2 * UnicodeEncoding.cs - Implementation of the
3 * "System.Text.UnicodeEncoding" class.
5 * Copyright (c) 2001, 2002 Southern Storm Software, Pty Ltd
6 * Copyright (C) 2003, 2004 Novell, Inc.
7 * Copyright (C) 2006 Kornél Pál <http://www.kornelpal.hu/>
9 * Permission is hereby granted, free of charge, to any person obtaining
10 * a copy of this software and associated documentation files (the "Software"),
11 * to deal in the Software without restriction, including without limitation
12 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
13 * and/or sell copies of the Software, and to permit persons to whom the
14 * Software is furnished to do so, subject to the following conditions:
16 * The above copyright notice and this permission notice shall be included
17 * in all copies or substantial portions of the Software.
19 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
20 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
21 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
22 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
23 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
24 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
25 * OTHER DEALINGS IN THE SOFTWARE.
32 using System.Runtime.InteropServices;
36 [MonoLimitation ("Serialization format not compatible with .NET")]
37 public class UnicodeEncoding : Encoding
39 // Magic numbers used by Windows for Unicode.
40 internal const int UNICODE_CODE_PAGE = 1200;
41 internal const int BIG_UNICODE_CODE_PAGE = 1201;
44 // Size of characters in this encoding.
45 public const int CharSize = 2;
49 private bool bigEndian;
50 private bool byteOrderMark;
53 public UnicodeEncoding () : this (false, true)
58 public UnicodeEncoding (bool bigEndian, bool byteOrderMark)
59 : this (bigEndian, byteOrderMark, false)
63 public UnicodeEncoding (bool bigEndian, bool byteOrderMark, bool throwOnInvalidBytes)
64 : base ((bigEndian ? BIG_UNICODE_CODE_PAGE : UNICODE_CODE_PAGE))
66 if (throwOnInvalidBytes)
67 SetFallbackInternal (null, new DecoderExceptionFallback ());
69 SetFallbackInternal (null, new DecoderReplacementFallback ("\uFFFD"));
71 this.bigEndian = bigEndian;
72 this.byteOrderMark = byteOrderMark;
75 body_name = "unicodeFFFE";
76 encoding_name = "Unicode (Big-Endian)";
77 header_name = "unicodeFFFE";
78 is_browser_save = false;
79 web_name = "unicodeFFFE";
82 encoding_name = "Unicode";
83 header_name = "utf-16";
84 is_browser_save = true;
88 // Windows reports the same code page number for
89 // both the little-endian and big-endian forms.
90 windows_code_page = UNICODE_CODE_PAGE;
93 // Get the number of bytes needed to encode a character buffer.
94 public override int GetByteCount (char[] chars, int index, int count)
97 throw new ArgumentNullException ("chars");
99 if (index < 0 || index > chars.Length) {
100 throw new ArgumentOutOfRangeException ("index", _("ArgRange_Array"));
102 if (count < 0 || count > (chars.Length - index)) {
103 throw new ArgumentOutOfRangeException ("count", _("ArgRange_Array"));
108 public override int GetByteCount (String s)
111 throw new ArgumentNullException ("s");
116 [CLSCompliantAttribute (false)]
118 public unsafe override int GetByteCount (char* chars, int count)
121 throw new ArgumentNullException ("chars");
123 throw new ArgumentOutOfRangeException ("count");
128 // Get the bytes that result from encoding a character buffer.
