| 1 |
756 |
jeremybenn |
/* Copyright (C) 1999, 2000, 2001, 2005 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 |
|
|
import java.nio.charset.Charset;
|
| 12 |
|
|
|
| 13 |
|
|
public abstract class BytesToUnicode extends IOConverter
|
| 14 |
|
|
{
|
| 15 |
|
|
/** Buffer to read bytes from.
|
| 16 |
|
|
* The characters inbuffer[inpos] ... inbuffer[inlength-1] are available. */
|
| 17 |
|
|
public byte[] inbuffer;
|
| 18 |
|
|
/** Starting index in buffer to read bytes from. */
|
| 19 |
|
|
public int inpos;
|
| 20 |
|
|
/** End of valid bytes in buffer. */
|
| 21 |
|
|
public int inlength;
|
| 22 |
|
|
|
| 23 |
|
|
// The name of the default encoding.
|
| 24 |
|
|
static String defaultEncoding;
|
| 25 |
|
|
|
| 26 |
|
|
/* These keep a small cache of decoders for reuse. The array holds
|
| 27 |
|
|
the actual decoders. The currCachePos is the next value we are
|
| 28 |
|
|
going to replace in the cache. We don't just throw the data away
|
| 29 |
|
|
if the cache is full, because if the cache filled up with stuff
|
| 30 |
|
|
we don't need then the cache would be worthless. We instead
|
| 31 |
|
|
circulate through the cache the implement kind of an LRU
|
| 32 |
|
|
algorithm. */
|
| 33 |
|
|
private static final int CACHE_SIZE = 4; // A power of 2 for speed
|
| 34 |
|
|
private static BytesToUnicode[] decoderCache
|
| 35 |
|
|
= new BytesToUnicode[CACHE_SIZE];
|
| 36 |
|
|
private static int currCachePos = 0;
|
| 37 |
|
|
|
| 38 |
|
|
public abstract String getName();
|
| 39 |
|
|
|
| 40 |
|
|
public static BytesToUnicode getDefaultDecoder()
|
| 41 |
|
|
{
|
| 42 |
|
|
try
|
| 43 |
|
|
{
|
| 44 |
|
|
synchronized (BytesToUnicode.class)
|
| 45 |
|
|
{
|
| 46 |
|
|
if (defaultEncoding == null)
|
| 47 |
|
|
{
|
| 48 |
|
|
String encoding
|
| 49 |
|
|
= canonicalize (System.getProperty("file.encoding",
|
| 50 |
|
|
"8859_1"));
|
| 51 |
|
|
String className = "gnu.gcj.convert.Input_" + encoding;
|
| 52 |
|
|
try
|
| 53 |
|
|
{
|
| 54 |
|
|
Class defaultDecodingClass = Class.forName(className);
|
| 55 |
|
|
defaultEncoding = encoding;
|
| 56 |
|
|
}
|
| 57 |
|
|
catch (ClassNotFoundException ex)
|
| 58 |
|
|
{
|
| 59 |
|
|
throw new NoClassDefFoundError("missing default encoding "
|
| 60 |
|
|
+ encoding + " (class "
|
| 61 |
|
|
+ className
|
| 62 |
|
|
+ " not found)");
|
| 63 |
|
|
}
|
| 64 |
|
|
}
|
| 65 |
|
|
}
|
| 66 |
|
|
return getDecoder (defaultEncoding);
|
| 67 |
|
|
}
|
| 68 |
|
|
catch (Throwable ex)
|
| 69 |
|
|
{
|
| 70 |
|
|
return new Input_8859_1();
|
| 71 |
|
|
}
|
| 72 |
|
|
}
|
| 73 |
|
|
|
| 74 |
|
|
/** Get a byte-stream->char-stream converter given an encoding name. */
|
| 75 |
|
|
public static BytesToUnicode getDecoder (String encoding)
|
| 76 |
|
|
throws java.io.UnsupportedEncodingException
|
| 77 |
|
|
{
|
| 78 |
|
|
/* First hunt in our cache to see if we have a decoder that is
|
| 79 |
|
|
already allocated. */
|
| 80 |
|
|
String canonicalEncoding = canonicalize(encoding);
|
| 81 |
|
|
synchronized (BytesToUnicode.class)
|
| 82 |
|
|
{
|
| 83 |
|
|
int i;
|
| 84 |
|
|
for (i = 0; i < decoderCache.length; ++i)
|
| 85 |
|
|
{
|
| 86 |
|
|
if (decoderCache[i] != null
|
| 87 |
|
|
&& canonicalEncoding.equals(decoderCache[i].getName ()))
|
| 88 |
|
|
{
|
| 89 |
|
|
BytesToUnicode rv = decoderCache[i];
|
| 90 |
|
|
decoderCache[i] = null;
|
| 91 |
|
|
return rv;
|
| 92 |
|
|
}
|
| 93 |
|
|
}
|
| 94 |
|
|
}
|
| 95 |
|
|
|
| 96 |
|
|
// It's not in the cache, so now we have to do real work.
