI
	// is also specified, for example), but we're just following the
	// RFC specifications
	if ( baseURI )
    {
        // check to see if this is the current doc - RFC 2396 5.2 #2
        // note that this is slightly different from the RFC spec in that
        // we don't include the check for query string being null
        // - this handles cases where the urispec is just a query
        // string or a fragment (e.g. "?y" or "#s") -
        // see <http://www.ics.uci.edu/~fielding/url/test1.html> which
        // identified this as a bug in the RFC
        if ((!fPath || !*fPath) &&
            fScheme == 0 &&
            fHost == 0 && fRegAuth == 0)
        {
            fScheme = XMLString::replicate(baseURI->getScheme(), fMemoryManager);
            fMemoryManager->deallocate(fUserInfo);//delete [] fUserInfo;
            fUserInfo = XMLString::replicate(baseURI->getUserInfo(), fMemoryManager);
            fHost = XMLString::replicate(baseURI->getHost(), fMemoryManager);
            fPort = baseURI->getPort();
            fRegAuth = XMLString::replicate(baseURI->getRegBasedAuthority(), fMemoryManager);
            fMemoryManager->deallocate(fPath);//delete [] fPath;
            fPath = XMLString::replicate(baseURI->getPath(), fMemoryManager);

            if ( !fQueryString )
            {
                fQueryString = XMLString::replicate(baseURI->getQueryString(), fMemoryManager);
            }
            return;
        }

        // check for scheme - RFC 2396 5.2 #3
        // if we found a scheme, it means absolute URI, so we're done
        if (fScheme == 0)
        {
            fScheme = XMLString::replicate(baseURI->getScheme(), fMemoryManager);
        }
        else
        {
            return;
        }

        // check for authority - RFC 2396 5.2 #4
        // if we found a host, then we've got a network path, so we're done
        if (fHost == 0 && fRegAuth == 0)
        {
            fMemoryManager->deallocate(fUserInfo);//delete [] fUserInfo;
            fUserInfo = XMLString::replicate(baseURI->getUserInfo(), fMemoryManager);
            fHost = XMLString::replicate(baseURI->getHost(), fMemoryManager);
            fPort = baseURI->getPort();
            fRegAuth = XMLString::replicate(baseURI->getRegBasedAuthority(), fMemoryManager);
        }
        else
        {
            return;
        }

        // check for absolute path - RFC 2396 5.2 #5
        if ((fPath && *fPath) &&
            XMLString::startsWith(fPath, SINGLE_SLASH))
        {
            return;
        }

        // if we get to this point, we need to resolve relative path
        // RFC 2396 5.2 #6

        XMLCh* basePath = XMLString::replicate(baseURI->getPath(), fMemoryManager);
        ArrayJanitor<XMLCh> basePathName(basePath, fMemoryManager);

        int bufLen = trimmedUriSpecLen+XMLString::stringLen(fPath)+XMLString::stringLen(basePath)+1;
        XMLCh* path = (XMLCh*) fMemoryManager->allocate(bufLen * sizeof(XMLCh));//new XMLCh[bufLen];
        ArrayJanitor<XMLCh> pathName(path, fMemoryManager);
        path[0] = 0;

        XMLCh* tmp1 = (XMLCh*) fMemoryManager->allocate(bufLen * sizeof(XMLCh));//new XMLCh[bufLen];
        ArrayJanitor<XMLCh> tmp1Name(tmp1, fMemoryManager);
        XMLCh* tmp2 = (XMLCh*) fMemoryManager->allocate(bufLen * sizeof(XMLCh));//new XMLCh[bufLen];
        ArrayJanitor<XMLCh> tmp2Name(tmp2, fMemoryManager);

        // 6a - get all but the last segment of the base URI path
        if (basePath)
        {
            int lastSlash = XMLString::lastIndexOf(basePath, chForwardSlash);
            if (lastSlash != -1)
            {
                XMLString::subString(path, basePath, 0, lastSlash+1, fMemoryManager);
            }
        }

        // 6b - append the relative URI path
        XMLString::catString(path, fPath);

        // 6c - remove all "./" where "." is a complete path segment
        index = -1;
        while ((index = XMLString::patternMatch(path, SLASH_DOT_SLASH)) != -1)
        {
            XMLString::subString(tmp1, path, 0, index, fMemoryManager);
            XMLString::subString(tmp2, path, index+2, XMLString::stringLen(path), fMemoryManager);

            path[0] = 0;
            XMLString::catString(path, tmp1);
            XMLString::catString(path, tmp2);
        }

