1 /*
2 * Copyright (c) 1996, 2019, Oracle and/or its affiliates. All rights reserved.
3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4 *
5 * This code is free software; you can redistribute it and/or modify it
6 * under the terms of the GNU General Public License version 2 only, as
7 * published by the Free Software Foundation. Oracle designates this
8 * particular file as subject to the "Classpath" exception as provided
9 * by Oracle in the LICENSE file that accompanied this code.
10 *
11 * This code is distributed in the hope that it will be useful, but WITHOUT
12 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
13 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
14 * version 2 for more details (a copy is included in the LICENSE file that
15 * accompanied this code).
16 *
17 * You should have received a copy of the GNU General Public License version
18 * 2 along with this work; if not, write to the Free Software Foundation,
19 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
20 *
21 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
22 * or visit www.oracle.com if you need additional information or have any
23 * questions.
24 */
25
26 /*
27 * (C) Copyright Taligent, Inc. 1996 - All Rights Reserved
28 * (C) Copyright IBM Corp. 1996-1998 - All Rights Reserved
29 *
30 * The original version of this source code and documentation is copyrighted
31 * and owned by Taligent, Inc., a wholly-owned subsidiary of IBM. These
32 * materials are provided under terms of a License Agreement between Taligent
33 * and Sun. This technology is protected by multiple US and International
34 * patents. This notice and attribution to Taligent may not be removed.
35 * Taligent is a registered trademark of Taligent, Inc.
36 *
37 */
38
39 package java.text;
40
41 import java.io.IOException;
42 import java.io.InvalidObjectException;
43 import java.io.ObjectInputStream;
44 import static java.text.DateFormatSymbols.*;
45 import java.util.Calendar;
46 import java.util.Date;
47 import java.util.GregorianCalendar;
48 import java.util.Locale;
49 import java.util.Map;
50 import java.util.SimpleTimeZone;
51 import java.util.SortedMap;
52 import java.util.TimeZone;
53 import java.util.concurrent.ConcurrentHashMap;
54 import java.util.concurrent.ConcurrentMap;
55 import sun.util.calendar.CalendarUtils;
56 import sun.util.calendar.ZoneInfoFile;
57 import sun.util.locale.provider.LocaleProviderAdapter;
58 import sun.util.locale.provider.TimeZoneNameUtility;
59
60 /**
61 * <code>SimpleDateFormat</code> is a concrete class for formatting and
62 * parsing dates in a locale-sensitive manner. It allows for formatting
63 * (date → text), parsing (text → date), and normalization.
64 *
65 * <p>
66 * <code>SimpleDateFormat</code> allows you to start by choosing
67 * any user-defined patterns for date-time formatting. However, you
68 * are encouraged to create a date-time formatter with either
69 * <code>getTimeInstance</code>, <code>getDateInstance</code>, or
70 * <code>getDateTimeInstance</code> in <code>DateFormat</code>. Each
71 * of these class methods can return a date/time formatter initialized
72 * with a default format pattern. You may modify the format pattern
73 * using the <code>applyPattern</code> methods as desired.
74 * For more information on using these methods, see
75 * {@link DateFormat}.
76 *
77 * <h3>Date and Time Patterns</h3>
78 * <p>
79 * Date and time formats are specified by <em>date and time pattern</em>
80 * strings.
81 * Within date and time pattern strings, unquoted letters from
82 * <code>'A'</code> to <code>'Z'</code> and from <code>'a'</code> to
83 * <code>'z'</code> are interpreted as pattern letters representing the
84 * components of a date or time string.
85 * Text can be quoted using single quotes (<code>'</code>) to avoid
86 * interpretation.
87 * <code>"''"</code> represents a single quote.
88 * All other characters are not interpreted; they're simply copied into the
89 * output string during formatting or matched against the input string
90 * during parsing.
91 * <p>
92 * The following pattern letters are defined (all other characters from
93 * <code>'A'</code> to <code>'Z'</code> and from <code>'a'</code> to
94 * <code>'z'</code> are reserved):
95 * <blockquote>
96 * <table class="striped">
97 * <caption style="display:none">Chart shows pattern letters, date/time component, presentation, and examples.</caption>
98 * <thead>
99 * <tr>
100 * <th scope="col" style="text-align:left">Letter
101 * <th scope="col" style="text-align:left">Date or Time Component
102 * <th scope="col" style="text-align:left">Presentation
103 * <th scope="col" style="text-align:left">Examples
104 * </thead>
105 * <tbody>
106 * <tr>
107 * <th scope="row"><code>G</code>
108 * <td>Era designator
109 * <td><a href="#text">Text</a>
110 * <td><code>AD</code>
111 * <tr>
112 * <th scope="row"><code>y</code>
113 * <td>Year
114 * <td><a href="#year">Year</a>
115 * <td><code>1996</code>; <code>96</code>
116 * <tr>
117 * <th scope="row"><code>Y</code>
118 * <td>Week year
119 * <td><a href="#year">Year</a>
120 * <td><code>2009</code>; <code>09</code>
121 * <tr>
122 * <th scope="row"><code>M</code>
123 * <td>Month in year (context sensitive)
124 * <td><a href="#month">Month</a>
125 * <td><code>July</code>; <code>Jul</code>; <code>07</code>
126 * <tr>
127 * <th scope="row"><code>L</code>
128 * <td>Month in year (standalone form)
129 * <td><a href="#month">Month</a>
130 * <td><code>July</code>; <code>Jul</code>; <code>07</code>
131 * <tr>
132 * <th scope="row"><code>w</code>
133 * <td>Week in year
134 * <td><a href="#number">Number</a>
135 * <td><code>27</code>
136 * <tr>
137 * <th scope="row"><code>W</code>
138 * <td>Week in month
139 * <td><a href="#number">Number</a>
140 * <td><code>2</code>
141 * <tr>
142 * <th scope="row"><code>D</code>
143 * <td>Day in year
144 * <td><a href="#number">Number</a>
145 * <td><code>189</code>
146 * <tr>
147 * <th scope="row"><code>d</code>
148 * <td>Day in month
149 * <td><a href="#number">Number</a>
150 * <td><code>10</code>
151 * <tr>
152 * <th scope="row"><code>F</code>
153 * <td>Day of week in month
154 * <td><a href="#number">Number</a>
155 * <td><code>2</code>
156 * <tr>
157 * <th scope="row"><code>E</code>
158 * <td>Day name in week
159 * <td><a href="#text">Text</a>
160 * <td><code>Tuesday</code>; <code>Tue</code>
161 * <tr>
162 * <th scope="row"><code>u</code>
163 * <td>Day number of week (1 = Monday, ..., 7 = Sunday)
164 * <td><a href="#number">Number</a>
165 * <td><code>1</code>
166 * <tr>
167 * <th scope="row"><code>a</code>
168 * <td>Am/pm marker
169 * <td><a href="#text">Text</a>
170 * <td><code>PM</code>
171 * <tr>
172 * <th scope="row"><code>H</code>
173 * <td>Hour in day (0-23)
174 * <td><a href="#number">Number</a>
175 * <td><code>0</code>
176 * <tr>
177 * <th scope="row"><code>k</code>
178 * <td>Hour in day (1-24)
179 * <td><a href="#number">Number</a>
180 * <td><code>24</code>
181 * <tr>
182 * <th scope="row"><code>K</code>
183 * <td>Hour in am/pm (0-11)
184 * <td><a href="#number">Number</a>
185 * <td><code>0</code>
186 * <tr>
187 * <th scope="row"><code>h</code>
188 * <td>Hour in am/pm (1-12)
189 * <td><a href="#number">Number</a>
190 * <td><code>12</code>
191 * <tr>
192 * <th scope="row"><code>m</code>
193 * <td>Minute in hour
194 * <td><a href="#number">Number</a>
195 * <td><code>30</code>
196 * <tr>
197 * <th scope="row"><code>s</code>
198 * <td>Second in minute
199 * <td><a href="#number">Number</a>
200 * <td><code>55</code>
201 * <tr>
202 * <th scope="row"><code>S</code>
203 * <td>Millisecond
204 * <td><a href="#number">Number</a>
205 * <td><code>978</code>
206 * <tr>
207 * <th scope="row"><code>z</code>
208 * <td>Time zone
209 * <td><a href="#timezone">General time zone</a>
210 * <td><code>Pacific Standard Time</code>; <code>PST</code>; <code>GMT-08:00</code>
211 * <tr>
212 * <th scope="row"><code>Z</code>
213 * <td>Time zone
214 * <td><a href="#rfc822timezone">RFC 822 time zone</a>
215 * <td><code>-0800</code>
216 * <tr>
217 * <th scope="row"><code>X</code>
218 * <td>Time zone
219 * <td><a href="#iso8601timezone">ISO 8601 time zone</a>
220 * <td><code>-08</code>; <code>-0800</code>; <code>-08:00</code>
221 * </tbody>
222 * </table>
223 * </blockquote>
224 * Pattern letters are usually repeated, as their number determines the
225 * exact presentation:
226 * <ul>
227 * <li><strong><a id="text">Text:</a></strong>
228 * For formatting, if the number of pattern letters is 4 or more,
229 * the full form is used; otherwise a short or abbreviated form
230 * is used if available.
231 * For parsing, both forms are accepted, independent of the number
232 * of pattern letters.<br><br></li>
233 * <li><strong><a id="number">Number:</a></strong>
234 * For formatting, the number of pattern letters is the minimum
235 * number of digits, and shorter numbers are zero-padded to this amount.
236 * For parsing, the number of pattern letters is ignored unless
237 * it's needed to separate two adjacent fields.<br><br></li>
238 * <li><strong><a id="year">Year:</a></strong>
239 * If the formatter's {@link #getCalendar() Calendar} is the Gregorian
240 * calendar, the following rules are applied.<br>
241 * <ul>
242 * <li>For formatting, if the number of pattern letters is 2, the year
243 * is truncated to 2 digits; otherwise it is interpreted as a
244 * <a href="#number">number</a>.
245 * <li>For parsing, if the number of pattern letters is more than 2,
246 * the year is interpreted literally, regardless of the number of
247 * digits. So using the pattern "MM/dd/yyyy", "01/11/12" parses to
248 * Jan 11, 12 A.D.
249 * <li>For parsing with the abbreviated year pattern ("y" or "yy"),
250 * <code>SimpleDateFormat</code> must interpret the abbreviated year
251 * relative to some century. It does this by adjusting dates to be
252 * within 80 years before and 20 years after the time the <code>SimpleDateFormat</code>
253 * instance is created. For example, using a pattern of "MM/dd/yy" and a
254 * <code>SimpleDateFormat</code> instance created on Jan 1, 1997, the string
255 * "01/11/12" would be interpreted as Jan 11, 2012 while the string "05/04/64"
256 * would be interpreted as May 4, 1964.
257 * During parsing, only strings consisting of exactly two digits, as defined by
258 * {@link Character#isDigit(char)}, will be parsed into the default century.
259 * Any other numeric string, such as a one digit string, a three or more digit
260 * string, or a two digit string that isn't all digits (for example, "-1"), is
261 * interpreted literally. So "01/02/3" or "01/02/003" are parsed, using the
262 * same pattern, as Jan 2, 3 AD. Likewise, "01/02/-3" is parsed as Jan 2, 4 BC.
263 * </ul>
264 * Otherwise, calendar system specific forms are applied.
