2 **********************************************************************
\r
3 * Copyright (c) 2004-2010, International Business Machines
\r
4 * Corporation and others. All Rights Reserved.
\r
5 **********************************************************************
\r
7 * Created: April 6, 2004
\r
9 **********************************************************************
\r
11 package com.ibm.icu.text;
\r
13 import java.io.IOException;
\r
14 import java.io.InvalidObjectException;
\r
15 import java.io.ObjectInputStream;
\r
16 import java.text.AttributedCharacterIterator;
\r
17 import java.text.AttributedString;
\r
18 import java.text.CharacterIterator;
\r
19 import java.text.ChoiceFormat;
\r
20 import java.text.FieldPosition;
\r
21 import java.text.Format;
\r
22 import java.text.ParseException;
\r
23 import java.text.ParsePosition;
\r
24 import java.text.AttributedCharacterIterator.Attribute;
\r
25 import java.util.ArrayList;
\r
26 import java.util.Date;
\r
27 import java.util.HashMap;
\r
28 import java.util.HashSet;
\r
29 import java.util.List;
\r
30 import java.util.Locale;
\r
31 import java.util.Map;
\r
32 import java.util.Set;
\r
34 import com.ibm.icu.impl.Utility;
\r
35 import com.ibm.icu.lang.UCharacter;
\r
36 import com.ibm.icu.util.ULocale;
\r
39 * {@icuenhanced java.text.MessageFormat}.{@icu _usage_}
\r
41 * <p>MessageFormat produces concatenated messages in a language-neutral
\r
42 * way. Use this whenever concatenating strings that are displayed to
\r
45 * <p>A MessageFormat contains an array of <em>subformats</em> arranged
\r
46 * within a <em>template string</em>. Together, the subformats and
\r
47 * template string determine how the MessageFormat will operate during
\r
48 * formatting and parsing.
\r
50 * <p>Typically, both the subformats and the template string are
\r
51 * specified at once in a <em>pattern</em>. By using different
\r
52 * patterns for different locales, messages may be localized.
\r
54 * <p>When formatting, MessageFormat takes a collection of arguments
\r
55 * and produces a user-readable string. The arguments may be passed
\r
56 * as an array or as a Map. Each argument is matched up with its
\r
57 * corresponding subformat, which then formats it into a string. The
\r
58 * resulting strings are then assembled within the string template of
\r
59 * the MessageFormat to produce the final output string.
\r
61 * <p><strong>Note:</strong>
\r
62 * <code>MessageFormat</code> differs from the other <code>Format</code>
\r
63 * classes in that you create a <code>MessageFormat</code> object with one
\r
64 * of its constructors (not with a <code>getInstance</code> style factory
\r
65 * method). The factory methods aren't necessary because <code>MessageFormat</code>
\r
66 * itself doesn't implement locale-specific behavior. Any locale-specific
\r
67 * behavior is defined by the pattern that you provide and the
\r
68 * subformats used for inserted arguments.
\r
70 * <p><strong>Note:</strong>
\r
71 * In ICU 3.8 MessageFormat supports named arguments. If a named argument
\r
72 * is used, all arguments must be named. Names start with a character in
\r
73 * <code>:ID_START:</code> and continue with characters in <code>:ID_CONTINUE:</code>,
\r
74 * in particular they do not start with a digit. If named arguments
\r
75 * are used, {@link #usesNamedArguments()} will return true.
\r
77 * <p>The other new methods supporting named arguments are
\r
78 * {@link #setFormatsByArgumentName(Map)},
\r
79 * {@link #setFormatByArgumentName(String, Format)},
\r
80 * {@link #format(Map, StringBuffer, FieldPosition)},
\r
81 * {@link #format(String, Map)}, {@link #parseToMap(String, ParsePosition)},
\r
82 * and {@link #parseToMap(String)}. These methods are all compatible
\r
83 * with patterns that do not used named arguments-- in these cases
\r
84 * the keys in the input or output <code>Map</code>s use
\r
85 * <code>String</code>s that name the argument indices, e.g. "0",
\r
88 * <p>When named arguments are used, certain methods on MessageFormat that take or
\r
89 * return arrays will throw an exception, since it is not possible to
\r
90 * identify positions in an array using a name. These methods are
\r
91 * {@link #setFormatsByArgumentIndex(Format[])},
\r
92 * {@link #setFormatByArgumentIndex(int, Format)},
\r
93 * {@link #getFormatsByArgumentIndex()},
\r
94 * {@link #getFormats()},
\r
95 * {@link #format(Object[], StringBuffer, FieldPosition)},
\r
96 * {@link #format(String, Object[])},
\r
97 * {@link #parse(String, ParsePosition)}, and
\r
98 * {@link #parse(String)}.
