001/* 002 * Licensed to the Apache Software Foundation (ASF) under one or more 003 * contributor license agreements. See the NOTICE file distributed with 004 * this work for additional information regarding copyright ownership. 005 * The ASF licenses this file to You under the Apache License, Version 2.0 006 * (the "License"); you may not use this file except in compliance with 007 * the License. You may obtain a copy of the License at 008 * 009 * http://www.apache.org/licenses/LICENSE-2.0 010 * 011 * Unless required by applicable law or agreed to in writing, software 012 * distributed under the License is distributed on an "AS IS" BASIS, 013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 014 * See the License for the specific language governing permissions and 015 * limitations under the License. 016 */ 017 018package org.apache.commons.codec.binary; 019 020/** 021 * Provides Base32 encoding and decoding as defined by <a href="http://www.ietf.org/rfc/rfc4648.txt">RFC 4648</a>. 022 * 023 * <p> 024 * The class can be parameterized in the following manner with various constructors: 025 * </p> 026 * <ul> 027 * <li>Whether to use the "base32hex" variant instead of the default "base32"</li> 028 * <li>Line length: Default 76. Line length that aren't multiples of 8 will still essentially end up being multiples of 029 * 8 in the encoded data. 030 * <li>Line separator: Default is CRLF ("\r\n")</li> 031 * </ul> 032 * <p> 033 * This class operates directly on byte streams, and not character streams. 034 * </p> 035 * <p> 036 * This class is thread-safe. 037 * </p> 038 * 039 * @see <a href="http://www.ietf.org/rfc/rfc4648.txt">RFC 4648</a> 040 * 041 * @since 1.5 042 */ 043public class Base32 extends BaseNCodec { 044 045 /** 046 * BASE32 characters are 5 bits in length. 047 * They are formed by taking a block of five octets to form a 40-bit string, 048 * which is converted into eight BASE32 characters. 049 */ 050 private static final int BITS_PER_ENCODED_BYTE = 5; 051 private static final int BYTES_PER_ENCODED_BLOCK = 8; 052 private static final int BYTES_PER_UNENCODED_BLOCK = 5; 053 054 /** 055 * Chunk separator per RFC 2045 section 2.1. 056 * 057 * @see <a href="http://www.ietf.org/rfc/rfc2045.txt">RFC 2045 section 2.1</a> 058 */ 059 private static final byte[] CHUNK_SEPARATOR = {'\r', '\n'}; 060 061 /** 062 * This array is a lookup table that translates Unicode characters drawn from the "Base32 Alphabet" (as specified 063 * in Table 3 of RFC 4648) into their 5-bit positive integer equivalents. Characters that are not in the Base32 064 * alphabet but fall within the bounds of the array are translated to -1. 065 */ 066 private static final byte[] DECODE_TABLE = { 067 // 0 1 2 3 4 5 6 7 8 9 A B C D E F 068 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, // 00-0f 069 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, // 10-1f 070 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, // 20-2f 071 -1, -1, 26, 27, 28, 29, 30, 31, -1, -1, -1, -1, -1, -1, -1, -1, // 30-3f 2-7 072 -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, // 40-4f A-O 073 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, // 50-5a P-Z 074 -1, -1, -1, -1, -1, // 5b - 5f 075 -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, // 60 - 6f a-o 076 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, // 70 - 7a p-z/**/ 077 }; 078 079 /** 080 * This array is a lookup table that translates 5-bit positive integer index values into their "Base32 Alphabet" 081 * equivalents as specified in Table 3 of RFC 4648. 082 */ 083 private static final byte[] ENCODE_TABLE = { 084 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 085 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', 086 '2', '3', '4', '5', '6', '7', 087 }; 088 089 /** 090 * This array is a lookup table that translates Unicode characters drawn from the "Base32 Hex Alphabet" (as 091 * specified in Table 4 of RFC 4648) into their 5-bit positive integer equivalents. Characters that are not in the 092 * Base32 Hex alphabet but fall within the bounds of the array are translated to -1. 