265 * For both formatting and parsing, if the number of pattern
266 * letters is 4 or more, a calendar specific {@linkplain
267 * Calendar#LONG long form} is used. Otherwise, a calendar
268 * specific {@linkplain Calendar#SHORT short or abbreviated form}
269 * is used.<br>
270 * <br>
271 * If week year {@code 'Y'} is specified and the {@linkplain
272 * #getCalendar() calendar} doesn't support any <a
273 * href="../util/GregorianCalendar.html#week_year"> week
274 * years</a>, the calendar year ({@code 'y'}) is used instead. The
275 * support of week years can be tested with a call to {@link
276 * DateFormat#getCalendar() getCalendar()}.{@link
277 * java.util.Calendar#isWeekDateSupported()
278 * isWeekDateSupported()}.<br><br></li>
279 * <li><strong><a id="month">Month:</a></strong>
280 * If the number of pattern letters is 3 or more, the month is
281 * interpreted as <a href="#text">text</a>; otherwise,
282 * it is interpreted as a <a href="#number">number</a>.<br>
283 * <ul>
284 * <li>Letter <em>M</em> produces context-sensitive month names, such as the
285 * embedded form of names. Letter <em>M</em> is context-sensitive in the
286 * sense that when it is used in the standalone pattern, for example,
287 * "MMMM", it gives the standalone form of a month name and when it is
288 * used in the pattern containing other field(s), for example, "d MMMM",
289 * it gives the format form of a month name. For example, January in the
290 * Catalan language is "de gener" in the format form while it is "gener"
291 * in the standalone form. In this case, "MMMM" will produce "gener" and
292 * the month part of the "d MMMM" will produce "de gener". If a
293 * {@code DateFormatSymbols} has been set explicitly with constructor
294 * {@link #SimpleDateFormat(String,DateFormatSymbols)} or method {@link
295 * #setDateFormatSymbols(DateFormatSymbols)}, the month names given by
296 * the {@code DateFormatSymbols} are used.</li>
297 * <li>Letter <em>L</em> produces the standalone form of month names.</li>
298 * </ul>
299 * <br></li>
300 * <li><strong><a id="timezone">General time zone:</a></strong>
301 * Time zones are interpreted as <a href="#text">text</a> if they have
302 * names. For time zones representing a GMT offset value, the
303 * following syntax is used:
304 * <pre>
305 * <a id="GMTOffsetTimeZone"><i>GMTOffsetTimeZone:</i></a>
306 * <code>GMT</code> <i>Sign</i> <i>Hours</i> <code>:</code> <i>Minutes</i>
307 * <i>Sign:</i> one of
308 * <code>+ -</code>
309 * <i>Hours:</i>
310 * <i>Digit</i>
311 * <i>Digit</i> <i>Digit</i>
312 * <i>Minutes:</i>
313 * <i>Digit</i> <i>Digit</i>
314 * <i>Digit:</i> one of
315 * <code>0 1 2 3 4 5 6 7 8 9</code></pre>
316 * <i>Hours</i> must be between 0 and 23, and <i>Minutes</i> must be between
317 * 00 and 59. The format is locale independent and digits must be taken
318 * from the Basic Latin block of the Unicode standard.
319 * <p>For parsing, <a href="#rfc822timezone">RFC 822 time zones</a> are also
320 * accepted.<br><br></li>
321 * <li><strong><a id="rfc822timezone">RFC 822 time zone:</a></strong>
322 * For formatting, the RFC 822 4-digit time zone format is used:
323 *
324 * <pre>
325 * <i>RFC822TimeZone:</i>
326 * <i>Sign</i> <i>TwoDigitHours</i> <i>Minutes</i>
327 * <i>TwoDigitHours:</i>
328 * <i>Digit Digit</i></pre>
329 * <i>TwoDigitHours</i> must be between 00 and 23. Other definitions
330 * are as for <a href="#timezone">general time zones</a>.
331 *
332 * <p>For parsing, <a href="#timezone">general time zones</a> are also
333 * accepted.
334 * <li><strong><a id="iso8601timezone">ISO 8601 Time zone:</a></strong>
335 * The number of pattern letters designates the format for both formatting
336 * and parsing as follows:
337 * <pre>
338 * <i>ISO8601TimeZone:</i>
339 * <i>OneLetterISO8601TimeZone</i>
340 * <i>TwoLetterISO8601TimeZone</i>
341 * <i>ThreeLetterISO8601TimeZone</i>
342 * <i>OneLetterISO8601TimeZone:</i>
343 * <i>Sign</i> <i>TwoDigitHours</i>
344 * {@code Z}
345 * <i>TwoLetterISO8601TimeZone:</i>
346 * <i>Sign</i> <i>TwoDigitHours</i> <i>Minutes</i>
347 * {@code Z}
348 * <i>ThreeLetterISO8601TimeZone:</i>
349 * <i>Sign</i> <i>TwoDigitHours</i> {@code :} <i>Minutes</i>
350 * {@code Z}</pre>
351 * Other definitions are as for <a href="#timezone">general time zones</a> or
352 * <a href="#rfc822timezone">RFC 822 time zones</a>.
353 *
354 * <p>For formatting, if the offset value from GMT is 0, {@code "Z"} is
355 * produced. If the number of pattern letters is 1, any fraction of an hour
356 * is ignored. For example, if the pattern is {@code "X"} and the time zone is
357 * {@code "GMT+05:30"}, {@code "+05"} is produced.
358 *
359 * <p>For parsing, {@code "Z"} is parsed as the UTC time zone designator.
360 * <a href="#timezone">General time zones</a> are <em>not</em> accepted.
361 *
362 * <p>If the number of pattern letters is 4 or more, {@link
363 * IllegalArgumentException} is thrown when constructing a {@code
364 * SimpleDateFormat} or {@linkplain #applyPattern(String) applying a
365 * pattern}.
366 * </ul>
367 * <code>SimpleDateFormat</code> also supports <em>localized date and time
368 * pattern</em> strings. In these strings, the pattern letters described above
369 * may be replaced with other, locale dependent, pattern letters.
370 * <code>SimpleDateFormat</code> does not deal with the localization of text
371 * other than the pattern letters; that's up to the client of the class.
372 *
373 * <h4>Examples</h4>
374 *
375 * The following examples show how date and time patterns are interpreted in
376 * the U.S. locale. The given date and time are 2001-07-04 12:08:56 local time
377 * in the U.S. Pacific Time time zone.
378 * <blockquote>
379 * <table class="striped">
380 * <caption style="display:none">Examples of date and time patterns interpreted in the U.S. locale</caption>
381 * <thead>
382 * <tr>
383 * <th scope="col" style="text-align:left">Date and Time Pattern
384 * <th scope="col" style="text-align:left">Result
385 * </thead>
386 * <tbody>
387 * <tr>
388 * <th scope="row"><code>"yyyy.MM.dd G 'at' HH:mm:ss z"</code>
389 * <td><code>2001.07.04 AD at 12:08:56 PDT</code>
390 * <tr>
391 * <th scope="row"><code>"EEE, MMM d, ''yy"</code>
392 * <td><code>Wed, Jul 4, '01</code>
393 * <tr>
394 * <th scope="row"><code>"h:mm a"</code>
395 * <td><code>12:08 PM</code>
396 * <tr>
397 * <th scope="row"><code>"hh 'o''clock' a, zzzz"</code>
398 * <td><code>12 o'clock PM, Pacific Daylight Time</code>
399 * <tr>
400 * <th scope="row"><code>"K:mm a, z"</code>
401 * <td><code>0:08 PM, PDT</code>
402 * <tr>
403 * <th scope="row"><code>"yyyyy.MMMMM.dd GGG hh:mm aaa"</code>
404 * <td><code>02001.July.04 AD 12:08 PM</code>
405 * <tr>
406 * <th scope="row"><code>"EEE, d MMM yyyy HH:mm:ss Z"</code>
407 * <td><code>Wed, 4 Jul 2001 12:08:56 -0700</code>
408 * <tr>
409 * <th scope="row"><code>"yyMMddHHmmssZ"</code>
410 * <td><code>010704120856-0700</code>
411 * <tr>
412 * <th scope="row"><code>"yyyy-MM-dd'T'HH:mm:ss.SSSZ"</code>
413 * <td><code>2001-07-04T12:08:56.235-0700</code>
414 * <tr>
415 * <th scope="row"><code>"yyyy-MM-dd'T'HH:mm:ss.SSSXXX"</code>
416 * <td><code>2001-07-04T12:08:56.235-07:00</code>
417 * <tr>
418 * <th scope="row"><code>"YYYY-'W'ww-u"</code>
419 * <td><code>2001-W27-3</code>
420 * </tbody>
421 * </table>
422 * </blockquote>
423 *
424 * <h4><a id="synchronization">Synchronization</a></h4>
425 *
426 * <p>
427 * Date formats are not synchronized.
428 * It is recommended to create separate format instances for each thread.
429 * If multiple threads access a format concurrently, it must be synchronized
430 * externally.
431 *
432 * @see <a href="http://docs.oracle.com/javase/tutorial/i18n/format/simpleDateFormat.html">Java Tutorial</a>
433 * @see java.util.Calendar
434 * @see java.util.TimeZone
435 * @see DateFormat
436 * @see DateFormatSymbols
437 * @author Mark Davis, Chen-Lieh Huang, Alan Liu
438 * @since 1.1
439 */
440 public class SimpleDateFormat extends DateFormat {
441
442 // the official serial version ID which says cryptically
443 // which version we're compatible with
444 static final long serialVersionUID = 4774881970558875024L;
445
446 // the internal serial version which says which version was written
447 // - 0 (default) for version up to JDK 1.1.3
448 // - 1 for version from JDK 1.1.4, which includes a new field
449 static final int currentSerialVersion = 1;
450
451 /**
452 * The version of the serialized data on the stream. Possible values:
453 * <ul>
454 * <li><b>0</b> or not present on stream: JDK 1.1.3. This version
455 * has no <code>defaultCenturyStart</code> on stream.
456 * <li><b>1</b> JDK 1.1.4 or later. This version adds
457 * <code>defaultCenturyStart</code>.
458 * </ul>
459 * When streaming out this class, the most recent format
460 * and the highest allowable <code>serialVersionOnStream</code>
461 * is written.
462 * @serial
463 * @since 1.1.4
464 */
465 private int serialVersionOnStream = currentSerialVersion;
466
467 /**
468 * The pattern string of this formatter. This is always a non-localized
469 * pattern. May not be null. See class documentation for details.
470 * @serial
471 */
472 private String pattern;
473
474 /**
475 * Saved numberFormat and pattern.
476 * @see SimpleDateFormat#checkNegativeNumberExpression
477 */
478 private transient NumberFormat originalNumberFormat;
479 private transient String originalNumberPattern;
480
481 /**
482 * The minus sign to be used with format and parse.
483 */
484 private transient char minusSign = '-';
485
486 /**
487 * True when a negative sign follows a number.
488 * (True as default in Arabic.)
489 */
490 private transient boolean hasFollowingMinusSign = false;
491
492 /**
493 * True if standalone form needs to be used.
494 */
495 private transient boolean forceStandaloneForm = false;
496
497 /**
498 * The compiled pattern.
499 */
500 private transient char[] compiledPattern;
501
502 /**
503 * Tags for the compiled pattern.
504 */
505 private static final int TAG_QUOTE_ASCII_CHAR = 100;
506 private static final int TAG_QUOTE_CHARS = 101;
507
508 /**
509 * Locale dependent digit zero.
510 * @see #zeroPaddingNumber
511 * @see java.text.DecimalFormatSymbols#getZeroDigit
512 */
513 private transient char zeroDigit;
514
515 /**
516 * The symbols used by this formatter for week names, month names,
517 * etc. May not be null.
518 * @serial
519 * @see java.text.DateFormatSymbols
520 */
521 private DateFormatSymbols formatData;
522
523 /**
524 * We map dates with two-digit years into the century starting at
525 * <code>defaultCenturyStart</code>, which may be any date. May
526 * not be null.
527 * @serial
528 * @since 1.1.4
529 */
530 private Date defaultCenturyStart;
531
532 private transient int defaultCenturyStartYear;
533
534 private static final int MILLIS_PER_MINUTE = 60 * 1000;
535
536 private static final int STANDALONE_MASK = 0x8000; // Calendar.STANDALONE_MASK
537
538 // For time zones that have no names, use strings GMT+minutes and
539 // GMT-minutes. For instance, in France the time zone is GMT+60.
540 private static final String GMT = "GMT";
541
542 /**
543 * Cache NumberFormat instances with Locale key.
544 */
545 private static final ConcurrentMap<Locale, NumberFormat> cachedNumberFormatData
546 = new ConcurrentHashMap<>(3);
547
548 /**
549 * The Locale used to instantiate this
550 * <code>SimpleDateFormat</code>. The value may be null if this object
551 * has been created by an older <code>SimpleDateFormat</code> and
552 * deserialized.
553 *
554 * @serial
555 * @since 1.6
556 */
557 private Locale locale;
558
559 /**
560 * Indicates whether this <code>SimpleDateFormat</code> should use
561 * the DateFormatSymbols. If true, the format and parse methods
562 * use the DateFormatSymbols values. If false, the format and
563 * parse methods call Calendar.getDisplayName or
564 * Calendar.getDisplayNames.
565 */
566 transient boolean useDateFormatSymbols;
567
568 /**
569 * Constructs a <code>SimpleDateFormat</code> using the default pattern and
570 * date format symbols for the default
571 * {@link java.util.Locale.Category#FORMAT FORMAT} locale.
572 * <b>Note:</b> This constructor may not support all locales.