\r
99 * These APIs all have corresponding new versions as listed above.
\r
101 * <p>The API {@link #format(Object, StringBuffer, FieldPosition)} has
\r
102 * been modified so that the <code>Object</code> argument can be
\r
103 * either an <code>Object</code> array or a <code>Map</code>. If this
\r
104 * format uses named arguments, this argument must not be an
\r
105 * <code>Object</code> array otherwise an exception will be thrown.
\r
106 * If the argument is a <code>Map</code> it can be used with Strings that
\r
107 * represent indices as described above.
\r
109 * <h4><a name="patterns">Patterns and Their Interpretation</a></h4>
\r
111 * <code>MessageFormat</code> uses patterns of the following form:
\r
112 * <blockquote><pre>
\r
113 * <i>MessageFormatPattern:</i>
\r
115 * <i>MessageFormatPattern</i> <i>FormatElement</i> <i>String</i>
\r
117 * <i>FormatElement:</i>
\r
118 * { <i>ArgumentIndexOrName</i> }
\r
119 * { <i>ArgumentIndexOrName</i> , <i>FormatType</i> }
\r
120 * { <i>ArgumentIndexOrName</i> , <i>FormatType</i> , <i>FormatStyle</i> }
\r
122 * <i>ArgumentIndexOrName: one of </i>
\r
124 * [:ID_START:][:ID_CONTINUE:]*
\r
126 * <i>FormatType: one of </i>
\r
127 * number date time choice spellout ordinal duration plural
\r
129 * <i>FormatStyle:</i>
\r
137 * <i>SubformatPattern</i>
\r
138 * <i>RulesetName</i>
\r
141 * <i>StringPart<sub>opt</sub></i>
\r
142 * <i>String</i> <i>StringPart</i>
\r
144 * <i>StringPart:</i>
\r
146 * ' <i>QuotedString</i> '
\r
147 * <i>UnquotedString</i>
\r
149 * <i>SubformatPattern:</i>
\r
150 * <i>SubformatPatternPart<sub>opt</sub></i>
\r
151 * <i>SubformatPattern</i> <i>SubformatPatternPart</i>
\r
153 * <i>SubFormatPatternPart:</i>
\r
154 * ' <i>QuotedPattern</i> '
\r
155 * <i>UnquotedPattern</i>
\r
156 * </pre></blockquote>
\r
158 * <i>RulesetName:</i>
\r
159 * <i>UnquotedString</i>
\r
161 * <p>Within a <i>String</i>, <code>"''"</code> represents a single
\r
162 * quote. A <i>QuotedString</i> can contain arbitrary characters
\r
163 * except single quotes; the surrounding single quotes are removed.
\r
164 * An <i>UnquotedString</i> can contain arbitrary characters
\r
165 * except single quotes and left curly brackets. Thus, a string that
\r
166 * should result in the formatted message "'{0}'" can be written as
\r
167 * <code>"'''{'0}''"</code> or <code>"'''{0}'''"</code>.
\r
169 * <p>Within a <i>SubformatPattern</i>, different rules apply.
\r
170 * A <i>QuotedPattern</i> can contain arbitrary characters
\r
171 * except single quotes; but the surrounding single quotes are
\r
172 * <strong>not</strong> removed, so they may be interpreted by the
\r
173 * subformat. For example, <code>"{1,number,$'#',##}"</code> will
\r
174 * produce a number format with the pound-sign quoted, with a result
\r
175 * such as: "$#31,45".
\r
176 * An <i>UnquotedPattern</i> can contain arbitrary characters
\r
177 * except single quotes, but curly braces within it must be balanced.
\r
178 * For example, <code>"ab {0} de"</code> and <code>"ab '}' de"</code>
\r
179 * are valid subformat patterns, but <code>"ab {0'}' de"</code> and
\r
180 * <code>"ab } de"</code> are not.
\r
182 * <p><dl><dt><b>Warning:</b><dd>The rules for using quotes within message
\r
183 * format patterns unfortunately have shown to be somewhat confusing.
\r
184 * In particular, it isn't always obvious to localizers whether single
\r
185 * quotes need to be doubled or not. Make sure to inform localizers about
\r
186 * the rules, and tell them (for example, by using comments in resource
\r
187 * bundle source files) which strings will be processed by MessageFormat.