093 */ 094 private static final byte[] HEX_DECODE_TABLE = { 095 // 0 1 2 3 4 5 6 7 8 9 A B C D E F 096 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, // 00-0f 097 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, // 10-1f 098 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, // 20-2f 099 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, -1, -1, -1, -1, -1, -1, // 30-3f 2-7 100 -1, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, // 40-4f A-O 101 25, 26, 27, 28, 29, 30, 31, // 50-56 P-V 102 -1, -1, -1, -1, -1, -1, -1, -1, -1, // 57-5f Z-_ 103 -1, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, // 60-6f `-o 104 25, 26, 27, 28, 29, 30, 31 // 70-76 p-v 105 }; 106 107 /** 108 * This array is a lookup table that translates 5-bit positive integer index values into their 109 * "Base32 Hex Alphabet" equivalents as specified in Table 4 of RFC 4648. 110 */ 111 private static final byte[] HEX_ENCODE_TABLE = { 112 '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 113 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', 114 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 115 }; 116 117 /** Mask used to extract 7 bits, used when decoding final trailing character. */ 118 private static final long MASK_7BITS = 0x7fL; 119 /** Mask used to extract 6 bits, used when decoding final trailing character. */ 120 private static final long MASK_6BITS = 0x3fL; 121 /** Mask used to extract 5 bits, used when encoding Base32 bytes */ 122 private static final int MASK_5BITS = 0x1f; 123 /** Mask used to extract 4 bits, used when decoding final trailing character. */ 124 private static final long MASK_4BITS = 0x0fL; 125 /** Mask used to extract 3 bits, used when decoding final trailing character. */ 126 private static final long MASK_3BITS = 0x07L; 127 /** Mask used to extract 2 bits, used when decoding final trailing character. */ 128 private static final long MASK_2BITS = 0x03L; 129 /** Mask used to extract 1 bits, used when decoding final trailing character. */ 130 private static final long MASK_1BITS = 0x01L; 131 132 // The static final fields above are used for the original static byte[] methods on Base32. 133 // The private member fields below are used with the new streaming approach, which requires 134 // some state be preserved between calls of encode() and decode(). 135 136 /** 137 * Place holder for the bytes we're dealing with for our based logic. 138 * Bitwise operations store and extract the encoding or decoding from this variable. 139 */ 140 141 /** 142 * Convenience variable to help us determine when our buffer is going to run out of room and needs resizing. 143 * {@code decodeSize = {@link #BYTES_PER_ENCODED_BLOCK} - 1 + lineSeparator.length;} 144 */ 145 private final int decodeSize; 146 147 /** 148 * Decode table to use. 149 */ 150 private final byte[] decodeTable; 151 152 /** 153 * Convenience variable to help us determine when our buffer is going to run out of room and needs resizing. 154 * {@code encodeSize = {@link #BYTES_PER_ENCODED_BLOCK} + lineSeparator.length;} 155 */ 156 private final int encodeSize; 157 158 /** 159 * Encode table to use. 160 */ 161 private final byte[] encodeTable; 162 163 /** 164 * Line separator for encoding. Not used when decoding. Only used if lineLength > 0. 165 */ 166 private final byte[] lineSeparator; 167 168 /** 169 * Creates a Base32 codec used for decoding and encoding. 170 * <p> 171 * When encoding the line length is 0 (no chunking). 172 * </p> 173 * 174 */ 175 public Base32() { 176 this(false); 177 } 178 179 /** 180 * Creates a Base32 codec used for decoding and encoding. 181 * <p> 182 * When encoding the line length is 0 (no chunking). 183 * </p> 184 * @param pad byte used as padding byte. 185 */ 186 public Base32(final byte pad) { 187 this(false, pad); 188 } 189 190 /** 191 * Creates a Base32 codec used for decoding and encoding. 192 * <p> 193 * When encoding the line length is 0 (no chunking). 194 * </p> 195 * @param useHex if {@code true} then use Base32 Hex alphabet 196 */ 197 public Base32(final boolean useHex) { 198 this(0, null, useHex, PAD_DEFAULT); 199 } 200 201 /** 202 * Creates a Base32 codec used for decoding and encoding. 203 * <p> 204 * When encoding the line length is 0 (no chunking). 