class/System.XML/System.Xml/XmlReaderBinarySupport.cs

   1 //
   2 // System.Xml.XmlReaderBinarySupport.cs
   3 //
   4 // Author:
   5 //   Atsushi Enomoto  (ginga@kit.hi-ho.ne.jp)
   6 //
   7 // (C)2004 Novell Inc,
   8 //
   9
  10 //
  11 // Permission is hereby granted, free of charge, to any person obtaining
  12 // a copy of this software and associated documentation files (the
  13 // "Software"), to deal in the Software without restriction, including
  14 // without limitation the rights to use, copy, modify, merge, publish,
  15 // distribute, sublicense, and/or sell copies of the Software, and to
  16 // permit persons to whom the Software is furnished to do so, subject to
  17 // the following conditions:
  18 //
  19 // The above copyright notice and this permission notice shall be
  20 // included in all copies or substantial portions of the Software.
  21 //
  22 // THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
  23 // EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  24 // MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
  25 // NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
  26 // LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
  27 // OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
  28 // WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  29 //
  30
  31 using System;
  32 using System.Collections;
  33 using System.Text;
  34
  35 namespace System.Xml
  36 {
  37         internal class XmlReaderBinarySupport
  38         {
  39                 public delegate int CharGetter (
  40                         char [] buffer, int offset, int length);
  41
  42                 public enum CommandState {
  43                         None,
  44                         ReadElementContentAsBase64,
  45                         ReadContentAsBase64,
  46                         ReadElementContentAsBinHex,
  47                         ReadContentAsBinHex
  48                 }
  49
  50                 public XmlReaderBinarySupport (XmlReader reader)
  51                 {
  52                         this.reader = reader;
  53                         Reset ();
  54                 }
  55
  56                 XmlReader reader;
  57                 CharGetter getter;
  58                 byte [] base64Cache = new byte [3];
  59                 int base64CacheStartsAt;
  60                 CommandState state;
  61                 StringBuilder textCache;
  62                 bool hasCache;
  63                 bool dontReset;
  64
  65                 public CharGetter Getter {
  66                         get { return getter; }
  67                         set { getter = value; }
  68                 }
  69
  70                 public void Reset ()
  71                 {
  72                         if (!dontReset) {
  73                                 dontReset = true;
  74                                 if (hasCache) {
  75                                         switch (reader.NodeType) {
  76                                         case XmlNodeType.Text:
  77                                         case XmlNodeType.CDATA:
  78                                         case XmlNodeType.SignificantWhitespace:
  79                                         case XmlNodeType.Whitespace:
  80                                                 reader.Read ();
  81                                                 break;
  82                                         }
  83                                         switch (state) {
  84                                         case CommandState.ReadElementContentAsBase64:
  85                                         case CommandState.ReadElementContentAsBinHex:
  86                                                 reader.Read ();
  87                                                 break;
  88                                         }
  89                                 }
  90                                 base64CacheStartsAt = -1;
  91                                 state = CommandState.None;
  92                                 hasCache = false;
  93                                 dontReset = false;
  94                         }
  95                 }
  96
  97                 InvalidOperationException StateError (CommandState action)
  98                 {
  99                         return new InvalidOperationException (
 100                                 String.Format ("Invalid attempt to read binary content by {0}, while once binary reading was started by {1}", action, state));
 101                 }
 102
 103                 private void CheckState (bool element, CommandState action)
 104                 {
 105                         if (state == CommandState.None) {
 106                                 if (textCache == null)
 107                                         textCache = new StringBuilder ();
 108                                 else
 109                                         textCache.Length = 0;
 110                                 if (action == CommandState.None)
 111                                         return; // for ReadValueChunk()
 112                                 if (reader.ReadState != ReadState.Interactive)
 113                                         return;
 114                                 switch (reader.NodeType) {
 115                                 case XmlNodeType.Text:
 116                                 case XmlNodeType.CDATA:
 117                                 case XmlNodeType.SignificantWhitespace:
 118                                 case XmlNodeType.Whitespace:
 119                                         if (!element) {
 120                                                 state = action;
 121                                                 return;
 122                                         }
 123                                         break;
 124                                 case XmlNodeType.Element:
 125                                         if (element) {
 126                                                 if (!reader.IsEmptyElement)
 127                                                         reader.Read ();
 128                                                 state = action;
 129                                                 return;
 130                                         }
 131                                         break;
 132                                 }
 133                                 throw new XmlException ((element ?