\r
188 * Note that localizers may need to use single quotes in translated
\r
189 * strings where the original version doesn't have them.
\r
191 * <br>Note also that the simplest way to avoid the problem is to
\r
192 * use the real apostrophe (single quote) character \u2019 (') for
\r
193 * human-readable text, and to use the ASCII apostrophe (\u0027 ' )
\r
194 * only in program syntax, like quoting in MessageFormat.
\r
195 * See the annotations for U+0027 Apostrophe in The Unicode Standard.</p>
\r
198 * <p>The <i>ArgumentIndex</i> value is a non-negative integer written
\r
199 * using the digits '0' through '9', and represents an index into the
\r
200 * <code>arguments</code> array passed to the <code>format</code> methods
\r
201 * or the result array returned by the <code>parse</code> methods.
\r
203 * <p>The <i>FormatType</i> and <i>FormatStyle</i> values are used to create
\r
204 * a <code>Format</code> instance for the format element. The following
\r
205 * table shows how the values map to Format instances. Combinations not
\r
206 * shown in the table are illegal. A <i>SubformatPattern</i> must
\r
207 * be a valid pattern string for the Format subclass used.
\r
209 * <p><table border=1>
\r
213 * <th>Subformat Created
\r
215 * <td colspan=2><i>(none)</i>
\r
216 * <td><code>null</code>
\r
218 * <td rowspan=5><code>number</code>
\r
219 * <td><i>(none)</i>
\r
220 * <td><code>NumberFormat.getInstance(getLocale())</code>
\r
222 * <td><code>integer</code>
\r
223 * <td><code>NumberFormat.getIntegerInstance(getLocale())</code>
\r
225 * <td><code>currency</code>
\r
226 * <td><code>NumberFormat.getCurrencyInstance(getLocale())</code>
\r
228 * <td><code>percent</code>
\r
229 * <td><code>NumberFormat.getPercentInstance(getLocale())</code>
\r
231 * <td><i>SubformatPattern</i>
\r
232 * <td><code>new DecimalFormat(subformatPattern, new DecimalFormatSymbols(getLocale()))</code>
\r
234 * <td rowspan=6><code>date</code>
\r
235 * <td><i>(none)</i>
\r
236 * <td><code>DateFormat.getDateInstance(DateFormat.DEFAULT, getLocale())</code>
\r
238 * <td><code>short</code>
\r
239 * <td><code>DateFormat.getDateInstance(DateFormat.SHORT, getLocale())</code>
\r
241 * <td><code>medium</code>
\r
242 * <td><code>DateFormat.getDateInstance(DateFormat.DEFAULT, getLocale())</code>
\r
244 * <td><code>long</code>
\r
245 * <td><code>DateFormat.getDateInstance(DateFormat.LONG, getLocale())</code>
\r
247 * <td><code>full</code>
\r
248 * <td><code>DateFormat.getDateInstance(DateFormat.FULL, getLocale())</code>
\r
250 * <td><i>SubformatPattern</i>
\r
251 * <td><code>new SimpleDateFormat(subformatPattern, getLocale())
\r
253 * <td rowspan=6><code>time</code>
\r
254 * <td><i>(none)</i>
\r
255 * <td><code>DateFormat.getTimeInstance(DateFormat.DEFAULT, getLocale())</code>
\r
257 * <td><code>short</code>
\r
258 * <td><code>DateFormat.getTimeInstance(DateFormat.SHORT, getLocale())</code>
\r
260 * <td><code>medium</code>
\r
261 * <td><code>DateFormat.getTimeInstance(DateFormat.DEFAULT, getLocale())</code>
\r
263 * <td><code>long</code>
\r
264 * <td><code>DateFormat.getTimeInstance(DateFormat.LONG, getLocale())</code>
\r
266 * <td><code>full</code>
\r
267 * <td><code>DateFormat.getTimeInstance(DateFormat.FULL, getLocale())</code>
\r
269 * <td><i>SubformatPattern</i>
\r
270 * <td><code>new SimpleDateFormat(subformatPattern, getLocale())
\r
272 * <td><code>choice</code>
\r
273 * <td><i>SubformatPattern</i>
\r
274 * <td><code>new ChoiceFormat(subformatPattern)</code>
