1 /* 2 * Copyright (C) 2007 The Android Open Source Project 3 * 4 * Licensed under the Apache License, Version 2.0 (the "License"); 5 * you may not use this file except in compliance with the License. 6 * You may obtain a copy of the License at 7 * 8 * http://www.apache.org/licenses/LICENSE-2.0 9 * 10 * Unless required by applicable law or agreed to in writing, software 11 * distributed under the License is distributed on an "AS IS" BASIS, 12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 13 * See the License for the specific language governing permissions and 14 * limitations under the License. 15 */ 16 17 package java.util.regex; 18 19 /** 20 * The result of applying a {@code Pattern} to a given input. See {@link Pattern} for 21 * example uses. 22 */ 23 public final class Matcher implements MatchResult { 24 25 /** 26 * Holds the pattern, that is, the compiled regular expression. 27 */ 28 private Pattern pattern; 29 30 /** 31 * The address of the native peer. 32 * Uses of this must be manually synchronized to avoid native crashes. 33 */ 34 private long address; 35 36 /** 37 * Holds the input text. 38 */ 39 private String input; 40 41 /** 42 * Holds the start of the region, or 0 if the matching should start at the 43 * beginning of the text. 44 */ 45 private int regionStart; 46 47 /** 48 * Holds the end of the region, or input.length() if the matching should 49 * go until the end of the input. 50 */ 51 private int regionEnd; 52 53 /** 54 * Holds the position where the next append operation will take place. 55 */ 56 private int appendPos; 57 58 /** 59 * Reflects whether a match has been found during the most recent find 60 * operation. 61 */ 62 private boolean matchFound; 63 64 /** 65 * Holds the offsets for the most recent match. 66 */ 67 private int[] matchOffsets; 68 69 /** 70 * Reflects whether the bounds of the region are anchoring. 71 */ 72 private boolean anchoringBounds = true; 73 74 /** 75 * Reflects whether the bounds of the region are transparent. 76 */ 77 private boolean transparentBounds; 78 79 /** 80 * Creates a matcher for a given combination of pattern and input. Both 81 * elements can be changed later on. 82 * 83 * @param pattern 84 * the pattern to use. 85 * @param input 86 * the input to use. 87 */ 88 Matcher(Pattern pattern, CharSequence input) { 89 usePattern(pattern); 90 reset(input); 91 } 92 93 /** 94 * Appends a literal part of the input plus a replacement for the current 95 * match to a given {@link StringBuffer}. The literal part is exactly the 96 * part of the input between the previous match and the current match. The 97 * method can be used in conjunction with {@link #find()} and 98 * {@link #appendTail(StringBuffer)} to walk through the input and replace 99 * all occurrences of the {@code Pattern} with something else. 100 * 101 * @param buffer 102 * the {@code StringBuffer} to append to. 103 * @param replacement 104 * the replacement text. 105 * @return the {@code Matcher} itself. 106 * @throws IllegalStateException 107 * if no successful match has been made. 108 */ 109 public Matcher appendReplacement(StringBuffer buffer, String replacement) { 110 buffer.append(input.substring(appendPos, start())); 111 appendEvaluated(buffer, replacement); 112 appendPos = end(); 113 114 return this; 115 } 116 117 /** 118 * Internal helper method to append a given string to a given string buffer. 119 * If the string contains any references to groups, these are replaced by 120 * the corresponding group's contents. 121 * 122 * @param buffer 123 * the string buffer. 124 * @param s 125 * the string to append. 126 */ 127 private void appendEvaluated(StringBuffer buffer, String s) { 128 boolean escape = false; 129 boolean dollar = false; 130 131 for (int i = 0; i < s.length(); i++) { 132 char c = s.charAt(i); 133 if (c == '\\' && !escape) { 134 escape = true; 135 } else if (c == '$' && !escape) { 136 dollar = true; 137 } else if (c >= '0' && c <= '9' && dollar) { 138 buffer.append(group(c - '0')); 139 dollar = false; 140 } else { 141 buffer.append(c); 142 dollar = false; 143 escape = false; 144 } 145 } 146 147 // This seemingly stupid piece of code reproduces a JDK bug. 148 if (escape) { 149 throw new ArrayIndexOutOfBoundsException(s.length()); 150 } 151 } 152 153 /** 154 * Resets the {@code Matcher}. This results in the region being set to the 155 * whole input. Results of a previous find get lost. The next attempt to 156 * find an occurrence of the {@link Pattern} in the string will start at the 157 * beginning of the input. 158 * 159 * @return the {@code Matcher} itself. 