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 1.16 2004/01/12 22:01:02 cargilld Exp $
00059  * $Log: XMLUri.hpp,v $
00060  * Revision 1.16  2004/01/12 22:01:02  cargilld
00061  * Minor performance change for handling reserved and unreserved characters.
00062  *
00063  * Revision 1.15  2003/12/17 00:18:35  cargilld
00064  * Update to memory management so that the static memory manager (one used to call Initialize) is only for static data.
00065  *
00066  * Revision 1.14  2003/12/11 22:21:25  neilg
00067  * fixes for the URI implementation to take registry names into account; much thanks to Michael Glavassevich
00068  *
00069  * Revision 1.13  2003/12/02 17:50:21  neilg
00070  * additional fix for bug 25118; once again, thanks to Jeroen Whitmond
00071  *
00072  * Revision 1.12  2003/10/01 00:20:41  knoaman
00073  * Add a static method to check whether a given string is a valid URI.
00074  *
00075  * Revision 1.11  2003/09/25 22:23:25  peiyongz
00076  * Implementation of Serialization/Deserialization
00077  *
00078  * Revision 1.10  2003/07/25 10:15:16  gareth
00079  * Patch by Michael Glavassevich
00080  *
00081  * The patch fixes Bugzilla #19787, #20006, #20009, #20010 and #20287, and
00082  * several other issues. A summary of the changes is listed below:
00083  *
00084  * 1. Added '[' and ']' to reserved characters as per RFC 2732.
00085  * 2. '[' and ']' added in RFC 2732, are not allowed in path segments, but
00086  * may appear in the opaque part.
00087  * 3. No URI can begin with a ':'.
00088  * 4. URI has no scheme if ':' occurs in a URI after '?' or '#', it's part of
00089  * the query string or fragment.
00090  * 5. Whitespace (even escaped as %20) is not permitted in the authority
00091  * portion of a URI.
00092  * 6. IPv4 addresses must match 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT "."
00093  * 1*3DIGIT. Since RFC 2732.
00094  * 7. IPv4 addresses are 32-bit, therefore no segment may be larger than 255.
00095  * This isn't expressed by the grammar.
00096  * 8. Hostnames cannot end with a '-'.
00097  * 9. Labels in a hostname must be 63 bytes or less [RFC 1034].
00098  * 10. Hostnames may be no longer than 255 bytes [RFC 1034]. (That
00099  * restriction was already there. I just moved it inwards.
00100  * 11. Added support for IPv6 references added in RFC 2732. URIs such as
00101  * http://[::ffff:1.2.3.4] are valid. The BNF in RFC 2373 isn't correct. IPv6
00102  * addresses are read according to section 2.2 of RFC 2373.
00103  *
00104  * Revision 1.9  2003/05/16 06:01:53  knoaman
00105  * Partial implementation of the configurable memory manager.
00106  *
00107  * Revision 1.8  2003/05/15 19:07:46  knoaman
00108  * Partial implementation of the configurable memory manager.
00109  *
00110  * Revision 1.7  2003/01/06 19:43:18  tng
00111  * New feature StandardUriConformant to force strict standard uri conformance.
00112  *
00113  * Revision 1.6  2002/11/21 15:42:39  gareth
00114  * Implemented copy constructor and operator =. Patch by Jennifer Schachter.
00115  *
00116  * Revision 1.5  2002/11/04 15:22:05  tng
00117  * C++ Namespace Support.
00118  *
00119  * Revision 1.4  2002/09/23 18:41:00  tng
00120  * DOM L3: Support baseURI.   Add fURIText to XMLUri.   Added by Gareth Reakes and Thomas Ford.
00121  *
00122  * Revision 1.3  2002/08/23 20:45:24  tng
00123  * .Memory leak fix: XMLUri data not deleted if constructor failed.
00124  *
00125  * Revision 1.2  2002/02/20 18:17:02  tng
00126  * [Bug 5977] Warnings on generating apiDocs.
00127  *
00128  * Revision 1.1.1.1  2002/02/01 22:22:17  peiyongz
00129  * sane_include
00130  *
00131  * Revision 1.3  2001/08/29 19:03:03  peiyongz
00132  * Bugzilla# 2816:on AIX 4.2, xlC 3 r ev.1, Compilation error on inline method
00133  *
00134  * Revision 1.2  2001/08/16 14:09:44  peiyongz
00135  * Removed unused ctors and methods
00136  *
00137  * Revision 1.1  2001/08/10 16:23:41  peiyongz
00138  * XMLUri: creation
00139  *
00140  *
00141  */
00142 
00143 #if !defined(XMLURI_HPP)
00144 #define XMLURI_HPP
00145 
00146 #include <xercesc/util/XMemory.hpp>
00147 #include <xercesc/util/XMLString.hpp>
00148 
00149 #include <xercesc/internal/XSerializable.hpp>
00150 
00151 XERCES_CPP_NAMESPACE_BEGIN
00152 
00153 /*
00154  * This class is a direct port of Java's URI class, to distinguish
00155  * itself from the XMLURL, we use the name XMLUri instead of
00156  * XMLURI.