573 * For full coverage, use the factory methods in the {@link DateFormat}
574 * class.
575 */
576 public SimpleDateFormat() {
577 this("", Locale.getDefault(Locale.Category.FORMAT));
578 applyPatternImpl(LocaleProviderAdapter.getResourceBundleBased().getLocaleResources(locale)
579 .getDateTimePattern(SHORT, SHORT, calendar));
580 }
581
582 /**
583 * Constructs a <code>SimpleDateFormat</code> using the given pattern and
584 * the default date format symbols for the default
585 * {@link java.util.Locale.Category#FORMAT FORMAT} locale.
586 * <b>Note:</b> This constructor may not support all locales.
587 * For full coverage, use the factory methods in the {@link DateFormat}
588 * class.
589 * <p>This is equivalent to calling
590 * {@link #SimpleDateFormat(String, Locale)
591 * SimpleDateFormat(pattern, Locale.getDefault(Locale.Category.FORMAT))}.
592 *
593 * @see java.util.Locale#getDefault(java.util.Locale.Category)
594 * @see java.util.Locale.Category#FORMAT
595 * @param pattern the pattern describing the date and time format
596 * @exception NullPointerException if the given pattern is null
597 * @exception IllegalArgumentException if the given pattern is invalid
598 */
599 public SimpleDateFormat(String pattern)
600 {
601 this(pattern, Locale.getDefault(Locale.Category.FORMAT));
602 }
603
604 /**
605 * Constructs a <code>SimpleDateFormat</code> using the given pattern and
606 * the default date format symbols for the given locale.
607 * <b>Note:</b> This constructor may not support all locales.
608 * For full coverage, use the factory methods in the {@link DateFormat}
609 * class.
610 *
611 * @param pattern the pattern describing the date and time format
612 * @param locale the locale whose date format symbols should be used
613 * @exception NullPointerException if the given pattern or locale is null
614 * @exception IllegalArgumentException if the given pattern is invalid
615 */
616 public SimpleDateFormat(String pattern, Locale locale)
617 {
618 if (pattern == null || locale == null) {
619 throw new NullPointerException();
620 }
621
622 initializeCalendar(locale);
623 this.pattern = pattern;
624 this.formatData = DateFormatSymbols.getInstanceRef(locale);
625 this.locale = locale;
626 initialize(locale);
627 }
628
629 /**
630 * Constructs a <code>SimpleDateFormat</code> using the given pattern and
631 * date format symbols.
632 *
633 * @param pattern the pattern describing the date and time format
634 * @param formatSymbols the date format symbols to be used for formatting
635 * @exception NullPointerException if the given pattern or formatSymbols is null
636 * @exception IllegalArgumentException if the given pattern is invalid
637 */
638 public SimpleDateFormat(String pattern, DateFormatSymbols formatSymbols)
639 {
640 if (pattern == null || formatSymbols == null) {
641 throw new NullPointerException();
642 }
643
644 this.pattern = pattern;
645 this.formatData = (DateFormatSymbols) formatSymbols.clone();
646 this.locale = Locale.getDefault(Locale.Category.FORMAT);
647 initializeCalendar(this.locale);
648 initialize(this.locale);
649 useDateFormatSymbols = true;
650 }
651
652 /* Initialize compiledPattern and numberFormat fields */
653 private void initialize(Locale loc) {
654 // Verify and compile the given pattern.
655 compiledPattern = compile(pattern);
656
657 /* try the cache first */
658 numberFormat = cachedNumberFormatData.get(loc);
659 if (numberFormat == null) { /* cache miss */
660 numberFormat = NumberFormat.getIntegerInstance(loc);
661 numberFormat.setGroupingUsed(false);
662
663 /* update cache */
664 cachedNumberFormatData.putIfAbsent(loc, numberFormat);
665 }
666 numberFormat = (NumberFormat) numberFormat.clone();
667
668 initializeDefaultCentury();
669 }
670
671 private void initializeCalendar(Locale loc) {
672 if (calendar == null) {
673 assert loc != null;
674 // The format object must be constructed using the symbols for this zone.
675 // However, the calendar should use the current default TimeZone.
676 // If this is not contained in the locale zone strings, then the zone
677 // will be formatted using generic GMT+/-H:MM nomenclature.
678 calendar = Calendar.getInstance(loc);
679 }
680 }
681
682 /**
683 * Returns the compiled form of the given pattern. The syntax of
684 * the compiled pattern is:
685 * <blockquote>
686 * CompiledPattern:
687 * EntryList
688 * EntryList:
689 * Entry
690 * EntryList Entry
691 * Entry:
692 * TagField
693 * TagField data
694 * TagField:
695 * Tag Length
696 * TaggedData
697 * Tag:
698 * pattern_char_index
699 * TAG_QUOTE_CHARS
700 * Length:
701 * short_length
702 * long_length
703 * TaggedData:
704 * TAG_QUOTE_ASCII_CHAR ascii_char
705 *
706 * </blockquote>
707 *
708 * where `short_length' is an 8-bit unsigned integer between 0 and
709 * 254. `long_length' is a sequence of an 8-bit integer 255 and a
710 * 32-bit signed integer value which is split into upper and lower
711 * 16-bit fields in two char's. `pattern_char_index' is an 8-bit
712 * integer between 0 and 18. `ascii_char' is an 7-bit ASCII
713 * character value. `data' depends on its Tag value.
714 * <p>
715 * If Length is short_length, Tag and short_length are packed in a
716 * single char, as illustrated below.
717 * <blockquote>
718 * char[0] = (Tag << 8) | short_length;
719 * </blockquote>
720 *
721 * If Length is long_length, Tag and 255 are packed in the first
722 * char and a 32-bit integer, as illustrated below.
723 * <blockquote>
724 * char[0] = (Tag << 8) | 255;
725 * char[1] = (char) (long_length >>> 16);
726 * char[2] = (char) (long_length & 0xffff);
727 * </blockquote>
728 * <p>
729 * If Tag is a pattern_char_index, its Length is the number of
730 * pattern characters. For example, if the given pattern is
731 * "yyyy", Tag is 1 and Length is 4, followed by no data.
732 * <p>
733 * If Tag is TAG_QUOTE_CHARS, its Length is the number of char's
734 * following the TagField. For example, if the given pattern is
735 * "'o''clock'", Length is 7 followed by a char sequence of
736 * <code>o&nbs;'&nbs;c&nbs;l&nbs;o&nbs;c&nbs;k</code>.
737 * <p>
738 * TAG_QUOTE_ASCII_CHAR is a special tag and has an ASCII
739 * character in place of Length. For example, if the given pattern
740 * is "'o'", the TaggedData entry is
741 * <code>((TAG_QUOTE_ASCII_CHAR&nbs;<<&nbs;8)&nbs;|&nbs;'o')</code>.
742 *
743 * @exception NullPointerException if the given pattern is null
744 * @exception IllegalArgumentException if the given pattern is invalid
745 */
746 private char[] compile(String pattern) {
747 int length = pattern.length();
748 boolean inQuote = false;
749 StringBuilder compiledCode = new StringBuilder(length * 2);
750 StringBuilder tmpBuffer = null;
751 int count = 0, tagcount = 0;
752 int lastTag = -1, prevTag = -1;
753
754 for (int i = 0; i < length; i++) {
755 char c = pattern.charAt(i);
756
757 if (c == '\'') {
758 // '' is treated as a single quote regardless of being
759 // in a quoted section.
760 if ((i + 1) < length) {
761 c = pattern.charAt(i + 1);
762 if (c == '\'') {
763 i++;
764 if (count != 0) {
765 encode(lastTag, count, compiledCode);
766 tagcount++;
767 prevTag = lastTag;
768 lastTag = -1;
769 count = 0;
770 }
771 if (inQuote) {
772 tmpBuffer.append(c);
773 } else {
774 compiledCode.append((char)(TAG_QUOTE_ASCII_CHAR << 8 | c));
775 }
776 continue;
777 }
778 }
779 if (!inQuote) {
780 if (count != 0) {
781 encode(lastTag, count, compiledCode);
782 tagcount++;
783 prevTag = lastTag;
784 lastTag = -1;
785 count = 0;
786 }
787 if (tmpBuffer == null) {
788 tmpBuffer = new StringBuilder(length);
789 } else {
790 tmpBuffer.setLength(0);
791 }
792 inQuote = true;
793 } else {
794 int len = tmpBuffer.length();
795 if (len == 1) {
796 char ch = tmpBuffer.charAt(0);
797 if (ch < 128) {
798 compiledCode.append((char)(TAG_QUOTE_ASCII_CHAR << 8 | ch));
799 } else {
800 compiledCode.append((char)(TAG_QUOTE_CHARS << 8 | 1));
801 compiledCode.append(ch);
802 }
803 } else {
804 encode(TAG_QUOTE_CHARS, len, compiledCode);
805 compiledCode.append(tmpBuffer);
806 }
807 inQuote = false;
808 }
809 continue;
810 }
811 if (inQuote) {
812 tmpBuffer.append(c);
813 continue;
814 }
815 if (!(c >= 'a' && c <= 'z' || c >= 'A' && c <= 'Z')) {
816 if (count != 0) {
817 encode(lastTag, count, compiledCode);
818 tagcount++;
819 prevTag = lastTag;
820 lastTag = -1;
821 count = 0;
822 }
823 if (c < 128) {
824 // In most cases, c would be a delimiter, such as ':'.
825 compiledCode.append((char)(TAG_QUOTE_ASCII_CHAR << 8 | c));
826 } else {
827 // Take any contiguous non-ASCII alphabet characters and
828 // put them in a single TAG_QUOTE_CHARS.
829 int j;
830 for (j = i + 1; j < length; j++) {
831 char d = pattern.charAt(j);
832 if (d == '\'' || (d >= 'a' && d <= 'z' || d >= 'A' && d <= 'Z')) {
833 break;
834 }
835 }
836 encode(TAG_QUOTE_CHARS, j - i, compiledCode);
837 for (; i < j; i++) {
838 compiledCode.append(pattern.charAt(i));
839 }
840 i--;
841 }
842 continue;
843 }
844
845 int tag;
846 if ((tag = DateFormatSymbols.patternChars.indexOf(c)) == -1) {
847 throw new IllegalArgumentException("Illegal pattern character " +
848 "'" + c + "'");
849 }
850 if (lastTag == -1 || lastTag == tag) {
851 lastTag = tag;
852 count++;
853 continue;
854 }
855 encode(lastTag, count, compiledCode);
856 tagcount++;
857 prevTag = lastTag;
858 lastTag = tag;
859 count = 1;
860 }
861
862 if (inQuote) {
863 throw new IllegalArgumentException("Unterminated quote");
864 }
865
866 if (count != 0) {
867 encode(lastTag, count, compiledCode);
868 tagcount++;
869 prevTag = lastTag;
870 }
871
872 forceStandaloneForm = (tagcount == 1 && prevTag == PATTERN_MONTH);
873
874 // Copy the compiled pattern to a char array
875 int len = compiledCode.length();
876 char[] r = new char[len];
877 compiledCode.getChars(0, len, r, 0);
878 return r;
879 }
880
881 /**
882 * Encodes the given tag and length and puts encoded char(s) into buffer.
883 */
884 private static void encode(int tag, int length, StringBuilder buffer) {
885 if (tag == PATTERN_ISO_ZONE && length >= 4) {
886 throw new IllegalArgumentException("invalid ISO 8601 format: length=" + length);
887 }
888 if (length < 255) {
889 buffer.append((char)(tag << 8 | length));
890 } else {
891 buffer.append((char)((tag << 8) | 0xff));
892 buffer.append((char)(length >>> 16));
893 buffer.append((char)(length & 0xffff));
894 }
895 }
896
897 /* Initialize the fields we use to disambiguate ambiguous years. Separate
898 * so we can call it from readObject().
899 */
900 private void initializeDefaultCentury() {
901 calendar.setTimeInMillis(System.currentTimeMillis());
902 calendar.add( Calendar.YEAR, -80 );
903 parseAmbiguousDatesAsAfter(calendar.getTime());
904 }
905
906 /* Define one-century window into which to disambiguate dates using
907 * two-digit years.
908 */
909 private void parseAmbiguousDatesAsAfter(Date startDate) {
910 defaultCenturyStart = startDate;
911 calendar.setTime(startDate);
912 defaultCenturyStartYear = calendar.get(Calendar.YEAR);
913 }
914
915 /**
916 * Sets the 100-year period 2-digit years will be interpreted as being in
917 * to begin on the date the user specifies.
