XMLUri.hpp Source File

00001 /*
00002  * The Apache Software License, Version 1.1
00003  *
00004  * Copyright (c) 2001-2003 The Apache Software Foundation.  All rights
00005  * reserved.
00006  *
00007  * Redistribution and use in source and binary forms, with or without
00008  * modification, are permitted provided that the following conditions
00009  * are met:
00010  *
00011  * 1. Redistributions of source code must retain the above copyright
00012  *    notice, this list of conditions and the following disclaimer.
00013  *
00014  * 2. Redistributions in binary form must reproduce the above copyright
00015  *    notice, this list of conditions and the following disclaimer in
00016  *    the documentation and/or other materials provided with the
00017  *    distribution.
00018  *
00019  * 3. The end-user documentation included with the redistribution,
00020  *    if any, must include the following acknowledgment:
00021  *       "This product includes software developed by the
00022  *        Apache Software Foundation (http://www.apache.org/)."
00023  *    Alternately, this acknowledgment may appear in the software itself,
00024  *    if and wherever such third-party acknowledgments normally appear.
00025  *
00026  * 4. The names "Xerces" and "Apache Software Foundation" must
00027  *    not be used to endorse or promote products derived from this
00028  *    software without prior written permission. For written
00029  *    permission, please contact apache\@apache.org.
00030  *
00031  * 5. Products derived from this software may not be called "Apache",
00032  *    nor may "Apache" appear in their name, without prior written
00033  *    permission of the Apache Software Foundation.
00034  *
00035  * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED OR IMPLIED
00036  * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
00037  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
00038  * DISCLAIMED.  IN NO EVENT SHALL THE APACHE SOFTWARE FOUNDATION OR
00039  * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
00040  * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
00041  * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
00042  * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
00043  * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
00044  * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
00045  * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
00046  * SUCH DAMAGE.
00047  * ====================================================================
00048  *
00049  * This software consists of voluntary contributions made by many
00050  * individuals on behalf of the Apache Software Foundation, and was
00051  * originally based on software copyright (c) 1999, International
00052  * Business Machines, Inc., http://www.ibm.com .  For more information
00053  * on the Apache Software Foundation, please see
00054  * <http://www.apache.org/>.
00055  */
00056 
00057 /*
00058  * $Id: XMLUri.hpp,v 1.13 2003/12/02 17:50:21 neilg Exp $
00059  * $Log: XMLUri.hpp,v $
00060  * Revision 1.13  2003/12/02 17:50:21  neilg
00061  * additional fix for bug 25118; once again, thanks to Jeroen Whitmond
00062  *
00063  * Revision 1.12  2003/10/01 00:20:41  knoaman
00064  * Add a static method to check whether a given string is a valid URI.
00065  *
00066  * Revision 1.11  2003/09/25 22:23:25  peiyongz
00067  * Implementation of Serialization/Deserialization
00068  *
00069  * Revision 1.10  2003/07/25 10:15:16  gareth
00070  * Patch by Michael Glavassevich
00071  *
00072  * The patch fixes Bugzilla #19787, #20006, #20009, #20010 and #20287, and
00073  * several other issues. A summary of the changes is listed below:
00074  *
00075  * 1. Added '[' and ']' to reserved characters as per RFC 2732.
00076  * 2. '[' and ']' added in RFC 2732, are not allowed in path segments, but
00077  * may appear in the opaque part.
00078  * 3. No URI can begin with a ':'.
00079  * 4. URI has no scheme if ':' occurs in a URI after '?' or '#', it's part of
00080  * the query string or fragment.
00081  * 5. Whitespace (even escaped as %20) is not permitted in the authority
00082  * portion of a URI.
00083  * 6. IPv4 addresses must match 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT "."
00084  * 1*3DIGIT. Since RFC 2732.
00085  * 7. IPv4 addresses are 32-bit, therefore no segment may be larger than 255.
00086  * This isn't expressed by the grammar.
00087  * 8. Hostnames cannot end with a '-'.
00088  * 9. Labels in a hostname must be 63 bytes or less [RFC 1034].
00089  * 10. Hostnames may be no longer than 255 bytes [RFC 1034]. (That
00090  * restriction was already there. I just moved it inwards.
00091  * 11. Added support for IPv6 references added in RFC 2732. URIs such as
00092  * http://[::ffff:1.2.3.4] are valid. The BNF in RFC 2373 isn't correct. IPv6
00093  * addresses are read according to section 2.2 of RFC 2373.
00094  *
00095  * Revision 1.9  2003/05/16 06:01:53  knoaman
00096  * Partial implementation of the configurable memory manager.
00097  *
00098  * Revision 1.8  2003/05/15 19:07:46  knoaman
00099  * Partial implementation of the configurable memory manager.
00100  *
00101  * Revision 1.7  2003/01/06 19:43:18  tng
00102  * New feature StandardUriConformant to force strict standard uri conformance.
