Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 1 | /* Jackson JSON-processor. |
| 2 | * |
| 3 | * Copyright (c) 2007- Tatu Saloranta, tatu.saloranta@iki.fi |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 4 | */ |
| 5 | package com.fasterxml.jackson.core; |
| 6 | |
| 7 | import java.util.Arrays; |
| 8 | |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 9 | import com.fasterxml.jackson.core.util.ByteArrayBuilder; |
| 10 | |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 11 | /** |
| 12 | * Abstract base class used to define specific details of which |
| 13 | * variant of Base64 encoding/decoding is to be used. Although there is |
| 14 | * somewhat standard basic version (so-called "MIME Base64"), other variants |
| 15 | * exists, see <a href="http://en.wikipedia.org/wiki/Base64">Base64 Wikipedia entry</a> for details. |
| 16 | * |
| 17 | * @author Tatu Saloranta |
| 18 | */ |
| 19 | public final class Base64Variant |
Tatu Saloranta | 41bd0eb | 2012-10-05 15:01:05 -0700 | [diff] [blame] | 20 | implements java.io.Serializable |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 21 | { |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 22 | private final static int INT_SPACE = 0x20; |
| 23 | |
Tatu Saloranta | 41bd0eb | 2012-10-05 15:01:05 -0700 | [diff] [blame] | 24 | // We'll only serialize name |
| 25 | private static final long serialVersionUID = 1L; |
| 26 | |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 27 | /** |
| 28 | * Placeholder used by "no padding" variant, to be used when a character |
| 29 | * value is needed. |
| 30 | */ |
| 31 | final static char PADDING_CHAR_NONE = '\0'; |
| 32 | |
| 33 | /** |
| 34 | * Marker used to denote ascii characters that do not correspond |
| 35 | * to a 6-bit value (in this variant), and is not used as a padding |
| 36 | * character. |
| 37 | */ |
| 38 | public final static int BASE64_VALUE_INVALID = -1; |
| 39 | |
| 40 | /** |
| 41 | * Marker used to denote ascii character (in decoding table) that |
| 42 | * is the padding character using this variant (if any). |
| 43 | */ |
| 44 | public final static int BASE64_VALUE_PADDING = -2; |
| 45 | |
| 46 | /* |
| 47 | /********************************************************** |
| 48 | /* Encoding/decoding tables |
| 49 | /********************************************************** |
| 50 | */ |
| 51 | |
| 52 | /** |
| 53 | * Decoding table used for base 64 decoding. |
| 54 | */ |
Tatu Saloranta | 41bd0eb | 2012-10-05 15:01:05 -0700 | [diff] [blame] | 55 | private final transient int[] _asciiToBase64 = new int[128]; |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 56 | |
| 57 | /** |
| 58 | * Encoding table used for base 64 decoding when output is done |
| 59 | * as characters. |
| 60 | */ |
Tatu Saloranta | 41bd0eb | 2012-10-05 15:01:05 -0700 | [diff] [blame] | 61 | private final transient char[] _base64ToAsciiC = new char[64]; |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 62 | |
| 63 | /** |
| 64 | * Alternative encoding table used for base 64 decoding when output is done |
| 65 | * as ascii bytes. |
| 66 | */ |
Tatu Saloranta | 41bd0eb | 2012-10-05 15:01:05 -0700 | [diff] [blame] | 67 | private final transient byte[] _base64ToAsciiB = new byte[64]; |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 68 | |
| 69 | /* |
| 70 | /********************************************************** |
| 71 | /* Other configuration |
| 72 | /********************************************************** |
| 73 | */ |
| 74 | |
| 75 | /** |
| 76 | * Symbolic name of variant; used for diagnostics/debugging. |
