TransService.hpp Source File

00001 /*
00002  * The Apache Software License, Version 1.1

00003  * 

00004  * Copyright (c) 1999-2000 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  * $Log: TransService.hpp,v $

00059  * Revision 1.12  2001/01/25 19:19:32  tng

00060  * Let user add their encoding to the intrinsic mapping table.  Added by Khaled Noaman.

00061  *
00062  * Revision 1.11  2000/04/12 22:57:45  roddey

00063  * A couple of fixes to comments and parameter names to make them

00064  * more correct.

00065  *
00066  * Revision 1.10  2000/03/28 19:43:19  roddey

00067  * Fixes for signed/unsigned warnings. New work for two way transcoding

00068  * stuff.

00069  *
00070  * Revision 1.9  2000/03/17 23:59:54  roddey

00071  * Initial updates for two way transcoding support

00072  *
00073  * Revision 1.8  2000/03/02 19:54:46  roddey

00074  * This checkin includes many changes done while waiting for the

00075  * 1.1.0 code to be finished. I can't list them all here, but a list is

00076  * available elsewhere.

00077  *
00078  * Revision 1.7  2000/02/24 20:05:25  abagchi

00079  * Swat for removing Log from API docs

00080  *
00081  * Revision 1.6  2000/02/06 07:48:04  rahulj

00082  * Year 2K copyright swat.

00083  *
00084  * Revision 1.5  2000/01/25 22:49:55  roddey

00085  * Moved the supportsSrcOfs() method from the individual transcoder to the

00086  * transcoding service, where it should have been to begin with.

00087  *
00088  * Revision 1.4  2000/01/25 19:19:07  roddey

00089  * Simple addition of a getId() method to the xcode and netacess abstractions to

00090  * allow each impl to give back an id string.

00091  *
00092  * Revision 1.3  1999/12/18 00:18:10  roddey

00093  * More changes to support the new, completely orthagonal support for

00094  * intrinsic encodings.

00095  *
00096  * Revision 1.2  1999/12/15 19:41:28  roddey

00097  * Support for the new transcoder system, where even intrinsic encodings are

00098  * done via the same transcoder abstraction as external ones.

00099  *
00100  * Revision 1.1.1.1  1999/11/09 01:05:16  twl

00101  * Initial checkin

00102  *
00103  * Revision 1.2  1999/11/08 20:45:16  rahul

00104  * Swat for adding in Product name and CVS comment log variable.