|
| 97 |
|
|
String className = "gnu.gcj.convert.Input_" + canonicalEncoding;
|
| 98 |
|
|
Class decodingClass;
|
| 99 |
|
|
try
|
| 100 |
|
|
{
|
| 101 |
|
|
decodingClass = Class.forName(className);
|
| 102 |
|
|
return (BytesToUnicode) decodingClass.newInstance();
|
| 103 |
|
|
}
|
| 104 |
|
|
catch (Throwable ex)
|
| 105 |
|
|
{
|
| 106 |
|
|
try
|
| 107 |
|
|
{
|
| 108 |
|
|
// We pass the original name to iconv and let it handle
|
| 109 |
|
|
// its own aliasing. Note that we intentionally prefer
|
| 110 |
|
|
// iconv over nio.
|
| 111 |
|
|
return new Input_iconv (encoding);
|
| 112 |
|
|
}
|
| 113 |
|
|
catch (Throwable _)
|
| 114 |
|
|
{
|
| 115 |
|
|
// Ignore, and try the next method.
|
| 116 |
|
|
}
|
| 117 |
|
|
try
|
| 118 |
|
|
{
|
| 119 |
|
|
return new BytesToCharsetAdaptor(Charset.forName(encoding));
|
| 120 |
|
|
}
|
| 121 |
|
|
catch (Throwable _)
|
| 122 |
|
|
{
|
| 123 |
|
|
throw new java.io.UnsupportedEncodingException(encoding
|
| 124 |
|
|
+ " (" + ex + ')');
|
| 125 |
|
|
}
|
| 126 |
|
|
}
|
| 127 |
|
|
}
|
| 128 |
|
|
|
| 129 |
|
|
/** Make input bytes available to the conversion.
|
| 130 |
|
|
* @param buffer source of input bytes
|
| 131 |
|
|
* @param pos index of first available byte
|
| 132 |
|
|
* @param length one more than index of last available byte
|
| 133 |
|
|
*/
|
| 134 |
|
|
public final void setInput(byte[] buffer, int pos, int length)
|
| 135 |
|
|
{
|
| 136 |
|
|
inbuffer = buffer;
|
| 137 |
|
|
inpos = pos;
|
| 138 |
|
|
inlength = length;
|
| 139 |
|
|
}
|
| 140 |
|
|
|
| 141 |
|
|
/** Convert bytes to chars.
|
| 142 |
|
|
* Input bytes are taken from this.inbuffer. The available input
|
| 143 |
|
|
* bytes start at inbuffer[inpos], and end at inbuffer[inlength-1].
|
| 144 |
|
|
* @param outbuffer buffer for the converted character
|
| 145 |
|
|
* @param outpos position in buffer to start putting converted characters
|
| 146 |
|
|
* @param count the maximum number of characters to convert
|
| 147 |
|
|
* @return number of chars placed in outbuffer.
|
| 148 |
|
|
* Also, this.inpos is incremented by the number of bytes consumed.
|
| 149 |
|
|
*
|
| 150 |
|
|
* (Note the asymmetry in that the input upper bound is inbuffer[inlength-1],
|
| 151 |
|
|
* while the output upper bound is outbuffer[outpos+count-1]. The
|
| 152 |
|
|
* justification is that inlength is like the count field of a
|
| 153 |
|
|
* BufferedInputStream, while the count parameter is like the
|
| 154 |
|
|
* length parameter of a read request.) The count parameter is
|
| 155 |
|
|
* also defined to be <= outbuffer.length - outpos (per the specification
|
| 156 |
|
|
* of the length parameter for a read request).
|
| 157 |
|
|
*/
|
| 158 |
|
|
public abstract int read (char[] outbuffer, int outpos, int count);
|
| 159 |
|
|
|
| 160 |
|
|
/** Indicate that the converter is resuable.
|
| 161 |
|
|
* This class keeps track of converters on a per-encoding basis.
|
| 162 |
|
|
* When done with an encoder you may call this method to indicate
|
| 163 |
|
|
* that it can be reused later.
|
| 164 |
|
|
*/
|
| 165 |
|
|
public void done ()
|
| 166 |
|
|
{
|
| 167 |
|
|
synchronized (BytesToUnicode.class)
|
| 168 |
|
|
{
|
| 169 |
|
|
this.inbuffer = null;
|
| 170 |
|
|
this.inpos = 0;
|
| 171 |
|
|
this.inlength = 0;
|
| 172 |
|
|
|
| 173 |
|
|
decoderCache[currCachePos] = this;
|
| 174 |
|
|
currCachePos = (currCachePos + 1) % CACHE_SIZE;
|
| 175 |
|
|
}
|
| 176 |
|
|
}
|
| 177 |
|
|
}
|