\r
276 * <td><code>spellout</code>
\r
277 * <td><i>RulesetName (optional)</i>
\r
278 * <td><code>new RuleBasedNumberFormat(getLocale(), RuleBasedNumberFormat.SPELLOUT)
\r
279 * <br/> .setDefaultRuleset(ruleset);</code>
\r
281 * <td><code>ordinal</code>
\r
282 * <td><i>RulesetName (optional)</i>
\r
283 * <td><code>new RuleBasedNumberFormat(getLocale(), RuleBasedNumberFormat.ORDINAL)
\r
284 * <br/> .setDefaultRuleset(ruleset);</code>
\r
286 * <td><code>duration</code>
\r
287 * <td><i>RulesetName (optional)</i>
\r
288 * <td><code>new RuleBasedNumberFormat(getLocale(), RuleBasedNumberFormat.DURATION)
\r
289 * <br/> .setDefaultRuleset(ruleset);</code>
\r
291 * <td><code>plural</code>
\r
292 * <td><i>SubformatPattern</i>
\r
293 * <td><code>new PluralFormat(subformatPattern)</code>
\r
295 * <td><code>select</code>
\r
296 * <td><i>SubformatPattern</i>
\r
297 * <td><code>new SelectFormat(subformatPattern)</code>
\r
301 * <h4>Usage Information</h4>
\r
303 * <p>Here are some examples of usage:
\r
306 * Object[] arguments = {
\r
308 * new Date(System.currentTimeMillis()),
\r
309 * "a disturbance in the Force"
\r
312 * String result = MessageFormat.format(
\r
313 * "At {1,time} on {1,date}, there was {2} on planet {0,number,integer}.",
\r
316 * <em>output</em>: At 12:30 PM on Jul 3, 2053, there was a disturbance
\r
317 * in the Force on planet 7.
\r
321 * Typically, the message format will come from resources, and the
\r
322 * arguments will be dynamically set at runtime.
\r
327 * Object[] testArgs = {new Long(3), "MyDisk"};
\r
329 * MessageFormat form = new MessageFormat(
\r
330 * "The disk \"{1}\" contains {0} file(s).");
\r
332 * System.out.println(form.format(testArgs));
\r
334 * // output, with different testArgs
\r
335 * <em>output</em>: The disk "MyDisk" contains 0 file(s).
\r
336 * <em>output</em>: The disk "MyDisk" contains 1 file(s).
\r
337 * <em>output</em>: The disk "MyDisk" contains 1,273 file(s).
\r
341 * <p>For more sophisticated patterns, you can use a <code>ChoiceFormat</code> to get
\r
345 * MessageFormat form = new MessageFormat("The disk \"{1}\" contains {0}.");
\r
346 * double[] filelimits = {0,1,2};
\r
347 * String[] filepart = {"no files","one file","{0,number} files"};
\r
348 * ChoiceFormat fileform = new ChoiceFormat(filelimits, filepart);
\r
349 * form.setFormatByArgumentIndex(0, fileform);
\r
351 * Object[] testArgs = {new Long(12373), "MyDisk"};
\r
353 * System.out.println(form.format(testArgs));
\r
355 * // output, with different testArgs
\r
356 * output: The disk "MyDisk" contains no files.
\r
357 * output: The disk "MyDisk" contains one file.
\r
358 * output: The disk "MyDisk" contains 1,273 files.
\r
361 * You can either do this programmatically, as in the above example,
\r
362 * or by using a pattern (see
\r
363 * {@link ChoiceFormat}
\r
364 * for more information) as in:
\r
367 * form.applyPattern(
\r
368 * "There {0,choice,0#are no files|1#is one file|1<are {0,number,integer} files}.");
\r
372 * <p><strong>Note:</strong> As we see above, the string produced
\r
373 * by a <code>ChoiceFormat</code> in <code>MessageFormat</code> is treated specially;
\r
374 * occurances of '{' are used to indicated subformats, and cause recursion.
\r
375 * If you create both a <code>MessageFormat</code> and <code>ChoiceFormat</code>
\r
376 * programmatically (instead of using the string patterns), then be careful not to
\r
377 * produce a format that recurses on itself, which will cause an infinite loop.