00103  *
00104  * Revision 1.6  2002/11/21 15:42:39  gareth
00105  * Implemented copy constructor and operator =. Patch by Jennifer Schachter.
00106  *
00107  * Revision 1.5  2002/11/04 15:22:05  tng
00108  * C++ Namespace Support.
00109  *
00110  * Revision 1.4  2002/09/23 18:41:00  tng
00111  * DOM L3: Support baseURI.   Add fURIText to XMLUri.   Added by Gareth Reakes and Thomas Ford.
00112  *
00113  * Revision 1.3  2002/08/23 20:45:24  tng
00114  * .Memory leak fix: XMLUri data not deleted if constructor failed.
00115  *
00116  * Revision 1.2  2002/02/20 18:17:02  tng
00117  * [Bug 5977] Warnings on generating apiDocs.
00118  *
00119  * Revision 1.1.1.1  2002/02/01 22:22:17  peiyongz
00120  * sane_include
00121  *
00122  * Revision 1.3  2001/08/29 19:03:03  peiyongz
00123  * Bugzilla# 2816:on AIX 4.2, xlC 3 r ev.1, Compilation error on inline method
00124  *
00125  * Revision 1.2  2001/08/16 14:09:44  peiyongz
00126  * Removed unused ctors and methods
00127  *
00128  * Revision 1.1  2001/08/10 16:23:41  peiyongz
00129  * XMLUri: creation
00130  *
00131  *
00132  */
00133 
00134 #if !defined(XMLURI_HPP)
00135 #define XMLURI_HPP
00136 
00137 #include <xercesc/util/XMemory.hpp>
00138 #include <xercesc/util/XMLString.hpp>
00139 
00140 #include <xercesc/internal/XSerializable.hpp>
00141 
00142 XERCES_CPP_NAMESPACE_BEGIN
00143 
00144 /*
00145  * This class is a direct port of Java's URI class, to distinguish
00146  * itself from the XMLURL, we use the name XMLUri instead of
00147  * XMLURI.
00148  *
00149  * TODO: how to relate XMLUri and XMLURL since URL is part of URI.
00150  *
00151  */
00152 
00153 class  XMLUri : public XSerializable, public XMemory
00154 {
00155 public:
00156 
00157     // -----------------------------------------------------------------------
00158     //  Constructors and Destructor
00159     // -----------------------------------------------------------------------
00160 
00189     XMLUri(const XMLCh* const    uriSpec,
00190            MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager);
00191 
00208     XMLUri(const XMLUri* const  baseURI
00209          , const XMLCh* const   uriSpec
00210          , MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager);
00211 
00215     XMLUri(const XMLUri& toCopy);
00216     XMLUri& operator=(const XMLUri& toAssign);
00217 
00218     virtual ~XMLUri();
00219 
00220     // -----------------------------------------------------------------------
00221     //  Getter methods
00222     // -----------------------------------------------------------------------
00228     const XMLCh* getUriText() const;
00229 
00235      const XMLCh* getScheme() const;
00236 
00242      const XMLCh* getUserInfo() const;
00243 
00244 
00250      const XMLCh* getHost() const;
00251 
00257      int getPort() const;
00258 
00265      const XMLCh* getPath() const;
00266 
00274      const XMLCh* getQueryString() const;
00275 
00283      const XMLCh* getFragment() const;
00284 
00285     // -----------------------------------------------------------------------
00286     //  Setter methods
00287     // -----------------------------------------------------------------------
00288 
00296      void setScheme(const XMLCh* const newScheme);
00297 
00305      void setUserInfo(const XMLCh* const newUserInfo);
00306 
00314      void setHost(const XMLCh* const newHost);
00315 
00325      void setPort(int newPort);
00326 
00347      void setPath(const XMLCh* const newPath);
00348 
00357      void setQueryString(const XMLCh* const newQueryString);
00358 
00367      void setFragment(const XMLCh* const newFragment);
00368 
00369      // -----------------------------------------------------------------------
00370     //  Miscellaneous methods
00371     // -----------------------------------------------------------------------
00372 
00380     static bool isURIString(const XMLCh* const uric);
00381 
00385     static bool isValidURI( const XMLUri* const baseURI
00386                           , const XMLCh* const uriStr);
00387 
00388     /***
00389      * Support for Serialization/De-serialization
00390      ***/
00391     DECL_XSERIALIZABLE(XMLUri)
00392 
00393     XMLUri(MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager);
00394 
00395 private:
00396 
00397     static const XMLCh RESERVED_CHARACTERS[];
00398     static const XMLCh MARK_CHARACTERS[];
00399     static const XMLCh SCHEME_CHARACTERS[];
00400     static const XMLCh USERINFO_CHARACTERS[];
00401     static const XMLCh PATH_CHARACTERS[];
00402 
00403     //helper method for getUriText
00404     void buildFullText();
00405 
00406     // -----------------------------------------------------------------------
00407     //  Private helper methods
00408     // -----------------------------------------------------------------------