160 */ 161 public Matcher reset() { 162 return reset(input, 0, input.length()); 163 } 164 165 /** 166 * Provides a new input and resets the {@code Matcher}. This results in the 167 * region being set to the whole input. Results of a previous find get lost. 168 * The next attempt to find an occurrence of the {@link Pattern} in the 169 * string will start at the beginning of the input. 170 * 171 * @param input 172 * the new input sequence. 173 * 174 * @return the {@code Matcher} itself. 175 */ 176 public Matcher reset(CharSequence input) { 177 return reset(input, 0, input.length()); 178 } 179 180 /** 181 * Resets the Matcher. A new input sequence and a new region can be 182 * specified. Results of a previous find get lost. The next attempt to find 183 * an occurrence of the Pattern in the string will start at the beginning of 184 * the region. This is the internal version of reset() to which the several 185 * public versions delegate. 186 * 187 * @param input 188 * the input sequence. 189 * @param start 190 * the start of the region. 191 * @param end 192 * the end of the region. 193 * 194 * @return the matcher itself. 195 */ 196 private Matcher reset(CharSequence input, int start, int end) { 197 if (input == null) { 198 throw new IllegalArgumentException("input == null"); 199 } 200 201 if (start < 0 || end < 0 || start > input.length() || end > input.length() || start > end) { 202 throw new IndexOutOfBoundsException(); 203 } 204 205 this.input = input.toString(); 206 this.regionStart = start; 207 this.regionEnd = end; 208 resetForInput(); 209 210 matchFound = false; 211 appendPos = 0; 212 213 return this; 214 } 215 216 /** 217 * Sets a new pattern for the {@code Matcher}. Results of a previous find 218 * get lost. The next attempt to find an occurrence of the {@link Pattern} 219 * in the string will start at the beginning of the input. 220 * 221 * @param pattern 222 * the new {@code Pattern}. 223 * 224 * @return the {@code Matcher} itself. 225 */ 226 public Matcher usePattern(Pattern pattern) { 227 if (pattern == null) { 228 throw new IllegalArgumentException("pattern == null"); 229 } 230 231 this.pattern = pattern; 232 233 synchronized (this) { 234 if (address != 0) { 235 closeImpl(address); 236 address = 0; // In case openImpl throws. 237 } 238 address = openImpl(pattern.address); 239 } 240 241 if (input != null) { 242 resetForInput(); 243 } 244 245 matchOffsets = new int[(groupCount() + 1) * 2]; 246 matchFound = false; 247 return this; 248 } 249 250 private void resetForInput() { 251 synchronized (this) { 252 setInputImpl(address, input, regionStart, regionEnd); 253 useAnchoringBoundsImpl(address, anchoringBounds); 254 useTransparentBoundsImpl(address, transparentBounds); 255 } 256 } 257 258 /** 259 * Resets this matcher and sets a region. Only characters inside the region 260 * are considered for a match. 261 * 262 * @param start 263 * the first character of the region. 264 * @param end 265 * the first character after the end of the region. 266 * @return the {@code Matcher} itself. 267 */ 268 public Matcher region(int start, int end) { 269 return reset(input, start, end); 270 } 271 272 /** 273 * Appends the (unmatched) remainder of the input to the given 274 * {@link StringBuffer}. The method can be used in conjunction with 275 * {@link #find()} and {@link #appendReplacement(StringBuffer, String)} to 276 * walk through the input and replace all matches of the {@code Pattern} 277 * with something else. 278 * 279 * @return the {@code StringBuffer}. 280 * @throws IllegalStateException 281 * if no successful match has been made. 282 */ 283 public StringBuffer appendTail(StringBuffer buffer) { 284 if (appendPos < regionEnd) { 285 buffer.append(input.substring(appendPos, regionEnd)); 286 } 287 return buffer; 288 } 289 290 /** 291 * Replaces the first occurrence of this matcher's pattern in the input with 292 * a given string. 293 * 294 * @param replacement 295 * the replacement text. 296 * @return the modified input string. 297 */ 298 public String replaceFirst(String replacement) { 299 reset(); 300 StringBuffer buffer = new StringBuffer(input.length()); 301 if (find()) { 302 appendReplacement(buffer, replacement); 303 } 304 return appendTail(buffer).toString(); 305 } 306 307 /** 308 * Replaces all occurrences of this matcher's pattern in the input with a 309 * given string. 310 * 311 * @param replacement 312 * the replacement text. 313 * @return the modified input string. 314 */ 315 public String replaceAll(String replacement) { 316 reset(); 317 StringBuffer buffer = new StringBuffer(input.length()); 318 while (find()) { 319 appendReplacement(buffer, replacement); 320 } 321 return appendTail(buffer).toString(); 322 } 323 324 /** 325 * Returns the {@link Pattern} instance used inside this matcher. 