\r
379 * <p>When a single argument is parsed more than once in the string, the last match
\r
380 * will be the final result of the parsing. For example,
\r
382 * MessageFormat mf = new MessageFormat("{0,number,#.##}, {0,number,#.#}");
\r
383 * Object[] objs = {new Double(3.1415)};
\r
384 * String result = mf.format( objs );
\r
385 * // result now equals "3.14, 3.1"
\r
387 * objs = mf.parse(result, new ParsePosition(0));
\r
388 * // objs now equals {new Double(3.1)}
\r
391 * <p>Likewise, parsing with a MessageFormat object using patterns containing
\r
392 * multiple occurances of the same argument would return the last match. For
\r
395 * MessageFormat mf = new MessageFormat("{0}, {0}, {0}");
\r
396 * String forParsing = "x, y, z";
\r
397 * Object[] objs = mf.parse(forParsing, new ParsePosition(0));
\r
398 * // result now equals {new String("z")}
\r
401 * <h4><a name="synchronization">Synchronization</a></h4>
\r
403 * <p>Message formats are not synchronized.
\r
404 * It is recommended to create separate format instances for each thread.
\r
405 * If multiple threads access a format concurrently, it must be synchronized
\r
408 * @see java.util.Locale
\r
410 * @see NumberFormat
\r
411 * @see DecimalFormat
\r
412 * @see ChoiceFormat
\r
413 * @see PluralFormat
\r
414 * @see SelectFormat
\r
415 * @author Mark Davis
\r
418 public class MessageFormat extends UFormat {
\r
420 // Generated by serialver from JDK 1.4.1_01
\r
421 static final long serialVersionUID = 7136212545847378651L;
\r
424 * Constructs a MessageFormat for the default locale and the
\r
425 * specified pattern.
\r
426 * The constructor first sets the locale, then parses the pattern and
\r
427 * creates a list of subformats for the format elements contained in it.
\r
428 * Patterns and their interpretation are specified in the
\r
429 * <a href="#patterns">class description</a>.
\r
431 * @param pattern the pattern for this message format
\r
432 * @exception IllegalArgumentException if the pattern is invalid
\r
435 public MessageFormat(String pattern) {
\r
436 this.ulocale = ULocale.getDefault();
\r
437 applyPattern(pattern);
\r
441 * Constructs a MessageFormat for the specified locale and
\r
443 * The constructor first sets the locale, then parses the pattern and
\r
444 * creates a list of subformats for the format elements contained in it.
\r
445 * Patterns and their interpretation are specified in the
\r
446 * <a href="#patterns">class description</a>.
\r
448 * @param pattern the pattern for this message format
\r
449 * @param locale the locale for this message format
\r
450 * @exception IllegalArgumentException if the pattern is invalid
\r
453 public MessageFormat(String pattern, Locale locale) {
\r
454 this(pattern, ULocale.forLocale(locale));
\r
458 * Constructs a MessageFormat for the specified locale and
\r
460 * The constructor first sets the locale, then parses the pattern and
\r
461 * creates a list of subformats for the format elements contained in it.
\r
462 * Patterns and their interpretation are specified in the
\r
463 * <a href="#patterns">class description</a>.
\r
465 * @param pattern the pattern for this message format
\r
466 * @param locale the locale for this message format
\r
467 * @exception IllegalArgumentException if the pattern is invalid
\r
470 public MessageFormat(String pattern, ULocale locale) {
\r
471 this.ulocale = locale;
\r
472 applyPattern(pattern);
\r
476 * Sets the locale to be used when creating or comparing subformats.
\r
477 * This affects subsequent calls to the {@link #applyPattern applyPattern}
\r
478 * and {@link #toPattern toPattern} methods as well as to the
\r
479 * <code>format</code> and
\r
480 * {@link #formatToCharacterIterator formatToCharacterIterator} methods.
\r
482 * @param locale the locale to be used when creating or comparing subformats
\r
485 public void setLocale(Locale locale) {
\r
486 setLocale(ULocale.forLocale(locale));
\r
490 * Sets the locale to be used when creating or comparing subformats.
\r
491 * This affects subsequent calls to the {@link #applyPattern applyPattern}
\r
492 * and {@link #toPattern toPattern} methods as well as to the
\r
493 * <code>format</code> and
\r
494 * {@link #formatToCharacterIterator formatToCharacterIterator} methods.
\r
496 * @param locale the locale to be used when creating or comparing subformats
\r
499 public void setLocale(ULocale locale) {
\r
500 /* Save the pattern, and then reapply so that */
\r
501 /* we pick up any changes in locale specific */
\r
503 String existingPattern = toPattern(); /*ibm.3550*/
\r
504 this.ulocale = locale;
\r
505 applyPattern(existingPattern); /*ibm.3550*/
\r
509 * Returns the locale that's used when creating or comparing subformats.
