1 /* 2 * Copyright (C) 2014 The Android Open Source Project 3 * Copyright (c) 1999, 2013, Oracle and/or its affiliates. All rights reserved. 4 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. 5 * 6 * This code is free software; you can redistribute it and/or modify it 7 * under the terms of the GNU General Public License version 2 only, as 8 * published by the Free Software Foundation. Oracle designates this 9 * particular file as subject to the "Classpath" exception as provided 10 * by Oracle in the LICENSE file that accompanied this code. 11 * 12 * This code is distributed in the hope that it will be useful, but WITHOUT 13 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or 14 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 15 * version 2 for more details (a copy is included in the LICENSE file that 16 * accompanied this code). 17 * 18 * You should have received a copy of the GNU General Public License version 19 * 2 along with this work; if not, write to the Free Software Foundation, 20 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. 21 * 22 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA 23 * or visit www.oracle.com if you need additional information or have any 24 * questions. 25 */ 26 27 package java.util.regex; 28 29 import com.android.icu.util.regex.PatternNative; 30 import dalvik.system.VMRuntime; 31 32 import java.util.Iterator; 33 import java.util.ArrayList; 34 import java.util.NoSuchElementException; 35 import java.util.Spliterator; 36 import java.util.Spliterators; 37 import java.util.function.Predicate; 38 import java.util.stream.Stream; 39 import java.util.stream.StreamSupport; 40 41 import libcore.util.EmptyArray; 42 43 // Android-changed: Document that named capturing is only available from API 26. 44 // Android-changed: Android always uses unicode character classes. 45 // UNICODE_CHARACTER_CLASS has no effect on Android. 46 /** 47 * A compiled representation of a regular expression. 48 * 49 * <p> A regular expression, specified as a string, must first be compiled into 50 * an instance of this class. The resulting pattern can then be used to create 51 * a {@link Matcher} object that can match arbitrary {@linkplain 52 * java.lang.CharSequence character sequences} against the regular 53 * expression. All of the state involved in performing a match resides in the 54 * matcher, so many matchers can share the same pattern. 55 * 56 * <p> A typical invocation sequence is thus 57 * 58 * <blockquote><pre> 59 * Pattern p = Pattern.{@link #compile compile}("a*b"); 60 * Matcher m = p.{@link #matcher matcher}("aaaaab"); 61 * boolean b = m.{@link Matcher#matches matches}();</pre></blockquote> 62 * 63 * <p> A {@link #matches matches} method is defined by this class as a 64 * convenience for when a regular expression is used just once. This method 65 * compiles an expression and matches an input sequence against it in a single 66 * invocation. The statement 67 * 68 * <blockquote><pre> 69 * boolean b = Pattern.matches("a*b", "aaaaab");</pre></blockquote> 70 * 71 * is equivalent to the three statements above, though for repeated matches it 72 * is less efficient since it does not allow the compiled pattern to be reused. 73 * 74 * <p> Instances of this class are immutable and are safe for use by multiple 75 * concurrent threads. Instances of the {@link Matcher} class are not safe for 76 * such use. 77 * 78 * 79 * <h3><a name="sum">Summary of regular-expression constructs</a></h3> 80 * 81 * <table border="0" cellpadding="1" cellspacing="0" 82 * summary="Regular expression constructs, and what they match"> 83 * 84 * <tr align="left"> 85 * <th align="left" id="construct">Construct</th> 86 * <th align="left" id="matches">Matches</th> 87 * </tr> 88 * 89 * <tr><th> </th></tr> 90 * <tr align="left"><th colspan="2" id="characters">Characters</th></tr> 91 * 92 * <tr><td valign="top" headers="construct characters"><i>x</i></td> 93 * <td headers="matches">The character <i>x</i></td></tr> 94 * <tr><td valign="top" headers="construct characters"><tt>\\</tt></td> 95 * <td headers="matches">The backslash character</td></tr> 96 * <tr><td valign="top" headers="construct characters"><tt>\0</tt><i>n</i></td> 97 * <td headers="matches">The character with octal value <tt>0</tt><i>n</i> 98 * (0 <tt><=</tt> <i>n</i> <tt><=</tt> 7)</td></tr> 99 * <tr><td valign="top" headers="construct characters"><tt>\0</tt><i>nn</i></td> 100 * <td headers="matches">The character with octal value <tt>0</tt><i>nn</i> 101 * (0 <tt><=</tt> <i>n</i> <tt><=</tt> 7)</td></tr> 102 * <tr><td valign="top" headers="construct characters"><tt>\0</tt><i>mnn</i></td> 103 * <td headers="matches">The character with octal value <tt>0</tt><i>mnn</i> 104 * (0 <tt><=</tt> <i>m</i> <tt><=</tt> 3, 105 * 0 <tt><=</tt> <i>n</i> <tt><=</tt> 7)</td></tr> 106 * <tr><td valign="top" headers="construct characters"><tt>\x</tt><i>hh</i></td> 107 * <td headers="matches">The character with hexadecimal value <tt>0x</tt><i>hh</i></td></tr> 108 * <tr><td valign="top" headers="construct characters"><tt>\u</tt><i>hhhh</i></td> 109 * <td headers="matches">The character with hexadecimal value <tt>0x</tt><i>hhhh</i></td></tr> 110 * <tr><td valign="top" headers="construct characters"><tt>\x</tt><i>{h...h}</i></td> 111 * <td headers="matches">The character with hexadecimal value <tt>0x</tt><i>h...h</i> 112 * ({@link java.lang.Character#MIN_CODE_POINT Character.MIN_CODE_POINT} 113 * <= <tt>0x</tt><i>h...h</i> <= 114 * {@link java.lang.Character#MAX_CODE_POINT Character.MAX_CODE_POINT})</td></tr> 115 * <tr><td valign="top" headers="matches"><tt>\t</tt></td> 116 * <td headers="matches">The tab character (<tt>'\u0009'</tt>)</td></tr> 117 * <tr><td valign="top" headers="construct characters"><tt>\n</tt></td> 118 * <td headers="matches">The newline (line feed) character (<tt>'\u000A'</tt>)</td></tr> 119 * <tr><td valign="top" headers="construct characters"><tt>\r</tt></td> 120 * <td headers="matches">The carriage-return character (<tt>'\u000D'</tt>)</td></tr> 121 * <tr><td valign="top" headers="construct characters"><tt>\f</tt></td> 122 * <td headers="matches">The form-feed character (<tt>'\u000C'</tt>)</td></tr> 123 * <tr><td valign="top" headers="construct characters"><tt>\a</tt></td> 124 * <td headers="matches">The alert (bell) character (<tt>'\u0007'</tt>)</td></tr> 125 * <tr><td valign="top" headers="construct characters"><tt>\e</tt></td> 126 * <td headers="matches">The escape character (<tt>'\u001B'</tt>)</td></tr> 127 * <tr><td valign="top" headers="construct characters"><tt>\c</tt><i>x</i></td> 128 * <td headers="matches">The control character corresponding to <i>x</i></td></tr> 129 * 130 * <tr><th> </th></tr> 131 * <tr align="left"><th colspan="2" id="classes">Character classes</th></tr> 132 * 133 * <tr><td valign="top" headers="construct classes">{@code [abc]}</td> 134 * <td headers="matches">{@code a}, {@code b}, or {@code c} (simple class)</td></tr> 135 * <tr><td valign="top" headers="construct classes">{@code [^abc]}</td> 136 * <td headers="matches">Any character except {@code a}, {@code b}, or {@code c} (negation)</td></tr> 137 * <tr><td valign="top" headers="construct classes">{@code [a-zA-Z]}</td> 138 * <td headers="matches">{@code a} through {@code z} 139 * or {@code A} through {@code Z}, inclusive (range)</td></tr> 140 * <tr><td valign="top" headers="construct classes">{@code [a-d[m-p]]}</td> 141 * <td headers="matches">{@code a} through {@code d}, 142 * or {@code m} through {@code p}: {@code [a-dm-p]} (union)</td></tr> 143 * <tr><td valign="top" headers="construct classes">{@code [a-z&&[def]]}</td> 144 * <td headers="matches">{@code d}, {@code e}, or {@code f} (intersection)</tr> 145 * <tr><td valign="top" headers="construct classes">{@code [a-z&&[^bc]]}</td> 146 * <td headers="matches">{@code a} through {@code z}, 147 * except for {@code b} and {@code c}: {@code [ad-z]} (subtraction)</td></tr> 148 * <tr><td valign="top" headers="construct classes">{@code [a-z&&[^m-p]]}</td> 149 * <td headers="matches">{@code a} through {@code z}, 150 * and not {@code m} through {@code p}: {@code [a-lq-z]}(subtraction)</td></tr> 151 * <tr><th> </th></tr> 152 * 153 * <tr align="left"><th colspan="2" id="predef">Predefined character classes</th></tr> 154 * 155 * <tr><td valign="top" headers="construct predef"><tt>.