918 *
919 * @param startDate During parsing, two digit years will be placed in the range
920 * <code>startDate</code> to <code>startDate + 100 years</code>.
921 * @see #get2DigitYearStart
922 * @throws NullPointerException if {@code startDate} is {@code null}.
923 * @since 1.2
924 */
925 public void set2DigitYearStart(Date startDate) {
926 parseAmbiguousDatesAsAfter(new Date(startDate.getTime()));
927 }
928
929 /**
930 * Returns the beginning date of the 100-year period 2-digit years are interpreted
931 * as being within.
932 *
933 * @return the start of the 100-year period into which two digit years are
934 * parsed
935 * @see #set2DigitYearStart
936 * @since 1.2
937 */
938 public Date get2DigitYearStart() {
939 return (Date) defaultCenturyStart.clone();
940 }
941
942 /**
943 * Formats the given <code>Date</code> into a date/time string and appends
944 * the result to the given <code>StringBuffer</code>.
945 *
946 * @param date the date-time value to be formatted into a date-time string.
947 * @param toAppendTo where the new date-time text is to be appended.
948 * @param pos keeps track on the position of the field within
949 * the returned string. For example, given a date-time text
950 * {@code "1996.07.10 AD at 15:08:56 PDT"}, if the given {@code fieldPosition}
951 * is {@link DateFormat#YEAR_FIELD}, the begin index and end index of
952 * {@code fieldPosition} will be set to 0 and 4, respectively.
953 * Notice that if the same date-time field appears more than once in a
954 * pattern, the {@code fieldPosition} will be set for the first occurrence
955 * of that date-time field. For instance, formatting a {@code Date} to the
956 * date-time string {@code "1 PM PDT (Pacific Daylight Time)"} using the
957 * pattern {@code "h a z (zzzz)"} and the alignment field
958 * {@link DateFormat#TIMEZONE_FIELD}, the begin index and end index of
959 * {@code fieldPosition} will be set to 5 and 8, respectively, for the
960 * first occurrence of the timezone pattern character {@code 'z'}.
961 * @return the formatted date-time string.
962 * @exception NullPointerException if any of the parameters is {@code null}.
963 */
964 @Override
965 public StringBuffer format(Date date, StringBuffer toAppendTo,
966 FieldPosition pos)
967 {
968 pos.beginIndex = pos.endIndex = 0;
969 return format(date, toAppendTo, pos.getFieldDelegate());
970 }
971
972 // Called from Format after creating a FieldDelegate
973 private StringBuffer format(Date date, StringBuffer toAppendTo,
974 FieldDelegate delegate) {
975 // Convert input date to time field list
976 calendar.setTime(date);
977
978 boolean useDateFormatSymbols = useDateFormatSymbols();
979
980 for (int i = 0; i < compiledPattern.length; ) {
981 int tag = compiledPattern[i] >>> 8;
982 int count = compiledPattern[i++] & 0xff;
983 if (count == 255) {
984 count = compiledPattern[i++] << 16;
985 count |= compiledPattern[i++];
986 }
987
988 switch (tag) {
989 case TAG_QUOTE_ASCII_CHAR:
990 toAppendTo.append((char)count);
991 break;
992
993 case TAG_QUOTE_CHARS:
994 toAppendTo.append(compiledPattern, i, count);
995 i += count;
996 break;
997
998 default:
999 subFormat(tag, count, delegate, toAppendTo, useDateFormatSymbols);
1000 break;
1001 }
1002 }
1003 return toAppendTo;
1004 }
1005
1006 /**
1007 * Formats an Object producing an <code>AttributedCharacterIterator</code>.
1008 * You can use the returned <code>AttributedCharacterIterator</code>
1009 * to build the resulting String, as well as to determine information
1010 * about the resulting String.
1011 * <p>
1012 * Each attribute key of the AttributedCharacterIterator will be of type
1013 * <code>DateFormat.Field</code>, with the corresponding attribute value
1014 * being the same as the attribute key.
1015 *
1016 * @exception NullPointerException if obj is null.
1017 * @exception IllegalArgumentException if the Format cannot format the
1018 * given object, or if the Format's pattern string is invalid.
1019 * @param obj The object to format
1020 * @return AttributedCharacterIterator describing the formatted value.
1021 * @since 1.4
1022 */
1023 @Override
1024 public AttributedCharacterIterator formatToCharacterIterator(Object obj) {
1025 StringBuffer sb = new StringBuffer();
1026 CharacterIteratorFieldDelegate delegate = new
1027 CharacterIteratorFieldDelegate();
1028
1029 if (obj instanceof Date) {
1030 format((Date)obj, sb, delegate);
1031 }
1032 else if (obj instanceof Number) {
1033 format(new Date(((Number)obj).longValue()), sb, delegate);
1034 }
1035 else if (obj == null) {
1036 throw new NullPointerException(
1037 "formatToCharacterIterator must be passed non-null object");
1038 }
1039 else {
1040 throw new IllegalArgumentException(
1041 "Cannot format given Object as a Date");
1042 }
1043 return delegate.getIterator(sb.toString());
1044 }
1045
1046 // Map index into pattern character string to Calendar field number
1047 private static final int[] PATTERN_INDEX_TO_CALENDAR_FIELD = {
1048 Calendar.ERA,
1049 Calendar.YEAR,
1050 Calendar.MONTH,
1051 Calendar.DATE,
1052 Calendar.HOUR_OF_DAY,
1053 Calendar.HOUR_OF_DAY,
1054 Calendar.MINUTE,
1055 Calendar.SECOND,
1056 Calendar.MILLISECOND,
1057 Calendar.DAY_OF_WEEK,
1058 Calendar.DAY_OF_YEAR,
1059 Calendar.DAY_OF_WEEK_IN_MONTH,
1060 Calendar.WEEK_OF_YEAR,
1061 Calendar.WEEK_OF_MONTH,
1062 Calendar.AM_PM,
1063 Calendar.HOUR,
1064 Calendar.HOUR,
1065 Calendar.ZONE_OFFSET,
1066 Calendar.ZONE_OFFSET,
1067 CalendarBuilder.WEEK_YEAR, // Pseudo Calendar field
1068 CalendarBuilder.ISO_DAY_OF_WEEK, // Pseudo Calendar field
1069 Calendar.ZONE_OFFSET,
1070 Calendar.MONTH
1071 };
1072
1073 // Map index into pattern character string to DateFormat field number
1074 private static final int[] PATTERN_INDEX_TO_DATE_FORMAT_FIELD = {
1075 DateFormat.ERA_FIELD,
1076 DateFormat.YEAR_FIELD,
1077 DateFormat.MONTH_FIELD,
1078 DateFormat.DATE_FIELD,
1079 DateFormat.HOUR_OF_DAY1_FIELD,
1080 DateFormat.HOUR_OF_DAY0_FIELD,
1081 DateFormat.MINUTE_FIELD,
1082 DateFormat.SECOND_FIELD,
1083 DateFormat.MILLISECOND_FIELD,
1084 DateFormat.DAY_OF_WEEK_FIELD,
1085 DateFormat.DAY_OF_YEAR_FIELD,
1086 DateFormat.DAY_OF_WEEK_IN_MONTH_FIELD,
1087 DateFormat.WEEK_OF_YEAR_FIELD,
1088 DateFormat.WEEK_OF_MONTH_FIELD,
1089 DateFormat.AM_PM_FIELD,
1090 DateFormat.HOUR1_FIELD,
1091 DateFormat.HOUR0_FIELD,
1092 DateFormat.TIMEZONE_FIELD,
1093 DateFormat.TIMEZONE_FIELD,
1094 DateFormat.YEAR_FIELD,
1095 DateFormat.DAY_OF_WEEK_FIELD,
1096 DateFormat.TIMEZONE_FIELD,
1097 DateFormat.MONTH_FIELD
1098 };
1099
1100 // Maps from DecimalFormatSymbols index to Field constant
1101 private static final Field[] PATTERN_INDEX_TO_DATE_FORMAT_FIELD_ID = {
1102 Field.ERA,
1103 Field.YEAR,
1104 Field.MONTH,
1105 Field.DAY_OF_MONTH,
1106 Field.HOUR_OF_DAY1,
1107 Field.HOUR_OF_DAY0,
1108 Field.MINUTE,
1109 Field.SECOND,
1110 Field.MILLISECOND,
1111 Field.DAY_OF_WEEK,
1112 Field.DAY_OF_YEAR,
1113 Field.DAY_OF_WEEK_IN_MONTH,
1114 Field.WEEK_OF_YEAR,
1115 Field.WEEK_OF_MONTH,
1116 Field.AM_PM,
1117 Field.HOUR1,
1118 Field.HOUR0,
1119 Field.TIME_ZONE,
1120 Field.TIME_ZONE,
1121 Field.YEAR,
1122 Field.DAY_OF_WEEK,
1123 Field.TIME_ZONE,
1124 Field.MONTH
1125 };
1126
1127 /**
1128 * Private member function that does the real date/time formatting.
1129 */
1130 private void subFormat(int patternCharIndex, int count,
1131 FieldDelegate delegate, StringBuffer buffer,
1132 boolean useDateFormatSymbols)
1133 {
1134 int maxIntCount = Integer.MAX_VALUE;
1135 String current = null;
1136 int beginOffset = buffer.length();
1137
1138 int field = PATTERN_INDEX_TO_CALENDAR_FIELD[patternCharIndex];
1139 int value;
1140 if (field == CalendarBuilder.WEEK_YEAR) {
1141 if (calendar.isWeekDateSupported()) {
1142 value = calendar.getWeekYear();
1143 } else {
1144 // use calendar year 'y' instead
1145 patternCharIndex = PATTERN_YEAR;
1146 field = PATTERN_INDEX_TO_CALENDAR_FIELD[patternCharIndex];
1147 value = calendar.get(field);
1148 }
1149 } else if (field == CalendarBuilder.ISO_DAY_OF_WEEK) {
1150 value = CalendarBuilder.toISODayOfWeek(calendar.get(Calendar.DAY_OF_WEEK));
1151 } else {
1152 value = calendar.get(field);
1153 }
1154
1155 int style = (count >= 4) ? Calendar.LONG : Calendar.SHORT;
1156 if (!useDateFormatSymbols && field < Calendar.ZONE_OFFSET
1157 && patternCharIndex != PATTERN_MONTH_STANDALONE) {
1158 current = calendar.getDisplayName(field, style, locale);
1159 }
1160
1161 // Note: zeroPaddingNumber() assumes that maxDigits is either
1162 // 2 or maxIntCount. If we make any changes to this,
1163 // zeroPaddingNumber() must be fixed.