\r
511 * @return the locale used when creating or comparing subformats
\r
514 public Locale getLocale() {
\r
515 return ulocale.toLocale();
\r
519 * {@icu} Returns the locale that's used when creating or comparing subformats.
\r
521 * @return the locale used when creating or comparing subformats
\r
524 public ULocale getULocale() {
\r
529 * Sets the pattern used by this message format.
\r
530 * The method parses the pattern and creates a list of subformats
\r
531 * for the format elements contained in it.
\r
532 * Patterns and their interpretation are specified in the
\r
533 * <a href="#patterns">class description</a>.
\r
535 * The pattern must contain only named or only numeric arguments,
\r
536 * mixing them is not allowed.
\r
538 * @param pttrn the pattern for this message format
\r
539 * @throws IllegalArgumentException if the pattern is invalid
\r
542 @SuppressWarnings("fallthrough")
\r
543 public void applyPattern(String pttrn) {
\r
544 StringBuilder[] segments = new StringBuilder[4];
\r
545 for (int i = 0; i < segments.length; ++i) {
\r
546 segments[i] = new StringBuilder();
\r
549 int formatNumber = 0;
\r
550 boolean inQuote = false;
\r
551 int braceStack = 0;
\r
553 for (int i = 0; i < pttrn.length(); ++i) {
\r
554 char ch = pttrn.charAt(i);
\r
557 if (i + 1 < pttrn.length()
\r
558 && pttrn.charAt(i+1) == '\'') {
\r
559 segments[part].append(ch); // handle doubles
\r
562 inQuote = !inQuote;
\r
564 } else if (ch == '{' && !inQuote) {
\r
567 segments[part].append(ch);
\r
569 } else if (inQuote) { // just copy quotes in parts
\r
570 segments[part].append(ch);
\r
580 segments[part].append(ch);
\r
584 segments[part].append(ch);
\r
587 if (braceStack == 0) {
\r
589 makeFormat(i, formatNumber, segments);
\r
593 segments[part].append(ch);
\r
598 // fall through, so we keep quotes in other parts
\r
600 segments[part].append(ch);
\r
605 if (braceStack == 0 && part != 0) {
\r
607 throw new IllegalArgumentException("Unmatched braces in the pattern.");
\r
609 this.pattern = segments[0].toString();
\r
613 * Returns a pattern representing the current state of the message format.
\r
614 * The string is constructed from internal information and therefore
\r
615 * does not necessarily equal the previously applied pattern.
\r
617 * @return a pattern representing the current state of the message format
\r
620 public String toPattern() {
\r
621 // later, make this more extensible
\r
622 int lastOffset = 0;
\r
623 StringBuilder result = new StringBuilder();
\r
624 for (int i = 0; i <= maxOffset; ++i) {
\r
625 copyAndFixQuotes(pattern, lastOffset, offsets[i],result);
\r
626 lastOffset = offsets[i];
\r
627 result.append('{');
\r
628 result.append(argumentNames[i]);
\r
629 if (formats[i] == null) {
\r
630 // do nothing, string format
\r
631 } else if (formats[i] instanceof DecimalFormat) {
\r
632 if (formats[i].equals(NumberFormat.getInstance(ulocale))) {
\r
633 result.append(",number");
\r
634 } else if (formats[i].equals(NumberFormat.getCurrencyInstance(ulocale))) {
\r
635 result.append(",number,currency");
\r
636 } else if (formats[i].equals(NumberFormat.getPercentInstance(ulocale))) {
\r
637 result.append(",number,percent");
\r
638 } else if (formats[i].equals(NumberFormat.getIntegerInstance(ulocale))) {
\r
639 result.append(",number,integer");
\r
641 result.append(",number," +
\r
642 ((DecimalFormat)formats[i]).toPattern());
\r
644 } else if (formats[i] instanceof SimpleDateFormat) {
\r
645 if (formats[i].equals(DateFormat.getDateInstance(DateFormat.DEFAULT,ulocale))) {
\r
646 result.append(",date");
\r
647 } else if (formats[i].equals(DateFormat.getDateInstance(DateFormat.SHORT,ulocale))) {
\r
648 result.append(",date,short");
\r
649 // This code will never be executed [alan]
\r
650 // } else if (formats[i].equals(DateFormat.getDateInstance(DateFormat.DEFAULT,ulocale))) {
\r
651 // result.append(",date,medium");
\r
652 } else if (formats[i].equals(DateFormat.getDateInstance(DateFormat.LONG,ulocale))) {
\r
653 result.append(",date,long");
\r
654 } else if (formats[i].equals(DateFormat.getDateInstance(DateFormat.FULL,ulocale))) {
\r
655 result.append(",date,full");
\r