</tt></td> 156 * <td headers="matches">Any character (may or may not match <a href="#lt">line terminators</a>)</td></tr> 157 * <tr><td valign="top" headers="construct predef"><tt>\d</tt></td> 158 * <td headers="matches">A digit: <tt>[0-9]</tt></td></tr> 159 * <tr><td valign="top" headers="construct predef"><tt>\D</tt></td> 160 * <td headers="matches">A non-digit: <tt>[^0-9]</tt></td></tr> 161 * <tr><td valign="top" headers="construct predef"><tt>\h</tt></td> 162 * <td headers="matches">A horizontal whitespace character: 163 * <tt>[ \t\xA0\u1680\u180e\u2000-\u200a\u202f\u205f\u3000]</tt></td></tr> 164 * <tr><td valign="top" headers="construct predef"><tt>\H</tt></td> 165 * <td headers="matches">A non-horizontal whitespace character: <tt>[^\h]</tt></td></tr> 166 * <tr><td valign="top" headers="construct predef"><tt>\s</tt></td> 167 * <td headers="matches">A whitespace character: <tt>[ \t\n\x0B\f\r]</tt></td></tr> 168 * <tr><td valign="top" headers="construct predef"><tt>\S</tt></td> 169 * <td headers="matches">A non-whitespace character: <tt>[^\s]</tt></td></tr> 170 * <tr><td valign="top" headers="construct predef"><tt>\v</tt></td> 171 * <td headers="matches">A vertical whitespace character: <tt>[\n\x0B\f\r\x85\u2028\u2029]</tt> 172 * </td></tr> 173 * <tr><td valign="top" headers="construct predef"><tt>\V</tt></td> 174 * <td headers="matches">A non-vertical whitespace character: <tt>[^\v]</tt></td></tr> 175 * <tr><td valign="top" headers="construct predef"><tt>\w</tt></td> 176 * <td headers="matches">A word character: <tt>[a-zA-Z_0-9]</tt></td></tr> 177 * <tr><td valign="top" headers="construct predef"><tt>\W</tt></td> 178 * <td headers="matches">A non-word character: <tt>[^\w]</tt></td></tr> 179 * <tr><th> </th></tr> 180 * <tr align="left"><th colspan="2" id="posix"><b>POSIX character classes (US-ASCII only)</b></th></tr> 181 * 182 * <tr><td valign="top" headers="construct posix">{@code \p{Lower}}</td> 183 * <td headers="matches">A lower-case alphabetic character: {@code [a-z]}</td></tr> 184 * <tr><td valign="top" headers="construct posix">{@code \p{Upper}}</td> 185 * <td headers="matches">An upper-case alphabetic character:{@code [A-Z]}</td></tr> 186 * <tr><td valign="top" headers="construct posix">{@code \p{ASCII}}</td> 187 * <td headers="matches">All ASCII:{@code [\x00-\x7F]}</td></tr> 188 * <tr><td valign="top" headers="construct posix">{@code \p{Alpha}}</td> 189 * <td headers="matches">An alphabetic character:{@code [\p{Lower}\p{Upper}]}</td></tr> 190 * <tr><td valign="top" headers="construct posix">{@code \p{Digit}}</td> 191 * <td headers="matches">A decimal digit: {@code [0-9]}</td></tr> 192 * <tr><td valign="top" headers="construct posix">{@code \p{Alnum}}</td> 193 * <td headers="matches">An alphanumeric character:{@code [\p{Alpha}\p{Digit}]}</td></tr> 194 * <tr><td valign="top" headers="construct posix">{@code \p{Punct}}</td> 195 * <td headers="matches">Punctuation: One of {@code !"#$%&'()*+,-./:;<=>?@[\]^_`{|}~}</td></tr> 196 * <!-- {@code [\!"#\$%&'\(\)\*\+,\-\./:;\<=\>\?@\[\\\]\^_`\{\|\}~]} 197 * {@code [\X21-\X2F\X31-\X40\X5B-\X60\X7B-\X7E]} --> 198 * <tr><td valign="top" headers="construct posix">{@code \p{Graph}}</td> 199 * <td headers="matches">A visible character: {@code [\p{Alnum}\p{Punct}]}</td></tr> 200 * <tr><td valign="top" headers="construct posix">{@code \p{Print}}</td> 201 * <td headers="matches">A printable character: {@code [\p{Graph}\x20]}</td></tr> 202 * <tr><td valign="top" headers="construct posix">{@code \p{Blank}}</td> 203 * <td headers="matches">A space or a tab: {@code [ \t]}</td></tr> 204 * <tr><td valign="top" headers="construct posix">{@code \p{Cntrl}}</td> 205 * <td headers="matches">A control character: {@code [\x00-\x1F\x7F]}</td></tr> 206 * <tr><td valign="top" headers="construct posix">{@code \p{XDigit}}</td> 207 * <td headers="matches">A hexadecimal digit: {@code [0-9a-fA-F]}</td></tr> 208 * <tr><td valign="top" headers="construct posix">{@code \p{Space}}</td> 209 * <td headers="matches">A whitespace character: {@code [ \t\n\x0B\f\r]}</td></tr> 210 * 211 * <tr><th> </th></tr> 212 * <tr align="left"><th colspan="2">java.lang.Character classes (simple <a href="#jcc">java character type</a>)</th></tr> 213 * 214 * <tr><td valign="top"><tt>\p{javaLowerCase}</tt></td> 215 * <td>Equivalent to java.lang.Character.isLowerCase()</td></tr> 216 * <tr><td valign="top"><tt>\p{javaUpperCase}</tt></td> 217 * <td>Equivalent to java.lang.Character.isUpperCase()</td></tr> 218 * <tr><td valign="top"><tt>\p{javaWhitespace}</tt></td> 219 * <td>Equivalent to java.lang.Character.isWhitespace()</td></tr> 220 * <tr><td valign="top"><tt>\p{javaMirrored}</tt></td> 221 * <td>Equivalent to java.lang.Character.isMirrored()</td></tr> 222 * 223 * <tr><th> </th></tr> 224 * <tr align="left"><th colspan="2" id="unicode">Classes for Unicode scripts, blocks, categories and binary properties</th></tr> 225 * <tr><td valign="top" headers="construct unicode">{@code \p{IsLatin}}</td> 226 * <td headers="matches">A Latin script character (<a href="#usc">script</a>)</td></tr> 227 * <tr><td valign="top" headers="construct unicode">{@code \p{InGreek}}</td> 228 * <td headers="matches">A character in the Greek block (<a href="#ubc">block</a>)</td></tr> 229 * <tr><td valign="top" headers="construct unicode">{@code \p{Lu}}</td> 230 * <td headers="matches">An uppercase letter (<a href="#ucc">category</a>)</td></tr> 231 * <tr><td valign="top" headers="construct unicode">{@code \p{IsAlphabetic}}</td> 232 * <td headers="matches">An alphabetic character (<a href="#ubpc">binary property</a>)</td></tr> 233 * <tr><td valign="top" headers="construct unicode">{@code \p{Sc}}</td> 234 * <td headers="matches">A currency symbol</td></tr> 235 * <tr><td valign="top" headers="construct unicode">{@code \P{InGreek}}</td> 236 * <td headers="matches">Any character except one in the Greek block (negation)</td></tr> 237 * <tr><td valign="top" headers="construct unicode">{@code [\p{L}&&[^\p{Lu}]]}</td> 238 * <td headers="matches">Any letter except an uppercase letter (subtraction)</td></tr> 239 * 240 * <tr><th> </th></tr> 241 * <tr align="left"><th colspan="2" id="bounds">Boundary matchers</th></tr> 242 * 243 * <tr><td valign="top" headers="construct bounds"><tt>^</tt></td> 244 * <td headers="matches">The beginning of a line</td></tr> 245 * <tr><td valign="top" headers="construct bounds"><tt>$</tt></td> 246 * <td headers="matches">The end of a line</td></tr> 247 * <tr><td valign="top" headers="construct bounds"><tt>\b</tt></td> 248 * <td headers="matches">A word boundary</td></tr> 249 * <tr><td valign="top" headers="construct bounds"><tt>\B</tt></td> 250 * <td headers="matches">A non-word boundary</td></tr> 251 * <tr><td valign="top" headers="construct bounds"><tt>\A</tt></td> 252 * <td headers="matches">The beginning of the input</td></tr> 253 * <tr><td valign="top" headers="construct bounds"><tt>\G</tt></td> 254 * <td headers="matches">The end of the previous match</td></tr> 255 * <tr><td valign="top" headers="construct bounds"><tt>\Z</tt></td> 256 * <td headers="matches">The end of the input but for the final 257 * <a href="#lt">terminator</a>, if any</td></tr> 258 * <tr><td valign="top" headers="construct bounds"><tt>\z</tt></td> 259 * <td headers="matches">The end of the input</td></tr> 260 * 261 * <tr><th> </th></tr> 262 * <tr align="left"><th colspan="2" id="lineending">Linebreak matcher</th></tr> 263 * <tr><td valign="top" headers="construct lineending"><tt>\R</tt></td> 264 * <td headers="matches">Any Unicode linebreak sequence, is equivalent to 265 * <tt>\u000D\u000A|[\u000A\u000B\u000C\u000D\u0085\u2028\u2029] 266 * </tt></td></tr> 267 * 268 * <tr><th> </th></tr> 269 * <tr align="left"><th colspan="2" id="greedy">Greedy quantifiers</th></tr> 270 * 271 * <tr><td valign="top" headers="construct greedy"><i>X</i><tt>?</tt></td> 272 * <td headers="matches"><i>X</i>, once or not at all</td></tr> 273 * <tr><td valign="top" headers="construct greedy"><i>X</i><tt>*</tt></td> 274 * <td headers="matches"><i>X</i>, zero or more times</td></tr> 275 * <tr><td valign="top" headers="construct greedy"><i>X</i><tt>+</tt></td> 276 * <td headers="matches"><i>X</i>, one or more times</td></tr> 277 * <tr><td valign="top" headers="construct greedy"><i>X</i><tt>{</tt><i>n</i><tt>}</tt></td> 278 * <td headers="matches"><i>X</i>, exactly <i>n</i> times</td></tr> 279 * <tr><td valign="top" headers="construct greedy"><i>X</i><tt>{</tt><i>n</i><tt>,}</tt></td> 280 * <td headers="matches"><i>X</i>, at least <i>n</i> times</td></tr> 281 * <tr><td valign="top" headers="construct greedy"><i>X</i><tt>{</tt><i>n</i><tt>,</tt><i>m</i><tt>}</tt></td> 282 * <td headers="matches"><i>X</i>, at least <i>n</i> but not more than <i>m</i> times</td></tr> 283 * 284 * <tr><th> </th></tr> 285 * <tr align="left"><th colspan="2" id="reluc">Reluctant quantifiers</th></tr> 286 * 287 * <tr><td valign="top" headers="construct reluc"><i>X</i><tt>??</tt></td> 288 * <td headers="matches"><i>X</i>, once or not at all</td></tr> 289 * <tr><td valign="top" headers="construct reluc"><i>X</i><tt>*?