205 * </p> 206 * @param useHex if {@code true} then use Base32 Hex alphabet 207 * @param pad byte used as padding byte. 208 */ 209 public Base32(final boolean useHex, final byte pad) { 210 this(0, null, useHex, pad); 211 } 212 213 /** 214 * Creates a Base32 codec used for decoding and encoding. 215 * <p> 216 * When encoding the line length is given in the constructor, the line separator is CRLF. 217 * </p> 218 * 219 * @param lineLength 220 * Each line of encoded data will be at most of the given length (rounded down to nearest multiple of 221 * 8). If lineLength <= 0, then the output will not be divided into lines (chunks). Ignored when 222 * decoding. 223 */ 224 public Base32(final int lineLength) { 225 this(lineLength, CHUNK_SEPARATOR); 226 } 227 228 /** 229 * Creates a Base32 codec used for decoding and encoding. 230 * <p> 231 * When encoding the line length and line separator are given in the constructor. 232 * </p> 233 * <p> 234 * Line lengths that aren't multiples of 8 will still essentially end up being multiples of 8 in the encoded data. 235 * </p> 236 * 237 * @param lineLength 238 * Each line of encoded data will be at most of the given length (rounded down to nearest multiple of 239 * 8). If lineLength <= 0, then the output will not be divided into lines (chunks). Ignored when 240 * decoding. 241 * @param lineSeparator 242 * Each line of encoded data will end with this sequence of bytes. 243 * @throws IllegalArgumentException 244 * The provided lineSeparator included some Base32 characters. That's not going to work! 245 */ 246 public Base32(final int lineLength, final byte[] lineSeparator) { 247 this(lineLength, lineSeparator, false, PAD_DEFAULT); 248 } 249 250 /** 251 * Creates a Base32 / Base32 Hex codec used for decoding and encoding. 252 * <p> 253 * When encoding the line length and line separator are given in the constructor. 254 * </p> 255 * <p> 256 * Line lengths that aren't multiples of 8 will still essentially end up being multiples of 8 in the encoded data. 257 * </p> 258 * 259 * @param lineLength 260 * Each line of encoded data will be at most of the given length (rounded down to nearest multiple of 261 * 8). If lineLength <= 0, then the output will not be divided into lines (chunks). Ignored when 262 * decoding. 263 * @param lineSeparator 264 * Each line of encoded data will end with this sequence of bytes. 265 * @param useHex 266 * if {@code true}, then use Base32 Hex alphabet, otherwise use Base32 alphabet 267 * @throws IllegalArgumentException 268 * The provided lineSeparator included some Base32 characters. That's not going to work! Or the 269 * lineLength > 0 and lineSeparator is null. 270 */ 271 public Base32(final int lineLength, final byte[] lineSeparator, final boolean useHex) { 272 this(lineLength, lineSeparator, useHex, PAD_DEFAULT); 273 } 274 275 /** 276 * Creates a Base32 / Base32 Hex codec used for decoding and encoding. 277 * <p> 278 * When encoding the line length and line separator are given in the constructor. 279 * </p> 280 * <p> 281 * Line lengths that aren't multiples of 8 will still essentially end up being multiples of 8 in the encoded data. 282 * </p> 283 * 284 * @param lineLength 285 * Each line of encoded data will be at most of the given length (rounded down to nearest multiple of 286 * 8). If lineLength <= 0, then the output will not be divided into lines (chunks). Ignored when 287 * decoding. 288 * @param lineSeparator 289 * Each line of encoded data will end with this sequence of bytes. 290 * @param useHex 291 * if {@code true}, then use Base32 Hex alphabet, otherwise use Base32 alphabet 292 * @param pad byte used as padding byte. 293 * @throws IllegalArgumentException 294 * The provided lineSeparator included some Base32 characters. That's not going to work! Or the 295 * lineLength > 0 and lineSeparator is null. 296 */ 297 public Base32(final int lineLength, final byte[] lineSeparator, final boolean useHex, final byte pad) { 298 super(BYTES_PER_UNENCODED_BLOCK, BYTES_PER_ENCODED_BLOCK, lineLength, 299 lineSeparator == null ? 