ICUTransService.cpp

/*
 * The Apache Software License, Version 1.1
 *
 * Copyright (c) 1999-2000 The Apache Software Foundation.  All rights
 * reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 *
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in
 *    the documentation and/or other materials provided with the
 *    distribution.
 *
 * 3. The end-user documentation included with the redistribution,
 *    if any, must include the following acknowledgment:
 *       "This product includes software developed by the
 *        Apache Software Foundation (http://www.apache.org/)."
 *    Alternately, this acknowledgment may appear in the software itself,
 *    if and wherever such third-party acknowledgments normally appear.
 *
 * 4. The names "Xerces" and "Apache Software Foundation" must
 *    not be used to endorse or promote products derived from this
 *    software without prior written permission. For written
 *    permission, please contact apache\@apache.org.
 *
 * 5. Products derived from this software may not be called "Apache",
 *    nor may "Apache" appear in their name, without prior written
 *    permission of the Apache Software Foundation.
 *
 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED OR IMPLIED
 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED.  IN NO EVENT SHALL THE APACHE SOFTWARE FOUNDATION OR
 * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
 * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 * ====================================================================
 *
 * This software consists of voluntary contributions made by many
 * individuals on behalf of the Apache Software Foundation, and was
 * originally based on software copyright (c) 1999, International
 * Business Machines, Inc., http://www.ibm.com .  For more information
 * on the Apache Software Foundation, please see
 * <http://www.apache.org/>.
 */

/*
 * $Id$
 */


// ---------------------------------------------------------------------------
//  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
};



// ---------------------------------------------------------------------------
//  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)
{
    const unsigned int srcLen = u_strlen(toConvert);
    XMLCh* retBuf = 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)
{
    //
    //  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*)encodingName;
    }
     else
    {
        tmpName = convertToUChar(encodingName);
        actualName = tmpName;
    }

    ArrayJanitor<UChar> janTmp(tmpName);

    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);
}




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

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

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

ICUTranscoder::~ICUTranscoder()
{
    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 = 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)
        , false
        , &err
    );

    if ((err != U_ZERO_ERROR) && (err != U_BUFFER_OVERFLOW_ERROR))
    {
        if (orgTarget != (UChar*)toFill)
            delete [] orgTarget;

        if (fFixed)
        {
            XMLCh tmpBuf[16];
            XMLString::binToText((unsigned int)(*startTarget), tmpBuf, 16, 16);
            ThrowXML2
            (
                TranscodingException
                , XMLExcepts::Trans_BadSrcCP
                , tmpBuf
                , getEncodingName()
            );
        }
         else
        {
            ThrowXML(TranscodingException, XMLExcepts::Trans_BadSrcSeq);
        }
    }

    // Calculate the bytes eaten and store in caller's param
    bytesEaten = startSrc - srcData;

    // And the characters decoded
    const unsigned int charsDecoded = startTarget - orgTarget;

    //
    //  Translate the array of char offsets into an array of character
    //  sizes, which is what the transcoder interface semantics requires.
    //  If its fixed, then we can optimize it.
    //
    if (fFixed)
    {
        const unsigned char fillSize = (unsigned char)ucnv_getMaxCharSize(fConverter);;
        memset(charSizes, fillSize, maxChars);
    }
     else
    {
        //
        //  We have to convert the series of offsets into a series of
        //  sizes. If just one char was decoded, then its the total bytes
        //  eaten. Otherwise, do a loop and subtract out each element from
        //  its previous element.
        //
        if (charsDecoded == 1)
        {
            charSizes[0] = (unsigned char)bytesEaten;
        }
         else
        {
            //  ICU does not return an extra element to allow us to figure
            //  out the last char size, so we have to compute it from the
            //  total bytes used.
            unsigned int index;
            for (index = 0; index < charsDecoded - 1; index++)
            {
                charSizes[index] = (unsigned char)(fSrcOffsets[index + 1]
                                                    - fSrcOffsets[index]);
            }
            if( charsDecoded > 0 ) {
                charSizes[charsDecoded - 1] = (unsigned char)(bytesEaten
                                              - fSrcOffsets[charsDecoded - 1]);
            }
        }
    }