129 public unsafe override int GetBytes (char [] chars, int charIndex, int charCount,
130 byte [] bytes, int byteIndex)
133 throw new ArgumentNullException ("chars");
136 throw new ArgumentNullException ("bytes");
138 if (charIndex < 0 || charIndex > chars.Length) {
139 throw new ArgumentOutOfRangeException ("charIndex", _("ArgRange_Array"));
141 if (charCount < 0 || charCount > (chars.Length - charIndex)) {
142 throw new ArgumentOutOfRangeException ("charCount", _("ArgRange_Array"));
144 if (byteIndex < 0 || byteIndex > bytes.Length) {
145 throw new ArgumentOutOfRangeException ("byteIndex", _("ArgRange_Array"));
151 int byteCount = bytes.Length - byteIndex;
152 if (bytes.Length == 0)
153 bytes = new byte [1];
155 fixed (char* charPtr = chars)
156 fixed (byte* bytePtr = bytes)
157 return GetBytesInternal (charPtr + charIndex, charCount, bytePtr + byteIndex, byteCount);
160 public unsafe override int GetBytes (String s, int charIndex, int charCount,
161 byte [] bytes, int byteIndex)
164 throw new ArgumentNullException ("s");
167 throw new ArgumentNullException ("bytes");
169 if (charIndex < 0 || charIndex > s.Length) {
170 throw new ArgumentOutOfRangeException ("charIndex", _("ArgRange_StringIndex"));
172 if (charCount < 0 || charCount > (s.Length - charIndex)) {
173 throw new ArgumentOutOfRangeException ("charCount", _("ArgRange_StringRange"));
175 if (byteIndex < 0 || byteIndex > bytes.Length) {
176 throw new ArgumentOutOfRangeException ("byteIndex", _("ArgRange_Array"));
183 int byteCount = bytes.Length - byteIndex;
184 if (bytes.Length == 0)
185 bytes = new byte [1];
187 fixed (char* charPtr = s)
188 fixed (byte* bytePtr = bytes)
189 return GetBytesInternal (charPtr + charIndex, charCount, bytePtr + byteIndex, byteCount);
192 [CLSCompliantAttribute (false)]
194 public unsafe override int GetBytes (char* chars, int charCount,
195 byte* bytes, int byteCount)
198 throw new ArgumentNullException ("bytes");
200 throw new ArgumentNullException ("chars");
202 throw new ArgumentOutOfRangeException ("charCount");
204 throw new ArgumentOutOfRangeException ("byteCount");
206 return GetBytesInternal (chars, charCount, bytes, byteCount);
209 private unsafe int GetBytesInternal (char* chars, int charCount,
210 byte* bytes, int byteCount)
212 int count = charCount * 2;
214 if (byteCount < count)
215 throw new ArgumentException (_("Arg_InsufficientSpace"));
217 CopyChars ((byte*) chars, bytes, count, bigEndian);
221 // Get the number of characters needed to decode a byte buffer.
222 public override int GetCharCount (byte[] bytes, int index, int count)
225 throw new ArgumentNullException ("bytes");
227 if (index < 0 || index > bytes.Length) {
228 throw new ArgumentOutOfRangeException ("index", _("ArgRange_Array"));
230 if (count < 0 || count > (bytes.Length - index)) {
231 throw new ArgumentOutOfRangeException ("count", _("ArgRange_Array"));
236 [CLSCompliantAttribute (false)]
238 public unsafe override int GetCharCount (byte* bytes, int count)
241 throw new ArgumentNullException ("bytes");
243 throw new ArgumentOutOfRangeException ("count");
248 // Get the characters that result from decoding a byte buffer.
249 public unsafe override int GetChars (byte [] bytes, int byteIndex, int byteCount,
250 char [] chars, int charIndex)
253 throw new ArgumentNullException ("bytes");
256 throw new ArgumentNullException ("chars");
258 if (byteIndex < 0 || byteIndex > bytes.Length) {
259 throw new ArgumentOutOfRangeException ("byteIndex", _("ArgRange_Array"));
261 if (byteCount < 0 || byteCount > (bytes.Length - byteIndex)) {
262 throw new ArgumentOutOfRangeException ("byteCount", _("ArgRange_Array"));
264 if (charIndex < 0 || charIndex > chars.Length) {
265 throw new ArgumentOutOfRangeException ("charIndex", _("ArgRange_Array"));
271 int charCount = chars.Length - charIndex;
272 if (chars.Length == 0)
273 chars = new char [1];
275 fixed (byte* bytePtr = bytes)
276 fixed (char* charPtr = chars)
277 return GetCharsInternal (bytePtr + byteIndex, byteCount, charPtr + charIndex, charCount);
280 [CLSCompliantAttribute (false)]
282 public unsafe override int GetChars (byte* bytes, int byteCount,
283 char* chars, int charCount)
286 throw new ArgumentNullException ("bytes");
288 throw new ArgumentNullException ("chars");
290 throw new ArgumentOutOfRangeException ("charCount");
292 throw new ArgumentOutOfRangeException ("byteCount");
294 return GetCharsInternal (bytes, byteCount, chars, charCount);
297 // Decode a buffer of bytes into a string.
299 public unsafe override String GetString (byte [] bytes, int index, int count)
302 throw new ArgumentNullException ("bytes");
303 if (index < 0 || index > bytes.Length)
304 throw new ArgumentOutOfRangeException ("index", _("ArgRange_Array"));
305 if (count < 0 || count > (bytes.Length - index))
306 throw new ArgumentOutOfRangeException ("count", _("ArgRange_Array"));
311 // GetCharCountInternal
312 int charCount = count / 2;
313 string s = string.InternalAllocateStr (charCount);
315 fixed (byte* bytePtr = bytes)
316 fixed (char* charPtr = s)
317 GetCharsInternal (bytePtr + index, count, charPtr, charCount);
322 private unsafe int GetCharsInternal (byte* bytes, int byteCount,
323 char* chars, int charCount)
325 int count = byteCount / 2;
327 // Validate that we have sufficient space in "chars".