656 } else if (formats[i].equals(DateFormat.getTimeInstance(DateFormat.DEFAULT,ulocale))) {
\r
657 result.append(",time");
\r
658 } else if (formats[i].equals(DateFormat.getTimeInstance(DateFormat.SHORT,ulocale))) {
\r
659 result.append(",time,short");
\r
660 // This code will never be executed [alan]
\r
661 // } else if (formats[i].equals(DateFormat.getTimeInstance(DateFormat.DEFAULT,ulocale))) {
\r
662 // result.append(",time,medium");
\r
663 } else if (formats[i].equals(DateFormat.getTimeInstance(DateFormat.LONG,ulocale))) {
\r
664 result.append(",time,long");
\r
665 } else if (formats[i].equals(DateFormat.getTimeInstance(DateFormat.FULL,ulocale))) {
\r
666 result.append(",time,full");
\r
668 result.append(",date," + ((SimpleDateFormat)formats[i]).toPattern());
\r
670 } else if (formats[i] instanceof ChoiceFormat) {
\r
671 result.append(",choice,"
\r
672 + ((ChoiceFormat) formats[i]).toPattern());
\r
673 } else if (formats[i] instanceof PluralFormat ){
\r
674 String pat = ((PluralFormat)formats[i]).toPattern();
\r
675 // TODO: PluralFormat does not do the single quote thing, just reapply
\r
676 pat = duplicateSingleQuotes(pat);
\r
677 result.append(",plural," + pat);
\r
678 } else if (formats[i] instanceof SelectFormat) {
\r
679 String pat = ((SelectFormat)formats[i]).toPattern();
\r
680 //TODO: SelectFormat does not do the single quote thing, just reapply
\r
681 pat = duplicateSingleQuotes(pat);
\r
682 result.append(",select," + pat);
\r
684 //result.append(", unknown");
\r
686 result.append('}');
\r
688 copyAndFixQuotes(pattern, lastOffset, pattern.length(), result);
\r
689 return result.toString();
\r
693 * Double every single quote
\r
695 private String duplicateSingleQuotes(String pat) {
\r
696 String result = pat;
\r
697 if (pat.indexOf('\'') != 0) {
\r
698 StringBuilder buf = new StringBuilder();
\r
699 for (int j = 0; j < pat.length(); ++j) {
\r
700 char ch = pat.charAt(j);
\r
702 buf.append(ch); // double it
\r
706 result = buf.toString();
\r
712 * Sets the formats to use for the values passed into
\r
713 * <code>format</code> methods or returned from <code>parse</code>
\r
714 * methods. The indices of elements in <code>newFormats</code>
\r
715 * correspond to the argument indices used in the previously set
\r
717 * The order of formats in <code>newFormats</code> thus corresponds to
\r
718 * the order of elements in the <code>arguments</code> array passed
\r
719 * to the <code>format</code> methods or the result array returned
\r
720 * by the <code>parse</code> methods.
\r
722 * If an argument index is used for more than one format element
\r
723 * in the pattern string, then the corresponding new format is used
\r
724 * for all such format elements. If an argument index is not used
\r
725 * for any format element in the pattern string, then the
\r
726 * corresponding new format is ignored. If fewer formats are provided
\r
727 * than needed, then only the formats for argument indices less
\r
728 * than <code>newFormats.length</code> are replaced.
\r
730 * This method is only supported if the format does not use
\r
731 * named arguments, otherwise an IllegalArgumentException is thrown.
\r
733 * @param newFormats the new formats to use
\r
734 * @throws NullPointerException if <code>newFormats</code> is null
\r
735 * @throws IllegalArgumentException if this formatter uses named arguments
\r
738 public void setFormatsByArgumentIndex(Format[] newFormats) {
\r
739 if (!argumentNamesAreNumeric) {
\r
740 throw new IllegalArgumentException(
\r
741 "This method is not available in MessageFormat objects " +
\r
742 "that use alphanumeric argument names.");
\r
744 for (int i = 0; i <= maxOffset; i++) {
\r
745 int j = Integer.parseInt(argumentNames[i]);
\r
746 if (j < newFormats.length) {
\r
747 formats[i] = newFormats[j];
\r
753 * {@icu} Sets the formats to use for the values passed into
\r
754 * <code>format</code> methods or returned from <code>parse</code>
\r
755 * methods. The keys in <code>newFormats</code> are the argument
\r
756 * names in the previously set pattern string, and the values
\r
759 * Only argument names from the pattern string are considered.