0 : lineSeparator.length, pad); 300 if (useHex) { 301 this.encodeTable = HEX_ENCODE_TABLE; 302 this.decodeTable = HEX_DECODE_TABLE; 303 } else { 304 this.encodeTable = ENCODE_TABLE; 305 this.decodeTable = DECODE_TABLE; 306 } 307 if (lineLength > 0) { 308 if (lineSeparator == null) { 309 throw new IllegalArgumentException("lineLength " + lineLength + " > 0, but lineSeparator is null"); 310 } 311 // Must be done after initializing the tables 312 if (containsAlphabetOrPad(lineSeparator)) { 313 final String sep = StringUtils.newStringUtf8(lineSeparator); 314 throw new IllegalArgumentException("lineSeparator must not contain Base32 characters: [" + sep + "]"); 315 } 316 this.encodeSize = BYTES_PER_ENCODED_BLOCK + lineSeparator.length; 317 this.lineSeparator = new byte[lineSeparator.length]; 318 System.arraycopy(lineSeparator, 0, this.lineSeparator, 0, lineSeparator.length); 319 } else { 320 this.encodeSize = BYTES_PER_ENCODED_BLOCK; 321 this.lineSeparator = null; 322 } 323 this.decodeSize = this.encodeSize - 1; 324 325 if (isInAlphabet(pad) || isWhiteSpace(pad)) { 326 throw new IllegalArgumentException("pad must not be in alphabet or whitespace"); 327 } 328 } 329 330 /** 331 * <p> 332 * Decodes all of the provided data, starting at inPos, for inAvail bytes. Should be called at least twice: once 333 * with the data to decode, and once with inAvail set to "-1" to alert decoder that EOF has been reached. The "-1" 334 * call is not necessary when decoding, but it doesn't hurt, either. 335 * </p> 336 * <p> 337 * Ignores all non-Base32 characters. This is how chunked (e.g. 76 character) data is handled, since CR and LF are 338 * silently ignored, but has implications for other bytes, too. This method subscribes to the garbage-in, 339 * garbage-out philosophy: it will not check the provided data for validity. 340 * </p> 341 * <p> 342 * Output is written to {@link org.apache.commons.codec.binary.BaseNCodec.Context#buffer Context#buffer} as 8-bit 343 * octets, using {@link org.apache.commons.codec.binary.BaseNCodec.Context#pos Context#pos} as the buffer position 344 * </p> 345 * 346 * @param input byte[] array of ascii data to Base32 decode. 347 * @param inPos Position to start reading data from. 348 * @param inAvail Amount of bytes available from input for decoding. 349 * @param context the context to be used 350 * 351 */ 352 @Override 353 void decode(final byte[] input, int inPos, final int inAvail, final Context context) { 354 // package protected for access from I/O streams 355 356 if (context.eof) { 357 return; 358 } 359 if (inAvail < 0) { 360 context.eof = true; 361 } 362 for (int i = 0; i < inAvail; i++) { 363 final byte b = input[inPos++]; 364 if (b == pad) { 365 // We're done. 366 context.eof = true; 367 break; 368 } 369 final byte[] buffer = ensureBufferSize(decodeSize, context); 370 if (b >= 0 && b < this.decodeTable.length) { 371 final int result = this.decodeTable[b]; 372 if (result >= 0) { 373 context.modulus = (context.modulus+1) % BYTES_PER_ENCODED_BLOCK; 374 // collect decoded bytes 375 context.lbitWorkArea = (context.lbitWorkArea << BITS_PER_ENCODED_BYTE) + result; 376 if (context.modulus == 0) { // we can output the 5 bytes 377 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 32) & MASK_8BITS); 378 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 24) & MASK_8BITS); 379 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 16) & MASK_8BITS); 380 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 8) & MASK_8BITS); 381 buffer[context.pos++] = (byte) (context.lbitWorkArea & MASK_8BITS); 382 } 383 } 384 } 385 } 386 387 // Two forms of EOF as far as Base32 decoder is concerned: actual 388 // EOF (-1) and first time '=' character is encountered in stream. 389 // This approach makes the '=' padding characters completely optional. 