 134                                         "Reader is not positioned on an element."
 135                                         : "Reader is not positioned on a text node."));
 136                         }
 137                         if (state == action)
 138                                 return;
 139                         throw StateError (action);
 140                 }
 141
 142                 public int ReadElementContentAsBase64 (
 143                         byte [] buffer, int offset, int length)
 144                 {
 145                         CheckState (true, CommandState.ReadElementContentAsBase64);
 146                         return ReadBase64 (buffer, offset, length);
 147                 }
 148
 149                 public int ReadContentAsBase64 (
 150                         byte [] buffer, int offset, int length)
 151                 {
 152                         CheckState (false, CommandState.ReadContentAsBase64);
 153                         return ReadBase64 (buffer, offset, length);
 154                 }
 155
 156                 public int ReadElementContentAsBinHex (
 157                         byte [] buffer, int offset, int length)
 158                 {
 159                         CheckState (true, CommandState.ReadElementContentAsBinHex);
 160                         return ReadBinHex (buffer, offset, length);
 161                 }
 162
 163                 public int ReadContentAsBinHex (
 164                         byte [] buffer, int offset, int length)
 165                 {
 166                         CheckState (false, CommandState.ReadContentAsBinHex);
 167                         return ReadBinHex (buffer, offset, length);
 168                 }
 169
 170                 public int ReadBase64 (byte [] buffer, int offset, int length)
 171                 {
 172                         if (offset < 0)
 173                                 throw CreateArgumentOutOfRangeException ("offset", offset, "Offset must be non-negative integer.");
 174                         else if (length < 0)
 175                                 throw CreateArgumentOutOfRangeException ("length", length, "Length must be non-negative integer.");
 176                         else if (buffer.Length < offset + length)
 177                                 throw new ArgumentOutOfRangeException ("buffer length is smaller than the sum of offset and length.");
 178
 179                         if (reader.IsEmptyElement)
 180                                 return 0;
 181                         if (length == 0)        // It does not raise an error.
 182                                 return 0;
 183
 184                         int bufIndex = offset;
 185                         int bufLast = offset + length;
 186
 187                         if (base64CacheStartsAt >= 0) {
 188                                 for (int i = base64CacheStartsAt; i < 3; i++) {
 189                                         buffer [bufIndex++] = base64Cache [base64CacheStartsAt++];
 190                                         if (bufIndex == bufLast)
 191                                                 return bufLast - offset;
 192                                 }
 193                         }
 194
 195                         for (int i = 0; i < 3; i++)
 196                                 base64Cache [i] = 0;
 197                         base64CacheStartsAt = -1;
 198
 199                         int max = (int) System.Math.Ceiling (4.0 / 3 * length);
 200                         int additional = max % 4;
 201                         if (additional > 0)
 202                                 max += 4 - additional;
 203                         char [] chars = new char [max];
 204                         int charsLength = getter != null ?
 205                                 getter (chars, 0, max) :
 206                                 ReadValueChunk (chars, 0, max);
 207
 208                         byte b = 0;
 209                         byte work = 0;
 210                         for (int i = 0; i < charsLength - 3; i++) {
 211                                 if ((i = SkipIgnorableBase64Chars (chars, charsLength, i)) == charsLength)
 212                                         break;
 213                                 b = (byte) (GetBase64Byte (chars [i]) << 2);
 214                                 if (bufIndex < bufLast)
 215                                         buffer [bufIndex] = b;
 216                                 else if (b != 0) {
 217                                         if (base64CacheStartsAt < 0)
 218                                                 base64CacheStartsAt = 0;
 219                                         base64Cache [0] = b;
 220                                 }
 221                                 // charsLength mod 4 might not equals to 0.