326 */ 327 public Pattern pattern() { 328 return pattern; 329 } 330 331 /** 332 * Returns true if there is another match in the input, starting 333 * from the given position. The region is ignored. 334 * 335 * @throws IndexOutOfBoundsException if {@code start < 0 || start > input.length()} 336 */ 337 public boolean find(int start) { 338 if (start < 0 || start > input.length()) { 339 throw new IndexOutOfBoundsException("start=" + start + "; length=" + input.length()); 340 } 341 342 synchronized (this) { 343 matchFound = findImpl(address, input, start, matchOffsets); 344 } 345 return matchFound; 346 } 347 348 /** 349 * Moves to the next occurrence of the pattern in the input. If a 350 * previous match was successful, the method continues the search from the 351 * first character following that match in the input. Otherwise it searches 352 * either from the region start (if one has been set), or from position 0. 353 * 354 * @return true if (and only if) a match has been found. 355 */ 356 public boolean find() { 357 synchronized (this) { 358 matchFound = findNextImpl(address, input, matchOffsets); 359 } 360 return matchFound; 361 } 362 363 /** 364 * Tries to match the {@link Pattern}, starting from the beginning of the 365 * region (or the beginning of the input, if no region has been set). 366 * Doesn't require the {@code Pattern} to match against the whole region. 367 * 368 * @return true if (and only if) the {@code Pattern} matches. 369 */ 370 public boolean lookingAt() { 371 synchronized (this) { 372 matchFound = lookingAtImpl(address, input, matchOffsets); 373 } 374 return matchFound; 375 } 376 377 /** 378 * Tries to match the {@link Pattern} against the entire region (or the 379 * entire input, if no region has been set). 380 * 381 * @return true if (and only if) the {@code Pattern} matches the entire 382 * region. 383 */ 384 public boolean matches() { 385 synchronized (this) { 386 matchFound = matchesImpl(address, input, matchOffsets); 387 } 388 return matchFound; 389 } 390 391 /** 392 * Returns a replacement string for the given one that has all backslashes 393 * and dollar signs escaped. 394 */ 395 public static String quoteReplacement(String s) { 396 StringBuilder result = new StringBuilder(s.length()); 397 for (int i = 0; i < s.length(); i++) { 398 char c = s.charAt(i); 399 if (c == '\\' || c == '$') { 400 result.append('\\'); 401 } 402 result.append(c); 403 } 404 return result.toString(); 405 } 406 407 /** 408 * Converts the current match into a separate {@link MatchResult} instance 409 * that is independent from this matcher. The new object is unaffected when 410 * the state of this matcher changes. 411 * 412 * @throws IllegalStateException 413 * if no successful match has been made. 414 */ 415 public MatchResult toMatchResult() { 416 ensureMatch(); 417 return new MatchResultImpl(input, matchOffsets); 418 } 419 420 /** 421 * Determines whether this matcher has anchoring bounds enabled or not. When 422 * anchoring bounds are enabled, the start and end of the input match the 423 * '^' and '$' meta-characters, otherwise not. Anchoring bounds are enabled 424 * by default. 425 * 426 * @return the {@code Matcher} itself. 427 */ 428 public Matcher useAnchoringBounds(boolean value) { 429 synchronized (this) { 430 anchoringBounds = value; 431 useAnchoringBoundsImpl(address, value); 432 } 433 return this; 434 } 435 436 /** 437 * Returns true if this matcher has anchoring bounds enabled. When 438 * anchoring bounds are enabled, the start and end of the input match the 439 * '^' and '$' meta-characters, otherwise not. Anchoring bounds are enabled 440 * by default. 441 */ 442 public boolean hasAnchoringBounds() { 443 return anchoringBounds; 444 } 445 446 /** 447 * Determines whether this matcher has transparent bounds enabled or not. 448 * When transparent bounds are enabled, the parts of the input outside the 449 * region are subject to lookahead and lookbehind, otherwise they are not. 450 * Transparent bounds are disabled by default. 451 * 452 * @return the {@code Matcher} itself. 453 */ 454 public Matcher useTransparentBounds(boolean value) { 455 synchronized (this) { 456 transparentBounds = value; 457 useTransparentBoundsImpl(address, value); 458 } 459 return this; 460 } 461 462 /** 463 * Makes sure that a successful match has been made. Is invoked internally 464 * from various places in the class. 