Tatu Saloranta | 41bd0eb | 2012-10-05 15:01:05 -0700 | [diff] [blame] | 77 | *<p> |
| 78 | * Note that this is the only non-transient field; used when reading |
Tatu Saloranta | e962fb9 | 2016-07-21 14:14:26 -0700 | [diff] [blame] | 79 | * back from serialized state. |
| 80 | *<p> |
| 81 | * Also: must not be private, accessed from `BaseVariants` |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 82 | */ |
Tatu Saloranta | e962fb9 | 2016-07-21 14:14:26 -0700 | [diff] [blame] | 83 | final String _name; |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 84 | |
| 85 | /** |
| 86 | * Whether this variant uses padding or not. |
| 87 | */ |
Tatu Saloranta | e962fb9 | 2016-07-21 14:14:26 -0700 | [diff] [blame] | 88 | private final transient boolean _usesPadding; |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 89 | |
| 90 | /** |
| 91 | * Characted used for padding, if any ({@link #PADDING_CHAR_NONE} if not). |
| 92 | */ |
Tatu Saloranta | e962fb9 | 2016-07-21 14:14:26 -0700 | [diff] [blame] | 93 | private final transient char _paddingChar; |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 94 | |
| 95 | /** |
| 96 | * Maximum number of encoded base64 characters to output during encoding |
| 97 | * before adding a linefeed, if line length is to be limited |
| 98 | * ({@link java.lang.Integer#MAX_VALUE} if not limited). |
| 99 | *<p> |
| 100 | * Note: for some output modes (when writing attributes) linefeeds may |
| 101 | * need to be avoided, and this value ignored. |
| 102 | */ |
Tatu Saloranta | e962fb9 | 2016-07-21 14:14:26 -0700 | [diff] [blame] | 103 | private final transient int _maxLineLength; |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 104 | |
| 105 | /* |
| 106 | /********************************************************** |
| 107 | /* Life-cycle |
| 108 | /********************************************************** |
| 109 | */ |
| 110 | |
| 111 | public Base64Variant(String name, String base64Alphabet, boolean usesPadding, char paddingChar, int maxLineLength) |
| 112 | { |
| 113 | _name = name; |
| 114 | _usesPadding = usesPadding; |
| 115 | _paddingChar = paddingChar; |
| 116 | _maxLineLength = maxLineLength; |
| 117 | |
| 118 | // Ok and then we need to create codec tables. |
| 119 | |
| 120 | // First the main encoding table: |
| 121 | int alphaLen = base64Alphabet.length(); |
| 122 | if (alphaLen != 64) { |
| 123 | throw new IllegalArgumentException("Base64Alphabet length must be exactly 64 (was "+alphaLen+")"); |
| 124 | } |
| 125 | |
| 126 | // And then secondary encoding table and decoding table: |
| 127 | base64Alphabet.getChars(0, alphaLen, _base64ToAsciiC, 0); |
| 128 | Arrays.fill(_asciiToBase64, BASE64_VALUE_INVALID); |
| 129 | for (int i = 0; i < alphaLen; ++i) { |
| 130 | char alpha = _base64ToAsciiC[i]; |
| 131 | _base64ToAsciiB[i] = (byte) alpha; |
| 132 | _asciiToBase64[alpha] = i; |
| 133 | } |
| 134 | |
| 135 | // Plus if we use padding, add that in too |
| 136 | if (usesPadding) { |
| 137 | _asciiToBase64[(int) paddingChar] = BASE64_VALUE_PADDING; |
| 138 | } |
| 139 | } |
| 140 | |
| 141 | /** |
| 142 | * "Copy constructor" that can be used when the base alphabet is identical |
| 143 | * to one used by another variant except for the maximum line length |
| 144 | * (and obviously, name). |
| 145 | */ |
| 146 | public Base64Variant(Base64Variant base, String name, int maxLineLength) |
| 147 | { |
| 148 | this(base, name, base._usesPadding, base._paddingChar, maxLineLength); |
| 149 | } |
| 150 | |
| 151 | /** |