00409 
00415     static bool isReservedCharacter(const XMLCh theChar);
00416     
00422     static bool isPathCharacter(const XMLCh theChar);
00423 
00429     static bool isUnreservedCharacter(const XMLCh theChar);
00430 
00438     static bool isConformantSchemeName(const XMLCh* const scheme);
00439 
00445     static void isConformantUserInfo(const XMLCh* const userInfo);
00463      static bool isWellFormedAddress(const XMLCh* const addr);
00464      
00474      static bool isWellFormedIPv4Address(const XMLCh* const addr, const int length);
00475      
00489      static bool isWellFormedIPv6Reference(const XMLCh* const addr, const int length);
00490      
00506      static int scanHexSequence (const XMLCh* const addr, int index, int end, int& counter);
00507 
00515      bool isGenericURI();
00516 
00517     // -----------------------------------------------------------------------
00518     //  Miscellaneous methods
00519     // -----------------------------------------------------------------------
00520 
00526      void initialize(const XMLUri& toCopy);
00527 
00542      void initialize(const XMLUri* const baseURI
00543                    , const XMLCh*  const uriSpec);
00544 
00551      void initializeScheme(const XMLCh* const uriSpec);
00552 
00560      void initializeAuthority(const XMLCh* const uriSpec);
00561 
00568      void initializePath(const XMLCh* const uriSpec);
00569 
00574      void cleanUp();
00575 
00576     static bool isConformantSchemeName(const XMLCh* const scheme,
00577                                        const int schemeLen);
00578     static bool processScheme(const XMLCh* const uriStr, int& index);
00579     static bool processAuthority(const XMLCh* const uriStr, const int authLen);
00580     static bool isWellFormedAddress(const XMLCh* const addr, const int addrLen);
00581     static bool processPath(const XMLCh* const pathStr, const int pathStrLen,
00582                             const bool isSchemePresent);
00583 
00584     // -----------------------------------------------------------------------
00585     //  Data members
00586     //
00587     //  for all the data member, we own it,
00588     //  responsible for the creation and/or deletion for
00589     //  the memory allocated.
00590     //
00591     // -----------------------------------------------------------------------
00592     XMLCh*          fScheme;
00593     XMLCh*          fUserInfo;
00594     XMLCh*          fHost;
00595     int             fPort;
00596     XMLCh*          fPath;
00597     XMLCh*          fQueryString;
00598     XMLCh*          fFragment;
00599     XMLCh*          fURIText;
00600     MemoryManager*  fMemoryManager;
00601 };
00602 
00603 // ---------------------------------------------------------------------------
00604 //  XMLUri: Getter methods
00605 // ---------------------------------------------------------------------------
00606 inline const XMLCh* XMLUri::getScheme() const
00607 {
00608     return fScheme;
00609 }
00610 
00611 inline const XMLCh* XMLUri::getUserInfo() const
00612 {
00613     return fUserInfo;
00614 }
00615 
00616 inline const XMLCh* XMLUri::getHost() const
00617 {
00618     return fHost;
00619 }
00620 
00621 inline int XMLUri::getPort() const
00622 {
00623     return fPort;
00624 }
00625 
00626 inline const XMLCh* XMLUri::getPath() const
00627 {
00628     return fPath;
00629 }
00630 
00631 inline const XMLCh* XMLUri::getQueryString() const
00632 {
00633     return fQueryString;
00634 }
00635 
00636 inline const XMLCh* XMLUri::getFragment() const
00637 {
00638     return fFragment;
00639 }
00640 
00641 inline const XMLCh* XMLUri::getUriText() const
00642 {
00643     //
00644     //  Fault it in if not already. Since this is a const method and we
00645     //  can't use mutable members due the compilers we have to support,
00646     //  we have to cast off the constness.
00647     //
00648     if (!fURIText)
00649         ((XMLUri*)this)->buildFullText();
00650 
00651     return fURIText;
00652 }
00653 
00654 // ---------------------------------------------------------------------------
00655 //  XMLUri: Helper methods
00656 // ---------------------------------------------------------------------------
00657 inline bool XMLUri::isReservedCharacter(const XMLCh theChar)
00658 {
00659     return (XMLString::indexOf(RESERVED_CHARACTERS, theChar) != -1);
00660 }
00661 
00662 inline bool XMLUri::isPathCharacter(const XMLCh theChar)
00663 {
00664     return (XMLString::indexOf(PATH_CHARACTERS, theChar) != -1);
00665 }
00666 
00667 inline bool XMLUri::isUnreservedCharacter(const XMLCh theChar)
00668 {
00669     return (XMLString::isAlphaNum(theChar) ||
00670             XMLString::indexOf(MARK_CHARACTERS, theChar) != -1);
00671 }
00672 
00673 XERCES_CPP_NAMESPACE_END
00674 
00675 #endif