</tt></td> 290 * <td headers="matches"><i>X</i>, zero or more times</td></tr> 291 * <tr><td valign="top" headers="construct reluc"><i>X</i><tt>+?</tt></td> 292 * <td headers="matches"><i>X</i>, one or more times</td></tr> 293 * <tr><td valign="top" headers="construct reluc"><i>X</i><tt>{</tt><i>n</i><tt>}?</tt></td> 294 * <td headers="matches"><i>X</i>, exactly <i>n</i> times</td></tr> 295 * <tr><td valign="top" headers="construct reluc"><i>X</i><tt>{</tt><i>n</i><tt>,}?</tt></td> 296 * <td headers="matches"><i>X</i>, at least <i>n</i> times</td></tr> 297 * <tr><td valign="top" headers="construct reluc"><i>X</i><tt>{</tt><i>n</i><tt>,</tt><i>m</i><tt>}?</tt></td> 298 * <td headers="matches"><i>X</i>, at least <i>n</i> but not more than <i>m</i> times</td></tr> 299 * 300 * <tr><th> </th></tr> 301 * <tr align="left"><th colspan="2" id="poss">Possessive quantifiers</th></tr> 302 * 303 * <tr><td valign="top" headers="construct poss"><i>X</i><tt>?+</tt></td> 304 * <td headers="matches"><i>X</i>, once or not at all</td></tr> 305 * <tr><td valign="top" headers="construct poss"><i>X</i><tt>*+</tt></td> 306 * <td headers="matches"><i>X</i>, zero or more times</td></tr> 307 * <tr><td valign="top" headers="construct poss"><i>X</i><tt>++</tt></td> 308 * <td headers="matches"><i>X</i>, one or more times</td></tr> 309 * <tr><td valign="top" headers="construct poss"><i>X</i><tt>{</tt><i>n</i><tt>}+</tt></td> 310 * <td headers="matches"><i>X</i>, exactly <i>n</i> times</td></tr> 311 * <tr><td valign="top" headers="construct poss"><i>X</i><tt>{</tt><i>n</i><tt>,}+</tt></td> 312 * <td headers="matches"><i>X</i>, at least <i>n</i> times</td></tr> 313 * <tr><td valign="top" headers="construct poss"><i>X</i><tt>{</tt><i>n</i><tt>,</tt><i>m</i><tt>}+</tt></td> 314 * <td headers="matches"><i>X</i>, at least <i>n</i> but not more than <i>m</i> times</td></tr> 315 * 316 * <tr><th> </th></tr> 317 * <tr align="left"><th colspan="2" id="logical">Logical operators</th></tr> 318 * 319 * <tr><td valign="top" headers="construct logical"><i>XY</i></td> 320 * <td headers="matches"><i>X</i> followed by <i>Y</i></td></tr> 321 * <tr><td valign="top" headers="construct logical"><i>X</i><tt>|</tt><i>Y</i></td> 322 * <td headers="matches">Either <i>X</i> or <i>Y</i></td></tr> 323 * <tr><td valign="top" headers="construct logical"><tt>(</tt><i>X</i><tt>)</tt></td> 324 * <td headers="matches">X, as a <a href="#cg">capturing group</a></td></tr> 325 * 326 * <tr><th> </th></tr> 327 * <tr align="left"><th colspan="2" id="backref">Back references</th></tr> 328 * 329 * <tr><td valign="bottom" headers="construct backref"><tt>\</tt><i>n</i></td> 330 * <td valign="bottom" headers="matches">Whatever the <i>n</i><sup>th</sup> 331 * <a href="#cg">capturing group</a> matched</td></tr> 332 * 333 * <tr><td valign="bottom" headers="construct backref"><tt>\</tt><i>k</i><<i>name</i>></td> 334 * <td valign="bottom" headers="matches">Whatever the 335 * <a href="#groupname">named-capturing group</a> "name" matched. Only available for API 26 or above</td></tr> 336 * 337 * <tr><th> </th></tr> 338 * <tr align="left"><th colspan="2" id="quot">Quotation</th></tr> 339 * 340 * <tr><td valign="top" headers="construct quot"><tt>\</tt></td> 341 * <td headers="matches">Nothing, but quotes the following character</td></tr> 342 * <tr><td valign="top" headers="construct quot"><tt>\Q</tt></td> 343 * <td headers="matches">Nothing, but quotes all characters until <tt>\E</tt></td></tr> 344 * <tr><td valign="top" headers="construct quot"><tt>\E</tt></td> 345 * <td headers="matches">Nothing, but ends quoting started by <tt>\Q</tt></td></tr> 346 * <!-- Metachars: !$()*+.<>?[\]^{|} --> 347 * 348 * <tr><th> </th></tr> 349 * <tr align="left"><th colspan="2" id="special">Special constructs (named-capturing and non-capturing)</th></tr> 350 * 351 * <tr><td valign="top" headers="construct special"><tt>(?<<a href="#groupname">name</a>></tt><i>X</i><tt>)</tt></td> 352 * <td headers="matches"><i>X</i>, as a named-capturing group. Only available for API 26 or above.</td></tr> 353 * <tr><td valign="top" headers="construct special"><tt>(?:</tt><i>X</i><tt>)</tt></td> 354 * <td headers="matches"><i>X</i>, as a non-capturing group</td></tr> 355 * <tr><td valign="top" headers="construct special"><tt>(?idmsuxU-idmsuxU) </tt></td> 356 * <td headers="matches">Nothing, but turns match flags <a href="#CASE_INSENSITIVE">i</a> 357 * <a href="#UNIX_LINES">d</a> <a href="#MULTILINE">m</a> <a href="#DOTALL">s</a> 358 * <a href="#UNICODE_CASE">u</a> <a href="#COMMENTS">x</a> <a href="#UNICODE_CHARACTER_CLASS">U</a> 359 * on - off</td></tr> 360 * <tr><td valign="top" headers="construct special"><tt>(?idmsux-idmsux:</tt><i>X</i><tt>)</tt> </td> 361 * <td headers="matches"><i>X</i>, as a <a href="#cg">non-capturing group</a> with the 362 * given flags <a href="#CASE_INSENSITIVE">i</a> <a href="#UNIX_LINES">d</a> 363 * <a href="#MULTILINE">m</a> <a href="#DOTALL">s</a> <a href="#UNICODE_CASE">u</a > 364 * <a href="#COMMENTS">x</a> on - off</td></tr> 365 * <tr><td valign="top" headers="construct special"><tt>(?=</tt><i>X</i><tt>)</tt></td> 366 * <td headers="matches"><i>X</i>, via zero-width positive lookahead</td></tr> 367 * <tr><td valign="top" headers="construct special"><tt>(?!</tt><i>X</i><tt>)</tt></td> 368 * <td headers="matches"><i>X</i>, via zero-width negative lookahead</td></tr> 369 * <tr><td valign="top" headers="construct special"><tt>(?<=</tt><i>X</i><tt>)</tt></td> 370 * <td headers="matches"><i>X</i>, via zero-width positive lookbehind</td></tr> 371 * <tr><td valign="top" headers="construct special"><tt>(?<!</tt><i>X</i><tt>)</tt></td> 372 * <td headers="matches"><i>X</i>, via zero-width negative lookbehind</td></tr> 373 * <tr><td valign="top" headers="construct special"><tt>(?></tt><i>X</i><tt>)</tt></td> 374 * <td headers="matches"><i>X</i>, as an independent, non-capturing group</td></tr> 375 * 376 * </table> 377 * 378 * <hr> 379 * 380 * 381 * <h3><a name="bs">Backslashes, escapes, and quoting</a></h3> 382 * 383 * <p> The backslash character (<tt>'\'</tt>) serves to introduce escaped 384 * constructs, as defined in the table above, as well as to quote characters 385 * that otherwise would be interpreted as unescaped constructs. Thus the 386 * expression <tt>\\</tt> matches a single backslash and <tt>\{</tt> matches a 387 * left brace. 388 * 389 * <p> It is an error to use a backslash prior to any alphabetic character that 390 * does not denote an escaped construct; these are reserved for future 391 * extensions to the regular-expression language. A backslash may be used 392 * prior to a non-alphabetic character regardless of whether that character is 393 * part of an unescaped construct. 394 * 395 * <p> Backslashes within string literals in Java source code are interpreted 396 * as required by 397 * <cite>The Java™ Language Specification</cite> 398 * as either Unicode escapes (section 3.3) or other character escapes (section 3.10.6) 399 * It is therefore necessary to double backslashes in string 400 * literals that represent regular expressions to protect them from 401 * interpretation by the Java bytecode compiler. The string literal 402 * <tt>"\b"</tt>, for example, matches a single backspace character when 403 * interpreted as a regular expression, while <tt>"\\b"</tt> matches a 404 * word boundary. The string literal <tt>"\(hello\)"</tt> is illegal 405 * and leads to a compile-time error; in order to match the string 406 * <tt>(hello)</tt> the string literal <tt>"\\(hello\\)"</tt> 407 * must be used. 408 * 409 * <h3><a name="cc">Character Classes</a></h3> 410 * 411 * <p> Character classes may appear within other character classes, and 412 * may be composed by the union operator (implicit) and the intersection 413 * operator (<tt>&&</tt>). 414 * The union operator denotes a class that contains every character that is 415 * in at least one of its operand classes. The intersection operator 416 * denotes a class that contains every character that is in both of its 417 * operand classes. 418 * 419 * <p> The precedence of character-class operators is as follows, from 420 * highest to lowest: 421 * 422 * <blockquote><table border="0" cellpadding="1" cellspacing="0" 423 * summary="Precedence of character class operators."> 424 * <tr><th>1 </th> 425 * <td>Literal escape </td> 426 * <td><tt>\x</tt></td></tr> 427 * <tr><th>2 </th> 428 * <td>Grouping</td> 429 * <td><tt>[...]</tt></td></tr> 430 * <tr><th>3 </th> 431 * <td>Range</td> 432 * <td><tt>a-z</tt></td></tr> 433 * <tr><th>4 </th> 434 * <td>Union</td> 435 * <td><tt>[a-e][i-u]</tt></td></tr> 436 * <tr><th>5 </th> 437 * <td>Intersection</td> 438 * <td>{@code [a-z&&[aeiou]]}</td></tr> 439 * </table></blockquote> 440 * 441 * <p> Note that a different set of metacharacters are in effect inside 442 * a character class than outside a character class. For instance, the 443 * regular expression <tt>.