| 152 | * "Copy constructor" that can be used when the base alphabet is identical |
| 153 | * to one used by another variant, but other details (padding, maximum |
| 154 | * line length) differ |
| 155 | */ |
| 156 | public Base64Variant(Base64Variant base, String name, boolean usesPadding, char paddingChar, int maxLineLength) |
| 157 | { |
| 158 | _name = name; |
| 159 | byte[] srcB = base._base64ToAsciiB; |
| 160 | System.arraycopy(srcB, 0, this._base64ToAsciiB, 0, srcB.length); |
| 161 | char[] srcC = base._base64ToAsciiC; |
| 162 | System.arraycopy(srcC, 0, this._base64ToAsciiC, 0, srcC.length); |
| 163 | int[] srcV = base._asciiToBase64; |
| 164 | System.arraycopy(srcV, 0, this._asciiToBase64, 0, srcV.length); |
| 165 | |
| 166 | _usesPadding = usesPadding; |
| 167 | _paddingChar = paddingChar; |
| 168 | _maxLineLength = maxLineLength; |
| 169 | } |
| 170 | |
| 171 | /* |
| 172 | /********************************************************** |
Tatu Saloranta | 41bd0eb | 2012-10-05 15:01:05 -0700 | [diff] [blame] | 173 | /* Serializable overrides |
| 174 | /********************************************************** |
| 175 | */ |
| 176 | |
| 177 | /** |
| 178 | * Method used to "demote" deserialized instances back to |
| 179 | * canonical ones |
| 180 | */ |
| 181 | protected Object readResolve() { |
| 182 | return Base64Variants.valueOf(_name); |
| 183 | } |
| 184 | |
| 185 | /* |
| 186 | /********************************************************** |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 187 | /* Public accessors |
| 188 | /********************************************************** |
| 189 | */ |
| 190 | |
| 191 | public String getName() { return _name; } |
| 192 | |
| 193 | public boolean usesPadding() { return _usesPadding; } |
| 194 | public boolean usesPaddingChar(char c) { return c == _paddingChar; } |
| 195 | public boolean usesPaddingChar(int ch) { return ch == (int) _paddingChar; } |
| 196 | public char getPaddingChar() { return _paddingChar; } |
| 197 | public byte getPaddingByte() { return (byte)_paddingChar; } |
| 198 | |
| 199 | public int getMaxLineLength() { return _maxLineLength; } |
| 200 | |
| 201 | /* |
| 202 | /********************************************************** |
| 203 | /* Decoding support |
| 204 | /********************************************************** |
| 205 | */ |
| 206 | |
| 207 | /** |
| 208 | * @return 6-bit decoded value, if valid character; |
| 209 | */ |
| 210 | public int decodeBase64Char(char c) |
| 211 | { |
| 212 | int ch = (int) c; |
| 213 | return (ch <= 127) ? _asciiToBase64[ch] : BASE64_VALUE_INVALID; |
| 214 | } |
| 215 | |
| 216 | public int decodeBase64Char(int ch) |
| 217 | { |
| 218 | return (ch <= 127) ? _asciiToBase64[ch] : BASE64_VALUE_INVALID; |
| 219 | } |
| 220 | |
| 221 | public int decodeBase64Byte(byte b) |
| 222 | { |
| 223 | int ch = (int) b; |
Tatu Saloranta | e962fb9 | 2016-07-21 14:14:26 -0700 | [diff] [blame] | 224 | // note: cast retains sign, so it's from -128 to +127 |
| 225 | if (ch < 0) { |
| 226 | return BASE64_VALUE_INVALID; |
| 227 | } |
| 228 | return _asciiToBase64[ch]; |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 229 | } |
| 230 | |
| 231 | /* |
| 232 | /********************************************************** |
| 233 | /* Encoding support |
| 234 | /********************************************************** |
| 235 | */ |
| 236 | |
| 237 | public char encodeBase64BitsAsChar(int value) |
| 238 | { |
| 239 | /* Let's assume caller has done necessary checks; this |
| 240 | * method must be fast and inlinable |
| 241 | */ |
