avc_mbyte.c

用于读取TAB、MIF、SHP文件的类

        psDBCSInfo->nDBCSBufSize = nMaxOutputLen+2;
        psDBCSInfo->pszDBCSBuf = 
            (unsigned char *)CPLRealloc(psDBCSInfo->pszDBCSBuf,
                                        psDBCSInfo->nDBCSBufSize*
                                        sizeof(unsigned char));
    }

    /* Do the conversion according to current code page 
     */
    switch (psDBCSInfo->nDBCSCodePage)
    {
      case AVC_DBCS_JAPANESE:
        pszOutBuf = (char*)_AVCArcDBCS2JapaneseShiftJIS(psDBCSInfo,
                                                (const unsigned char *)pszLine,
                                                        nMaxOutputLen);
        break;
      default:
        /* We should never get here anyways, but just in case return pszLine 
         */
        pszOutBuf = pszLine;
    }
    
    return pszOutBuf;
}

/*=====================================================================
 *=====================================================================
 * Functions Specific to Japanese encoding (CodePage 932).  
 *
 * For now we assume that we can receive only Katakana, Shift-JIS, or EUC
 * encoding as input.  Coverages use EUC encoding in most cases, except
 * for Katakana characters that are prefixed with a 0x8e byte.
 *
 * Most of the Japanese conversion functions are based on information and
 * algorithms found at:
 *  http://www.mars.dti.ne.jp/~torao/program/appendix/japanese-en.html
 *=====================================================================
 *====================================================================*/

/**********************************************************************
 *                          _AVCDetectJapaneseEncoding()
 *
 * Scan a line of text to try to establish the type of japanese encoding
 *
 * Returns the encoding number (AVC_CODE_JAP_*), or AVC_CODE_UNKNOWN if no
 * specific encoding was detected.
 **********************************************************************/

#define IS_JAP_SHIFTJIS_1(c)  ((c) >= 0x81 && (c) <= 0x9f)
#define IS_JAP_SHIFTJIS_2(c)  (((c) >= 0x40 && (c) <= 0x7e) ||   \
                               ((c) >= 0x80 && (c) <= 0xA0) )
#define IS_JAP_EUC_1(c)       ((c) >= 0xF0 && (c) <= 0xFE)
#define IS_JAP_EUC_2(c)       ((c) >= 0xFD && (c) <= 0xFE)
#define IS_JAP_KANA(c)        ((c) >= 0xA1 && (c) <= 0xDF)

static int _AVCDetectJapaneseEncoding(const unsigned char *pszLine)
{
    int nEncoding = AVC_CODE_UNKNOWN;

    for( ; nEncoding == AVC_CODE_UNKNOWN && pszLine && *pszLine; pszLine++)
    {
        if (IS_ASCII(*pszLine))
            continue;
        else if (IS_JAP_SHIFTJIS_1(*pszLine))
        {
            nEncoding = AVC_CODE_JAP_SHIFTJIS;
            break;
        }
        else if (IS_JAP_KANA(*pszLine) && *(pszLine+1) &&
                 (IS_ASCII(*(pszLine+1)) || 
                  (*(pszLine+1)>=0x80 && *(pszLine+1)<=0xA0) ) )
        {
            nEncoding = AVC_CODE_JAP_SHIFTJIS; /* SHIFT-JIS + Kana */
            break;
        }
        else if (IS_JAP_EUC_1(*pszLine))
        {
            nEncoding = AVC_CODE_JAP_EUC;
            break;
        }

        if (*(++pszLine) == '\0')
            break;

        if (IS_JAP_SHIFTJIS_2(*pszLine))
        {
            nEncoding = AVC_CODE_JAP_SHIFTJIS;
            break;
        }
        else if (IS_JAP_EUC_2(*pszLine))
        {
            nEncoding = AVC_CODE_JAP_EUC;
            break;
        }
    }

    return nEncoding;
}


/**********************************************************************
 *                          _AVCJapanese2ArcDBCS()
 *
 * Try to detect type of Japanese encoding if not done yet, and convert
 * string from Japanese to proper coverage DBCS encoding.
 **********************************************************************/
static const char *_AVCJapanese2ArcDBCS(AVCDBCSInfo *psDBCSInfo, 
                                        const unsigned char *pszLine,
                                        int nMaxOutputLen)
{
    unsigned char *pszOut;
    int iDst;

    pszOut = psDBCSInfo->pszDBCSBuf;

    if (psDBCSInfo->nDBCSEncoding == AVC_CODE_UNKNOWN)
    {
        /* Type of encoding (Shift-JIS or EUC) not known yet... try to
         * detect it now.
         */
        psDBCSInfo->nDBCSEncoding = _AVCDetectJapaneseEncoding(pszLine);