    //
    //  If XMLCh and UChar are not the same size, then we need to copy over
    //  the temp buffer to the new one.
    //
    if (sizeof(UChar) != sizeof(XMLCh))
    {
        XMLCh* outPtr = toFill;
        startTarget = orgTarget;
        for (unsigned int index = 0; index < charsDecoded; index++)
            *outPtr++ = XMLCh(*startTarget++);

        // And delete the temp buffer
        delete [] orgTarget;
    }

    // Return the chars we put into the target buffer
    return charsDecoded;
}


unsigned int
ICUTranscoder::transcodeTo( const   XMLCh* const    srcData
                            , const unsigned int    srcCount
                            ,       XMLByte* const  toFill
                            , const unsigned int    maxBytes
                            ,       unsigned int&   charsEaten
                            , const UnRepOpts       options)
{
    //
    //  Get a pointer to the buffer to transcode. If UChar and XMLCh are
    //  the same size here, then use the original. Else, create a temp
    //  one and put a janitor on it.
    //
    const UChar* srcPtr;
    UChar* tmpBufPtr = 0;
    if (sizeof(XMLCh) == sizeof(UChar))
    {
        srcPtr = (const UChar*)srcData;
    }
     else
    {
        tmpBufPtr = convertToUChar(srcData, srcCount);
        srcPtr = tmpBufPtr;
    }
    ArrayJanitor<UChar> janTmpBuf(tmpBufPtr);

    //
    //  Set the appropriate callback so that it will either fail or use
    //  the rep char. Remember the old one so we can put it back.
    //
    UErrorCode  err = U_ZERO_ERROR;
    UConverterFromUCallback oldCB = NULL;
    #if (U_ICU_VERSION_MAJOR_NUM < 2)
    void* orgContent;
    #else
    const void* orgContent;
    #endif
    ucnv_setFromUCallBack
    (
        fConverter
        , (options == UnRep_Throw) ? UCNV_FROM_U_CALLBACK_STOP
                                   : UCNV_FROM_U_CALLBACK_SUBSTITUTE
        , NULL
        , &oldCB
        , &orgContent
        , &err
    );

    //
    //  Ok, lets transcode as many chars as we we can in one shot. The
    //  ICU API gives enough info not to have to do this one char by char.
    //
    XMLByte*        startTarget = toFill;
    const UChar*    startSrc = srcPtr;
    err = U_ZERO_ERROR;
    ucnv_fromUnicode
    (
        fConverter
        , (char**)&startTarget
        , (char*)(startTarget + maxBytes)
        , &startSrc
        , srcPtr + srcCount
        , 0
        , false
        , &err
    );

    // Rememember the status before we possibly overite the error code
    const bool res = (err == U_ZERO_ERROR);

    // Put the old handler back
    err = U_ZERO_ERROR;
    UConverterFromUCallback orgAction = NULL;

    ucnv_setFromUCallBack(fConverter, oldCB, NULL, &orgAction, &orgContent, &err);

    if (!res)
    {
        XMLCh tmpBuf[16];
        XMLString::binToText((unsigned int)*startSrc, tmpBuf, 16, 16);
        ThrowXML2
        (
            TranscodingException
            , XMLExcepts::Trans_Unrepresentable
            , tmpBuf
            , getEncodingName()
        );
    }

    // Fill in the chars we ate from the input
    charsEaten = startSrc - srcPtr;

    // Return the chars we stored
    return startTarget - toFill;
}


bool ICUTranscoder::canTranscodeTo(const unsigned int toCheck) const
{
    //
    //  If the passed value is really a surrogate embedded together, then
    //  we need to break it out into its two chars. Else just one. While
    //  we are ate it, convert them to UChar format if required.
    //
    UChar           srcBuf[2];
    unsigned int    srcCount = 1;
    if (toCheck & 0xFFFF0000)
    {
        srcBuf[0] = UChar((toCheck >> 10) + 0xD800);
        srcBuf[1] = UChar(toCheck & 0x3FF) + 0xDC00;
        srcCount++;
    }
     else
    {
        srcBuf[0] = UChar(toCheck);
    }