</tt> loses its special meaning inside a 444 * character class, while the expression <tt>-</tt> becomes a range 445 * forming metacharacter. 446 * 447 * <h3><a name="lt">Line terminators</a></h3> 448 * 449 * <p> A <i>line terminator</i> is a one- or two-character sequence that marks 450 * the end of a line of the input character sequence. The following are 451 * recognized as line terminators: 452 * 453 * <ul> 454 * 455 * <li> A newline (line feed) character (<tt>'\n'</tt>), 456 * 457 * <li> A carriage-return character followed immediately by a newline 458 * character (<tt>"\r\n"</tt>), 459 * 460 * <li> A standalone carriage-return character (<tt>'\r'</tt>), 461 * 462 * <li> A next-line character (<tt>'\u0085'</tt>), 463 * 464 * <li> A line-separator character (<tt>'\u2028'</tt>), or 465 * 466 * <li> A paragraph-separator character (<tt>'\u2029</tt>). 467 * 468 * </ul> 469 * <p>If {@link #UNIX_LINES} mode is activated, then the only line terminators 470 * recognized are newline characters. 471 * 472 * <p> The regular expression <tt>.</tt> matches any character except a line 473 * terminator unless the {@link #DOTALL} flag is specified. 474 * 475 * <p> By default, the regular expressions <tt>^</tt> and <tt>$</tt> ignore 476 * line terminators and only match at the beginning and the end, respectively, 477 * of the entire input sequence. If {@link #MULTILINE} mode is activated then 478 * <tt>^</tt> matches at the beginning of input and after any line terminator 479 * except at the end of input. When in {@link #MULTILINE} mode <tt>$</tt> 480 * matches just before a line terminator or the end of the input sequence. 481 * 482 * <h3><a name="cg">Groups and capturing</a></h3> 483 * 484 * <h4><a name="gnumber">Group number</a></h4> 485 * <p> Capturing groups are numbered by counting their opening parentheses from 486 * left to right. In the expression <tt>((A)(B(C)))</tt>, for example, there 487 * are four such groups: </p> 488 * 489 * <blockquote><table cellpadding=1 cellspacing=0 summary="Capturing group numberings"> 490 * <tr><th>1 </th> 491 * <td><tt>((A)(B(C)))</tt></td></tr> 492 * <tr><th>2 </th> 493 * <td><tt>(A)</tt></td></tr> 494 * <tr><th>3 </th> 495 * <td><tt>(B(C))</tt></td></tr> 496 * <tr><th>4 </th> 497 * <td><tt>(C)</tt></td></tr> 498 * </table></blockquote> 499 * 500 * <p> Group zero always stands for the entire expression. 501 * 502 * <p> Capturing groups are so named because, during a match, each subsequence 503 * of the input sequence that matches such a group is saved. The captured 504 * subsequence may be used later in the expression, via a back reference, and 505 * may also be retrieved from the matcher once the match operation is complete. 506 * 507 * <h4><a name="groupname">Group name</a></h4> 508 * <p>The constructs and APIs are available since API level 26. A capturing group 509 * can also be assigned a "name", a <tt>named-capturing group</tt>, 510 * and then be back-referenced later by the "name". Group names are composed of 511 * the following characters. The first character must be a <tt>letter</tt>. 512 * 513 * <ul> 514 * <li> The uppercase letters <tt>'A'</tt> through <tt>'Z'</tt> 515 * (<tt>'\u0041'</tt> through <tt>'\u005a'</tt>), 516 * <li> The lowercase letters <tt>'a'</tt> through <tt>'z'</tt> 517 * (<tt>'\u0061'</tt> through <tt>'\u007a'</tt>), 518 * <li> The digits <tt>'0'</tt> through <tt>'9'</tt> 519 * (<tt>'\u0030'</tt> through <tt>'\u0039'</tt>), 520 * </ul> 521 * 522 * <p> A <tt>named-capturing group</tt> is still numbered as described in 523 * <a href="#gnumber">Group number</a>. 524 * 525 * <p> The captured input associated with a group is always the subsequence 526 * that the group most recently matched. If a group is evaluated a second time 527 * because of quantification then its previously-captured value, if any, will 528 * be retained if the second evaluation fails. Matching the string 529 * <tt>"aba"</tt> against the expression <tt>(a(b)?)+</tt>, for example, leaves 530 * group two set to <tt>"b"</tt>. All captured input is discarded at the 531 * beginning of each match. 532 * 533 * <p> Groups beginning with <tt>(?</tt> are either pure, <i>non-capturing</i> groups 534 * that do not capture text and do not count towards the group total, or 535 * <i>named-capturing</i> group. 536 * 537 * <h3> Unicode support </h3> 538 * 539 * <p> This class is in conformance with Level 1 of <a 540 * href="http://www.unicode.org/reports/tr18/"><i>Unicode Technical 541 * Standard #18: Unicode Regular Expression</i></a>, plus RL2.1 542 * Canonical Equivalents. 543 * <p> 544 * <b>Unicode escape sequences</b> such as <tt>\u2014</tt> in Java source code 545 * are processed as described in section 3.3 of 546 * <cite>The Java™ Language Specification</cite>. 547 * Such escape sequences are also implemented directly by the regular-expression 548 * parser so that Unicode escapes can be used in expressions that are read from 549 * files or from the keyboard. Thus the strings <tt>"\u2014"</tt> and 550 * <tt>"\\u2014"</tt>, while not equal, compile into the same pattern, which 551 * matches the character with hexadecimal value <tt>0x2014</tt>. 552 * <p> 553 * A Unicode character can also be represented in a regular-expression by 554 * using its <b>Hex notation</b>(hexadecimal code point value) directly as described in construct 555 * <tt>\x{...}</tt>, for example a supplementary character U+2011F 556 * can be specified as <tt>\x{2011F}</tt>, instead of two consecutive 557 * Unicode escape sequences of the surrogate pair 558 * <tt>\uD840</tt><tt>\uDD1F</tt>. 559 * <p> 560 * Unicode scripts, blocks, categories and binary properties are written with 561 * the <tt>\p</tt> and <tt>\P</tt> constructs as in Perl. 562 * <tt>\p{</tt><i>prop</i><tt>}</tt> matches if 563 * the input has the property <i>prop</i>, while <tt>\P{</tt><i>prop</i><tt>}</tt> 564 * does not match if the input has that property. 565 * <p> 566 * Scripts, blocks, categories and binary properties can be used both inside 567 * and outside of a character class. 568 * 569 * <p> 570 * <b><a name="usc">Scripts</a></b> are specified either with the prefix {@code Is}, as in 571 * {@code IsHiragana}, or by using the {@code script} keyword (or its short 572 * form {@code sc})as in {@code script=Hiragana} or {@code sc=Hiragana}. 573 * <p> 574 * The script names supported by <code>Pattern</code> are the valid script names 575 * accepted and defined by 576 * {@link java.lang.Character.UnicodeScript#forName(String) UnicodeScript.forName}. 577 * 578 * <p> 579 * <b><a name="ubc">Blocks</a></b> are specified with the prefix {@code In}, as in 580 * {@code InMongolian}, or by using the keyword {@code block} (or its short 581 * form {@code blk}) as in {@code block=Mongolian} or {@code blk=Mongolian}. 582 * <p> 583 * The block names supported by <code>Pattern</code> are the valid block names 584 * accepted and defined by 585 * {@link java.lang.Character.UnicodeBlock#forName(String) UnicodeBlock.forName}. 586 * <p> 587 * 588 * <b><a name="ucc">Categories</a></b> may be specified with the optional prefix {@code Is}: 589 * Both {@code \p{L}} and {@code \p{IsL}} denote the category of Unicode 590 * letters. Same as scripts and blocks, categories can also be specified 591 * by using the keyword {@code general_category} (or its short form 592 * {@code gc}) as in {@code general_category=Lu} or {@code gc=Lu}. 593 * <p> 594 * The supported categories are those of 595 * <a href="http://www.unicode.org/unicode/standard/standard.html"> 596 * <i>The Unicode Standard</i></a> in the version specified by the 597 * {@link java.lang.Character Character} class. The category names are those 598 * defined in the Standard, both normative and informative. 599 * <p> 600 * 601 * <b><a name="ubpc">Binary properties</a></b> are specified with the prefix {@code Is}, as in 602 * {@code IsAlphabetic}. The supported binary properties by <code>Pattern</code> 603 * are 604 * <ul> 605 * <li> Alphabetic 606 * <li> Ideographic 607 * <li> Letter 608 * <li> Lowercase 609 * <li> Uppercase 610 * <li> Titlecase 611 * <li> Punctuation 612 * <Li> Control 613 * <li> White_Space 614 * <li> Digit 615 * <li> Hex_Digit 616 * <li> Join_Control 617 * <li> Noncharacter_Code_Point 618 * <li> Assigned 619 * </ul> 620 * <p> 621 * The following <b>Predefined Character classes</b> and <b>POSIX character classes</b> 622 * are in conformance with the recommendation of <i>Annex C: Compatibility Properties</i> 623 * of <a href="http://www.unicode.org/reports/tr18/"><i>Unicode Regular Expression 624 * </i></a>. 