00157  *
00158  * TODO: how to relate XMLUri and XMLURL since URL is part of URI.
00159  *
00160  */
00161 
00162 class  XMLUri : public XSerializable, public XMemory
00163 {
00164 public:
00165 
00166     // -----------------------------------------------------------------------
00167     //  Constructors and Destructor
00168     // -----------------------------------------------------------------------
00169 
00198     XMLUri(const XMLCh* const    uriSpec,
00199            MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager);
00200 
00217     XMLUri(const XMLUri* const  baseURI
00218          , const XMLCh* const   uriSpec
00219          , MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager);
00220 
00224     XMLUri(const XMLUri& toCopy);
00225     XMLUri& operator=(const XMLUri& toAssign);
00226 
00227     virtual ~XMLUri();
00228 
00229     // -----------------------------------------------------------------------
00230     //  Getter methods
00231     // -----------------------------------------------------------------------
00237     const XMLCh* getUriText() const;
00238 
00244      const XMLCh* getScheme() const;
00245 
00251      const XMLCh* getUserInfo() const;
00252 
00253 
00259      const XMLCh* getHost() const;
00260 
00266      int getPort() const;
00267      
00273      const XMLCh* getRegBasedAuthority() const;
00274 
00281      const XMLCh* getPath() const;
00282 
00290      const XMLCh* getQueryString() const;
00291 
00299      const XMLCh* getFragment() const;
00300 
00301     // -----------------------------------------------------------------------
00302     //  Setter methods
00303     // -----------------------------------------------------------------------
00304 
00312      void setScheme(const XMLCh* const newScheme);
00313 
00321      void setUserInfo(const XMLCh* const newUserInfo);
00322 
00333      void setHost(const XMLCh* const newHost);
00334 
00344      void setPort(int newPort);
00345      
00354      void setRegBasedAuthority(const XMLCh* const newRegAuth);
00355 
00376      void setPath(const XMLCh* const newPath);
00377 
00386      void setQueryString(const XMLCh* const newQueryString);
00387 
00396      void setFragment(const XMLCh* const newFragment);
00397 
00398      // -----------------------------------------------------------------------
00399     //  Miscellaneous methods
00400     // -----------------------------------------------------------------------
00401 
00409     static bool isURIString(const XMLCh* const uric);
00410 
00414     static bool isValidURI( const XMLUri* const baseURI
00415                           , const XMLCh* const uriStr);
00419     static bool isValidURI( bool haveBaseURI
00420                           , const XMLCh* const uriStr);
00421 
00422     /***
00423      * Support for Serialization/De-serialization
00424      ***/
00425     DECL_XSERIALIZABLE(XMLUri)
00426 
00427     XMLUri(MemoryManager* const manager = XMLPlatformUtils::fgMemoryManager);
00428 
00429 private:
00430 
00431     static const XMLCh MARK_OR_RESERVED_CHARACTERS[];
00432     static const XMLCh RESERVED_CHARACTERS[];
00433     static const XMLCh MARK_CHARACTERS[];
00434     static const XMLCh SCHEME_CHARACTERS[];
00435     static const XMLCh USERINFO_CHARACTERS[];
00436     static const XMLCh REG_NAME_CHARACTERS[];
00437     static const XMLCh PATH_CHARACTERS[];
00438 
00439     //helper method for getUriText
00440     void buildFullText();
00441 
00442     // -----------------------------------------------------------------------
00443     //  Private helper methods
00444     // -----------------------------------------------------------------------
00445 
00451     static bool isReservedCharacter(const XMLCh theChar);
00452     
00458     static bool isPathCharacter(const XMLCh theChar);
00459 
00465     static bool isUnreservedCharacter(const XMLCh theChar);
00466 
00472     static bool isReservedOrUnreservedCharacter(const XMLCh theChar);
00473 
00481     static bool isConformantSchemeName(const XMLCh* const scheme);
00482 
00488     static void isConformantUserInfo(const XMLCh* const userInfo
00489         , MemoryManager* const manager);
00490     
00498     static bool isValidServerBasedAuthority(const XMLCh* const host
00499                                            , const int hostLen
00500                                            , const int port
00501                                            , const XMLCh* const userinfo
00502                                            , const int userLen);
00503                                            
00511     static bool isValidServerBasedAuthority(const XMLCh* const host
00512                                            , const int port
00513                                            , const XMLCh* const userinfo
00514                                            , MemoryManager* const manager);
00515       
00523     static bool isValidRegistryBasedAuthority(const XMLCh* const authority