    //
    //  Set the callback so that it will fail instead of using the rep char.
    //  Remember the old one so we can put it back.
    //
     UErrorCode  err = U_ZERO_ERROR;
     UConverterFromUCallback oldCB = NULL;
     #if (U_ICU_VERSION_MAJOR_NUM < 2)
     void* orgContent;
     #else
     const void* orgContent;
     #endif

     ucnv_setFromUCallBack
         (
         fConverter
         , UCNV_FROM_U_CALLBACK_STOP
         , NULL
         , &oldCB
         , &orgContent
         , &err
         );

    // Set upa temp buffer to format into. Make it more than big enough
    char            tmpBuf[64];
    char*           startTarget = tmpBuf;
    const UChar*    startSrc = srcBuf;

    err = U_ZERO_ERROR;
    ucnv_fromUnicode
    (
        fConverter
        , &startTarget
        , startTarget + 64
        , &startSrc
        , srcBuf + srcCount
        , 0
        , false
        , &err
    );

    // Save the result before we overight the error code
    const bool res = (err == U_ZERO_ERROR);

    // Put the old handler back
    err = U_ZERO_ERROR;
    UConverterFromUCallback orgAction = NULL;

    ucnv_setFromUCallBack(fConverter, oldCB, NULL, &orgAction, &orgContent, &err);

    return res;
}



// ---------------------------------------------------------------------------
//  ICULCPTranscoder: Constructors and Destructor
// ---------------------------------------------------------------------------
ICULCPTranscoder::ICULCPTranscoder(UConverter* const toAdopt) :

    fConverter(toAdopt)
{
}

ICULCPTranscoder::~ICULCPTranscoder()
{
    // 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;
    }
}


// ---------------------------------------------------------------------------
//  ICULCPTranscoder: Constructors and Destructor
// ---------------------------------------------------------------------------
unsigned int ICULCPTranscoder::calcRequiredSize(const XMLCh* const srcText)
{
    if (!srcText)
        return 0;

    //
    //  We do two different versions of this, according to whether XMLCh
    //  is the same size as UChar or not.
    //
    UErrorCode err = U_ZERO_ERROR;
    int32_t targetCap;
    if (sizeof(XMLCh) == sizeof(UChar))
    {
        // Use a faux scope to synchronize while we do this
        {
            XMLMutexLock lockConverter(&fMutex);

            targetCap = ucnv_fromUChars
            (
                fConverter
                , 0
                , 0
                , (const UChar*)srcText
                , -1
                , &err
            );
        }
    }
     else
    {
        // Copy the source to a local temp
        UChar* tmpBuf = convertToUChar(srcText);
        ArrayJanitor<UChar> janTmp(tmpBuf);

        // Use a faux scope to synchronize while we do this
        {
            XMLMutexLock lockConverter(&fMutex);

            targetCap = ucnv_fromUChars
            (
                fConverter
                , 0
                , 0
                , tmpBuf
                , -1
                , &err
            );
        }
    }

    if (err != U_BUFFER_OVERFLOW_ERROR)
        return 0;

    return (unsigned int)targetCap;
}

unsigned int ICULCPTranscoder::calcRequiredSize(const char* const srcText)
{
    if (!srcText)
        return 0;

    int32_t targetCap;
    UErrorCode err = U_ZERO_ERROR;

    // Use a faux scope to synchronize while we do this
    {
        XMLMutexLock lockConverter(&fMutex);
        targetCap = ucnv_toUChars
        (
            fConverter
            , 0
            , 0
            , srcText
            , strlen(srcText)
            , &err
        );
    }

    if (err != U_BUFFER_OVERFLOW_ERROR)
        return 0;

#if (U_ICU_VERSION_MAJOR_NUM < 2)
    // Subtract one since it includes the terminator space
    return (unsigned int)(targetCap - 1);
#else
    // Starting ICU 2.0, this is fixed and all ICU String functions have consistent NUL-termination behavior.
    // The returned length is always the number of output UChar's, not counting an additional, terminating NUL.
    return (unsigned int)(targetCap);
#endif
}


char* ICULCPTranscoder::transcode(const XMLCh* const toTranscode)
{
    char* retBuf = 0;