1164
1165 switch (patternCharIndex) {
1166 case PATTERN_ERA: // 'G'
1167 if (useDateFormatSymbols) {
1168 String[] eras = formatData.getEras();
1169 if (value < eras.length) {
1170 current = eras[value];
1171 }
1172 }
1173 if (current == null) {
1174 current = "";
1175 }
1176 break;
1177
1178 case PATTERN_WEEK_YEAR: // 'Y'
1179 case PATTERN_YEAR: // 'y'
1180 if (calendar instanceof GregorianCalendar) {
1181 if (count != 2) {
1182 zeroPaddingNumber(value, count, maxIntCount, buffer);
1183 } else {
1184 zeroPaddingNumber(value, 2, 2, buffer);
1185 } // clip 1996 to 96
1186 } else {
1187 if (current == null) {
1188 zeroPaddingNumber(value, style == Calendar.LONG ? 1 : count,
1189 maxIntCount, buffer);
1190 }
1191 }
1192 break;
1193
1194 case PATTERN_MONTH: // 'M' (context sensitive)
1195 if (useDateFormatSymbols) {
1196 String[] months;
1197 if (count >= 4) {
1198 months = formatData.getMonths();
1199 current = months[value];
1200 } else if (count == 3) {
1201 months = formatData.getShortMonths();
1202 current = months[value];
1203 }
1204 } else {
1205 if (count < 3) {
1206 current = null;
1207 } else if (forceStandaloneForm) {
1208 current = calendar.getDisplayName(field, style | STANDALONE_MASK, locale);
1209 if (current == null) {
1210 current = calendar.getDisplayName(field, style, locale);
1211 }
1212 }
1213 }
1214 if (current == null) {
1215 zeroPaddingNumber(value+1, count, maxIntCount, buffer);
1216 }
1217 break;
1218
1219 case PATTERN_MONTH_STANDALONE: // 'L'
1220 assert current == null;
1221 if (locale == null) {
1222 String[] months;
1223 if (count >= 4) {
1224 months = formatData.getMonths();
1225 current = months[value];
1226 } else if (count == 3) {
1227 months = formatData.getShortMonths();
1228 current = months[value];
1229 }
1230 } else {
1231 if (count >= 3) {
1232 current = calendar.getDisplayName(field, style | STANDALONE_MASK, locale);
1233 }
1234 }
1235 if (current == null) {
1236 zeroPaddingNumber(value+1, count, maxIntCount, buffer);
1237 }
1238 break;
1239
1240 case PATTERN_HOUR_OF_DAY1: // 'k' 1-based. eg, 23:59 + 1 hour =>> 24:59
1241 if (current == null) {
1242 if (value == 0) {
1243 zeroPaddingNumber(calendar.getMaximum(Calendar.HOUR_OF_DAY) + 1,
1244 count, maxIntCount, buffer);
1245 } else {
1246 zeroPaddingNumber(value, count, maxIntCount, buffer);
1247 }
1248 }
1249 break;
1250
1251 case PATTERN_DAY_OF_WEEK: // 'E'
1252 if (useDateFormatSymbols) {
1253 String[] weekdays;
1254 if (count >= 4) {
1255 weekdays = formatData.getWeekdays();
1256 current = weekdays[value];
1257 } else { // count < 4, use abbreviated form if exists
1258 weekdays = formatData.getShortWeekdays();
1259 current = weekdays[value];
1260 }
1261 }
1262 break;
1263
1264 case PATTERN_AM_PM: // 'a'
1265 if (useDateFormatSymbols) {
1266 String[] ampm = formatData.getAmPmStrings();
1267 current = ampm[value];
1268 }
1269 break;
1270
1271 case PATTERN_HOUR1: // 'h' 1-based. eg, 11PM + 1 hour =>> 12 AM
1272 if (current == null) {
1273 if (value == 0) {
1274 zeroPaddingNumber(calendar.getLeastMaximum(Calendar.HOUR) + 1,
1275 count, maxIntCount, buffer);
1276 } else {
1277 zeroPaddingNumber(value, count, maxIntCount, buffer);
1278 }
1279 }
1280 break;
1281
1282 case PATTERN_ZONE_NAME: // 'z'
1283 if (current == null) {
1284 if (formatData.locale == null || formatData.isZoneStringsSet) {
1285 int zoneIndex =
1286 formatData.getZoneIndex(calendar.getTimeZone().getID());
1287 if (zoneIndex == -1) {
1288 value = calendar.get(Calendar.ZONE_OFFSET) +
1289 calendar.get(Calendar.DST_OFFSET);
1290 buffer.append(ZoneInfoFile.toCustomID(value));
1291 } else {
1292 int index = (calendar.get(Calendar.DST_OFFSET) == 0) ? 1: 3;
1293 if (count < 4) {
1294 // Use the short name
1295 index++;
1296 }
1297 String[][] zoneStrings = formatData.getZoneStringsWrapper();
1298 buffer.append(zoneStrings[zoneIndex][index]);
1299 }
1300 } else {
1301 TimeZone tz = calendar.getTimeZone();
1302 boolean daylight = (calendar.get(Calendar.DST_OFFSET) != 0);
1303 int tzstyle = (count < 4 ? TimeZone.SHORT : TimeZone.LONG);
1304 buffer.append(tz.getDisplayName(daylight, tzstyle, formatData.locale));
1305 }
1306 }
1307 break;
1308
1309 case PATTERN_ZONE_VALUE: // 'Z' ("-/+hhmm" form)
1310 value = (calendar.get(Calendar.ZONE_OFFSET) +
1311 calendar.get(Calendar.DST_OFFSET)) / 60000;
1312
1313 int width = 4;
1314 if (value >= 0) {
1315 buffer.append('+');
1316 } else {
1317 width++;
1318 }
1319
1320 int num = (value / 60) * 100 + (value % 60);
1321 CalendarUtils.sprintf0d(buffer, num, width);
1322 break;
1323
1324 case PATTERN_ISO_ZONE: // 'X'
1325 value = calendar.get(Calendar.ZONE_OFFSET)
1326 + calendar.get(Calendar.DST_OFFSET);
1327
1328 if (value == 0) {
1329 buffer.append('Z');
1330 break;
1331 }
1332
1333 value /= 60000;
1334 if (value >= 0) {
1335 buffer.append('+');
1336 } else {
1337 buffer.append('-');
1338 value = -value;
1339 }
1340
1341 CalendarUtils.sprintf0d(buffer, value / 60, 2);
1342 if (count == 1) {
1343 break;
1344 }
1345
1346 if (count == 3) {
1347 buffer.append(':');
1348 }
1349 CalendarUtils.sprintf0d(buffer, value % 60, 2);
1350 break;
1351
1352 default:
1353 // case PATTERN_DAY_OF_MONTH: // 'd'
1354 // case PATTERN_HOUR_OF_DAY0: // 'H' 0-based. eg, 23:59 + 1 hour =>> 00:59
1355 // case PATTERN_MINUTE: // 'm'
1356 // case PATTERN_SECOND: // 's'
1357 // case PATTERN_MILLISECOND: // 'S'
1358 // case PATTERN_DAY_OF_YEAR: // 'D'
1359 // case PATTERN_DAY_OF_WEEK_IN_MONTH: // 'F'
1360 // case PATTERN_WEEK_OF_YEAR: // 'w'
1361 // case PATTERN_WEEK_OF_MONTH: // 'W'
1362 // case PATTERN_HOUR0: // 'K' eg, 11PM + 1 hour =>> 0 AM
1363 // case PATTERN_ISO_DAY_OF_WEEK: // 'u' pseudo field, Monday = 1, ..., Sunday = 7
1364 if (current == null) {
1365 zeroPaddingNumber(value, count, maxIntCount, buffer);
1366 }
1367 break;
1368 } // switch (patternCharIndex)
1369
1370 if (current != null) {
1371 buffer.append(current);
1372 }
1373
1374 int fieldID = PATTERN_INDEX_TO_DATE_FORMAT_FIELD[patternCharIndex];
1375 Field f = PATTERN_INDEX_TO_DATE_FORMAT_FIELD_ID[patternCharIndex];
1376
1377 delegate.formatted(fieldID, f, f, beginOffset, buffer.length(), buffer);
1378 }
1379
1380 /**
1381 * Formats a number with the specified minimum and maximum number of digits.
1382 */
1383 private void zeroPaddingNumber(int value, int minDigits, int maxDigits, StringBuffer buffer)
1384 {
1385 // Optimization for 1, 2 and 4 digit numbers. This should
1386 // cover most cases of formatting date/time related items.
1387 // Note: This optimization code assumes that maxDigits is
1388 // either 2 or Integer.MAX_VALUE (maxIntCount in format()).
1389 try {
1390 if (zeroDigit == 0) {
1391 zeroDigit = ((DecimalFormat)numberFormat).getDecimalFormatSymbols().getZeroDigit();
1392 }
1393 if (value >= 0) {
1394 if (value < 100 && minDigits >= 1 && minDigits <= 2) {
1395 if (value < 10) {
1396 if (minDigits == 2) {
1397 buffer.append(zeroDigit);
1398 }
1399 buffer.append((char)(zeroDigit + value));
1400 } else {
1401 buffer.append((char)(zeroDigit + value / 10));
1402 buffer.append((char)(zeroDigit + value % 10));
1403 }
1404 return;
1405 } else if (value >= 1000 && value < 10000) {
1406 if (minDigits == 4) {
1407 buffer.append((char)(zeroDigit + value / 1000));
1408 value %= 1000;
1409 buffer.append((char)(zeroDigit + value / 100));
1410 value %= 100;
1411 buffer.append((char)(zeroDigit + value / 10));
1412 buffer.append((char)(zeroDigit + value % 10));
1413 return;
1414 }
1415 if (minDigits == 2 && maxDigits == 2) {
1416 zeroPaddingNumber(value % 100, 2, 2, buffer);
1417 return;
1418 }
1419 }
1420 }
1421 } catch (Exception e) {
1422 }
1423
1424 numberFormat.setMinimumIntegerDigits(minDigits);
1425 numberFormat.setMaximumIntegerDigits(maxDigits);
1426 numberFormat.format((long)value, buffer, DontCareFieldPosition.INSTANCE);
1427 }
1428
1429
1430 /**
1431 * Parses text from a string to produce a <code>Date</code>.
1432 * <p>
1433 * The method attempts to parse text starting at the index given by
1434 * <code>pos</code>.
1435 * If parsing succeeds, then the index of <code>pos</code> is updated
1436 * to the index after the last character used (parsing does not necessarily
1437 * use all characters up to the end of the string), and the parsed
1438 * date is returned. The updated <code>pos</code> can be used to
1439 * indicate the starting point for the next call to this method.
1440 * If an error occurs, then the index of <code>pos</code> is not
1441 * changed, the error index of <code>pos</code> is set to the index of
1442 * the character where the error occurred, and null is returned.
1443 *
1444 * <p>This parsing operation uses the {@link DateFormat#calendar
1445 * calendar} to produce a {@code Date}. All of the {@code
1446 * calendar}'s date-time fields are {@linkplain Calendar#clear()
1447 * cleared} before parsing, and the {@code calendar}'s default
1448 * values of the date-time fields are used for any missing
1449 * date-time information. For example, the year value of the
1450 * parsed {@code Date} is 1970 with {@link GregorianCalendar} if
1451 * no year value is given from the parsing operation. The {@code
1452 * TimeZone} value may be overwritten, depending on the given
1453 * pattern and the time zone value in {@code text}. Any {@code
1454 * TimeZone} value that has previously been set by a call to
1455 * {@link #setTimeZone(java.util.TimeZone) setTimeZone} may need
1456 * to be restored for further operations.
1457 *
1458 * @param text A <code>String</code>, part of which should be parsed.
1459 * @param pos A <code>ParsePosition</code> object with index and error
1460 * index information as described above.
1461 * @return A <code>Date</code> parsed from the string. In case of
1462 * error, returns null.
1463 * @exception NullPointerException if <code>text</code> or <code>pos</code> is null.
1464 */
1465 @Override
1466 public Date parse(String text, ParsePosition pos)
1467 {
1468 checkNegativeNumberExpression();
1469
1470 int start = pos.index;
1471 int oldStart = start;
1472 int textLength = text.length();
1473
1474 boolean[] ambiguousYear = {false};
1475
1476 CalendarBuilder calb = new CalendarBuilder();
1477
1478 for (int i = 0; i < compiledPattern.length; ) {
1479 int tag = compiledPattern[i] >>> 8;
1480 int count = compiledPattern[i++] & 0xff;
1481 if (count == 255) {
1482 count = compiledPattern[i++] << 16;
1483 count |= compiledPattern[i++];
1484 }
1485
1486 switch (tag) {
1487 case TAG_QUOTE_ASCII_CHAR:
1488 if (start >= textLength || text.charAt(start) != (char)count) {
1489 pos.index = oldStart;
1490 pos.errorIndex = start;
1491 return null;
1492 }
1493 start++;
1494 break;
1495
1496 case TAG_QUOTE_CHARS:
1497 while (count-- > 0) {
1498 if (start >= textLength || text.charAt(start) != compiledPattern[i++]) {
1499 pos.index = oldStart;
1500 pos.errorIndex = start;
1501 return null;
1502 }
1503 start++;
1504 }
1505 break;
1506
1507 default:
1508 // Peek the next pattern to determine if we need to
1509 // obey the number of pattern letters for
1510 // parsing. It's required when parsing contiguous
1511 // digit text (e.g., "20010704") with a pattern which
1512 // has no delimiters between fields, like "yyyyMMdd".
1513 boolean obeyCount = false;
1514
1515 // In Arabic, a minus sign for a negative number is put after
1516 // the number. Even in another locale, a minus sign can be
1517 // put after a number using DateFormat.setNumberFormat().
1518 // If both the minus sign and the field-delimiter are '-',
1519 // subParse() needs to determine whether a '-' after a number
1520 // in the given text is a delimiter or is a minus sign for the
1521 // preceding number. We give subParse() a clue based on the
1522 // information in compiledPattern.