465 * 466 * @throws IllegalStateException 467 * if no successful match has been made. 468 */ 469 private void ensureMatch() { 470 if (!matchFound) { 471 throw new IllegalStateException("No successful match so far"); 472 } 473 } 474 475 /** 476 * Returns true if this matcher has transparent bounds enabled. When 477 * transparent bounds are enabled, the parts of the input outside the region 478 * are subject to lookahead and lookbehind, otherwise they are not. 479 * Transparent bounds are disabled by default. 480 */ 481 public boolean hasTransparentBounds() { 482 return transparentBounds; 483 } 484 485 /** 486 * Returns this matcher's region start, that is, the index of the first character that is 487 * considered for a match. 488 */ 489 public int regionStart() { 490 return regionStart; 491 } 492 493 /** 494 * Returns this matcher's region end, that is, the index of the first character that is 495 * not considered for a match. 496 */ 497 public int regionEnd() { 498 return regionEnd; 499 } 500 501 /** 502 * Returns true if the most recent match succeeded and additional input could cause 503 * it to fail. If this method returns false and a match was found, then more input 504 * might change the match but the match won't be lost. If a match was not found, 505 * then requireEnd has no meaning. 506 */ 507 public boolean requireEnd() { 508 synchronized (this) { 509 return requireEndImpl(address); 510 } 511 } 512 513 /** 514 * Returns true if the most recent matching operation attempted to access 515 * additional text beyond the available input, meaning that additional input 516 * could change the results of the match. 517 */ 518 public boolean hitEnd() { 519 synchronized (this) { 520 return hitEndImpl(address); 521 } 522 } 523 524 @Override protected void finalize() throws Throwable { 525 try { 526 synchronized (this) { 527 closeImpl(address); 528 } 529 } finally { 530 super.finalize(); 531 } 532 } 533 534 /** 535 * Returns a string representing this {@code Matcher}. 536 * The format of this string is unspecified. 537 */ 538 @Override public String toString() { 539 return getClass().getName() + "[pattern=" + pattern() + 540 " region=" + regionStart() + "," + regionEnd() + 541 " lastmatch=" + (matchFound ? group() : "") + "]"; 542 } 543 544 /** 545 * {@inheritDoc} 546 * 547 * @throws IllegalStateException if no successful match has been made. 548 */ 549 public int end() { 550 return end(0); 551 } 552 553 /** 554 * {@inheritDoc} 555 * 556 * @throws IllegalStateException if no successful match has been made. 557 */ 558 public int end(int group) { 559 ensureMatch(); 560 return matchOffsets[(group * 2) + 1]; 561 } 562 563 /** 564 * {@inheritDoc} 565 * 566 * @throws IllegalStateException if no successful match has been made. 567 */ 568 public String group() { 569 return group(0); 570 } 571 572 /** 573 * {@inheritDoc} 574 * 575 * @throws IllegalStateException if no successful match has been made. 576 */ 577 public String group(int group) { 578 ensureMatch(); 579 int from = matchOffsets[group * 2]; 580 int to = matchOffsets[(group * 2) + 1]; 581 if (from == -1 || to == -1) { 582 return null; 583 } else { 584 return input.substring(from, to); 585 } 586 } 587 588 /** 589 * {@inheritDoc} 590 */ 591 public int groupCount() { 592 synchronized (this) { 593 return groupCountImpl(address); 594 } 595 } 596 597 /** 598 * {@inheritDoc} 599 * 600 * @throws IllegalStateException if no successful match has been made. 601 */ 602 public int start() { 603 return start(0); 604 } 605 606 /** 607 * {@inheritDoc} 608 * 609 * @throws IllegalStateException if no successful match has been made. 610 */ 611 public int start(int group) throws IllegalStateException { 612 ensureMatch(); 613 return matchOffsets[group * 2]; 614 } 615 616 private static native void closeImpl(long addr); 617 private static native boolean findImpl(long addr, String s, int startIndex, int[] offsets); 618 private static native boolean findNextImpl(long addr, String s, int[] offsets); 619 private static native int groupCountImpl(long addr); 620 private static native boolean hitEndImpl(long addr); 621 private static native boolean lookingAtImpl(long addr, String s, int[] offsets); 622 private static native boolean matchesImpl(long addr, String s, int[] offsets); 623 private static native long openImpl(long patternAddr); 624 private static native boolean requireEndImpl(long addr); 625 private static native void setInputImpl(long addr, String s, int start, int end); 626 private static native void useAnchoringBoundsImpl(long addr, boolean value); 627 private static native void useTransparentBoundsImpl(long addr, boolean value); 628 } 629