| 242 | return _base64ToAsciiC[value]; |
| 243 | } |
| 244 | |
| 245 | /** |
| 246 | * Method that encodes given right-aligned (LSB) 24-bit value |
| 247 | * into 4 base64 characters, stored in given result buffer. |
| 248 | */ |
| 249 | public int encodeBase64Chunk(int b24, char[] buffer, int ptr) |
| 250 | { |
| 251 | buffer[ptr++] = _base64ToAsciiC[(b24 >> 18) & 0x3F]; |
| 252 | buffer[ptr++] = _base64ToAsciiC[(b24 >> 12) & 0x3F]; |
| 253 | buffer[ptr++] = _base64ToAsciiC[(b24 >> 6) & 0x3F]; |
| 254 | buffer[ptr++] = _base64ToAsciiC[b24 & 0x3F]; |
| 255 | return ptr; |
| 256 | } |
| 257 | |
| 258 | public void encodeBase64Chunk(StringBuilder sb, int b24) |
| 259 | { |
| 260 | sb.append(_base64ToAsciiC[(b24 >> 18) & 0x3F]); |
| 261 | sb.append(_base64ToAsciiC[(b24 >> 12) & 0x3F]); |
| 262 | sb.append(_base64ToAsciiC[(b24 >> 6) & 0x3F]); |
| 263 | sb.append(_base64ToAsciiC[b24 & 0x3F]); |
| 264 | } |
| 265 | |
| 266 | /** |
| 267 | * Method that outputs partial chunk (which only encodes one |
| 268 | * or two bytes of data). Data given is still aligned same as if |
| 269 | * it as full data; that is, missing data is at the "right end" |
| 270 | * (LSB) of int. |
| 271 | * |
| 272 | * @param outputBytes Number of encoded bytes included (either 1 or 2) |
| 273 | */ |
| 274 | public int encodeBase64Partial(int bits, int outputBytes, char[] buffer, int outPtr) |
| 275 | { |
| 276 | buffer[outPtr++] = _base64ToAsciiC[(bits >> 18) & 0x3F]; |
| 277 | buffer[outPtr++] = _base64ToAsciiC[(bits >> 12) & 0x3F]; |
| 278 | if (_usesPadding) { |
| 279 | buffer[outPtr++] = (outputBytes == 2) ? |
| 280 | _base64ToAsciiC[(bits >> 6) & 0x3F] : _paddingChar; |
| 281 | buffer[outPtr++] = _paddingChar; |
| 282 | } else { |
| 283 | if (outputBytes == 2) { |
| 284 | buffer[outPtr++] = _base64ToAsciiC[(bits >> 6) & 0x3F]; |
| 285 | } |
| 286 | } |
| 287 | return outPtr; |
| 288 | } |
| 289 | |
| 290 | public void encodeBase64Partial(StringBuilder sb, int bits, int outputBytes) |
| 291 | { |
| 292 | sb.append(_base64ToAsciiC[(bits >> 18) & 0x3F]); |
| 293 | sb.append(_base64ToAsciiC[(bits >> 12) & 0x3F]); |
| 294 | if (_usesPadding) { |
| 295 | sb.append((outputBytes == 2) ? |
| 296 | _base64ToAsciiC[(bits >> 6) & 0x3F] : _paddingChar); |
| 297 | sb.append(_paddingChar); |
| 298 | } else { |
| 299 | if (outputBytes == 2) { |
| 300 | sb.append(_base64ToAsciiC[(bits >> 6) & 0x3F]); |
| 301 | } |
| 302 | } |
| 303 | } |
| 304 | |
| 305 | public byte encodeBase64BitsAsByte(int value) |
| 306 | { |
| 307 | // As with above, assuming it is 6-bit value |
| 308 | return _base64ToAsciiB[value]; |
| 309 | } |
| 310 | |
| 311 | /** |
| 312 | * Method that encodes given right-aligned (LSB) 24-bit value |
| 313 | * into 4 base64 bytes (ascii), stored in given result buffer. |
| 314 | */ |
| 315 | public int encodeBase64Chunk(int b24, byte[] buffer, int ptr) |
| 316 | { |
| 317 | buffer[ptr++] = _base64ToAsciiB[(b24 >> 18) & 0x3F]; |
| 318 | buffer[ptr++] = _base64ToAsciiB[(b24 >> 12) & 0x3F]; |
| 319 | buffer[ptr++] = _base64ToAsciiB[(b24 >> 6) & 0x3F]; |
| 320 | buffer[ptr++] = _base64ToAsciiB[b24 & 0x3F]; |
| 321 | return ptr; |
| 322 | } |
| 323 | |
| 324 | /** |
| 325 | * Method that outputs partial chunk (which only encodes one |
| 326 | * or two bytes of data). Data given is still aligned same as if |
| 327 | * it as full data; that is, missing data is at the "right end" |
| 328 | * (LSB) of int. |
| 329 | * |
| 330 | * @param outputBytes Number of encoded bytes included (either 1 or 2) |
| 331 | */ |
