icutransservice.cpp

来自「IBM的解析xml的工具Xerces的源代码」· C++ 代码 · 共 1,358 行 · 第 1/3 页

/*
 * Copyright 1999-2004 The Apache Software Foundation.
 * 
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 * 
 *      http://www.apache.org/licenses/LICENSE-2.0
 * 
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

/*
 * $Id: ICUTransService.cpp,v 1.16 2004/09/08 13:56:44 peiyongz Exp $
 */


// ---------------------------------------------------------------------------
//  Includes
// ---------------------------------------------------------------------------
#include <xercesc/util/Janitor.hpp>
#include <xercesc/util/TranscodingException.hpp>
#include <xercesc/util/XMLString.hpp>
#include <xercesc/util/XMLUniDefs.hpp>
#include "ICUTransService.hpp"
#include <string.h>
#include <unicode/uloc.h>
#include <unicode/uchar.h>
#include <unicode/ucnv.h>
#include <unicode/ucnv_err.h>
#include <unicode/ustring.h>
#include <unicode/udata.h>
#if (U_ICU_VERSION_MAJOR_NUM >= 2)
    #include <unicode/uclean.h>
#endif

#if !defined(XML_OS390) && !defined(XML_AS400) && !defined(XML_HPUX) && !defined(XML_PTX)
// Forward reference the symbol which points to the ICU converter data.
#if (U_ICU_VERSION_MAJOR_NUM < 2)
extern "C" const uint8_t U_IMPORT icudata_dat[];
#endif
#endif

XERCES_CPP_NAMESPACE_BEGIN

// ---------------------------------------------------------------------------
//  Local, const data
// ---------------------------------------------------------------------------
static const XMLCh gMyServiceId[] =
{
    chLatin_I, chLatin_C, chLatin_U, chNull
};

static const XMLCh gS390Id[] =
{
    chLatin_S, chDigit_3, chDigit_9, chDigit_0, chNull
};

static const XMLCh gs390Id[] =
{
    chLatin_s, chDigit_3, chDigit_9, chDigit_0, chNull
};

static const XMLCh gswaplfnlId[] =
{
    chComma, chLatin_s, chLatin_w, chLatin_a, chLatin_p,
    chLatin_l, chLatin_f, chLatin_n, chLatin_l, chNull
};
// ---------------------------------------------------------------------------
//  Local functions
// ---------------------------------------------------------------------------

//
//  When XMLCh and ICU's UChar are not the same size, we have to do a temp
//  conversion of all strings. These local helper methods make that easier.
//
static UChar* convertToUChar( const   XMLCh* const    toConvert
                            , const unsigned int    srcLen = 0
                            , MemoryManager* const manager = 0)
{
    const unsigned int actualLen = srcLen
                                   ? srcLen : XMLString::stringLen(toConvert);

    UChar* tmpBuf = (manager)
        ? (UChar*) manager->allocate((actualLen + 1) * sizeof(UChar))
		: new UChar[actualLen + 1];
    const XMLCh* srcPtr = toConvert;
    UChar* outPtr = tmpBuf;
    while (*srcPtr)
        *outPtr++ = UChar(*srcPtr++);
    *outPtr = 0;

    return tmpBuf;
}


static XMLCh* convertToXMLCh( const UChar* const toConvert,
                            MemoryManager* const manager = 0)
{
    const unsigned int srcLen = u_strlen(toConvert);
    XMLCh* retBuf = (manager)
        ? (XMLCh*) manager->allocate((srcLen+1) * sizeof(XMLCh))
        : new XMLCh[srcLen + 1];

    XMLCh* outPtr = retBuf;
    const UChar* srcPtr = toConvert;
    while (*srcPtr)
        *outPtr++ = XMLCh(*srcPtr++);
    *outPtr = 0;

    return retBuf;
}




// ---------------------------------------------------------------------------
//  ICUTransService: Constructors and Destructor
// ---------------------------------------------------------------------------
ICUTransService::ICUTransService()
{
#if !defined(XML_OS390) && !defined(XML_AS400) && !defined(XML_HPUX) && !defined(XML_PTX)
#if (U_ICU_VERSION_MAJOR_NUM < 2)
    // Starting with ICU 2.0, ICU itself includes a static reference to the data
    // entrypoint symbol.
    //
    // ICU 1.8 (and previous) did not include a static reference, but would
    // dynamically load the data dll when it was first needed, however this dynamic
    // loading proved unreliable in some of the odd environments that Xerces needed
    // to run in.  Hence, the static reference.