 222                                 if (++i == charsLength)
 223                                         break;
 224                                 if ((i = SkipIgnorableBase64Chars (chars, charsLength, i))  == charsLength)
 225                                         break;
 226                                 b = GetBase64Byte (chars [i]);
 227                                 work = (byte) (b >> 4);
 228                                 if (bufIndex < bufLast) {
 229                                         buffer [bufIndex] += work;
 230                                         bufIndex++;
 231                                 }
 232                                 else if (work != 0) {
 233                                         if (base64CacheStartsAt < 0)
 234                                                 base64CacheStartsAt = 0;
 235                                         base64Cache [0] += work;
 236                                 }
 237
 238                                 work = (byte) ((b & 0xf) << 4);
 239                                 if (bufIndex < bufLast) {
 240                                         buffer [bufIndex] = work;
 241                                 }
 242                                 else if (work != 0) {
 243                                         if (base64CacheStartsAt < 0)
 244                                                 base64CacheStartsAt = 1;
 245                                         base64Cache [1] = work;
 246                                 }
 247
 248                                 if (++i == charsLength)
 249                                         break;
 250                                 if ((i = SkipIgnorableBase64Chars (chars, charsLength, i)) == charsLength)
 251                                         break;
 252                                 b = GetBase64Byte (chars [i]);
 253                                 work = (byte) (b >> 2);
 254                                 if (bufIndex < bufLast) {
 255                                         buffer [bufIndex] += work;
 256                                         bufIndex++;
 257                                 }
 258                                 else if (work != 0) {
 259                                         if (base64CacheStartsAt < 0)
 260                                                 base64CacheStartsAt = 1;
 261                                         base64Cache [1] += work;
 262                                 }
 263
 264                                 work = (byte) ((b & 3) << 6);
 265                                 if (bufIndex < bufLast)
 266                                         buffer [bufIndex] = work;
 267                                 else if (work != 0) {
 268                                         if (base64CacheStartsAt < 0)
 269                                                 base64CacheStartsAt = 2;
 270                                         base64Cache [2] = work;
 271                                 }
 272                                 if (++i == charsLength)
 273                                         break;
 274                                 if ((i = SkipIgnorableBase64Chars (chars, charsLength, i)) == charsLength)
 275                                         break;
 276                                 work = GetBase64Byte (chars [i]);
 277                                 if (bufIndex < bufLast) {
 278                                         buffer [bufIndex] += work;
 279                                         bufIndex++;
 280                                 }
 281                                 else if (work != 0) {
 282                                         if (base64CacheStartsAt < 0)
 283                                                 base64CacheStartsAt = 2;
 284                                         base64Cache [2] += work;
 285                                 }
 286                         }
 287                         int ret = System.Math.Min (bufLast - offset, bufIndex - offset);
 288                         if (ret < length && charsLength > 0)
 289                                 return ret + ReadBase64 (buffer, offset + ret, length - ret);
 290                         else
 291                                 return ret;
 292                 }
 293
 294                 // Since ReadBase64() is processed for every 4 chars, it does
 295                 // not handle '=' here.