625 * 626 * <table border="0" cellpadding="1" cellspacing="0" 627 * summary="predefined and posix character classes in Unicode mode"> 628 * <tr align="left"> 629 * <th align="left" id="predef_classes">Classes</th> 630 * <th align="left" id="predef_matches">Matches</th> 631 *</tr> 632 * <tr><td><tt>\p{Lower}</tt></td> 633 * <td>A lowercase character:<tt>\p{IsLowercase}</tt></td></tr> 634 * <tr><td><tt>\p{Upper}</tt></td> 635 * <td>An uppercase character:<tt>\p{IsUppercase}</tt></td></tr> 636 * <tr><td><tt>\p{ASCII}</tt></td> 637 * <td>All ASCII:<tt>[\x00-\x7F]</tt></td></tr> 638 * <tr><td><tt>\p{Alpha}</tt></td> 639 * <td>An alphabetic character:<tt>\p{IsAlphabetic}</tt></td></tr> 640 * <tr><td><tt>\p{Digit}</tt></td> 641 * <td>A decimal digit character:<tt>p{IsDigit}</tt></td></tr> 642 * <tr><td><tt>\p{Alnum}</tt></td> 643 * <td>An alphanumeric character:<tt>[\p{IsAlphabetic}\p{IsDigit}]</tt></td></tr> 644 * <tr><td><tt>\p{Punct}</tt></td> 645 * <td>A punctuation character:<tt>p{IsPunctuation}</tt></td></tr> 646 * <tr><td><tt>\p{Graph}</tt></td> 647 * <td>A visible character: <tt>[^\p{IsWhite_Space}\p{gc=Cc}\p{gc=Cs}\p{gc=Cn}]</tt></td></tr> 648 * <tr><td><tt>\p{Print}</tt></td> 649 * <td>A printable character: {@code [\p{Graph}\p{Blank}&&[^\p{Cntrl}]]}</td></tr> 650 * <tr><td><tt>\p{Blank}</tt></td> 651 * <td>A space or a tab: {@code [\p{IsWhite_Space}&&[^\p{gc=Zl}\p{gc=Zp}\x0a\x0b\x0c\x0d\x85]]}</td></tr> 652 * <tr><td><tt>\p{Cntrl}</tt></td> 653 * <td>A control character: <tt>\p{gc=Cc}</tt></td></tr> 654 * <tr><td><tt>\p{XDigit}</tt></td> 655 * <td>A hexadecimal digit: <tt>[\p{gc=Nd}\p{IsHex_Digit}]</tt></td></tr> 656 * <tr><td><tt>\p{Space}</tt></td> 657 * <td>A whitespace character:<tt>\p{IsWhite_Space}</tt></td></tr> 658 * <tr><td><tt>\d</tt></td> 659 * <td>A digit: <tt>\p{IsDigit}</tt></td></tr> 660 * <tr><td><tt>\D</tt></td> 661 * <td>A non-digit: <tt>[^\d]</tt></td></tr> 662 * <tr><td><tt>\s</tt></td> 663 * <td>A whitespace character: <tt>\p{IsWhite_Space}</tt></td></tr> 664 * <tr><td><tt>\S</tt></td> 665 * <td>A non-whitespace character: <tt>[^\s]</tt></td></tr> 666 * <tr><td><tt>\w</tt></td> 667 * <td>A word character: <tt>[\p{Alpha}\p{gc=Mn}\p{gc=Me}\p{gc=Mc}\p{Digit}\p{gc=Pc}\p{IsJoin_Control}]</tt></td></tr> 668 * <tr><td><tt>\W</tt></td> 669 * <td>A non-word character: <tt>[^\w]</tt></td></tr> 670 * </table> 671 * <p> 672 * <a name="jcc"> 673 * Categories that behave like the java.lang.Character 674 * boolean is<i>methodname</i> methods (except for the deprecated ones) are 675 * available through the same <tt>\p{</tt><i>prop</i><tt>}</tt> syntax where 676 * the specified property has the name <tt>java<i>methodname</i></tt></a>. 677 * 678 * <h3> Behavior starting from API level 10 (Android 2.3) </h3> 679 * 680 * <p> Starting from Android 2.3 Gingerbread, ICU4C becomes the backend of the regular expression 681 * implementation. Android could behave differently compared with other regex implementation, e.g. 682 * literal right brace ('}') has to be escaped on Android.</p> 683 * 684 * <p> Some other behavior differences can be found in the 685 * <a href="https://unicode-org.github.io/icu/userguide/strings/regexp.html#differences-with-java-regular-expressions"> 686 * ICU documentation</a>. </p> 687 * 688 * <h3> Comparison to Perl 5 </h3> 689 * 690 * <p>The <code>Pattern</code> engine performs traditional NFA-based matching 691 * with ordered alternation as occurs in Perl 5. 692 * 693 * <p> Perl constructs not supported by this class: </p> 694 * 695 * <ul> 696 * <li><p> Predefined character classes (Unicode character) 697 * <p><tt>\X </tt>Match Unicode 698 * <a href="http://www.unicode.org/reports/tr18/#Default_Grapheme_Clusters"> 699 * <i>extended grapheme cluster</i></a> 700 * </p></li> 701 * 702 * <li><p> The backreference constructs, <tt>\g{</tt><i>n</i><tt>}</tt> for 703 * the <i>n</i><sup>th</sup><a href="#cg">capturing group</a> and 704 * <tt>\g{</tt><i>name</i><tt>}</tt> for 705 * <a href="#groupname">named-capturing group</a>. 706 * </p></li> 707 * 708 * <li><p> The named character construct, <tt>\N{</tt><i>name</i><tt>}</tt> 709 * for a Unicode character by its name. 710 * </p></li> 711 * 712 * <li><p> The conditional constructs 713 * <tt>(?(</tt><i>condition</i><tt>)</tt><i>X</i><tt>)</tt> and 714 * <tt>(?(</tt><i>condition</i><tt>)</tt><i>X</i><tt>|</tt><i>Y</i><tt>)</tt>, 715 * </p></li> 716 * 717 * <li><p> The embedded code constructs <tt>(?{</tt><i>code</i><tt>})</tt> 718 * and <tt>(??{</tt><i>code</i><tt>})</tt>,</p></li> 719 * 720 * <li><p> The embedded comment syntax <tt>(?#comment)</tt>, and </p></li> 721 * 722 * <li><p> The preprocessing operations <tt>\l</tt> <tt>\u</tt>, 723 * <tt>\L</tt>, and <tt>\U</tt>. </p></li> 724 * 725 * </ul> 726 * 727 * <p> Constructs supported by this class but not by Perl: </p> 728 * 729 * <ul> 730 * 731 * <li><p> Character-class union and intersection as described 732 * <a href="#cc">above</a>.</p></li> 733 * 734 * </ul> 735 * 736 * <p> Notable differences from Perl: </p> 737 * 738 * <ul> 739 * 740 * <li><p> In Perl, <tt>\1</tt> through <tt>\9</tt> are always interpreted 741 * as back references; a backslash-escaped number greater than <tt>9</tt> is 742 * treated as a back reference if at least that many subexpressions exist, 743 * otherwise it is interpreted, if possible, as an octal escape. In this 744 * class octal escapes must always begin with a zero. In this class, 745 * <tt>\1</tt> through <tt>\9</tt> are always interpreted as back 746 * references, and a larger number is accepted as a back reference if at 747 * least that many subexpressions exist at that point in the regular 748 * expression, otherwise the parser will drop digits until the number is 749 * smaller or equal to the existing number of groups or it is one digit. 750 * </p></li> 751 * 752 * <li><p> Perl uses the <tt>g</tt> flag to request a match that resumes 753 * where the last match left off. This functionality is provided implicitly 754 * by the {@link Matcher} class: Repeated invocations of the {@link 755 * Matcher#find find} method will resume where the last match left off, 756 * unless the matcher is reset. </p></li> 757 * 758 * <li><p> In Perl, embedded flags at the top level of an expression affect 759 * the whole expression. In this class, embedded flags always take effect 760 * at the point at which they appear, whether they are at the top level or 761 * within a group; in the latter case, flags are restored at the end of the 762 * group just as in Perl. </p></li> 763 * 764 * </ul> 765 * 766 * 767 * <p> For a more precise description of the behavior of regular expression 768 * constructs, please see <a href="http://www.oreilly.com/catalog/regex3/"> 769 * <i>Mastering Regular Expressions, 3nd Edition</i>, Jeffrey E. F. Friedl, 770 * O'Reilly and Associates, 2006.</a> 771 * </p> 772 * 773 * @see java.lang.String#split(String, int) 774 * @see java.lang.String#split(String) 775 * 776 * @author Mike McCloskey 777 * @author Mark Reinhold 778 * @author JSR-51 Expert Group 779 * @since 1.4 780 * @spec JSR-51 781 */ 782 783 public final class Pattern 784 implements java.io.Serializable 785 { 786 787 /** 788 * Regular expression modifier values. Instead of being passed as 789 * arguments, they can also be passed as inline modifiers. 790 * For example, the following statements have the same effect. 791 * <pre> 792 * RegExp r1 = RegExp.compile("abc", Pattern.I|Pattern.M); 793 * RegExp r2 = RegExp.compile("(?im)abc", 0); 794 * </pre> 795 * 796 * The flags are duplicated so that the familiar Perl match flag 797 * names are available. 798 */ 799 800 /** 801 * Enables Unix lines mode. 802 * 803 * <p> In this mode, only the <tt>'\n'</tt> line terminator is recognized 804 * in the behavior of <tt>.</tt>, <tt>^</tt>, and <tt>$</tt>. 805 * 806 * <p> Unix lines mode can also be enabled via the embedded flag 807 * expression <tt>(?d)</tt>. 808 */ 809 public static final int UNIX_LINES = 0x01; 810 811 /** 812 * Enables case-insensitive matching. 813 * 814 * <p> By default, case-insensitive matching assumes that only characters 815 * in the US-ASCII charset are being matched. Unicode-aware 816 * case-insensitive matching can be enabled by specifying the {@link 817 * #UNICODE_CASE} flag in conjunction with this flag. 818 * 819 * <p> Case-insensitive matching can also be enabled via the embedded flag 820 * expression <tt>(?i)</tt>. 821 * 822 * <p> Specifying this flag may impose a slight performance penalty. </p> 823 */ 824 public static final int CASE_INSENSITIVE = 0x02; 825 826 /** 827 * Permits whitespace and comments in pattern. 828 * 829 * <p> In this mode, whitespace is ignored, and embedded comments starting 830 * with <tt>#</tt> are ignored until the end of a line. 831 * 832 * <p> Comments mode can also be enabled via the embedded flag 833 * expression <tt>(?x)</tt>. 