old-autovect-branch/libjava/gnu/gcj/convert/natInput_EUCJIS.cc

   1 /* Copyright (C) 1999  Free Software Foundation
   2
   3    This file is part of libgcj.
   4
   5 This software is copyrighted work licensed under the terms of the
   6 Libgcj License.  Please consult the file "LIBGCJ_LICENSE" for
   7 details.  */
   8
   9 #include <config.h>
  10 #include <gcj/cni.h>
  11 #include <gnu/gcj/convert/Input_EUCJIS.h>
  12
  13 #define ERROR_CHAR 0xFFFD
  14
  15 extern unsigned short JIS0208_to_Unicode[84][94];
  16 extern unsigned short JIS0212_to_Unicode[76][94];
  17
  18 jint
  19 gnu::gcj::convert::Input_EUCJIS::read(jcharArray outbuffer, jint outpos,
  20                                       jint count)
  21 {
  22   jint start_outpos = outpos;
  23   for (;;)
  24     {
  25       if (outpos - start_outpos >= count)
  26         break;
  27       if (inpos >= inlength)
  28         break;
  29       int b = ((unsigned char*) elements(inbuffer))[inpos++];
  30       if (codeset == 0)  // ASCII or JIS-Roman
  31         {
  32           if (b < 128)
  33             {
  34 #if 1
  35               // Technically, we should translate 0x5c to Yen symbol;
  36               // in practice, it is not clear.
  37               if (b == 0x5c)
  38                 b = 0x00A5;  // Yen sign.
  39 #endif
  40               elements(outbuffer)[outpos++] = (char) b;
  41             }
  42           else
  43             {
  44               if (b == 0x8E) // SS2
  45                 codeset = 2;
  46               else if (b == 0x8F) // SS3
  47                 codeset = 3;
  48               else
  49                 {
  50                   codeset = 1;
  51                   first_byte = b;
  52                 }
  53             }
  54         }
  55       else if (codeset == 1) // JIS X 0208:1997
  56         {
  57           first_byte -= 0x80 + 33;
  58           b -= 0x80 + 33;
  59           if ((unsigned) first_byte >= 84 || (unsigned) b >= 94)
  60             b = ERROR_CHAR;
  61           else
  62             {
  63               b = JIS0208_to_Unicode[first_byte][b];
  64               if (b == 0)
  65                 b = ERROR_CHAR;
  66             }
  67           elements(outbuffer)[outpos++] = b;
  68           codeset = 0;
  69         }
  70       else if (codeset == 2) // Half-width katakana
  71         {
  72           if (b >= 0xA1 && b <= 0xDF)
  73             b += 0xFF61 - 0xA1;
  74           else
  75             b = ERROR_CHAR;
  76           elements(outbuffer)[outpos++] = b;
  77           codeset = 0;
  78         }
  79       else if (codeset == 3) // second byte of JIS X 0212-1990
  80         {
  81           first_byte = b;
  82           codeset = 4;
  83         }
  84       else // codeset == 4 // third byte of JIS X 0212-1990
  85         {
  86           first_byte -= 0x80 + 34;
  87           b -= 0x80 + 33;
  88           if ((unsigned) first_byte >= 76 || (unsigned) b >= 94)
  89             b = ERROR_CHAR;
  90           else
  91             {
  92               b = JIS0208_to_Unicode[first_byte][b];
  93               if (b == 0)
  94                 b = ERROR_CHAR;
  95             }
  96           elements(outbuffer)[outpos++] = b;
  97           codeset = 0;
  98         }
  99     }
 100   return outpos - start_outpos;
 101 }