| 332 | public int encodeBase64Partial(int bits, int outputBytes, byte[] buffer, int outPtr) |
| 333 | { |
| 334 | buffer[outPtr++] = _base64ToAsciiB[(bits >> 18) & 0x3F]; |
| 335 | buffer[outPtr++] = _base64ToAsciiB[(bits >> 12) & 0x3F]; |
| 336 | if (_usesPadding) { |
| 337 | byte pb = (byte) _paddingChar; |
| 338 | buffer[outPtr++] = (outputBytes == 2) ? |
| 339 | _base64ToAsciiB[(bits >> 6) & 0x3F] : pb; |
| 340 | buffer[outPtr++] = pb; |
| 341 | } else { |
| 342 | if (outputBytes == 2) { |
| 343 | buffer[outPtr++] = _base64ToAsciiB[(bits >> 6) & 0x3F]; |
| 344 | } |
| 345 | } |
| 346 | return outPtr; |
| 347 | } |
| 348 | |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 349 | /* |
| 350 | /********************************************************** |
| 351 | /* Convenience conversion methods for String to/from bytes |
| 352 | /* use case. |
| 353 | /********************************************************** |
| 354 | */ |
| 355 | |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 356 | /** |
| 357 | * Convenience method for converting given byte array as base64 encoded |
| 358 | * String using this variant's settings. |
| 359 | * Resulting value is "raw", that is, not enclosed in double-quotes. |
| 360 | * |
| 361 | * @param input Byte array to encode |
| 362 | */ |
| 363 | public String encode(byte[] input) |
| 364 | { |
| 365 | return encode(input, false); |
| 366 | } |
| 367 | |
| 368 | /** |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 369 | * Convenience method for converting given byte array as base64 encoded String |
| 370 | * using this variant's settings, |
| 371 | * optionally enclosed in double-quotes. |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 372 | * |
| 373 | * @param input Byte array to encode |
| 374 | * @param addQuotes Whether to surround resulting value in double quotes or not |
| 375 | */ |
| 376 | public String encode(byte[] input, boolean addQuotes) |
| 377 | { |
| 378 | int inputEnd = input.length; |
| 379 | StringBuilder sb; |
| 380 | { |
| 381 | // let's approximate... 33% overhead, ~= 3/8 (0.375) |
| 382 | int outputLen = inputEnd + (inputEnd >> 2) + (inputEnd >> 3); |
| 383 | sb = new StringBuilder(outputLen); |
| 384 | } |
| 385 | if (addQuotes) { |
| 386 | sb.append('"'); |
| 387 | } |
| 388 | |
| 389 | int chunksBeforeLF = getMaxLineLength() >> 2; |
| 390 | |
| 391 | // Ok, first we loop through all full triplets of data: |
| 392 | int inputPtr = 0; |
| 393 | int safeInputEnd = inputEnd-3; // to get only full triplets |
| 394 | |
| 395 | while (inputPtr <= safeInputEnd) { |
| 396 | // First, mash 3 bytes into lsb of 32-bit int |
| 397 | int b24 = ((int) input[inputPtr++]) << 8; |
| 398 | b24 |= ((int) input[inputPtr++]) & 0xFF; |
| 399 | b24 = (b24 << 8) | (((int) input[inputPtr++]) & 0xFF); |
| 400 | encodeBase64Chunk(sb, b24); |
| 401 | if (--chunksBeforeLF <= 0) { |
| 402 | // note: must quote in JSON value, so not really useful... |
| 403 | sb.append('\\'); |
| 404 | sb.append('n'); |
| 405 | chunksBeforeLF = getMaxLineLength() >> 2; |
| 406 | } |
| 407 | } |
| 408 | |
| 409 | // And then we may have 1 or 2 leftover bytes to encode |
| 410 | int inputLeft = inputEnd - inputPtr; // 0, 1 or 2 |
| 411 | if (inputLeft > 0) { // yes, but do we have room for output? |
| 412 | int b24 = ((int) input[inputPtr++]) << 16; |
| 413 | if (inputLeft == 2) { |
| 414 | b24 |= (((int) input[inputPtr++]) & 0xFF) << 8; |
| 415 | } |
| 416 | encodeBase64Partial(sb, b24, inputLeft); |
| 417 | } |
| 418 | |
| 419 | if (addQuotes) { |
| 420 | sb.append('"'); |
| 421 | } |
| 422 | return sb.toString(); |