390 if (context.eof && context.modulus >= 2) { // if modulus < 2, nothing to do 391 final byte[] buffer = ensureBufferSize(decodeSize, context); 392 393 // we ignore partial bytes, i.e. only multiples of 8 count 394 switch (context.modulus) { 395 case 2 : // 10 bits, drop 2 and output one byte 396 validateCharacter(MASK_2BITS, context); 397 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 2) & MASK_8BITS); 398 break; 399 case 3 : // 15 bits, drop 7 and output 1 byte 400 validateCharacter(MASK_7BITS, context); 401 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 7) & MASK_8BITS); 402 break; 403 case 4 : // 20 bits = 2*8 + 4 404 validateCharacter(MASK_4BITS, context); 405 context.lbitWorkArea = context.lbitWorkArea >> 4; // drop 4 bits 406 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 8) & MASK_8BITS); 407 buffer[context.pos++] = (byte) ((context.lbitWorkArea) & MASK_8BITS); 408 break; 409 case 5 : // 25bits = 3*8 + 1 410 validateCharacter(MASK_1BITS, context); 411 context.lbitWorkArea = context.lbitWorkArea >> 1; 412 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 16) & MASK_8BITS); 413 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 8) & MASK_8BITS); 414 buffer[context.pos++] = (byte) ((context.lbitWorkArea) & MASK_8BITS); 415 break; 416 case 6 : // 30bits = 3*8 + 6 417 validateCharacter(MASK_6BITS, context); 418 context.lbitWorkArea = context.lbitWorkArea >> 6; 419 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 16) & MASK_8BITS); 420 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 8) & MASK_8BITS); 421 buffer[context.pos++] = (byte) ((context.lbitWorkArea) & MASK_8BITS); 422 break; 423 case 7 : // 35 = 4*8 +3 424 validateCharacter(MASK_3BITS, context); 425 context.lbitWorkArea = context.lbitWorkArea >> 3; 426 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 24) & MASK_8BITS); 427 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 16) & MASK_8BITS); 428 buffer[context.pos++] = (byte) ((context.lbitWorkArea >> 8) & MASK_8BITS); 429 buffer[context.pos++] = (byte) ((context.lbitWorkArea) & MASK_8BITS); 430 break; 431 default: 432 // modulus can be 0-7, and we excluded 0,1 already 433 throw new IllegalStateException("Impossible modulus "+context.modulus); 434 } 435 } 436 } 437 438 /** 439 * <p> 440 * Encodes all of the provided data, starting at inPos, for inAvail bytes. Must be called at least twice: once with 441 * the data to encode, and once with inAvail set to "-1" to alert encoder that EOF has been reached, so flush last 442 * remaining bytes (if not multiple of 5). 443 * </p> 444 * 445 * @param input 446 * byte[] array of binary data to Base32 encode. 447 * @param inPos 448 * Position to start reading data from. 449 * @param inAvail 450 * Amount of bytes available from input for encoding. 451 * @param context the context to be used 452 */ 453 @Override 454 void encode(final byte[] input, int inPos, final int inAvail, final Context context) { 455 // package protected for access from I/O streams 456 457 if (context.eof) { 458 return; 459 } 460 // inAvail < 0 is how we're informed of EOF in the underlying data we're 461 // encoding. 462 if (inAvail < 0) { 463 context.eof = true; 464 if (0 == context.modulus && lineLength == 0) { 465 return; // no leftovers to process and not using chunking 466 } 467 final byte[] buffer = ensureBufferSize(encodeSize, context); 468 final int savedPos = context.pos; 469 switch (context.modulus) { // % 5 470 case 0 : 471 break; 472 case 1 : // Only 1 octet; take top 5 bits then remainder 473 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 3) & MASK_5BITS]; // 8-1*5 = 3 474 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea << 2) & MASK_5BITS]; // 5-3=2 475 buffer[context.pos++] = pad; 476 buffer[context.pos++] = pad; 477 buffer[context.pos++] = pad; 478 buffer[context.pos++] = pad; 479 buffer[context.pos++] = pad; 480 buffer[context.pos++] = pad; 481 break; 482 case 2 : // 2 octets = 16 bits to use 483 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 11) & MASK_5BITS]; // 16-1*5 = 11 484 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 6) & MASK_5BITS]; // 16-2*5 = 6 485 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 1) & MASK_5BITS]; // 16-3*5 = 1 486 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea << 4) & MASK_5BITS]; // 5-1 = 4 487 buffer[context.pos++] = pad; 488 buffer[context.pos++] = pad; 489 buffer[context.pos++] = pad; 490 buffer[context.pos++] = pad; 491 break; 492 case 3 : // 3 octets = 24 bits to use 493 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 19) & MASK_5BITS]; // 24-1*5 = 19 494 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 14) & MASK_5BITS]; // 24-2*5 = 14 495 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 9) & MASK_5BITS]; // 24-3*5 = 9 496 