        // 6d - remove "." if path ends with "." as a complete path segment
        if (XMLString::endsWith(path, SLASH_DOT))
        {
            path[XMLString::stringLen(path) - 1] = chNull;
        }

        // 6e - remove all "<segment>/../" where "<segment>" is a complete
        // path segment not equal to ".."
        index = -1;
        int segIndex = -1;
        int offset = 1;

        while ((index = XMLString::patternMatch(&(path[offset]), SLASH_DOTDOT_SLASH)) != -1)
        {
			// Undo offset
			index += offset;

			// Find start of <segment> within substring ending at found point.
			XMLString::subString(tmp1, path, 0, index-1, fMemoryManager);
			segIndex = XMLString::lastIndexOf(tmp1, chForwardSlash);

			// Ensure <segment> exists and != ".."
            if (segIndex != -1                &&
                (path[segIndex+1] != chPeriod ||
                 path[segIndex+2] != chPeriod ||
				 segIndex + 3 != index))
            {

                XMLString::subString(tmp1, path, 0, segIndex, fMemoryManager);
                XMLString::subString(tmp2, path, index+3, XMLString::stringLen(path), fMemoryManager);

                path[0] = 0;
                XMLString::catString(path, tmp1);
                XMLString::catString(path, tmp2);

                offset = (segIndex == 0 ? 1 : segIndex);
            }
            else
            {
                offset += 4;
            }
        }// while

        // 6f - remove ending "<segment>/.." where "<segment>" is a
        // complete path segment
        if (XMLString::endsWith(path, SLASH_DOTDOT))
        {
			// Find start of <segment> within substring ending at found point.
            index = XMLString::stringLen(path) - 3;
			XMLString::subString(tmp1, path, 0, index-1, fMemoryManager);
			segIndex = XMLString::lastIndexOf(tmp1, chForwardSlash);

            if (segIndex != -1                &&
                (path[segIndex+1] != chPeriod ||
                 path[segIndex+2] != chPeriod ||
				 segIndex + 3 != index))
            {
                path[segIndex+1] = chNull;
            }
        }

        if (getPath())
            fMemoryManager->deallocate(fPath);//delete [] fPath;

        fPath = XMLString::replicate(path, fMemoryManager);

    }
}

// ---------------------------------------------------------------------------
//  Components initialization
// ---------------------------------------------------------------------------

//
// authority     = server | reg_name
// server        = [ [ userinfo "@" ] hostport ]
// hostport      = host [ ":" port ]
//
// reg_name      = 1*( unreserved | escaped | "$" | "," |
//                    ";" | ":" | "@" | "&" | "=" | "+" )
//
// userinfo      = *( unreserved | escaped |
//                 ";" | ":" | "&" | "=" | "+" | "$" | "," )
//

void XMLUri::initializeAuthority(const XMLCh* const uriSpec)
{

    int index = 0;
    int start = 0;
    const int end = XMLString::stringLen(uriSpec);

    //
    // server = [ [ userinfo "@" ] hostport ]
    // userinfo is everything up @,
    //
    XMLCh* userinfo = (XMLCh*) fMemoryManager->allocate
    (
        (end+1) * sizeof(XMLCh)
    );//new XMLCh[end+1];
    ArrayJanitor<XMLCh> userName(userinfo, fMemoryManager);
    index = XMLString::indexOf(&(uriSpec[start]), chAt);

    if ( index != -1)
    {
        XMLString::subString(userinfo, &(uriSpec[start]), 0, index, fMemoryManager);
        index++; // skip the @
        start += index;
    }
    else
    {
        userinfo = 0;
    }

    //
    // hostport = host [ ":" port ]
    // host is everything up to ':', or up to 
    // and including ']' if followed by ':'.
    //
    XMLCh* host = (XMLCh*) fMemoryManager->allocate
    (
        (end+1) * sizeof(XMLCh)
    );//new XMLCh[end+1];
    ArrayJanitor<XMLCh> hostName(host, fMemoryManager);
    
