libjava/gnu/gcj/convert/BytesToUnicode.java

   1 /* Copyright (C) 1999, 2000, 2001  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 package gnu.gcj.convert;
  10
  11 public abstract class BytesToUnicode extends IOConverter
  12 {
  13   /** Buffer to read bytes from.
  14    * The characters inbuffer[inpos] ... inbuffer[inlength-1] are available. */
  15   public byte[] inbuffer;
  16   /** Starting index in buffer to read bytes from. */
  17   public int inpos;
  18   /** End of valid bytes in buffer. */
  19   public int inlength;
  20
  21   // The name of the default encoding.
  22   static String defaultEncoding;
  23
  24   /* These keep a small cache of decoders for reuse.  The array holds
  25      the actual decoders.  The currCachePos is the next value we are
  26      going to replace in the cache.  We don't just throw the data away
  27      if the cache is full, because if the cache filled up with stuff
  28      we don't need then the cache would be worthless.  We instead
  29      circulate through the cache the implement kind of an LRU
  30      algorithm. */
  31   private static final int CACHE_SIZE = 4;  // A power of 2 for speed
  32   private static BytesToUnicode[] decoderCache
  33     = new BytesToUnicode[CACHE_SIZE];
  34   private static int currCachePos = 0;
  35
  36   public abstract String getName();
  37
  38   public static BytesToUnicode getDefaultDecoder()
  39   {
  40     try
  41       {
  42         synchronized (BytesToUnicode.class)
  43           {
  44             if (defaultEncoding == null)
  45               {
  46                 String encoding
  47                   = canonicalize (System.getProperty("file.encoding",
  48                                                      "8859_1"));
  49                 String className = "gnu.gcj.convert.Input_" + encoding;
  50                 try
  51                   {
  52                     Class defaultDecodingClass = Class.forName(className);
  53                     defaultEncoding = encoding;
  54                   }
  55                 catch (ClassNotFoundException ex)
  56                   {
  57                     throw new NoClassDefFoundError("missing default encoding "
  58                                                    + encoding + " (class "
  59                                                    + className
  60                                                    + " not found)");
  61                   }
  62               }
  63           }
  64         return getDecoder (defaultEncoding);
  65       }
  66     catch (Throwable ex)
  67       {
  68         return new Input_8859_1();
  69       }
  70   }
  71
  72   /** Get a byte-stream->char-stream converter given an encoding name. */
  73   public static BytesToUnicode getDecoder (String encoding)
  74     throws java.io.UnsupportedEncodingException
  75   {
  76     /* First hunt in our cache to see if we have a decoder that is
  77        already allocated. */
  78     synchronized (BytesToUnicode.class)
  79       {
  80         int i;
  81         for (i = 0; i < decoderCache.length; ++i)
  82           {
  83             if (decoderCache[i] != null
  84                 && encoding.equals(decoderCache[i].getName ()))
  85               {
  86                 BytesToUnicode rv = decoderCache[i];
  87                 decoderCache[i] = null;
  88                 return rv;
  89             }
  90           }
  91       }
  92
  93     // It's not in the cache, so now we have to do real work.
  94     String className = "gnu.gcj.convert.Input_" + canonicalize (encoding);
  95     Class decodingClass;
  96     try
  97       {
  98         decodingClass = Class.forName(className);
  99         return (BytesToUnicode) decodingClass.newInstance();
 100       }
 101     catch (Throwable ex)
 102       {
 103         try
 104           {
 105             // We pass the original name to iconv and let it handle
 106             // its own aliasing.
 107             return new Input_iconv (encoding);
 108           }
 109         catch (Throwable _)
 110           {
 111             throw new java.io.UnsupportedEncodingException(encoding
 112                                                            + " (" + ex + ')');
 113           }
 114       }
 115   }
 116
 117   /** Make input bytes available to the conversion.
 118    * @param buffer source of input bytes
 119    * @param pos index of first available byte
 120    * @param length one more than index of last available byte
 121    */
 122   public final void setInput(byte[] buffer, int pos, int length)
 123   {
 124     inbuffer = buffer;
 125     inpos = pos;
 126     inlength = length;
 127   }
 128
 129   /** Convert bytes to chars.
 130    * Input bytes are taken from this.inbuffer.  The available input
 131    * bytes start at inbuffer[inpos], and end at inbuffer[inlength-1].
 132    * @param outbuffer buffer for the converted character
 133    * @param outpos position in buffer to start putting converted characters
 134    * @param count the maximum number of characters to convert
 135    * @return number of chars placed in outbuffer.
 136    * Also, this.inpos is incremented by the number of bytes consumed.
 137    *
 138    * (Note the asymmetry in that the input upper bound is inbuffer[inlength-1],
 139    * while the output upper bound is outbuffer[outpos+count-1].  The
 140    * justification is that inlength is like the count field of a
 141    * BufferedInputStream, while the count parameter is like the
 142    * length parameter of a read request.)  The count parameter is
 143    * also defined to be <= outbuffer.length - outpos (per the specification
 144    * of the length parameter for a read request).
 145    */
 146   public abstract int read (char[] outbuffer, int outpos, int count);
 147
 148   /** Indicate that the converter is resuable.
 149    * This class keeps track of converters on a per-encoding basis.
 150    * When done with an encoder you may call this method to indicate
 151    * that it can be reused later.
 152    */
 153   public void done ()
 154   {
 155     synchronized (BytesToUnicode.class)
 156       {
 157         this.inbuffer = null;
 158         this.inpos = 0;
 159         this.inlength = 0;
 160
 161         decoderCache[currCachePos] = this;
 162         currCachePos = (currCachePos + 1) % CACHE_SIZE;
 163       }
 164   }
 165 }