00524                                              , const int authLen);
00525 
00533     static bool isValidRegistryBasedAuthority(const XMLCh* const authority);
00534 
00552      static bool isWellFormedAddress(const XMLCh* const addr
00553          , MemoryManager* const manager);
00554      
00564      static bool isWellFormedIPv4Address(const XMLCh* const addr, const int length);
00565      
00579      static bool isWellFormedIPv6Reference(const XMLCh* const addr, const int length);
00580      
00596      static int scanHexSequence (const XMLCh* const addr, int index, int end, int& counter);
00597 
00605      bool isGenericURI();
00606 
00607     // -----------------------------------------------------------------------
00608     //  Miscellaneous methods
00609     // -----------------------------------------------------------------------
00610 
00616      void initialize(const XMLUri& toCopy);
00617 
00632      void initialize(const XMLUri* const baseURI
00633                    , const XMLCh*  const uriSpec);
00634 
00641      void initializeScheme(const XMLCh* const uriSpec);
00642 
00650      void initializeAuthority(const XMLCh* const uriSpec);
00651 
00658      void initializePath(const XMLCh* const uriSpec);
00659 
00664      void cleanUp();
00665 
00666     static bool isConformantSchemeName(const XMLCh* const scheme,
00667                                        const int schemeLen);
00668     static bool processScheme(const XMLCh* const uriStr, int& index);
00669     static bool processAuthority(const XMLCh* const uriStr, const int authLen);
00670     static bool isWellFormedAddress(const XMLCh* const addr, const int addrLen);
00671     static bool processPath(const XMLCh* const pathStr, const int pathStrLen,
00672                             const bool isSchemePresent);
00673 
00674     // -----------------------------------------------------------------------
00675     //  Data members
00676     //
00677     //  for all the data member, we own it,
00678     //  responsible for the creation and/or deletion for
00679     //  the memory allocated.
00680     //
00681     // -----------------------------------------------------------------------
00682     XMLCh*          fScheme;
00683     XMLCh*          fUserInfo;
00684     XMLCh*          fHost;
00685     int             fPort;
00686     XMLCh*          fRegAuth;
00687     XMLCh*          fPath;
00688     XMLCh*          fQueryString;
00689     XMLCh*          fFragment;
00690     XMLCh*          fURIText;
00691     MemoryManager*  fMemoryManager;
00692 };
00693 
00694 // ---------------------------------------------------------------------------
00695 //  XMLUri: Getter methods
00696 // ---------------------------------------------------------------------------
00697 inline const XMLCh* XMLUri::getScheme() const
00698 {
00699     return fScheme;
00700 }
00701 
00702 inline const XMLCh* XMLUri::getUserInfo() const
00703 {
00704     return fUserInfo;
00705 }
00706 
00707 inline const XMLCh* XMLUri::getHost() const
00708 {
00709     return fHost;
00710 }
00711 
00712 inline int XMLUri::getPort() const
00713 {
00714     return fPort;
00715 }
00716 
00717 inline const XMLCh* XMLUri::getRegBasedAuthority() const
00718 {
00719     return fRegAuth;
00720 }
00721 
00722 inline const XMLCh* XMLUri::getPath() const
00723 {
00724     return fPath;
00725 }
00726 
00727 inline const XMLCh* XMLUri::getQueryString() const
00728 {
00729     return fQueryString;
00730 }
00731 
00732 inline const XMLCh* XMLUri::getFragment() const
00733 {
00734     return fFragment;
00735 }
00736 
00737 inline const XMLCh* XMLUri::getUriText() const
00738 {
00739     //
00740     //  Fault it in if not already. Since this is a const method and we
00741     //  can't use mutable members due the compilers we have to support,
00742     //  we have to cast off the constness.
00743     //
00744     if (!fURIText)
00745         ((XMLUri*)this)->buildFullText();
00746 
00747     return fURIText;
00748 }
00749 
00750 // ---------------------------------------------------------------------------
00751 //  XMLUri: Helper methods
00752 // ---------------------------------------------------------------------------
00753 inline bool XMLUri::isReservedOrUnreservedCharacter(const XMLCh theChar)
00754 {
00755    return (XMLString::isAlphaNum(theChar) ||
00756            XMLString::indexOf(MARK_OR_RESERVED_CHARACTERS, theChar) != -1);
00757 }
00758 
00759 inline bool XMLUri::isReservedCharacter(const XMLCh theChar)
00760 {
00761     return (XMLString::indexOf(RESERVED_CHARACTERS, theChar) != -1);
00762 }
00763 
00764 inline bool XMLUri::isPathCharacter(const XMLCh theChar)
00765 {
00766     return (XMLString::indexOf(PATH_CHARACTERS, theChar) != -1);
00767 }
00768 
00769 inline bool XMLUri::isUnreservedCharacter(const XMLCh theChar)
00770 {
00771     return (XMLString::isAlphaNum(theChar) ||
00772             XMLString::indexOf(MARK_CHARACTERS, theChar) != -1);
00773 }
00774 
00775 XERCES_CPP_NAMESPACE_END
00776 
00777 #endif