/*
        if (psDBCSInfo->nDBCSEncoding == AVC_CODE_JAP_SHIFTJIS)
        {
            printf("Found Japanese Shift-JIS encoding\n");
        }
        else if (psDBCSInfo->nDBCSEncoding == AVC_CODE_JAP_EUC)
        {
            printf("Found Japanese EUC encoding\n");
        }
*/
    }

    for(iDst=0; *pszLine && iDst < nMaxOutputLen; pszLine++)
    {
        if (IS_ASCII(*pszLine))
        {
            /* No transformation required for ASCII */
            pszOut[iDst++] = *pszLine;
        }
        else if ( psDBCSInfo->nDBCSEncoding==AVC_CODE_JAP_EUC && *(pszLine+1) )
        {
            /* This must be a pair of EUC chars and both should be in
             * the range 0xA1-0xFE
             */
            pszOut[iDst++] = *(pszLine++);
            pszOut[iDst++] = *pszLine;
        }
        else if ( IS_JAP_KANA(*pszLine) )
        {
            /* Katakana char. prefix it with 0x8e */
            pszOut[iDst++] = 0x8e;
            pszOut[iDst++] = *pszLine;
        }
        else if ( *(pszLine+1) )
        {
            /* This must be a pair of Shift-JIS chars... convert them to EUC
             *
             * If we haven't been able to establish the encoding for sure
             * yet, then it is possible that a pair of EUC chars could be
             * treated as shift-JIS here... but there is not much we can do
             * about that unless we scan the whole E00 input before we
             * start the conversion.
             */
            unsigned char leader, trailer;
            leader = *(pszLine++);
            trailer = *pszLine;

            if(leader <= 0x9F)  leader -= 0x71;
            else                leader -= 0xB1;
            leader = (leader << 1) + 1;

            if(trailer > 0x7F)  trailer --;
            if(trailer >= 0x9E)
            {
                trailer -= 0x7D;
                leader ++;
            }
            else 
            {
                trailer -= 0x1F;
            }  

            pszOut[iDst++] = leader | 0x80;
            pszOut[iDst++] = trailer | 0x80;
        }
        else
        {
            /* We should never get here unless a double-byte pair was 
             * truncated... but just in case...
             */
            pszOut[iDst++] = *pszLine;
        }

    }

    pszOut[iDst] = '\0';

    return psDBCSInfo->pszDBCSBuf;
}


/**********************************************************************
 *                          _AVCArcDBCS2JapaneseShiftJIS()
 *
 * Convert string from coverage DBCS (EUC) to Japanese Shift-JIS.
 *
 * We know that binary coverages use a custom EUC encoding for japanese
 * which is EUC + all Katakana chars are prefixed with 0x8e.  So this
 * function just does a simple conversion.
 **********************************************************************/
static const char *_AVCArcDBCS2JapaneseShiftJIS(AVCDBCSInfo *psDBCSInfo, 
                                                const unsigned char *pszLine,
                                                int nMaxOutputLen)
{
    unsigned char *pszOut;
    int iDst;

    pszOut = psDBCSInfo->pszDBCSBuf;

    for(iDst=0; *pszLine && iDst < nMaxOutputLen; pszLine++)
    {
        if (IS_ASCII(*pszLine))
        {
            /* No transformation required for ASCII */
            pszOut[iDst++] = *pszLine;            
        }
        else if (*pszLine == 0x8e && *(pszLine+1))
        {
            pszLine++;  /* Flush the 0x8e */
            pszOut[iDst++] = *pszLine;
        }
        else if (*(pszLine+1))
        {
            /* This is a pair of EUC chars... convert them to Shift-JIS 
             */
            unsigned char leader, trailer;
            leader  = *(pszLine++) & 0x7F;
            trailer = *pszLine & 0x7F;

            if((leader & 0x01) != 0)    trailer += 0x1F;
            else                        trailer += 0x7D;
            if(trailer >= 0x7F)         trailer ++;
    
            leader = ((leader - 0x21) >> 1) + 0x81;
            if(leader > 0x9F)          leader += 0x40;

            pszOut[iDst++] = leader;
            pszOut[iDst++] = trailer;
        }
        else
        {
            /* We should never get here unless a double-byte pair was 
             * truncated... but just in case...
             */
            pszOut[iDst++] = *pszLine;
        }

    }

    pszOut[iDst] = '\0';

    return psDBCSInfo->pszDBCSBuf;
}