834 */ 835 public static final int COMMENTS = 0x04; 836 837 /** 838 * Enables multiline mode. 839 * 840 * <p> In multiline mode the expressions <tt>^</tt> and <tt>$</tt> match 841 * just after or just before, respectively, a line terminator or the end of 842 * the input sequence. By default these expressions only match at the 843 * beginning and the end of the entire input sequence. 844 * 845 * <p> Multiline mode can also be enabled via the embedded flag 846 * expression <tt>(?m)</tt>. </p> 847 */ 848 public static final int MULTILINE = 0x08; 849 850 /** 851 * Enables literal parsing of the pattern. 852 * 853 * <p> When this flag is specified then the input string that specifies 854 * the pattern is treated as a sequence of literal characters. 855 * Metacharacters or escape sequences in the input sequence will be 856 * given no special meaning. 857 * 858 * <p>The flags CASE_INSENSITIVE and UNICODE_CASE retain their impact on 859 * matching when used in conjunction with this flag. The other flags 860 * become superfluous. 861 * 862 * <p> There is no embedded flag character for enabling literal parsing. 863 * @since 1.5 864 */ 865 public static final int LITERAL = 0x10; 866 867 /** 868 * Enables dotall mode. 869 * 870 * <p> In dotall mode, the expression <tt>.</tt> matches any character, 871 * including a line terminator. By default this expression does not match 872 * line terminators. 873 * 874 * <p> Dotall mode can also be enabled via the embedded flag 875 * expression <tt>(?s)</tt>. (The <tt>s</tt> is a mnemonic for 876 * "single-line" mode, which is what this is called in Perl.) </p> 877 */ 878 public static final int DOTALL = 0x20; 879 880 /** 881 * Enables Unicode-aware case folding. 882 * 883 * <p> When this flag is specified then case-insensitive matching, when 884 * enabled by the {@link #CASE_INSENSITIVE} flag, is done in a manner 885 * consistent with the Unicode Standard. By default, case-insensitive 886 * matching assumes that only characters in the US-ASCII charset are being 887 * matched. 888 * 889 * <p> Unicode-aware case folding can also be enabled via the embedded flag 890 * expression <tt>(?u)</tt>. 891 * 892 * <p> Specifying this flag may impose a performance penalty. </p> 893 */ 894 public static final int UNICODE_CASE = 0x40; 895 896 // Android-changed: Android does not support CANON_EQ flag. 897 /** 898 * This flag is not supported on Android. 899 * 900 * Enables canonical equivalence. 901 * 902 * <p> When this flag is specified then two characters will be considered 903 * to match if, and only if, their full canonical decompositions match. 904 * The expression <tt>"a\u030A"</tt>, for example, will match the 905 * string <tt>"\u00E5"</tt> when this flag is specified. By default, 906 * matching does not take canonical equivalence into account. 907 * 908 * <p> There is no embedded flag character for enabling canonical 909 * equivalence. 910 * 911 * <p> Specifying this flag may impose a performance penalty. </p> 912 */ 913 public static final int CANON_EQ = 0x80; 914 915 // Android-changed: Android always uses unicode character classes. 916 /** 917 * This flag is not supported on Android, and Unicode character classes are always 918 * used. 919 * 920 * Enables the Unicode version of <i>Predefined character classes</i> and 921 * <i>POSIX character classes</i> as defined by <a href="http://www.unicode.org/reports/tr18/"><i>Unicode Technical 922 * Standard #18: Unicode Regular Expression</i></a> 923 * <i>Annex C: Compatibility Properties</i>. 924 * <p> 925 * 926 * @since 1.7 927 */ 928 public static final int UNICODE_CHARACTER_CLASS = 0x100; 929 930 /* Pattern has only two serialized components: The pattern string 931 * and the flags, which are all that is needed to recompile the pattern 932 * when it is deserialized. 933 */ 934 935 /** use serialVersionUID from Merlin b59 for interoperability */ 936 private static final long serialVersionUID = 5073258162644648461L; 937 938 /** 939 * The original regular-expression pattern string. 940 * 941 * @serial 942 */ 943 // Android-changed: reimplement matching logic natively via ICU. 944 // private String pattern; 945 private final String pattern; 946 947 /** 948 * The original pattern flags. 949 * 950 * @serial 951 */ 952 // Android-changed: reimplement matching logic natively via ICU. 953 // private int flags; 954 private final int flags; 955 956 // BEGIN Android-changed: reimplement matching logic natively via ICU. 957 // We only need some tie-ins to native memory, instead of a large number 958 // of fields on the .java side. 959 /* package */ transient PatternNative nativePattern; 960 // END Android-changed: reimplement matching logic natively via ICU. 961 962 /** 963 * Compiles the given regular expression into a pattern. 964 * 965 * @param regex 966 * The expression to be compiled 967 * @return the given regular expression compiled into a pattern 968 * @throws PatternSyntaxException 969 * If the expression's syntax is invalid 970 */ compile(String regex)971 public static Pattern compile(String regex) { 972 return new Pattern(regex, 0); 973 } 974 975 // Android-changed: Android doesn't support CANON_EQ and UNICODE_CHARACTER_CLASS flags. 976 /** 977 * Compiles the given regular expression into a pattern with the given 978 * flags. 979 * 980 * @param regex 981 * The expression to be compiled 982 * 983 * @param flags 984 * Match flags, a bit mask that may include 985 * {@link #CASE_INSENSITIVE}, {@link #MULTILINE}, {@link #DOTALL}, 986 * {@link #UNICODE_CASE}, {@link #UNIX_LINES}, {@link #LITERAL}, 987 * and {@link #COMMENTS} 988 * 989 * @return the given regular expression compiled into a pattern with the given flags 990 * @throws IllegalArgumentException 991 * If bit values other than those corresponding to the defined 992 * match flags are set in <tt>flags</tt> 993 * 994 * @throws PatternSyntaxException 995 * If the expression's syntax is invalid 996 */ compile(String regex, int flags)997 public static Pattern compile(String regex, int flags) { 998 return new Pattern(regex, flags); 999 } 1000 1001 /** 1002 * Returns the regular expression from which this pattern was compiled. 1003 * 1004 * @return The source of this pattern 1005 */ pattern()1006 public String pattern() { 1007 return pattern; 1008 } 1009 1010 /** 1011 * <p>Returns the string representation of this pattern. This 1012 * is the regular expression from which this pattern was 1013 * compiled.</p> 1014 * 1015 * @return The string representation of this pattern 1016 * @since 1.5 1017 */ toString()1018 public String toString() { 1019 return pattern; 1020 } 1021 1022 /** 1023 * Creates a matcher that will match the given input against this pattern. 1024 * 1025 * @param input 1026 * The character sequence to be matched 1027 * 1028 * @return A new matcher for this pattern 1029 */ matcher(CharSequence input)1030 public Matcher matcher(CharSequence input) { 1031 // Android-removed: Pattern is eagerly compiled() upon construction. 1032 /* 1033 if (!compiled) { 1034 synchronized(this) { 1035 if (!compiled) 1036 compile(); 1037 } 1038 } 1039 */ 1040 Matcher m = new Matcher(this, input); 1041 return m; 1042 } 1043 1044 /** 1045 * Returns this pattern's match flags. 1046 * 1047 * @return The match flags specified when this pattern was compiled 1048 */ flags()1049 public int flags() { 1050 return flags; 1051 } 1052 1053 /** 1054 * Compiles the given regular expression and attempts to match the given 1055 * input against it. 1056 * 1057 * <p> An invocation of this convenience method of the form 1058 * 1059 * <blockquote><pre> 1060 * Pattern.matches(regex, input);</pre></blockquote> 1061 * 1062 * behaves in exactly the same way as the expression 1063 * 1064 * <blockquote><pre> 1065 * Pattern.compile(regex).matcher(input).matches()</pre></blockquote> 1066 * 1067 * <p> If a pattern is to be used multiple times, compiling it once and reusing 1068 * it will be more efficient than invoking this method each time. </p> 1069 * 1070 * @param regex 1071 * The expression to be compiled 1072 * 1073 * @param input 1074 * The character sequence to be matched 1075 * @return whether or not the regular expression matches on the input 1076 * @throws PatternSyntaxException 1077 * If the expression's syntax is invalid 1078 */ matches(String regex, CharSequence input)1079 public static boolean matches(String regex, CharSequence input) { 1080 Pattern p = Pattern.compile(regex); 1081 Matcher m = p.matcher(input); 1082 return m.matches(); 1083 } 1084 1085 // Android-changed: Adopt split() behavior change only for apps targeting API > 28. 