 296                 private byte GetBase64Byte (char ch)
 297                 {
 298                         switch (ch) {
 299                         case '+':
 300                                 return 62;
 301                         case '/':
 302                                 return 63;
 303                         default:
 304                                 if (ch >= 'A' && ch <= 'Z')
 305                                         return (byte) (ch - 'A');
 306                                 else if (ch >= 'a' && ch <= 'z')
 307                                         return (byte) (ch - 'a' + 26);
 308                                 else if (ch >= '0' && ch <= '9')
 309                                         return (byte) (ch - '0' + 52);
 310                                 else
 311                                         throw new XmlException ("Invalid Base64 character was found.");
 312                         }
 313                 }
 314
 315                 private int SkipIgnorableBase64Chars (char [] chars, int charsLength, int i)
 316                 {
 317                         while (chars [i] == '=' || XmlChar.IsWhitespace (chars [i]))
 318                                 if (charsLength == ++i)
 319                                         break;
 320                         return i;
 321                 }
 322
 323                 static Exception CreateArgumentOutOfRangeException (string name, object value, string message)
 324                 {
 325                         return new ArgumentOutOfRangeException (
 326 #if !NET_2_1
 327                                 name, value,
 328 #endif
 329                                 message);
 330                 }
 331
 332                 public int ReadBinHex (byte [] buffer, int offset, int length)
 333                 {
 334                         if (offset < 0)
 335                                 throw CreateArgumentOutOfRangeException ("offset", offset, "Offset must be non-negative integer.");
 336                         else if (length < 0)
 337                                 throw CreateArgumentOutOfRangeException ("length", length, "Length must be non-negative integer.");
 338                         else if (buffer.Length < offset + length)
 339                                 throw new ArgumentOutOfRangeException ("buffer length is smaller than the sum of offset and length.");
 340
 341                         if (length == 0)
 342                                 return 0;
 343
 344                         char [] chars = new char [length * 2];
 345                         int charsLength = getter != null ?
 346                                 getter (chars, 0, length * 2) :
 347                                 ReadValueChunk (chars, 0, length * 2);
 348                         return XmlConvert.FromBinHexString (chars, offset, charsLength, buffer);
 349                 }
 350
 351                 public int ReadValueChunk (
 352                         char [] buffer, int offset, int length)
 353                 {
 354                         CommandState backup = state;
 355                         if (state == CommandState.None)
 356                                 CheckState (false, CommandState.None);
 357
 358                         if (offset < 0)
 359                                 throw CreateArgumentOutOfRangeException ("offset", offset, "Offset must be non-negative integer.");
 360                         else if (length < 0)
 361                                 throw CreateArgumentOutOfRangeException ("length", length, "Length must be non-negative integer.");
 362                         else if (buffer.Length < offset + length)
 363                                 throw new ArgumentOutOfRangeException ("buffer length is smaller than the sum of offset and length.");
 364
 365                         if (length == 0)
 366                                 return 0;
 367
 368                         if (!hasCache) {
 369                                 if (reader.IsEmptyElement)
 370                                         return 0;
 371                         }
 372
 373                         bool loop = true;
 374                         while (loop && textCache.Length < length) {
 375                                 switch (reader.NodeType) {
 376                                 case XmlNodeType.Text:
 377                                 case XmlNodeType.CDATA:
 378                                 case XmlNodeType.SignificantWhitespace:
 379                                 case XmlNodeType.Whitespace:
 380                                         if (hasCache) {
 381                                                 switch (reader.NodeType) {
 382                                                 case XmlNodeType.Text:
 383                                                 case XmlNodeType.CDATA:
 384                                                 case XmlNodeType.SignificantWhitespace:
 385                                                 case XmlNodeType.Whitespace:
 386                                                         Read ();
 387                                                         break;
 388                                                 default:
 389                                                         loop = false;
 390                                                         break;
 391                                                 }
 392                                         }
 393                                         textCache.Append (reader.Value);
 394                                         hasCache = true;
 395                                         break;
 396                                 default:
 397                                         loop = false;
 398                                         break;
 399                                 }
 400                         }
 401                         state = backup;
 402                         int min = textCache.Length;
 403                         if (min > length)
 404                                 min = length;
 405                         string str = textCache.ToString (0, min);
 406                         textCache.Remove (0, str.Length);
 407                         str.CopyTo (0, buffer, offset, str.Length);
 408                         if (min < length && loop)
 409                                 return min + ReadValueChunk (buffer, offset + min, length - min);
 410                         else
 411                                 return min;
 412                 }
 413
 414                 private bool Read ()
 415                 {
 416                         dontReset = true;
 417                         bool b = reader.Read ();
 418                         dontReset = false;
 419                         return b;
 420                 }
 421         }
 422 }