| 423 | } |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 424 | |
| 425 | /** |
| 426 | * Convenience method for decoding contents of a Base64-encoded String, |
| 427 | * using this variant's settings. |
| 428 | * |
| 429 | * @param input |
| 430 | * |
| 431 | * @since 2.2.3 |
| 432 | * |
| 433 | * @throws IllegalArgumentException if input is not valid base64 encoded data |
| 434 | */ |
Tatu Saloranta | 5963bc4 | 2013-08-03 20:07:13 -0700 | [diff] [blame] | 435 | @SuppressWarnings("resource") |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 436 | public byte[] decode(String input) throws IllegalArgumentException |
| 437 | { |
| 438 | ByteArrayBuilder b = new ByteArrayBuilder(); |
| 439 | decode(input, b); |
| 440 | return b.toByteArray(); |
| 441 | } |
| 442 | |
| 443 | /** |
| 444 | * Convenience method for decoding contents of a Base64-encoded String, |
| 445 | * using this variant's settings |
| 446 | * and appending decoded binary data using provided {@link ByteArrayBuilder}. |
| 447 | *<p> |
| 448 | * NOTE: builder will NOT be reset before decoding (nor cleared afterwards); |
| 449 | * assumption is that caller will ensure it is given in proper state, and |
| 450 | * used as appropriate afterwards. |
| 451 | * |
Tatu Saloranta | 7c3e56b | 2016-11-25 18:35:27 -0800 | [diff] [blame] | 452 | * @since 2.3 |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 453 | * |
| 454 | * @throws IllegalArgumentException if input is not valid base64 encoded data |
| 455 | */ |
| 456 | public void decode(String str, ByteArrayBuilder builder) throws IllegalArgumentException |
| 457 | { |
| 458 | int ptr = 0; |
| 459 | int len = str.length(); |
| 460 | |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 461 | while (ptr < len) { |
| 462 | // first, we'll skip preceding white space, if any |
| 463 | char ch; |
| 464 | do { |
| 465 | ch = str.charAt(ptr++); |
Tatu Saloranta | 7c3e56b | 2016-11-25 18:35:27 -0800 | [diff] [blame] | 466 | } while ((ptr < len) && (ch <= INT_SPACE)); |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 467 | int bits = decodeBase64Char(ch); |
| 468 | if (bits < 0) { |
| 469 | _reportInvalidBase64(ch, 0, null); |
| 470 | } |
| 471 | int decodedData = bits; |
| 472 | // then second base64 char; can't get padding yet, nor ws |
| 473 | if (ptr >= len) { |
| 474 | _reportBase64EOF(); |
| 475 | } |
| 476 | ch = str.charAt(ptr++); |
| 477 | bits = decodeBase64Char(ch); |
| 478 | if (bits < 0) { |
| 479 | _reportInvalidBase64(ch, 1, null); |
| 480 | } |
| 481 | decodedData = (decodedData << 6) | bits; |
| 482 | // third base64 char; can be padding, but not ws |
| 483 | if (ptr >= len) { |
| 484 | // but as per [JACKSON-631] can be end-of-input, iff not using padding |
| 485 | if (!usesPadding()) { |
| 486 | decodedData >>= 4; |
| 487 | builder.append(decodedData); |
| 488 | break; |
| 489 | } |
| 490 | _reportBase64EOF(); |
| 491 | } |
| 492 | ch = str.charAt(ptr++); |
| 493 | bits = decodeBase64Char(ch); |
| 494 | |
| 495 | // First branch: can get padding (-> 1 byte) |
| 496 | if (bits < 0) { |
| 497 | if (bits != Base64Variant.BASE64_VALUE_PADDING) { |
| 498 | _reportInvalidBase64(ch, 2, null); |
| 499 | } |
| 500 | // Ok, must get padding |
| 501 | if (ptr >= len) { |
| 502 | _reportBase64EOF(); |
| 503 | } |
| 504 | ch = str.charAt(ptr++); |
| 505 | if (!usesPaddingChar(ch)) { |
| 506 | _reportInvalidBase64(ch, 3, "expected padding character '"+getPaddingChar()+"'"); |
| 507 | } |
| 508 | // Got 12 bits, only need 8, need to shift |
| 509 | decodedData >>= 4; |
| 510 | builder.append(decodedData); |