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 4) & MASK_5BITS]; // 24-4*5 = 4 497 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea << 1) & MASK_5BITS]; // 5-4 = 1 498 buffer[context.pos++] = pad; 499 buffer[context.pos++] = pad; 500 buffer[context.pos++] = pad; 501 break; 502 case 4 : // 4 octets = 32 bits to use 503 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 27) & MASK_5BITS]; // 32-1*5 = 27 504 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 22) & MASK_5BITS]; // 32-2*5 = 22 505 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 17) & MASK_5BITS]; // 32-3*5 = 17 506 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 12) & MASK_5BITS]; // 32-4*5 = 12 507 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 7) & MASK_5BITS]; // 32-5*5 = 7 508 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 2) & MASK_5BITS]; // 32-6*5 = 2 509 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea << 3) & MASK_5BITS]; // 5-2 = 3 510 buffer[context.pos++] = pad; 511 break; 512 default: 513 throw new IllegalStateException("Impossible modulus "+context.modulus); 514 } 515 context.currentLinePos += context.pos - savedPos; // keep track of current line position 516 // if currentPos == 0 we are at the start of a line, so don't add CRLF 517 if (lineLength > 0 && context.currentLinePos > 0){ // add chunk separator if required 518 System.arraycopy(lineSeparator, 0, buffer, context.pos, lineSeparator.length); 519 context.pos += lineSeparator.length; 520 } 521 } else { 522 for (int i = 0; i < inAvail; i++) { 523 final byte[] buffer = ensureBufferSize(encodeSize, context); 524 context.modulus = (context.modulus+1) % BYTES_PER_UNENCODED_BLOCK; 525 int b = input[inPos++]; 526 if (b < 0) { 527 b += 256; 528 } 529 context.lbitWorkArea = (context.lbitWorkArea << 8) + b; // BITS_PER_BYTE 530 if (0 == context.modulus) { // we have enough bytes to create our output 531 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 35) & MASK_5BITS]; 532 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 30) & MASK_5BITS]; 533 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 25) & MASK_5BITS]; 534 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 20) & MASK_5BITS]; 535 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 15) & MASK_5BITS]; 536 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 10) & MASK_5BITS]; 537 buffer[context.pos++] = encodeTable[(int)(context.lbitWorkArea >> 5) & MASK_5BITS]; 538 buffer[context.pos++] = encodeTable[(int)context.lbitWorkArea & MASK_5BITS]; 539 context.currentLinePos += BYTES_PER_ENCODED_BLOCK; 540 if (lineLength > 0 && lineLength <= context.currentLinePos) { 541 System.arraycopy(lineSeparator, 0, buffer, context.pos, lineSeparator.length); 542 context.pos += lineSeparator.length; 543 context.currentLinePos = 0; 544 } 545 } 546 } 547 } 548 } 549 550 /** 551 * Returns whether or not the {@code octet} is in the Base32 alphabet. 552 * 553 * @param octet 554 * The value to test 555 * @return {@code true} if the value is defined in the the Base32 alphabet {@code false} otherwise. 556 */ 557 @Override 558 public boolean isInAlphabet(final byte octet) { 559 return octet >= 0 && octet < decodeTable.length && decodeTable[octet] != -1; 560 } 561 562 /** 563 * Validates whether decoding the final trailing character is possible in the context 564 * of the set of possible base 32 values. 565 * 566 * <p>The character is valid if the lower bits within the provided mask are zero. This 567 * is used to test the final trailing base-32 digit is zero in the bits that will be discarded. 568 * 569 * @param emptyBitsMask The mask of the lower bits that should be empty 570 * @param context the context to be used 571 * 572 * @throws IllegalArgumentException if the bits being checked contain any non-zero value 573 */ 574 private static void validateCharacter(final long emptyBitsMask, final Context context) { 575 // Use the long bit work area 576 if ((context.lbitWorkArea & emptyBitsMask) != 0) { 577 throw new IllegalArgumentException( 578 "Last encoded character (before the paddings if any) is a valid base 32 alphabet but not a possible value. " + 579 "Expected the discarded bits to be zero."); 580 } 581 } 582}