1523 boolean useFollowingMinusSignAsDelimiter = false;
1524
1525 if (i < compiledPattern.length) {
1526 int nextTag = compiledPattern[i] >>> 8;
1527 int nextCount = compiledPattern[i] & 0xff;
1528 obeyCount = shouldObeyCount(nextTag, nextCount);
1529
1530 if (hasFollowingMinusSign &&
1531 (nextTag == TAG_QUOTE_ASCII_CHAR ||
1532 nextTag == TAG_QUOTE_CHARS)) {
1533
1534 if (nextTag != TAG_QUOTE_ASCII_CHAR) {
1535 nextCount = compiledPattern[i+1];
1536 }
1537
1538 if (nextCount == minusSign) {
1539 useFollowingMinusSignAsDelimiter = true;
1540 }
1541 }
1542 }
1543 start = subParse(text, start, tag, count, obeyCount,
1544 ambiguousYear, pos,
1545 useFollowingMinusSignAsDelimiter, calb);
1546 if (start < 0) {
1547 pos.index = oldStart;
1548 return null;
1549 }
1550 }
1551 }
1552
1553 // At this point the fields of Calendar have been set. Calendar
1554 // will fill in default values for missing fields when the time
1555 // is computed.
1556
1557 pos.index = start;
1558
1559 Date parsedDate;
1560 try {
1561 parsedDate = calb.establish(calendar).getTime();
1562 // If the year value is ambiguous,
1563 // then the two-digit year == the default start year
1564 if (ambiguousYear[0]) {
1565 if (parsedDate.before(defaultCenturyStart)) {
1566 parsedDate = calb.addYear(100).establish(calendar).getTime();
1567 }
1568 }
1569 }
1570 // An IllegalArgumentException will be thrown by Calendar.getTime()
1571 // if any fields are out of range, e.g., MONTH == 17.
1572 catch (IllegalArgumentException e) {
1573 pos.errorIndex = start;
1574 pos.index = oldStart;
1575 return null;
1576 }
1577
1578 return parsedDate;
1579 }
1580
1581 /* If the next tag/pattern is a <Numeric_Field> then the parser
1582 * should consider the count of digits while parsing the contigous digits
1583 * for the current tag/pattern
1584 */
1585 private boolean shouldObeyCount(int tag, int count) {
1586 switch (tag) {
1587 case PATTERN_MONTH:
1588 case PATTERN_MONTH_STANDALONE:
1589 return count <= 2;
1590 case PATTERN_YEAR:
1591 case PATTERN_DAY_OF_MONTH:
1592 case PATTERN_HOUR_OF_DAY1:
1593 case PATTERN_HOUR_OF_DAY0:
1594 case PATTERN_MINUTE:
1595 case PATTERN_SECOND:
1596 case PATTERN_MILLISECOND:
1597 case PATTERN_DAY_OF_YEAR:
1598 case PATTERN_DAY_OF_WEEK_IN_MONTH:
1599 case PATTERN_WEEK_OF_YEAR:
1600 case PATTERN_WEEK_OF_MONTH:
1601 case PATTERN_HOUR1:
1602 case PATTERN_HOUR0:
1603 case PATTERN_WEEK_YEAR:
1604 case PATTERN_ISO_DAY_OF_WEEK:
1605 return true;
1606 default:
1607 return false;
1608 }
1609 }
1610
1611 /**
1612 * Private code-size reduction function used by subParse.
1613 * @param text the time text being parsed.
1614 * @param start where to start parsing.
1615 * @param field the date field being parsed.
1616 * @param data the string array to parsed.
1617 * @return the new start position if matching succeeded; a negative number
1618 * indicating matching failure, otherwise.
1619 */
1620 private int matchString(String text, int start, int field, String[] data, CalendarBuilder calb)
1621 {
1622 int i = 0;
1623 int count = data.length;
1624
1625 if (field == Calendar.DAY_OF_WEEK) {
1626 i = 1;
1627 }
1628
1629 // There may be multiple strings in the data[] array which begin with
1630 // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech).
1631 // We keep track of the longest match, and return that. Note that this
1632 // unfortunately requires us to test all array elements.
1633 int bestMatchLength = 0, bestMatch = -1;
1634 for (; i<count; ++i)
1635 {
1636 int length = data[i].length();
1637 // Always compare if we have no match yet; otherwise only compare
1638 // against potentially better matches (longer strings).
1639 if (length > bestMatchLength &&
1640 text.regionMatches(true, start, data[i], 0, length))
1641 {
1642 bestMatch = i;
1643 bestMatchLength = length;
1644 }
1645 }
1646 if (bestMatch >= 0)
1647 {
1648 calb.set(field, bestMatch);
1649 return start + bestMatchLength;
1650 }
1651 return -start;
1652 }
1653
1654 /**
1655 * Performs the same thing as matchString(String, int, int,
1656 * String[]). This method takes a Map<String, Integer> instead of
1657 * String[].
1658 */
1659 private int matchString(String text, int start, int field,
1660 Map<String,Integer> data, CalendarBuilder calb) {
1661 if (data != null) {
1662 // TODO: make this default when it's in the spec.
1663 if (data instanceof SortedMap) {
1664 for (String name : data.keySet()) {
1665 if (text.regionMatches(true, start, name, 0, name.length())) {
1666 calb.set(field, data.get(name));
1667 return start + name.length();
1668 }
1669 }
1670 return -start;
1671 }
1672
1673 String bestMatch = null;
1674
1675 for (String name : data.keySet()) {
1676 int length = name.length();
1677 if (bestMatch == null || length > bestMatch.length()) {
1678 if (text.regionMatches(true, start, name, 0, length)) {
1679 bestMatch = name;
1680 }
1681 }
1682 }
1683
1684 if (bestMatch != null) {
1685 calb.set(field, data.get(bestMatch));
1686 return start + bestMatch.length();
1687 }
1688 }
1689 return -start;
1690 }
1691
1692 private int matchZoneString(String text, int start, String[] zoneNames) {
1693 for (int i = 1; i <= 4; ++i) {
1694 // Checking long and short zones [1 & 2],
1695 // and long and short daylight [3 & 4].
1696 String zoneName = zoneNames[i];
1697 if (zoneName.isEmpty()) {
1698 // fill in by retrieving single name
1699 zoneName = TimeZoneNameUtility.retrieveDisplayName(
1700 zoneNames[0], i >= 3, i % 2, locale);
1701 zoneNames[i] = zoneName;
1702 }
1703 if (text.regionMatches(true, start,
1704 zoneName, 0, zoneName.length())) {
1705 return i;
1706 }
1707 }
1708 return -1;
1709 }
1710
1711 private boolean matchDSTString(String text, int start, int zoneIndex, int standardIndex,
1712 String[][] zoneStrings) {
1713 int index = standardIndex + 2;
1714 String zoneName = zoneStrings[zoneIndex][index];
1715 if (text.regionMatches(true, start,
1716 zoneName, 0, zoneName.length())) {
1717 return true;
1718 }
1719 return false;
1720 }
1721
1722 /**
1723 * find time zone 'text' matched zoneStrings and set to internal
1724 * calendar.
1725 */
1726 private int subParseZoneString(String text, int start, CalendarBuilder calb) {
1727 boolean useSameName = false; // true if standard and daylight time use the same abbreviation.
1728 TimeZone currentTimeZone = getTimeZone();
1729
1730 // At this point, check for named time zones by looking through
1731 // the locale data from the TimeZoneNames strings.
1732 // Want to be able to parse both short and long forms.
1733 int zoneIndex = formatData.getZoneIndex(currentTimeZone.getID());
1734 TimeZone tz = null;
1735 String[][] zoneStrings = formatData.getZoneStringsWrapper();
1736 String[] zoneNames = null;
1737 int nameIndex = 0;
1738 if (zoneIndex != -1) {
1739 zoneNames = zoneStrings[zoneIndex];
1740 if ((nameIndex = matchZoneString(text, start, zoneNames)) > 0) {
1741 if (nameIndex <= 2) {
1742 // Check if the standard name (abbr) and the daylight name are the same.
1743 useSameName = zoneNames[nameIndex].equalsIgnoreCase(zoneNames[nameIndex + 2]);
1744 }
1745 tz = TimeZone.getTimeZone(zoneNames[0]);
1746 }
1747 }
1748 if (tz == null) {
1749 zoneIndex = formatData.getZoneIndex(TimeZone.getDefault().getID());
1750 if (zoneIndex != -1) {
1751 zoneNames = zoneStrings[zoneIndex];
1752 if ((nameIndex = matchZoneString(text, start, zoneNames)) > 0) {
1753 if (nameIndex <= 2) {
1754 useSameName = zoneNames[nameIndex].equalsIgnoreCase(zoneNames[nameIndex + 2]);
1755 }
1756 tz = TimeZone.getTimeZone(zoneNames[0]);
1757 }
1758 }
1759 }
1760
1761 if (tz == null) {
1762 int len = zoneStrings.length;
1763 for (int i = 0; i < len; i++) {
1764 zoneNames = zoneStrings[i];
1765 if ((nameIndex = matchZoneString(text, start, zoneNames)) > 0) {
1766 if (nameIndex <= 2) {
1767 useSameName = zoneNames[nameIndex].equalsIgnoreCase(zoneNames[nameIndex + 2]);
1768 }
1769 tz = TimeZone.getTimeZone(zoneNames[0]);
1770 break;
1771 }
1772 }
1773 }
1774 if (tz != null) { // Matched any ?
1775 if (!tz.equals(currentTimeZone)) {
1776 setTimeZone(tz);
1777 }
1778 // If the time zone matched uses the same name
1779 // (abbreviation) for both standard and daylight time,
1780 // let the time zone in the Calendar decide which one.
1781 //
1782 // Also if tz.getDSTSaving() returns 0 for DST, use tz to
1783 // determine the local time. (6645292)
1784 int dstAmount = (nameIndex >= 3) ? tz.getDSTSavings() : 0;
1785 if (!(useSameName || (nameIndex >= 3 && dstAmount == 0))) {
1786 calb.clear(Calendar.ZONE_OFFSET).set(Calendar.DST_OFFSET, dstAmount);
1787 }
1788 return (start + zoneNames[nameIndex].length());
1789 }
1790 return -start;
1791 }
1792
1793 /**
1794 * Parses numeric forms of time zone offset, such as "hh:mm", and
1795 * sets calb to the parsed value.
1796 *
1797 * @param text the text to be parsed
1798 * @param start the character position to start parsing
1799 * @param sign 1: positive; -1: negative
1800 * @param count 0: 'Z' or "GMT+hh:mm" parsing; 1 - 3: the number of 'X's
1801 * @param colon true - colon required between hh and mm; false - no colon required
1802 * @param calb a CalendarBuilder in which the parsed value is stored
1803 * @return updated parsed position, or its negative value to indicate a parsing error
1804 */
1805 private int subParseNumericZone(String text, int start, int sign, int count,
1806 boolean colon, CalendarBuilder calb) {
1807 int index = start;
1808
1809 parse:
1810 try {
1811 char c = text.charAt(index++);
1812 // Parse hh
1813 int hours;
1814 if (!isDigit(c)) {
1815 break parse;
1816 }
1817 hours = c - '0';
1818 c = text.charAt(index++);
1819 if (isDigit(c)) {
1820 hours = hours * 10 + (c - '0');
1821 } else {
1822 // If no colon in RFC 822 or 'X' (ISO), two digits are
1823 // required.
1824 if (count > 0 || !colon) {
1825 break parse;
1826 }
1827 --index;
1828 }
1829 if (hours > 23) {
1830 break parse;
1831 }
1832 int minutes = 0;
1833 if (count != 1) {
1834 // Proceed with parsing mm
1835 c = text.charAt(index++);
1836 if (colon) {
1837 if (c != ':') {
1838 break parse;
1839 }
1840 c = text.charAt(index++);
1841 }
1842 if (!isDigit(c)) {
1843 break parse;
1844 }
1845 minutes = c - '0';
1846 c = text.charAt(index++);
1847 if (!isDigit(c)) {
1848 break parse;
1849 }
1850 minutes = minutes * 10 + (c - '0');
1851 if (minutes > 59) {
1852 break parse;
1853 }
1854 }
1855 minutes += hours * 60;
1856 calb.set(Calendar.ZONE_OFFSET, minutes * MILLIS_PER_MINUTE * sign)
1857 .set(Calendar.DST_OFFSET, 0);
1858 return index;
1859 } catch (IndexOutOfBoundsException e) {
1860 }
1861 return 1 - index; // -(index - 1)
1862 }
1863
1864 private boolean isDigit(char c) {
1865 return c >= '0' && c <= '9';
1866 }
1867
1868 /**
1869 * Private member function that converts the parsed date strings into
1870 * timeFields. Returns -start (for ParsePosition) if failed.
1871 * @param text the time text to be parsed.
1872 * @param start where to start parsing.