| 511 | continue; |
| 512 | } |
| 513 | // Nope, 2 or 3 bytes |
| 514 | decodedData = (decodedData << 6) | bits; |
| 515 | // fourth and last base64 char; can be padding, but not ws |
| 516 | if (ptr >= len) { |
| 517 | // but as per [JACKSON-631] can be end-of-input, iff not using padding |
| 518 | if (!usesPadding()) { |
| 519 | decodedData >>= 2; |
| 520 | builder.appendTwoBytes(decodedData); |
| 521 | break; |
| 522 | } |
| 523 | _reportBase64EOF(); |
| 524 | } |
| 525 | ch = str.charAt(ptr++); |
| 526 | bits = decodeBase64Char(ch); |
| 527 | if (bits < 0) { |
| 528 | if (bits != Base64Variant.BASE64_VALUE_PADDING) { |
| 529 | _reportInvalidBase64(ch, 3, null); |
| 530 | } |
| 531 | decodedData >>= 2; |
| 532 | builder.appendTwoBytes(decodedData); |
| 533 | } else { |
| 534 | // otherwise, our triple is now complete |
| 535 | decodedData = (decodedData << 6) | bits; |
| 536 | builder.appendThreeBytes(decodedData); |
| 537 | } |
| 538 | } |
| 539 | } |
| 540 | |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 541 | /* |
| 542 | /********************************************************** |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 543 | /* Overridden standard methods |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 544 | /********************************************************** |
| 545 | */ |
| 546 | |
| 547 | @Override |
| 548 | public String toString() { return _name; } |
Tatu Saloranta | 26894d0 | 2013-07-02 17:03:25 -0700 | [diff] [blame] | 549 | |
| 550 | @Override |
| 551 | public boolean equals(Object o) { |
| 552 | // identity comparison should be dine |
| 553 | return (o == this); |
| 554 | } |
| 555 | |
| 556 | @Override |
| 557 | public int hashCode() { |
| 558 | return _name.hashCode(); |
| 559 | } |
| 560 | |
| 561 | /* |
| 562 | /********************************************************** |
| 563 | /* Internal helper methods |
| 564 | /********************************************************** |
| 565 | */ |
| 566 | |
| 567 | /** |
| 568 | * @param bindex Relative index within base64 character unit; between 0 |
| 569 | * and 3 (as unit has exactly 4 characters) |
| 570 | */ |
| 571 | protected void _reportInvalidBase64(char ch, int bindex, String msg) |
| 572 | throws IllegalArgumentException |
| 573 | { |
| 574 | String base; |
| 575 | if (ch <= INT_SPACE) { |
| 576 | base = "Illegal white space character (code 0x"+Integer.toHexString(ch)+") as character #"+(bindex+1)+" of 4-char base64 unit: can only used between units"; |
| 577 | } else if (usesPaddingChar(ch)) { |
| 578 | base = "Unexpected padding character ('"+getPaddingChar()+"') as character #"+(bindex+1)+" of 4-char base64 unit: padding only legal as 3rd or 4th character"; |
| 579 | } else if (!Character.isDefined(ch) || Character.isISOControl(ch)) { |
| 580 | // Not sure if we can really get here... ? (most illegal xml chars are caught at lower level) |
| 581 | base = "Illegal character (code 0x"+Integer.toHexString(ch)+") in base64 content"; |
| 582 | } else { |
| 583 | base = "Illegal character '"+ch+"' (code 0x"+Integer.toHexString(ch)+") in base64 content"; |
| 584 | } |
| 585 | if (msg != null) { |
| 586 | base = base + ": " + msg; |
| 587 | } |
| 588 | throw new IllegalArgumentException(base); |
| 589 | } |
| 590 | |
| 591 | protected void _reportBase64EOF() throws IllegalArgumentException { |
| 592 | throw new IllegalArgumentException("Unexpected end-of-String in base64 content"); |
| 593 | } |
Tatu Saloranta | f15531c | 2011-12-22 23:00:40 -0800 | [diff] [blame] | 594 | } |
| 595 | |