00105  *
00106  */
00107 
00108 #ifndef TRANSSERVICE_HPP
00109 #define TRANSSERVICE_HPP
00110 
00111 #include <util/XercesDefs.hpp>
00112 
00113 
00114 // Forward references
00115 class XMLPlatformUtils;
00116 class XMLLCPTranscoder;
00117 class XMLTranscoder;
00118 class ENameMap;
00119 
00120 
00121 //
00122 //  This class is an abstract base class which are used to abstract the
00123 //  transcoding services that Xerces uses. The parser's actual transcoding
00124 //  needs are small so it is desirable to allow different implementations
00125 //  to be provided.
00126 //
00127 //  The transcoding service has to provide a couple of required string
00128 //  and character operations, but its most important service is the creation
00129 //  of transcoder objects. There are two types of transcoders, which are
00130 //  discussed below in the XMLTranscoder class' description.
00131 //
00132 class  XMLTransService
00133 {
00134 public :
00135     // -----------------------------------------------------------------------
00136     //  Class specific types
00137     // -----------------------------------------------------------------------
00138     enum Codes
00139     {
00140         Ok
00141         , UnsupportedEncoding
00142         , InternalFailure
00143         , SupportFilesNotFound
00144     };
00145 
00146     struct TransRec
00147     {
00148         XMLCh       intCh;
00149         XMLByte     extCh;
00150     };
00151 
00152 
00153     // -----------------------------------------------------------------------
00154     //  Public constructors and destructor
00155     // -----------------------------------------------------------------------
00156     virtual ~XMLTransService();
00157 
00158 
00159     // -----------------------------------------------------------------------
00160     //  Non-virtual API
00161     // -----------------------------------------------------------------------
00162     XMLTranscoder* makeNewTranscoderFor
00163     (
00164         const   XMLCh* const            encodingName
00165         ,       XMLTransService::Codes& resValue
00166         , const unsigned int            blockSize
00167     );
00168 
00169     XMLTranscoder* makeNewTranscoderFor
00170     (
00171         const   char* const             encodingName
00172         ,       XMLTransService::Codes& resValue
00173         , const unsigned int            blockSize
00174     );
00175 
00176 
00177     // -----------------------------------------------------------------------
00178     //  The virtual transcoding service API
00179     // -----------------------------------------------------------------------
00180     virtual int compareIString
00181     (
00182         const   XMLCh* const    comp1
00183         , const XMLCh* const    comp2
00184     ) = 0;
00185 
00186     virtual int compareNIString
00187     (
00188         const   XMLCh* const    comp1
00189         , const XMLCh* const    comp2
00190         , const unsigned int    maxChars
00191     ) = 0;
00192 
00193     virtual const XMLCh* getId() const = 0;
00194 
00195     virtual bool isSpace(const XMLCh toCheck) const = 0;
00196 
00197     virtual XMLLCPTranscoder* makeNewLCPTranscoder() = 0;
00198 
00199     virtual bool supportsSrcOfs() const = 0;
00200 
00201     virtual void upperCase(XMLCh* const toUpperCase) const = 0;
00202 
00203     // -----------------------------------------------------------------------
00204     //  Allow users to add their own encodings to the intrinsinc mapping
00205     //  table
00206     //  Usage:
00207     //      XMLTransService::addEncoding (
00208     //          gMyEncodingNameString
00209     //          , new ENameMapFor<MyTransClassType>(gMyEncodingNameString)
00210     //      );
00211     // -----------------------------------------------------------------------
00212     static void addEncoding(const XMLCh* const encoding, ENameMap* const ownMapping);
00213 
00214 
00215 protected :
00216     // -----------------------------------------------------------------------
00217     //  Hidden constructors
00218     // -----------------------------------------------------------------------
00219     XMLTransService();
00220 
00221 
00222     // -----------------------------------------------------------------------
00223     //  Protected virtual methods.
00224     // -----------------------------------------------------------------------
00225     virtual XMLTranscoder* makeNewXMLTranscoder
00226     (
00227         const   XMLCh* const            encodingName
00228         ,       XMLTransService::Codes& resValue
00229         , const unsigned int            blockSize
00230     ) = 0;
00231 
00232 
00233 private :
00234     // -----------------------------------------------------------------------
00235     //  Unimplemented constructors and operators
00236     // -----------------------------------------------------------------------
00237     XMLTransService(const XMLTransService&);
00238     void operator=(const XMLTransService&);
00239 
00240 
00241     // -----------------------------------------------------------------------
00242     //  Hidden init method for platform utils to call
00243     // -----------------------------------------------------------------------
00244     friend class XMLPlatformUtils;
00245     void initTransService();
00246 };
00247 
00248 
00249 
00250 //
00251 //  This type of transcoder is for non-local code page encodings, i.e.
00252 //  named encodings. These are used internally by the scanner to internalize
00253 //  raw XML into the internal Unicode format, and by writer classes to
00254 //  convert that internal Unicode format (which comes out of the parser)
00255 //  back out to a format that the receiving client code wants to use.
00256 //
00257 class  XMLTranscoder
00258 {
00259 public :
00260     // -----------------------------------------------------------------------
00261     //  This enum is used by the transcodeTo() method to indicate how to
00262     //  react to unrepresentable characters. The transcodeFrom() method always
00263     //  works the same. It will consider any invalid data to be an error and
00264     //  throw.
00265     //
00266     //  The options mean:
00267     //      Throw   - Throw an exception
00268     //      RepChar - Use the replacement char
00269     // -----------------------------------------------------------------------
00270     enum UnRepOpts
00271     {
00272         UnRep_Throw
00273         , UnRep_RepChar
00274     };
00275 
00276 
00277     // -----------------------------------------------------------------------