1086 // http://b/109659282#comment7 1087 /** 1088 * Splits the given input sequence around matches of this pattern. 1089 * 1090 * <p> The array returned by this method contains each substring of the 1091 * input sequence that is terminated by another subsequence that matches 1092 * this pattern or is terminated by the end of the input sequence. The 1093 * substrings in the array are in the order in which they occur in the 1094 * input. If this pattern does not match any subsequence of the input then 1095 * the resulting array has just one element, namely the input sequence in 1096 * string form. 1097 * 1098 * <p> When there is a positive-width match at the beginning of the input 1099 * sequence then an empty leading substring is included at the beginning 1100 * of the resulting array. A zero-width match at the beginning however 1101 * can only produce such an empty leading substring for apps running on or 1102 * targeting API versions <= 28. 1103 * 1104 * <p> The <tt>limit</tt> parameter controls the number of times the 1105 * pattern is applied and therefore affects the length of the resulting 1106 * array. If the limit <i>n</i> is greater than zero then the pattern 1107 * will be applied at most <i>n</i> - 1 times, the array's 1108 * length will be no greater than <i>n</i>, and the array's last entry 1109 * will contain all input beyond the last matched delimiter. If <i>n</i> 1110 * is non-positive then the pattern will be applied as many times as 1111 * possible and the array can have any length. If <i>n</i> is zero then 1112 * the pattern will be applied as many times as possible, the array can 1113 * have any length, and trailing empty strings will be discarded. 1114 * 1115 * <p> The input <tt>"boo:and:foo"</tt>, for example, yields the following 1116 * results with these parameters: 1117 * 1118 * <blockquote><table cellpadding=1 cellspacing=0 1119 * summary="Split examples showing regex, limit, and result"> 1120 * <tr><th align="left"><i>Regex </i></th> 1121 * <th align="left"><i>Limit </i></th> 1122 * <th align="left"><i>Result </i></th></tr> 1123 * <tr><td align=center>:</td> 1124 * <td align=center>2</td> 1125 * <td><tt>{ "boo", "and:foo" }</tt></td></tr> 1126 * <tr><td align=center>:</td> 1127 * <td align=center>5</td> 1128 * <td><tt>{ "boo", "and", "foo" }</tt></td></tr> 1129 * <tr><td align=center>:</td> 1130 * <td align=center>-2</td> 1131 * <td><tt>{ "boo", "and", "foo" }</tt></td></tr> 1132 * <tr><td align=center>o</td> 1133 * <td align=center>5</td> 1134 * <td><tt>{ "b", "", ":and:f", "", "" }</tt></td></tr> 1135 * <tr><td align=center>o</td> 1136 * <td align=center>-2</td> 1137 * <td><tt>{ "b", "", ":and:f", "", "" }</tt></td></tr> 1138 * <tr><td align=center>o</td> 1139 * <td align=center>0</td> 1140 * <td><tt>{ "b", "", ":and:f" }</tt></td></tr> 1141 * </table></blockquote> 1142 * 1143 * @param input 1144 * The character sequence to be split 1145 * 1146 * @param limit 1147 * The result threshold, as described above 1148 * 1149 * @return The array of strings computed by splitting the input 1150 * around matches of this pattern 1151 */ split(CharSequence input, int limit)1152 public String[] split(CharSequence input, int limit) { 1153 // BEGIN Android-added: fastSplit() to speed up simple cases. 1154 String[] fast = fastSplit(pattern, input.toString(), limit); 1155 if (fast != null) { 1156 return fast; 1157 } 1158 // END Android-added: fastSplit() to speed up simple cases. 1159 int index = 0; 1160 boolean matchLimited = limit > 0; 1161 ArrayList<String> matchList = new ArrayList<>(); 1162 Matcher m = matcher(input); 1163 1164 // Add segments before each match found 1165 while(m.find()) { 1166 if (!matchLimited || matchList.size() < limit - 1) { 1167 if (index == 0 && index == m.start() && m.start() == m.end()) { 1168 // no empty leading substring included for zero-width match 1169 // at the beginning of the input char sequence. 1170 // BEGIN Android-changed: split() compat behavior for apps targeting <= 28. 1171 // continue; 1172 int targetSdkVersion = VMRuntime.getRuntime().getTargetSdkVersion(); 1173 if (targetSdkVersion > 28) { 1174 continue; 1175 } 1176 // END Android-changed: split() compat behavior for apps targeting <= 28. 1177 } 1178 String match = input.subSequence(index, m.start()).toString(); 1179 matchList.add(match); 1180 index = m.end(); 1181 } else if (matchList.size() == limit - 1) { // last one 1182 String match = input.subSequence(index, 1183 input.length()).toString(); 1184 matchList.add(match); 1185 index = m.end(); 1186 } 1187 } 1188 1189 // If no match was found, return this 1190 if (index == 0) 1191 return new String[] {input.toString()}; 1192 1193 // Add remaining segment 1194 if (!matchLimited || matchList.size() < limit) 1195 matchList.add(input.subSequence(index, input.length()).toString()); 1196 1197 // Construct result 1198 int resultSize = matchList.size(); 1199 if (limit == 0) 1200 while (resultSize > 0 && matchList.get(resultSize-1).equals("")) 1201 resultSize--; 1202 String[] result = new String[resultSize]; 1203 return matchList.subList(0, resultSize).toArray(result); 1204 } 1205 1206 // BEGIN Android-added: fastSplit() to speed up simple cases. 1207 private static final String FASTSPLIT_METACHARACTERS = "\\?*+[](){}^$.|"; 1208 1209 /** 1210 * Returns a result equivalent to {@code s.split(separator, limit)} if it's able 1211 * to compute it more cheaply than native impl, or null if the caller should fall back to 1212 * using native impl. 1213 * 1214 * fastpath will work if the regex is a 1215 * (1)one-char String and this character is not one of the 1216 * RegEx's meta characters ".$|()[{^?*+\\", or 1217 * (2)two-char String and the first char is the backslash and 1218 * the second is one of regEx's meta characters ".$|()[{^?*+\\". 1219 * @hide 1220 */ fastSplit(String re, String input, int limit)1221 public static String[] fastSplit(String re, String input, int limit) { 1222 // Can we do it cheaply? 1223 int len = re.length(); 1224 if (len == 0) { 1225 return null; 1226 } 1227 char ch = re.charAt(0); 1228 if (len == 1) { 1229 if (Character.isSurrogate(ch)) { 1230 // Single surrogate is an invalid UTF-16 sequence. 1231 return null; 1232 } else if (FASTSPLIT_METACHARACTERS.indexOf(ch) != -1) { 1233 // We don't allow a single metacharacter. 1234 return null; 1235 } 1236 // pass through 1237 } else if (len == 2 && ch == '\\') { 1238 // We're looking for a quoted character. 1239 // Quoted metacharacters are effectively single non-metacharacters. 1240 ch = re.charAt(1); 1241 if (FASTSPLIT_METACHARACTERS.indexOf(ch) == -1) { 1242 return null; 1243 } 1244 } else { 1245 return null; 1246 } 1247 1248 // We can do this cheaply... 1249 1250 // Unlike Perl, which considers the result of splitting the empty string to be the empty 1251 // array, Java returns an array containing the empty string. 1252 if (input.isEmpty()) { 1253 return new String[] { "" }; 1254 } 1255 1256 // Count separators 1257 int separatorCount = 0; 1258 int begin = 0; 1259 int end; 1260 while (separatorCount + 1 != limit && (end = input.indexOf(ch, begin)) != -1) { 1261 ++separatorCount; 1262 begin = end + 1; 1263 } 1264 int lastPartEnd = input.length(); 1265 if (limit == 0 && begin == lastPartEnd) { 1266 // Last part is empty for limit == 0, remove all trailing empty matches. 1267 if (separatorCount == lastPartEnd) { 1268 // Input contains only separators. 1269 return EmptyArray.STRING; 1270 } 1271 // Find the beginning of trailing separators. 1272 do { 1273 --begin; 1274 } while (input.charAt(begin - 1) == ch); 1275 // Reduce separatorCount and fix lastPartEnd. 1276 separatorCount -= input.length() - begin; 1277 lastPartEnd = begin; 1278 } 1279 1280 // Collect the result parts. 1281 String[] result = new String[separatorCount + 1]; 1282 begin = 0; 1283 for (int i = 0; i != separatorCount; ++i) { 1284 end = input.indexOf(ch, begin); 1285 result[i] = input.substring(begin, end); 1286 begin = end + 1; 1287 } 1288 // Add last part. 1289 result[separatorCount] = input.substring(begin, lastPartEnd); 1290 return result; 1291 } 1292 // END Android-added: fastSplit() to speed up simple cases. 1293 1294 /** 1295 * Splits the given input sequence around matches of this pattern. 