    // Pass the location of the converter data to ICU. By doing so, we are
    // forcing the load of ICU converter data DLL, after the Xerces-C DLL is
    // loaded. This implies that Xerces-C, now has to explicitly link with the
    // ICU converter dll. However, the advantage is that we no longer depend
    // on the code which does demand dynamic loading of DLL's. The demand
    // loading is highly system dependent and was a constant source of support
    // calls.
    UErrorCode uerr = U_ZERO_ERROR;
    udata_setCommonData((void *) icudata_dat, &uerr);
#endif
#endif
}

ICUTransService::~ICUTransService()
{
    /*
     * commented out the following clean up code
     * in case users use ICU outside of the parser
     * if we clean up here, users' code may crash
     *
    #if (U_ICU_VERSION_MAJOR_NUM >= 2)
        // release all lasily allocated data
        u_cleanup();
    #endif
    */
}


// ---------------------------------------------------------------------------
//  ICUTransService: The virtual transcoding service API
// ---------------------------------------------------------------------------
int ICUTransService::compareIString(const   XMLCh* const    comp1
                                    , const XMLCh* const    comp2)
{
    const XMLCh* psz1 = comp1;
    const XMLCh* psz2 = comp2;

    unsigned int curCount = 0;
    while (true)
    {
        //
        //  If an inequality, then return the difference. Note that the XMLCh
        //  might be bigger physically than UChar, but it won't hold anything
        //  larger than 0xFFFF, so our cast here will work for both possible
        //  sizes of XMLCh.
        //
        if (u_toupper(UChar(*psz1)) != u_toupper(UChar(*psz2)))
            return int(*psz1) - int(*psz2);

        // If either has ended, then they both ended, so equal
        if (!*psz1 || !*psz2)
            break;

        // Move upwards for the next round
        psz1++;
        psz2++;
    }
    return 0;
}


int ICUTransService::compareNIString(const  XMLCh* const    comp1
                                    , const XMLCh* const    comp2
                                    , const unsigned int    maxChars)
{
    const XMLCh* psz1 = comp1;
    const XMLCh* psz2 = comp2;

    unsigned int curCount = 0;
    while (true)
    {
        //
        //  If an inequality, then return the difference. Note that the XMLCh
        //  might be bigger physically than UChar, but it won't hold anything
        //  larger than 0xFFFF, so our cast here will work for both possible
        //  sizes of XMLCh.
        //
        if (u_toupper(UChar(*psz1)) != u_toupper(UChar(*psz2)))
            return int(*psz1) - int(*psz2);

        // If either ended, then both ended, so equal
        if (!*psz1 || !*psz2)
            break;

        // Move upwards to next chars
        psz1++;
        psz2++;

        //
        //  Bump the count of chars done. If it equals the count then we
        //  are equal for the requested count, so break out and return
        //  equal.
        //
        curCount++;
        if (maxChars == curCount)
            break;
    }
    return 0;
}


const XMLCh* ICUTransService::getId() const
{
    return gMyServiceId;
}


bool ICUTransService::isSpace(const XMLCh toCheck) const
{
    //
    //  <TBD>
    //  For now, we short circuit some of the control chars because ICU
    //  is not correctly reporting them as space. Later, when they change
    //  this, we can get rid of this special case.
    //
    if ((toCheck == 0x09)
    ||  (toCheck == 0x0A)
    ||  (toCheck == 0x0D))
    {
        return true;
    }
    return (u_isspace(UChar(toCheck)) != 0);
}


XMLLCPTranscoder* ICUTransService::makeNewLCPTranscoder()
{
    //
    //  Try to create a default converter. If it fails, return a null
    //  pointer which will basically cause the system to give up because
    //  we really can't do anything without one.
    //
    UErrorCode uerr = U_ZERO_ERROR;
    UConverter* converter = ucnv_open(NULL, &uerr);
    if (!converter)
        return 0;

    // That went ok, so create an ICU LCP transcoder wrapper and return it
    return new ICULCPTranscoder(converter);
}


bool ICUTransService::supportsSrcOfs() const
{
    // This implementation supports source offset information
    return true;
}


void ICUTransService::upperCase(XMLCh* const toUpperCase) const
{
    XMLCh* outPtr = toUpperCase;
    while (*outPtr)
    {
        *outPtr = XMLCh(u_toupper(UChar(*outPtr)));
        outPtr++;
    }
}

void ICUTransService::lowerCase(XMLCh* const toLowerCase) const
{
    XMLCh* outPtr = toLowerCase;
    while (*outPtr)
    {
        *outPtr = XMLCh(u_tolower(UChar(*outPtr)));
        outPtr++;
    }
}