    // Check for a couple of special cases
    if (!toTranscode)
        return retBuf;

    if (!*toTranscode)
    {
        retBuf = new char[1];
        retBuf[0] = 0;
        return retBuf;
    }

    //
    //  Get the length of the source string since we'll have to use it in
    //  a couple places below.
    //
    const unsigned int srcLen = XMLString::stringLen(toTranscode);

    //
    //  If XMLCh and UChar are not the same size, then we have to make a
    //  temp copy of the text to pass to ICU.
    //
    const UChar* actualSrc;
    UChar* ncActual = 0;
    if (sizeof(XMLCh) == sizeof(UChar))
    {
        actualSrc = (const UChar*)toTranscode;
    }
     else
    {
        // Allocate a non-const temp buf, but store it also in the actual
        ncActual = convertToUChar(toTranscode);
        actualSrc = ncActual;
    }

    // Insure that the temp buffer, if any, gets cleaned up via the nc pointer
    ArrayJanitor<UChar> janTmp(ncActual);

    // Caculate a return buffer size not too big, but less likely to overflow
    int32_t targetLen = (int32_t)(srcLen * 1.25);

    // Allocate the return buffer
    retBuf = new char[targetLen + 1];

    //
    //  Lock now while we call the converter. Use a faux block to do the
    //  lock so that it unlocks immediately afterwards.
    //
    UErrorCode err = U_ZERO_ERROR;
    int32_t targetCap;
    {
        XMLMutexLock lockConverter(&fMutex);

        targetCap = ucnv_fromUChars
        (
            fConverter
            , retBuf
            , targetLen + 1
            , actualSrc
            , -1
            , &err
        );
    }

    // If targetLen is not enough then buffer overflow might occur
    if (err == U_BUFFER_OVERFLOW_ERROR)
    {
        //
        //  Reset the error, delete the old buffer, allocate a new one,
        //  and try again.
        //
        err = U_ZERO_ERROR;
        delete [] retBuf;
        retBuf = new char[targetCap + 1];

        // Lock again before we retry
        XMLMutexLock lockConverter(&fMutex);
        targetCap = ucnv_fromUChars
        (
            fConverter
            , retBuf
            , targetCap
            , actualSrc
            , -1
            , &err
        );
    }

    if (U_FAILURE(err))
    {
        delete [] retBuf;
        return 0;
    }

    // Cap it off and return
    retBuf[targetCap] = 0;
    return retBuf;
}

char* ICULCPTranscoder::transcode(const XMLCh* const toTranscode,
                                  MemoryManager* const manager)
{
    char* retBuf = 0;

    // Check for a couple of special cases
    if (!toTranscode)
        return retBuf;

    if (!*toTranscode)
    {
        retBuf = new char[1];
        retBuf[0] = 0;
        return retBuf;
    }

    //
    //  Get the length of the source string since we'll have to use it in
    //  a couple places below.
    //
    const unsigned int srcLen = XMLString::stringLen(toTranscode);

    //
    //  If XMLCh and UChar are not the same size, then we have to make a
    //  temp copy of the text to pass to ICU.
    //
    const UChar* actualSrc;
    UChar* ncActual = 0;
    if (sizeof(XMLCh) == sizeof(UChar))
    {
        actualSrc = (const UChar*)toTranscode;
    }
     else
    {
        // Allocate a non-const temp buf, but store it also in the actual
        ncActual = convertToUChar(toTranscode);
        actualSrc = ncActual;
    }

    // Insure that the temp buffer, if any, gets cleaned up via the nc pointer
    ArrayJanitor<UChar> janTmp(ncActual);

    // Caculate a return buffer size not too big, but less likely to overflow
    int32_t targetLen = (int32_t)(srcLen * 1.25);

    // Allocate the return buffer
    retBuf = new char[targetLen + 1];

    //
    //  Lock now while we call the converter. Use a faux block to do the
    //  lock so that it unlocks immediately afterwards.
    //
    UErrorCode err = U_ZERO_ERROR;
    int32_t targetCap;
    {
        XMLMutexLock lockConverter(&fMutex);

        targetCap = ucnv_fromUChars
        (
            fConverter
            , retBuf
            , targetLen + 1
            , actualSrc
            , -1
            , &err
        );
    }