328 if (charCount < count)
329 throw new ArgumentException (_("Arg_InsufficientSpace"));
331 CopyChars (bytes, (byte*) chars, byteCount, bigEndian);
336 public override Encoder GetEncoder ()
338 return(base.GetEncoder ());
341 // Get the maximum number of bytes needed to encode a
342 // specified number of characters.
343 public override int GetMaxByteCount (int charCount)
346 throw new ArgumentOutOfRangeException ("charCount", _("ArgRange_NonNegative"));
348 return charCount * 2;
351 // Get the maximum number of characters needed to decode a
352 // specified number of bytes.
353 public override int GetMaxCharCount (int byteCount)
356 throw new ArgumentOutOfRangeException
357 ("byteCount", _("ArgRange_NonNegative"));
359 return byteCount / 2;
362 // Get a Unicode-specific decoder that is attached to this instance.
363 public override Decoder GetDecoder ()
365 return new UnicodeDecoder (bigEndian);
368 // Get the Unicode preamble.
369 public override byte[] GetPreamble ()
372 byte[] preamble = new byte[2];
374 preamble[0] = (byte)0xFE;
375 preamble[1] = (byte)0xFF;
377 preamble[0] = (byte)0xFF;
378 preamble[1] = (byte)0xFE;
386 // Determine if this object is equal to another.
387 public override bool Equals (Object value)
389 UnicodeEncoding enc = (value as UnicodeEncoding);
391 return (codePage == enc.codePage &&
392 bigEndian == enc.bigEndian &&
393 byteOrderMark == enc.byteOrderMark);
399 // Get the hash code for this object.
400 public override int GetHashCode ()
402 return base.GetHashCode ();
405 private unsafe static void CopyChars (byte* src, byte* dest, int count, bool bigEndian)
407 if (BitConverter.IsLittleEndian != bigEndian) {
408 string.memcpy (dest, src, count & unchecked ((int) 0xFFFFFFFE));
458 dest [10] = src [11];
459 dest [11] = src [10];
460 dest [12] = src [13];
461 dest [13] = src [12];
462 dest [14] = src [15];
463 dest [15] = src [14];
467 } while ((count & unchecked ((int) 0xFFFFFFF0)) != 0);
500 if ((count & 4) == 0)
511 if ((count & 2) == 0)
518 // Unicode decoder implementation.
519 private sealed class UnicodeDecoder : Decoder
521 private bool bigEndian;
522 private int leftOverByte;
525 public UnicodeDecoder (bool bigEndian)
527 this.bigEndian = bigEndian;
531 // Override inherited methods.
532 public override int GetCharCount (byte[] bytes, int index, int count)
535 throw new ArgumentNullException ("bytes");
537 if (index < 0 || index > bytes.Length) {
538 throw new ArgumentOutOfRangeException ("index", _("ArgRange_Array"));
540 if (count < 0 || count > (bytes.Length - index)) {
541 throw new ArgumentOutOfRangeException ("count", _("ArgRange_Array"));
543 if (leftOverByte != -1) {
544 return (count + 1) / 2;
550 public unsafe override int GetChars (byte [] bytes, int byteIndex,
551 int byteCount, char [] chars,
555 throw new ArgumentNullException ("bytes");
558 throw new ArgumentNullException ("chars");
560 if (byteIndex < 0 || byteIndex > bytes.Length) {
561 throw new ArgumentOutOfRangeException ("byteIndex", _("ArgRange_Array"));
563 if (byteCount < 0 || byteCount > (bytes.Length - byteIndex)) {
564 throw new ArgumentOutOfRangeException ("byteCount", _("ArgRange_Array"));
566 if (charIndex < 0 || charIndex > chars.Length) {
567 throw new ArgumentOutOfRangeException ("charIndex", _("ArgRange_Array"));
573 int leftOver = leftOverByte;
577 count = (byteCount + 1) / 2;
579 count = byteCount / 2;
581 if (chars.Length - charIndex < count)
582 throw new ArgumentException (_("Arg_InsufficientSpace"));
584 if (leftOver != -1) {
586 chars [charIndex] = unchecked ((char) ((leftOver << 8) | (int) bytes [byteIndex]));
588 chars [charIndex] = unchecked ((char) (((int) bytes [byteIndex] << 8) | leftOver));
594 if ((byteCount & unchecked ((int) 0xFFFFFFFE)) != 0)
595 fixed (byte* bytePtr = bytes)
596 fixed (char* charPtr = chars)
597 CopyChars (bytePtr + byteIndex, (byte*) (charPtr + charIndex), byteCount, bigEndian);
599 if ((byteCount & 1) == 0)
602 leftOverByte = bytes [byteCount + byteIndex - 1];
607 } // class UnicodeDecoder
609 }; // class UnicodeEncoding
611 }; // namespace System.Text