1873 * @param patternCharIndex the index of the pattern character.
1874 * @param count the count of a pattern character.
1875 * @param obeyCount if true, then the next field directly abuts this one,
1876 * and we should use the count to know when to stop parsing.
1877 * @param ambiguousYear return parameter; upon return, if ambiguousYear[0]
1878 * is true, then a two-digit year was parsed and may need to be readjusted.
1879 * @param origPos origPos.errorIndex is used to return an error index
1880 * at which a parse error occurred, if matching failure occurs.
1881 * @return the new start position if matching succeeded; -1 indicating
1882 * matching failure, otherwise. In case matching failure occurred,
1883 * an error index is set to origPos.errorIndex.
1884 */
1885 private int subParse(String text, int start, int patternCharIndex, int count,
1886 boolean obeyCount, boolean[] ambiguousYear,
1887 ParsePosition origPos,
1888 boolean useFollowingMinusSignAsDelimiter, CalendarBuilder calb) {
1889 Number number;
1890 int value = 0;
1891 ParsePosition pos = new ParsePosition(0);
1892 pos.index = start;
1893 if (patternCharIndex == PATTERN_WEEK_YEAR && !calendar.isWeekDateSupported()) {
1894 // use calendar year 'y' instead
1895 patternCharIndex = PATTERN_YEAR;
1896 }
1897 int field = PATTERN_INDEX_TO_CALENDAR_FIELD[patternCharIndex];
1898
1899 // If there are any spaces here, skip over them. If we hit the end
1900 // of the string, then fail.
1901 for (;;) {
1902 if (pos.index >= text.length()) {
1903 origPos.errorIndex = start;
1904 return -1;
1905 }
1906 char c = text.charAt(pos.index);
1907 if (c != ' ' && c != '\t') {
1908 break;
1909 }
1910 ++pos.index;
1911 }
1912 // Remember the actual start index
1913 int actualStart = pos.index;
1914
1915 parsing:
1916 {
1917 // We handle a few special cases here where we need to parse
1918 // a number value. We handle further, more generic cases below. We need
1919 // to handle some of them here because some fields require extra processing on
1920 // the parsed value.
1921 if (patternCharIndex == PATTERN_HOUR_OF_DAY1 ||
1922 patternCharIndex == PATTERN_HOUR1 ||
1923 (patternCharIndex == PATTERN_MONTH && count <= 2) ||
1924 (patternCharIndex == PATTERN_MONTH_STANDALONE && count <= 2) ||
1925 patternCharIndex == PATTERN_YEAR ||
1926 patternCharIndex == PATTERN_WEEK_YEAR) {
1927 // It would be good to unify this with the obeyCount logic below,
1928 // but that's going to be difficult.
1929 if (obeyCount) {
1930 if ((start+count) > text.length()) {
1931 break parsing;
1932 }
1933 number = numberFormat.parse(text.substring(0, start+count), pos);
1934 } else {
1935 number = numberFormat.parse(text, pos);
1936 }
1937 if (number == null) {
1938 if (patternCharIndex != PATTERN_YEAR || calendar instanceof GregorianCalendar) {
1939 break parsing;
1940 }
1941 } else {
1942 value = number.intValue();
1943
1944 if (useFollowingMinusSignAsDelimiter && (value < 0) &&
1945 (((pos.index < text.length()) &&
1946 (text.charAt(pos.index) != minusSign)) ||
1947 ((pos.index == text.length()) &&
1948 (text.charAt(pos.index-1) == minusSign)))) {
1949 value = -value;
1950 pos.index--;
1951 }
1952 }
1953 }
1954
1955 boolean useDateFormatSymbols = useDateFormatSymbols();
1956
1957 int index;
1958 switch (patternCharIndex) {
1959 case PATTERN_ERA: // 'G'
1960 if (useDateFormatSymbols) {
1961 if ((index = matchString(text, start, Calendar.ERA, formatData.getEras(), calb)) > 0) {
1962 return index;
1963 }
1964 } else {
1965 Map<String, Integer> map = getDisplayNamesMap(field, locale);
1966 if ((index = matchString(text, start, field, map, calb)) > 0) {
1967 return index;
1968 }
1969 }
1970 break parsing;
1971
1972 case PATTERN_WEEK_YEAR: // 'Y'
1973 case PATTERN_YEAR: // 'y'
1974 if (!(calendar instanceof GregorianCalendar)) {
1975 // calendar might have text representations for year values,
1976 // such as "\u5143" in JapaneseImperialCalendar.
1977 int style = (count >= 4) ? Calendar.LONG : Calendar.SHORT;
1978 Map<String, Integer> map = calendar.getDisplayNames(field, style, locale);
1979 if (map != null) {
1980 if ((index = matchString(text, start, field, map, calb)) > 0) {
1981 return index;
1982 }
1983 }
1984 calb.set(field, value);
1985 return pos.index;
1986 }
1987
1988 // If there are 3 or more YEAR pattern characters, this indicates
1989 // that the year value is to be treated literally, without any
1990 // two-digit year adjustments (e.g., from "01" to 2001). Otherwise
1991 // we made adjustments to place the 2-digit year in the proper
1992 // century, for parsed strings from "00" to "99". Any other string
1993 // is treated literally: "2250", "-1", "1", "002".
1994 if (count <= 2 && (pos.index - actualStart) == 2
1995 && Character.isDigit(text.charAt(actualStart))
1996 && Character.isDigit(text.charAt(actualStart + 1))) {
1997 // Assume for example that the defaultCenturyStart is 6/18/1903.
1998 // This means that two-digit years will be forced into the range
1999 // 6/18/1903 to 6/17/2003. As a result, years 00, 01, and 02
2000 // correspond to 2000, 2001, and 2002. Years 04, 05, etc. correspond
2001 // to 1904, 1905, etc. If the year is 03, then it is 2003 if the
2002 // other fields specify a date before 6/18, or 1903 if they specify a
2003 // date afterwards. As a result, 03 is an ambiguous year. All other
2004 // two-digit years are unambiguous.
2005 int ambiguousTwoDigitYear = defaultCenturyStartYear % 100;
2006 ambiguousYear[0] = value == ambiguousTwoDigitYear;
2007 value += (defaultCenturyStartYear/100)*100 +
2008 (value < ambiguousTwoDigitYear ? 100 : 0);
2009 }
2010 calb.set(field, value);
2011 return pos.index;
2012
2013 case PATTERN_MONTH: // 'M'
2014 if (count <= 2) // i.e., M or MM.
2015 {
2016 // Don't want to parse the month if it is a string
2017 // while pattern uses numeric style: M or MM.
2018 // [We computed 'value' above.]
2019 calb.set(Calendar.MONTH, value - 1);
2020 return pos.index;
2021 }
2022
2023 if (useDateFormatSymbols) {
2024 // count >= 3 // i.e., MMM or MMMM
2025 // Want to be able to parse both short and long forms.
2026 // Try count == 4 first:
2027 int newStart;
2028 if ((newStart = matchString(text, start, Calendar.MONTH,
2029 formatData.getMonths(), calb)) > 0) {
2030 return newStart;
2031 }
2032 // count == 4 failed, now try count == 3
2033 if ((index = matchString(text, start, Calendar.MONTH,
2034 formatData.getShortMonths(), calb)) > 0) {
2035 return index;
2036 }
2037 } else {
2038 Map<String, Integer> map = getDisplayContextNamesMap(field, locale);
2039 if ((index = matchString(text, start, field, map, calb)) > 0) {
2040 return index;
2041 }
2042 }
2043 break parsing;
2044
2045 case PATTERN_MONTH_STANDALONE: // 'L'
2046 if (count <= 2) {
2047 // Don't want to parse the month if it is a string
2048 // while pattern uses numeric style: L or LL
2049 //[we computed 'value' above.]
2050 calb.set(Calendar.MONTH, value - 1);
2051 return pos.index;
2052 }
2053 Map<String, Integer> maps = getDisplayNamesMap(field, locale);
2054 if ((index = matchString(text, start, field, maps, calb)) > 0) {
2055 return index;
2056 }
2057 break parsing;
2058
2059 case PATTERN_HOUR_OF_DAY1: // 'k' 1-based. eg, 23:59 + 1 hour =>> 24:59
2060 if (!isLenient()) {
2061 // Validate the hour value in non-lenient
2062 if (value < 1 || value > 24) {
2063 break parsing;
2064 }
2065 }
2066 // [We computed 'value' above.]
2067 if (value == calendar.getMaximum(Calendar.HOUR_OF_DAY) + 1) {
2068 value = 0;
2069 }
2070 calb.set(Calendar.HOUR_OF_DAY, value);
2071 return pos.index;
2072
2073 case PATTERN_DAY_OF_WEEK: // 'E'
2074 {
2075 if (useDateFormatSymbols) {
2076 // Want to be able to parse both short and long forms.
2077 // Try count == 4 (DDDD) first:
2078 int newStart;
2079 if ((newStart=matchString(text, start, Calendar.DAY_OF_WEEK,
2080 formatData.getWeekdays(), calb)) > 0) {
2081 return newStart;
2082 }
2083 // DDDD failed, now try DDD
2084 if ((index = matchString(text, start, Calendar.DAY_OF_WEEK,
2085 formatData.getShortWeekdays(), calb)) > 0) {
2086 return index;
2087 }
2088 } else {
2089 int[] styles = { Calendar.LONG, Calendar.SHORT };
2090 for (int style : styles) {
2091 Map<String,Integer> map = calendar.getDisplayNames(field, style, locale);
2092 if ((index = matchString(text, start, field, map, calb)) > 0) {
2093 return index;
2094 }
2095 }
2096 }
2097 }
2098 break parsing;
2099
2100 case PATTERN_AM_PM: // 'a'
2101 if (useDateFormatSymbols) {
2102 if ((index = matchString(text, start, Calendar.AM_PM,
2103 formatData.getAmPmStrings(), calb)) > 0) {
2104 return index;
2105 }
2106 } else {
2107 Map<String,Integer> map = getDisplayNamesMap(field, locale);
2108 if ((index = matchString(text, start, field, map, calb)) > 0) {
2109 return index;
2110 }
2111 }
2112 break parsing;
2113
2114 case PATTERN_HOUR1: // 'h' 1-based. eg, 11PM + 1 hour =>> 12 AM
2115 if (!isLenient()) {
2116 // Validate the hour value in non-lenient
2117 if (value < 1 || value > 12) {
2118 break parsing;
2119 }
2120 }
2121 // [We computed 'value' above.]
2122 if (value == calendar.getLeastMaximum(Calendar.HOUR) + 1) {
2123 value = 0;
2124 }
2125 calb.set(Calendar.HOUR, value);
2126 return pos.index;
2127
2128 case PATTERN_ZONE_NAME: // 'z'
2129 case PATTERN_ZONE_VALUE: // 'Z'
2130 {
2131 int sign = 0;
2132 try {
2133 char c = text.charAt(pos.index);
2134 if (c == '+') {
2135 sign = 1;
2136 } else if (c == '-') {
2137 sign = -1;
2138 }
2139 if (sign == 0) {
2140 // Try parsing a custom time zone "GMT+hh:mm" or "GMT".
2141 if ((c == 'G' || c == 'g')
2142 && (text.length() - start) >= GMT.length()
2143 && text.regionMatches(true, start, GMT, 0, GMT.length())) {
2144 pos.index = start + GMT.length();
2145
2146 if ((text.length() - pos.index) > 0) {
2147 c = text.charAt(pos.index);
2148 if (c == '+') {
2149 sign = 1;
2150 } else if (c == '-') {
2151 sign = -1;
2152 }
2153 }
2154
2155 if (sign == 0) { /* "GMT" without offset */
2156 calb.set(Calendar.ZONE_OFFSET, 0)
2157 .set(Calendar.DST_OFFSET, 0);
2158 return pos.index;
2159 }
2160
2161 // Parse the rest as "hh:mm"
2162 int i = subParseNumericZone(text, ++pos.index,
2163 sign, 0, true, calb);
2164 if (i > 0) {
2165 return i;
2166 }
2167 pos.index = -i;
2168 } else {
2169 // Try parsing the text as a time zone
2170 // name or abbreviation.