    // Search for port boundary.
    if (start < end && uriSpec[start] == chOpenSquare)
    {
    	index = XMLString::indexOf(&(uriSpec[start]), chCloseSquare);
    	if (index != -1)
    	{
            // skip the ']'
            index = ((start + index + 1) < end 
              && uriSpec[start + index + 1] == chColon) ? index+1 : -1;
    	}
    }
    else
    {
        index = XMLString::indexOf(&(uriSpec[start]), chColon);
    }

    if ( index != -1 )
    {
        XMLString::subString(host, &(uriSpec[start]), 0, index, fMemoryManager);
        index++;  // skip the :
        start +=index;
    }
    else
    {
        XMLString::subString(host, &(uriSpec[start]), 0, end-start, fMemoryManager);
        start = end;
    }

    // port is everything after ":"

    XMLCh* portStr = (XMLCh*) fMemoryManager->allocate
    (
        (end+1) * sizeof(XMLCh)
    );//new XMLCh[end+1];
    ArrayJanitor<XMLCh> portName(portStr, fMemoryManager);
    int port = -1;

    if ((host && *host) &&   // non empty host
        (index != -1)                    &&   // ":" found
        (start < end)                     )   // ":" is not the last
    {
        XMLString::subString(portStr, &(uriSpec[start]), 0, end-start, fMemoryManager);

        if (portStr && *portStr)
        {
            try
            {
                port = XMLString::parseInt(portStr, fMemoryManager);
            }
            catch(const OutOfMemoryException&)
            {
                throw;
            }
            catch (...)
            {
                throw;
            }
        }
    } // if > 0

    // Check if we have server based authority.
    if (isValidServerBasedAuthority(host, port, userinfo, fMemoryManager))
    {
        if (fHost)
            fMemoryManager->deallocate(fHost);//delete [] fHost;
        
        if (fUserInfo)
            fMemoryManager->deallocate(fUserInfo);//delete[] fUserInfo;
            
        fHost = XMLString::replicate(host, fMemoryManager);
        fPort = port;
        fUserInfo = XMLString::replicate(userinfo, fMemoryManager);
        
        return;
    }
    // This must be registry based authority or the URI is malformed.
    setRegBasedAuthority(uriSpec);
}

// scheme = alpha *( alpha | digit | "+" | "-" | "." )
void XMLUri::initializeScheme(const XMLCh* const uriSpec)
{
    const XMLCh* tmpPtr = XMLString::findAny(uriSpec, SCHEME_SEPARATORS);

    if ( !tmpPtr )
    {
        ThrowXMLwithMemMgr(MalformedURLException, XMLExcepts::XMLNUM_URI_No_Scheme, fMemoryManager);
    }
	else
    {
        XMLCh* scheme = (XMLCh*) fMemoryManager->allocate
        (
            (XMLString::stringLen(uriSpec) + 1) * sizeof(XMLCh)
        );//new XMLCh[XMLString::stringLen(uriSpec)+1];
        ArrayJanitor<XMLCh> tmpName(scheme, fMemoryManager);
        XMLString::subString(scheme, uriSpec, 0, (tmpPtr - uriSpec), fMemoryManager);
        setScheme(scheme);
	}

}

void XMLUri::initializePath(const XMLCh* const uriSpec)
{
    if ( !uriSpec )
    {
        ThrowXMLwithMemMgr1(MalformedURLException
                , XMLExcepts::XMLNUM_URI_Component_Empty
                , errMsg_PATH
                , fMemoryManager);
    }

    int index = 0;
    int start = 0;
    int end = XMLString::stringLen(uriSpec);
    XMLCh testChar;

    // path - everything up to query string or fragment
    if (start < end)
    {
        // RFC 2732 only allows '[' and ']' to appear in the opaque part.
        if (!getScheme() || uriSpec[start] == chForwardSlash)
        {
            // Scan path.
            // abs_path = "/"  path_segments
            // rel_path = rel_segment [ abs_path ]
            while (index < end)
            {
                testChar = uriSpec[index];
                if (testChar == chQuestion || testChar == chPound)
                {
                    break;
                }

                // check for valid escape sequence
                if (testChar == chPercent)
                {
                    if (index+2 >= end ||
                        !XMLString::isHex(uriSpec[index+1]) ||
                        !XMLString::isHex(uriSpec[index+2]))
                    {
                        XMLCh value1[BUF_LEN+1];
                        XMLString::moveChars(value1, &(uriSpec[index]), 3);
                        value1[3] = chNull;
                        ThrowXMLwithMemMgr2(MalformedURLException
                                , XMLExcepts::XMLNUM_URI_Component_Invalid_EscapeSequence
                                , errMsg_PATH
                                , value1
                                , fMemoryManager);