1296 * 1297 * <p> This method works as if by invoking the two-argument {@link 1298 * #split(java.lang.CharSequence, int) split} method with the given input 1299 * sequence and a limit argument of zero. Trailing empty strings are 1300 * therefore not included in the resulting array. </p> 1301 * 1302 * <p> The input <tt>"boo:and:foo"</tt>, for example, yields the following 1303 * results with these expressions: 1304 * 1305 * <blockquote><table cellpadding=1 cellspacing=0 1306 * summary="Split examples showing regex and result"> 1307 * <tr><th align="left"><i>Regex </i></th> 1308 * <th align="left"><i>Result</i></th></tr> 1309 * <tr><td align=center>:</td> 1310 * <td><tt>{ "boo", "and", "foo" }</tt></td></tr> 1311 * <tr><td align=center>o</td> 1312 * <td><tt>{ "b", "", ":and:f" }</tt></td></tr> 1313 * </table></blockquote> 1314 * 1315 * 1316 * @param input 1317 * The character sequence to be split 1318 * 1319 * @return The array of strings computed by splitting the input 1320 * around matches of this pattern 1321 */ split(CharSequence input)1322 public String[] split(CharSequence input) { 1323 return split(input, 0); 1324 } 1325 1326 /** 1327 * Returns a literal pattern <code>String</code> for the specified 1328 * <code>String</code>. 1329 * 1330 * <p>This method produces a <code>String</code> that can be used to 1331 * create a <code>Pattern</code> that would match the string 1332 * <code>s</code> as if it were a literal pattern.</p> Metacharacters 1333 * or escape sequences in the input sequence will be given no special 1334 * meaning. 1335 * 1336 * @param s The string to be literalized 1337 * @return A literal string replacement 1338 * @since 1.5 1339 */ quote(String s)1340 public static String quote(String s) { 1341 int slashEIndex = s.indexOf("\\E"); 1342 if (slashEIndex == -1) 1343 return "\\Q" + s + "\\E"; 1344 1345 StringBuilder sb = new StringBuilder(s.length() * 2); 1346 sb.append("\\Q"); 1347 slashEIndex = 0; 1348 int current = 0; 1349 while ((slashEIndex = s.indexOf("\\E", current)) != -1) { 1350 sb.append(s.substring(current, slashEIndex)); 1351 current = slashEIndex + 2; 1352 sb.append("\\E\\\\E\\Q"); 1353 } 1354 sb.append(s.substring(current, s.length())); 1355 sb.append("\\E"); 1356 return sb.toString(); 1357 } 1358 1359 /** 1360 * Recompile the Pattern instance from a stream. The original pattern 1361 * string is read in and the object tree is recompiled from it. 1362 */ readObject(java.io.ObjectInputStream s)1363 private void readObject(java.io.ObjectInputStream s) 1364 throws java.io.IOException, ClassNotFoundException { 1365 1366 // Read in all fields 1367 s.defaultReadObject(); 1368 1369 // Android-removed: reimplement matching logic natively via ICU. 1370 // // Initialize counts 1371 // capturingGroupCount = 1; 1372 // localCount = 0; 1373 1374 // Android-changed: Pattern is eagerly compiled() upon construction. 1375 /* 1376 // if length > 0, the Pattern is lazily compiled 1377 compiled = false; 1378 if (pattern.length() == 0) { 1379 root = new Start(lastAccept); 1380 matchRoot = lastAccept; 1381 compiled = true; 1382 } 1383 */ 1384 compile(); 1385 } 1386 1387 // Android-changed: reimplement matching logic natively via ICU. 1388 // Dropped documentation reference to Start and LastNode implementation 1389 // details which do not apply on Android. 1390 /** 1391 * This private constructor is used to create all Patterns. The pattern 1392 * string and match flags are all that is needed to completely describe 1393 * a Pattern. 1394 */ Pattern(String p, int f)1395 private Pattern(String p, int f) { 1396 pattern = p; 1397 flags = f; 1398 1399 // BEGIN Android-changed: Only specific flags are supported. 1400 /* 1401 // to use UNICODE_CASE if UNICODE_CHARACTER_CLASS present 1402 if ((flags & UNICODE_CHARACTER_CLASS) != 0) 1403 flags |= UNICODE_CASE; 1404 1405 // Reset group index count 1406 capturingGroupCount = 1; 1407 localCount = 0; 1408 */ 1409 if ((f & CANON_EQ) != 0) { 1410 throw new UnsupportedOperationException("CANON_EQ flag not supported"); 1411 } 1412 int supportedFlags = CASE_INSENSITIVE | COMMENTS | DOTALL | LITERAL | MULTILINE | UNICODE_CASE | UNIX_LINES; 1413 if ((f & ~supportedFlags) != 0) { 1414 throw new IllegalArgumentException("Unsupported flags: " + (f & ~supportedFlags)); 1415 } 1416 // END Android-changed: Only specific flags are supported. 1417 1418 // BEGIN Android-removed: Pattern is eagerly compiled() upon construction. 1419 // if (pattern.length() > 0) { 1420 // END Android-removed: Pattern is eagerly compiled() upon construction. 1421 compile(); 1422 // Android-removed: reimplement matching logic natively via ICU. 1423 /* 1424 } else { 1425 root = new Start(lastAccept); 1426 matchRoot = lastAccept; 1427 } 1428 */ 1429 } 1430 1431 // BEGIN Android-changed: reimplement matching logic natively via ICU. 1432 // Use native implementation instead of > 3000 lines of helper methods. compile()1433 private void compile() throws PatternSyntaxException { 1434 if (pattern == null) { 1435 throw new NullPointerException("pattern == null"); 1436 } 1437 1438 String icuPattern = pattern; 1439 if ((flags & LITERAL) != 0) { 1440 icuPattern = quote(pattern); 1441 } 1442 1443 // These are the flags natively supported by ICU. 1444 // They even have the same value in native code. 1445 int icuFlags = flags & (CASE_INSENSITIVE | COMMENTS | MULTILINE | DOTALL | UNIX_LINES); 1446 nativePattern = PatternNative.create(icuPattern, icuFlags); 1447 } 1448 // END Android-changed: reimplement matching logic natively via ICU. 1449 1450 /** 1451 * Creates a predicate which can be used to match a string. 1452 * 1453 * @return The predicate which can be used for matching on a string 1454 * @since 1.8 1455 */ asPredicate()1456 public Predicate<String> asPredicate() { 1457 return s -> matcher(s).find(); 1458 } 1459 1460 /** 1461 * Creates a stream from the given input sequence around matches of this 1462 * pattern. 1463 * 1464 * <p> The stream returned by this method contains each substring of the 1465 * input sequence that is terminated by another subsequence that matches 1466 * this pattern or is terminated by the end of the input sequence. The 1467 * substrings in the stream are in the order in which they occur in the 1468 * input. Trailing empty strings will be discarded and not encountered in 1469 * the stream. 1470 * 1471 * <p> If this pattern does not match any subsequence of the input then 1472 * the resulting stream has just one element, namely the input sequence in 1473 * string form. 1474 * 1475 * <p> When there is a positive-width match at the beginning of the input 1476 * sequence then an empty leading substring is included at the beginning 1477 * of the stream. A zero-width match at the beginning however never produces 1478 * such empty leading substring. 1479 * 1480 * <p> If the input sequence is mutable, it must remain constant during the 1481 * execution of the terminal stream operation. Otherwise, the result of the 1482 * terminal stream operation is undefined. 1483 * 1484 * @param input 1485 * The character sequence to be split 1486 * 1487 * @return The stream of strings computed by splitting the input 1488 * around matches of this pattern 1489 * @see #split(CharSequence) 1490 * @since 1.8 1491 */ splitAsStream(final CharSequence input)1492 public Stream<String> splitAsStream(final CharSequence input) { 1493 class MatcherIterator implements Iterator<String> { 1494 private final Matcher matcher; 1495 // The start position of the next sub-sequence of input 1496 // when current == input.length there are no more elements 1497 private int current; 1498 // null if the next element, if any, needs to obtained 1499 private String nextElement; 1500 // > 0 if there are N next empty elements 1501 private int emptyElementCount; 1502 1503 MatcherIterator() { 1504 this.matcher = matcher(input); 1505 } 1506 1507 public String next() { 1508 if (!hasNext()) 1509 throw new NoSuchElementException(); 1510 1511 if (emptyElementCount == 0) { 1512 String n = nextElement; 1513 nextElement = null; 1514 return n; 1515 } else { 1516 emptyElementCount--; 1517 return ""; 1518 } 1519 } 1520 1521 public boolean hasNext() { 1522 if (nextElement != null || emptyElementCount > 0) 1523 return true; 1524 1525 if (current == input.length()) 1526 return false; 1527 1528 // Consume the next matching element 1529 // Count sequence of matching empty elements 1530 while (matcher.find()) { 1531 nextElement = input.subSequence(current, matcher.start()).toString(); 1532 current = matcher.end(); 1533 if (!nextElement.isEmpty()) { 1534 return true; 1535 } else if (current > 0) { // no empty leading substring for zero-width 1536 // match at the beginning of the input 1537 emptyElementCount++; 1538 } 1539 } 1540 1541 // Consume last matching element 1542 nextElement = input.subSequence(current, input.length()).toString(); 1543 current = input.length(); 1544 if (!nextElement.isEmpty()) { 1545 return true; 1546 } else { 1547 // Ignore a terminal sequence of matching empty elements 1548 emptyElementCount = 0; 1549 nextElement = null; 1550 return false; 1551 } 1552 } 1553 } 1554 return StreamSupport.stream(Spliterators.spliteratorUnknownSize( 1555 new MatcherIterator(), Spliterator.ORDERED | Spliterator.NONNULL), false); 1556 } 1557 } 1558