2171 int i = subParseZoneString(text, pos.index, calb);
2172 if (i > 0) {
2173 return i;
2174 }
2175 pos.index = -i;
2176 }
2177 } else {
2178 // Parse the rest as "hhmm" (RFC 822)
2179 int i = subParseNumericZone(text, ++pos.index,
2180 sign, 0, false, calb);
2181 if (i > 0) {
2182 return i;
2183 }
2184 pos.index = -i;
2185 }
2186 } catch (IndexOutOfBoundsException e) {
2187 }
2188 }
2189 break parsing;
2190
2191 case PATTERN_ISO_ZONE: // 'X'
2192 {
2193 if ((text.length() - pos.index) <= 0) {
2194 break parsing;
2195 }
2196
2197 int sign;
2198 char c = text.charAt(pos.index);
2199 if (c == 'Z') {
2200 calb.set(Calendar.ZONE_OFFSET, 0).set(Calendar.DST_OFFSET, 0);
2201 return ++pos.index;
2202 }
2203
2204 // parse text as "+/-hh[[:]mm]" based on count
2205 if (c == '+') {
2206 sign = 1;
2207 } else if (c == '-') {
2208 sign = -1;
2209 } else {
2210 ++pos.index;
2211 break parsing;
2212 }
2213 int i = subParseNumericZone(text, ++pos.index, sign, count,
2214 count == 3, calb);
2215 if (i > 0) {
2216 return i;
2217 }
2218 pos.index = -i;
2219 }
2220 break parsing;
2221
2222 default:
2223 // case PATTERN_DAY_OF_MONTH: // 'd'
2224 // case PATTERN_HOUR_OF_DAY0: // 'H' 0-based. eg, 23:59 + 1 hour =>> 00:59
2225 // case PATTERN_MINUTE: // 'm'
2226 // case PATTERN_SECOND: // 's'
2227 // case PATTERN_MILLISECOND: // 'S'
2228 // case PATTERN_DAY_OF_YEAR: // 'D'
2229 // case PATTERN_DAY_OF_WEEK_IN_MONTH: // 'F'
2230 // case PATTERN_WEEK_OF_YEAR: // 'w'
2231 // case PATTERN_WEEK_OF_MONTH: // 'W'
2232 // case PATTERN_HOUR0: // 'K' 0-based. eg, 11PM + 1 hour =>> 0 AM
2233 // case PATTERN_ISO_DAY_OF_WEEK: // 'u' (pseudo field);
2234
2235 // Handle "generic" fields
2236 if (obeyCount) {
2237 if ((start+count) > text.length()) {
2238 break parsing;
2239 }
2240 number = numberFormat.parse(text.substring(0, start+count), pos);
2241 } else {
2242 number = numberFormat.parse(text, pos);
2243 }
2244 if (number != null) {
2245 value = number.intValue();
2246
2247 if (useFollowingMinusSignAsDelimiter && (value < 0) &&
2248 (((pos.index < text.length()) &&
2249 (text.charAt(pos.index) != minusSign)) ||
2250 ((pos.index == text.length()) &&
2251 (text.charAt(pos.index-1) == minusSign)))) {
2252 value = -value;
2253 pos.index--;
2254 }
2255
2256 calb.set(field, value);
2257 return pos.index;
2258 }
2259 break parsing;
2260 }
2261 }
2262
2263 // Parsing failed.
2264 origPos.errorIndex = pos.index;
2265 return -1;
2266 }
2267
2268 /**
2269 * Returns true if the DateFormatSymbols has been set explicitly or locale
2270 * is null.
2271 */
2272 private boolean useDateFormatSymbols() {
2273 return useDateFormatSymbols || locale == null;
2274 }
2275
2276 /**
2277 * Translates a pattern, mapping each character in the from string to the
2278 * corresponding character in the to string.
2279 *
2280 * @exception IllegalArgumentException if the given pattern is invalid
2281 */
2282 private String translatePattern(String pattern, String from, String to) {
2283 StringBuilder result = new StringBuilder();
2284 boolean inQuote = false;
2285 for (int i = 0; i < pattern.length(); ++i) {
2286 char c = pattern.charAt(i);
2287 if (inQuote) {
2288 if (c == '\'') {
2289 inQuote = false;
2290 }
2291 }
2292 else {
2293 if (c == '\'') {
2294 inQuote = true;
2295 } else if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z')) {
2296 int ci = from.indexOf(c);
2297 if (ci >= 0) {
2298 // patternChars is longer than localPatternChars due
2299 // to serialization compatibility. The pattern letters
2300 // unsupported by localPatternChars pass through.
2301 if (ci < to.length()) {
2302 c = to.charAt(ci);
2303 }
2304 } else {
2305 throw new IllegalArgumentException("Illegal pattern " +
2306 " character '" +
2307 c + "'");
2308 }
2309 }
2310 }
2311 result.append(c);
2312 }
2313 if (inQuote) {
2314 throw new IllegalArgumentException("Unfinished quote in pattern");
2315 }
2316 return result.toString();
2317 }
2318
2319 /**
2320 * Returns a pattern string describing this date format.
2321 *
2322 * @return a pattern string describing this date format.
2323 */
2324 public String toPattern() {
2325 return pattern;
2326 }
2327
2328 /**
2329 * Returns a localized pattern string describing this date format.
2330 *
2331 * @return a localized pattern string describing this date format.
2332 */
2333 public String toLocalizedPattern() {
2334 return translatePattern(pattern,
2335 DateFormatSymbols.patternChars,
2336 formatData.getLocalPatternChars());
2337 }
2338
2339 /**
2340 * Applies the given pattern string to this date format.
2341 *
2342 * @param pattern the new date and time pattern for this date format
2343 * @exception NullPointerException if the given pattern is null
2344 * @exception IllegalArgumentException if the given pattern is invalid
2345 */
2346 public void applyPattern(String pattern)
2347 {
2348 applyPatternImpl(pattern);
2349 }
2350
2351 private void applyPatternImpl(String pattern) {
2352 compiledPattern = compile(pattern);
2353 this.pattern = pattern;
2354 }
2355
2356 /**
2357 * Applies the given localized pattern string to this date format.
2358 *
2359 * @param pattern a String to be mapped to the new date and time format
2360 * pattern for this format
2361 * @exception NullPointerException if the given pattern is null
2362 * @exception IllegalArgumentException if the given pattern is invalid
2363 */
2364 public void applyLocalizedPattern(String pattern) {
2365 String p = translatePattern(pattern,
2366 formatData.getLocalPatternChars(),
2367 DateFormatSymbols.patternChars);
2368 compiledPattern = compile(p);
2369 this.pattern = p;
2370 }
2371
2372 /**
2373 * Gets a copy of the date and time format symbols of this date format.
2374 *
2375 * @return the date and time format symbols of this date format
2376 * @see #setDateFormatSymbols
2377 */
2378 public DateFormatSymbols getDateFormatSymbols()
2379 {
2380 return (DateFormatSymbols)formatData.clone();
2381 }
2382
2383 /**
2384 * Sets the date and time format symbols of this date format.
2385 *
2386 * @param newFormatSymbols the new date and time format symbols
2387 * @exception NullPointerException if the given newFormatSymbols is null
2388 * @see #getDateFormatSymbols
2389 */
2390 public void setDateFormatSymbols(DateFormatSymbols newFormatSymbols)
2391 {
2392 this.formatData = (DateFormatSymbols)newFormatSymbols.clone();
2393 useDateFormatSymbols = true;
2394 }
2395
2396 /**
2397 * Creates a copy of this <code>SimpleDateFormat</code>. This also
2398 * clones the format's date format symbols.
2399 *
2400 * @return a clone of this <code>SimpleDateFormat</code>
2401 */
2402 @Override
2403 public Object clone() {
2404 SimpleDateFormat other = (SimpleDateFormat) super.clone();
2405 other.formatData = (DateFormatSymbols) formatData.clone();
2406 return other;
2407 }
2408
2409 /**
2410 * Returns the hash code value for this <code>SimpleDateFormat</code> object.
2411 *
2412 * @return the hash code value for this <code>SimpleDateFormat</code> object.
2413 */
2414 @Override
2415 public int hashCode()
2416 {
2417 return pattern.hashCode();
2418 // just enough fields for a reasonable distribution
2419 }
2420
2421 /**
2422 * Compares the given object with this <code>SimpleDateFormat</code> for
2423 * equality.
2424 *
2425 * @return true if the given object is equal to this
2426 * <code>SimpleDateFormat</code>
2427 */
2428 @Override
2429 public boolean equals(Object obj)
2430 {
2431 if (!super.equals(obj)) {
2432 return false; // super does class check
2433 }
2434 SimpleDateFormat that = (SimpleDateFormat) obj;
2435 return (pattern.equals(that.pattern)
2436 && formatData.equals(that.formatData));
2437 }
2438
2439 private static final int[] REST_OF_STYLES = {
2440 Calendar.SHORT_STANDALONE, Calendar.LONG_FORMAT, Calendar.LONG_STANDALONE,
2441 };
2442 private Map<String, Integer> getDisplayNamesMap(int field, Locale locale) {
2443 Map<String, Integer> map = calendar.getDisplayNames(field, Calendar.SHORT_FORMAT, locale);
2444 // Get all SHORT and LONG styles (avoid NARROW styles).
2445 for (int style : REST_OF_STYLES) {
2446 Map<String, Integer> m = calendar.getDisplayNames(field, style, locale);
2447 if (m != null) {
2448 map.putAll(m);
2449 }
2450 }
2451 return map;
2452 }
2453
2454 // for 'M' pattern only
2455 private Map<String, Integer> getDisplayContextNamesMap(int field, Locale locale) {
2456 Map<String, Integer> map = calendar.getDisplayNames(field,
2457 Calendar.SHORT_FORMAT | (forceStandaloneForm ? STANDALONE_MASK : 0), locale);
2458 // Get LONG styles
2459 Map<String, Integer> m = calendar.getDisplayNames(field,
2460 Calendar.LONG_FORMAT | (forceStandaloneForm ? STANDALONE_MASK : 0), locale);
2461 if (m != null) {
2462 map.putAll(m);
2463 }
2464 return map;
2465 }
2466
2467 /**
2468 * After reading an object from the input stream, the format
2469 * pattern in the object is verified.
2470 *
2471 * @exception InvalidObjectException if the pattern is invalid
2472 */
2473 private void readObject(ObjectInputStream stream)
2474 throws IOException, ClassNotFoundException {
2475 stream.defaultReadObject();
2476
2477 try {
2478 compiledPattern = compile(pattern);
2479 } catch (Exception e) {
2480 throw new InvalidObjectException("invalid pattern");
2481 }
2482
2483 if (serialVersionOnStream < 1) {
2484 // didn't have defaultCenturyStart field
2485 initializeDefaultCentury();
2486 }
2487 else {
2488 // fill in dependent transient field
2489 parseAmbiguousDatesAsAfter(defaultCenturyStart);
2490 }
2491 serialVersionOnStream = currentSerialVersion;
2492
2493 // If the deserialized object has a SimpleTimeZone, try
2494 // to replace it with a ZoneInfo equivalent in order to
2495 // be compatible with the SimpleTimeZone-based
2496 // implementation as much as possible.
2497 TimeZone tz = getTimeZone();
2498 if (tz instanceof SimpleTimeZone) {
2499 String id = tz.getID();
2500 TimeZone zi = TimeZone.getTimeZone(id);
2501 if (zi != null && zi.hasSameRules(tz) && zi.getID().equals(id)) {
2502 setTimeZone(zi);
2503 }
2504 }
2505 }
2506
2507 /**
2508 * Analyze the negative subpattern of DecimalFormat and set/update values
2509 * as necessary.
2510 */
2511 private void checkNegativeNumberExpression() {
2512 if ((numberFormat instanceof DecimalFormat) &&
2513 !numberFormat.equals(originalNumberFormat)) {
2514 String numberPattern = ((DecimalFormat)numberFormat).toPattern();
2515 if (!numberPattern.equals(originalNumberPattern)) {
2516 hasFollowingMinusSign = false;
2517
2518 int separatorIndex = numberPattern.indexOf(';');
2519 // If the negative subpattern is not absent, we have to analayze
2520 // it in order to check if it has a following minus sign.
2521 if (separatorIndex > -1) {
2522 int minusIndex = numberPattern.indexOf('-', separatorIndex);
2523 if ((minusIndex > numberPattern.lastIndexOf('0')) &&
2524 (minusIndex > numberPattern.lastIndexOf('#'))) {
2525 hasFollowingMinusSign = true;
2526 minusSign = ((DecimalFormat)numberFormat).getDecimalFormatSymbols().getMinusSign();
2527 }
2528 }
2529 originalNumberPattern = numberPattern;
2530 }
2531 originalNumberFormat = numberFormat;
2532 }
2533 }
2534
2535 }