// ---------------------------------------------------------------------------
//  ICUTransService: The protected virtual transcoding service API
// ---------------------------------------------------------------------------
XMLTranscoder* ICUTransService::
makeNewXMLTranscoder(const  XMLCh* const            encodingName
                    ,       XMLTransService::Codes& resValue
                    , const unsigned int            blockSize
                    ,       MemoryManager* const    manager)
{
    //  
    //  For encodings that end with "s390" we need to strip off the "s390" 
    //  from the encoding name and add ",swaplfnl" to the encoding name	
    //  that we pass into ICU on the ucnv_openU.  
    //  
    XMLCh* encodingNameToUse = (XMLCh*) encodingName;
    XMLCh* workBuffer = 0;

    if ( (XMLString::endsWith(encodingNameToUse, gs390Id)) ||
         (XMLString::endsWith(encodingNameToUse, gS390Id)) )
    {
       int workBufferSize = (XMLString::stringLen(encodingNameToUse) + XMLString::stringLen(gswaplfnlId) - XMLString::stringLen(gS390Id) + 1);
       workBuffer = (XMLCh*) manager->allocate(workBufferSize * sizeof(XMLCh));
       int moveSize = XMLString::stringLen(encodingNameToUse) - XMLString::stringLen(gS390Id);
       XMLString::moveChars(workBuffer, encodingNameToUse, moveSize);
       XMLString::moveChars((workBuffer + moveSize), gswaplfnlId, XMLString::stringLen(gswaplfnlId));
       encodingNameToUse = workBuffer;
    }

    //
    //  If UChar and XMLCh are not the same size, then we have premassage the
    //  encoding name into a UChar type string.
    //
    const UChar* actualName;
    UChar* tmpName = 0;
    if (sizeof(UChar) == sizeof(XMLCh))
    {
        actualName = (const UChar*)encodingNameToUse;
    }
    else
    {
        tmpName = convertToUChar(encodingNameToUse, 0, manager);
        actualName = tmpName;
    }

    ArrayJanitor<UChar> janTmp(tmpName, manager);
    ArrayJanitor<XMLCh> janTmp1(workBuffer, manager);

    UErrorCode uerr = U_ZERO_ERROR;
    UConverter* converter = ucnv_openU(actualName, &uerr);
    if (!converter)
    {
        resValue = XMLTransService::UnsupportedEncoding;
        return 0;
    }

    return new (manager) ICUTranscoder(encodingName, converter, blockSize, manager);
}




// ---------------------------------------------------------------------------
//  ICUTranscoder: Constructors and Destructor
// ---------------------------------------------------------------------------
ICUTranscoder::ICUTranscoder(const  XMLCh* const        encodingName
                            ,       UConverter* const   toAdopt
                            , const unsigned int        blockSize
                            , MemoryManager* const      manager) :

    XMLTranscoder(encodingName, blockSize, manager)
    , fConverter(toAdopt)
    , fFixed(false)
    , fSrcOffsets(0)
{
    // If there is a block size, then allocate our source offset array
    if (blockSize)
        fSrcOffsets = (XMLUInt32*) manager->allocate
        (
            blockSize * sizeof(XMLUInt32)
        );//new XMLUInt32[blockSize];

    // Remember if its a fixed size encoding
    fFixed = (ucnv_getMaxCharSize(fConverter) == ucnv_getMinCharSize(fConverter));
}

ICUTranscoder::~ICUTranscoder()
{
    getMemoryManager()->deallocate(fSrcOffsets);//delete [] fSrcOffsets;

    // If there is a converter, ask ICU to clean it up
    if (fConverter)
    {
        // <TBD> Does this actually delete the structure???
        ucnv_close(fConverter);
        fConverter = 0;
    }
}


// ---------------------------------------------------------------------------
//  ICUTranscoder: The virtual transcoder API
// ---------------------------------------------------------------------------
unsigned int
ICUTranscoder::transcodeFrom(const  XMLByte* const          srcData
                            , const unsigned int            srcCount
                            ,       XMLCh* const            toFill
                            , const unsigned int            maxChars
                            ,       unsigned int&           bytesEaten
                            ,       unsigned char* const    charSizes)
{
    // If debugging, insure the block size is legal
    #if defined(XERCES_DEBUG)
    checkBlockSize(maxChars);
    #endif

    // Set up pointers to the start and end of the source buffer
    const XMLByte*  startSrc = srcData;
    const XMLByte*  endSrc = srcData + srcCount;

    //
    //  And now do the target buffer. This works differently according to
    //  whether XMLCh and UChar are the same size or not.
    //
    UChar* startTarget;
    if (sizeof(XMLCh) == sizeof(UChar))
        startTarget = (UChar*)toFill;
     else
        startTarget = (UChar*) getMemoryManager()->allocate
        (
            maxChars * sizeof(UChar)
        );//new UChar[maxChars];
    UChar* orgTarget = startTarget;

    //
    //  Transoode the buffer.  Buffer overflow errors are normal, occuring
    //  when the raw input buffer holds more characters than will fit in
    //  the Unicode output buffer.
    //
    UErrorCode  err = U_ZERO_ERROR;
    ucnv_toUnicode
    (
        fConverter
        , &startTarget
        , startTarget + maxChars
        , (const char**)&startSrc
        , (const char*)endSrc
        , (fFixed ? 0 : (int32_t*)fSrcOffsets)