00278     //  Public constructors and destructor
00279     // -----------------------------------------------------------------------
00280     virtual ~XMLTranscoder();
00281 
00282 
00283     // -----------------------------------------------------------------------
00284     //  The virtual transcoding interface
00285     // -----------------------------------------------------------------------
00286     virtual unsigned int transcodeFrom
00287     (
00288         const   XMLByte* const          srcData
00289         , const unsigned int            srcCount
00290         ,       XMLCh* const            toFill
00291         , const unsigned int            maxChars
00292         ,       unsigned int&           bytesEaten
00293         ,       unsigned char* const    charSizes
00294     ) = 0;
00295 
00296     virtual unsigned int transcodeTo
00297     (
00298         const   XMLCh* const    srcData
00299         , const unsigned int    srcCount
00300         ,       XMLByte* const  toFill
00301         , const unsigned int    maxBytes
00302         ,       unsigned int&   charsEaten
00303         , const UnRepOpts       options
00304     ) = 0;
00305 
00306     virtual bool canTranscodeTo
00307     (
00308         const   unsigned int    toCheck
00309     )   const = 0;
00310 
00311 
00312     // -----------------------------------------------------------------------
00313     //  Getter methods
00314     // -----------------------------------------------------------------------
00315     unsigned int getBlockSize() const;
00316 
00317     const XMLCh* getEncodingName() const;
00318 
00319 
00320 protected :
00321     // -----------------------------------------------------------------------
00322     //  Hidden constructors
00323     // -----------------------------------------------------------------------
00324     XMLTranscoder
00325     (
00326         const   XMLCh* const    encodingName
00327         , const unsigned int    blockSize
00328     );
00329 
00330 
00331     // -----------------------------------------------------------------------
00332     //  Protected helper methods
00333     // -----------------------------------------------------------------------
00334     void checkBlockSize(const unsigned int toCheck);
00335 
00336 
00337 private :
00338     // -----------------------------------------------------------------------
00339     //  Unimplemented constructors and operators
00340     // -----------------------------------------------------------------------
00341     XMLTranscoder(const XMLTranscoder&);
00342     void operator=(const XMLTranscoder&);
00343 
00344 
00345     // -----------------------------------------------------------------------
00346     //  Private data members
00347     //
00348     //  fBlockSize
00349     //      This is the block size indicated in the constructor. This lets
00350     //      the derived class preallocate appopriately sized buffers. This
00351     //      sets the maximum number of characters which can be internalized
00352     //      per call to transcodeFrom() and transcodeTo().
00353     //
00354     //  fEncodingName
00355     //      This is the name of the encoding this encoder is for. All basic
00356     //      XML transcoder's are for named encodings.
00357     // -----------------------------------------------------------------------
00358     unsigned int    fBlockSize;
00359     XMLCh*          fEncodingName;
00360 };
00361 
00362 
00363 //
00364 //  This class is a specialized transcoder that only transcodes between
00365 //  the internal XMLCh format and the local code page. It is specialized
00366 //  for the very common job of translating data from the client app's
00367 //  native code page to the internal format and vice versa.
00368 //
00369 class  XMLLCPTranscoder
00370 {
00371 public :
00372     // -----------------------------------------------------------------------
00373     //  Public constructors and destructor
00374     // -----------------------------------------------------------------------
00375     virtual ~XMLLCPTranscoder();
00376 
00377 
00378     // -----------------------------------------------------------------------
00379     //  The virtual transcoder API
00380     //
00381     //  NOTE:   All these APIs don't include null terminator characters in
00382     //          their parameters. So calcRequiredSize() returns the number
00383     //          of actual chars, not including the null. maxBytes and maxChars
00384     //          parameters refer to actual chars, not including the null so
00385     //          its assumed that the buffer is physically one char or byte
00386     //          larger.
00387     // -----------------------------------------------------------------------
00388     virtual unsigned int calcRequiredSize(const char* const srcText) = 0;
00389 
00390     virtual unsigned int calcRequiredSize(const XMLCh* const srcText) = 0;
00391 
00392     virtual char* transcode(const XMLCh* const toTranscode) = 0;
00393 
00394     virtual XMLCh* transcode(const char* const toTranscode) = 0;
00395 
00396     virtual bool transcode
00397     (
00398         const   char* const     toTranscode
00399         ,       XMLCh* const    toFill
00400         , const unsigned int    maxChars
00401     ) = 0;
00402 
00403     virtual bool transcode
00404     (
00405         const   XMLCh* const    toTranscode
00406         ,       char* const     toFill
00407         , const unsigned int    maxBytes
00408     ) = 0;
00409 
00410 
00411 protected :
00412     // -----------------------------------------------------------------------
00413     //  Hidden constructors
00414     // -----------------------------------------------------------------------
00415     XMLLCPTranscoder();
00416 
00417 
00418 private :
00419     // -----------------------------------------------------------------------
00420     //  Unimplemented constructors and operators
00421     // -----------------------------------------------------------------------
00422     XMLLCPTranscoder(const XMLLCPTranscoder&);
00423     void operator=(const XMLLCPTranscoder&);
00424 };
00425 
00426 
00427 // ---------------------------------------------------------------------------
00428 //  XMLTranscoder: Protected helper methods
00429 // ---------------------------------------------------------------------------
00430 inline unsigned int XMLTranscoder::getBlockSize() const

00431 {
00432     return fBlockSize;
00433 }
00434 
00435 inline const XMLCh* XMLTranscoder::getEncodingName() const

00436 {
00437     return fEncodingName;
00438 }
00439 
00440 #endif