\r
760 * Extra keys in <code>newFormats</code> that do not correspond
\r
761 * to an argument name are ignored. Similarly, if there is no
\r
762 * format in newFormats for an argument name, the formatter
\r
763 * for that argument remains unchanged.
\r
765 * This may be called on formats that do not use named arguments.
\r
766 * In this case the map will be queried for key Strings that
\r
767 * represent argument indices, e.g. "0", "1", "2" etc.
\r
769 * @param newFormats a map from String to Format providing new
\r
770 * formats for named arguments.
\r
773 public void setFormatsByArgumentName(Map<String, Format> newFormats) {
\r
774 for (int i = 0; i <= maxOffset; i++) {
\r
775 if (newFormats.containsKey(argumentNames[i])) {
\r
776 Format f = newFormats.get(argumentNames[i]);
\r
783 * Sets the formats to use for the format elements in the
\r
784 * previously set pattern string.
\r
785 * The order of formats in <code>newFormats</code> corresponds to
\r
786 * the order of format elements in the pattern string.
\r
788 * If more formats are provided than needed by the pattern string,
\r
789 * the remaining ones are ignored. If fewer formats are provided
\r
790 * than needed, then only the first <code>newFormats.length</code>
\r
791 * formats are replaced.
\r
793 * Since the order of format elements in a pattern string often
\r
794 * changes during localization, it is generally better to use the
\r
795 * {@link #setFormatsByArgumentIndex setFormatsByArgumentIndex}
\r
796 * method, which assumes an order of formats corresponding to the
\r
797 * order of elements in the <code>arguments</code> array passed to
\r
798 * the <code>format</code> methods or the result array returned by
\r
799 * the <code>parse</code> methods.
\r
801 * @param newFormats the new formats to use
\r
802 * @exception NullPointerException if <code>newFormats</code> is null
\r
805 public void setFormats(Format[] newFormats) {
\r
806 int runsToCopy = newFormats.length;
\r
807 if (runsToCopy > maxOffset + 1) {
\r
808 runsToCopy = maxOffset + 1;
\r
810 for (int i = 0; i < runsToCopy; i++) {
\r
811 formats[i] = newFormats[i];
\r
816 * Sets the format to use for the format elements within the
\r
817 * previously set pattern string that use the given argument
\r
819 * The argument index is part of the format element definition and
\r
820 * represents an index into the <code>arguments</code> array passed
\r
821 * to the <code>format</code> methods or the result array returned
\r
822 * by the <code>parse</code> methods.
\r
824 * If the argument index is used for more than one format element
\r
825 * in the pattern string, then the new format is used for all such
\r
826 * format elements. If the argument index is not used for any format
\r
827 * element in the pattern string, then the new format is ignored.
\r
829 * This method is only supported when exclusively numbers are used for
\r
830 * argument names. Otherwise an IllegalArgumentException is thrown.
\r
832 * @param argumentIndex the argument index for which to use the new format
\r
833 * @param newFormat the new format to use
\r
834 * @throws IllegalArgumentException if this format uses named arguments
\r
837 public void setFormatByArgumentIndex(int argumentIndex, Format newFormat) {
\r
838 if (!argumentNamesAreNumeric) {
\r
839 throw new IllegalArgumentException(
\r
840 "This method is not available in MessageFormat objects " +
\r
841 "that use alphanumeric argument names.");
\r
843 for (int j = 0; j <= maxOffset; j++) {
\r
844 if (Integer.parseInt(argumentNames[j]) == argumentIndex) {
\r
845 formats[j] = newFormat;
\r
851 * {@icu} Sets the format to use for the format elements within the
\r
852 * previously set pattern string that use the given argument
\r
855 * If the argument name is used for more than one format element
\r
856 * in the pattern string, then the new format is used for all such
\r
857 * format elements. If the argument name is not used for any format
\r
858 * element in the pattern string, then the new format is ignored.
\r
860 * This API may be used on formats that do not use named arguments.
\r
861 * In this case <code>argumentName</code> should be a String that names
\r
862 * an argument index, e.g. "0", "1", "2"... etc. If it does not name
\r
863 * a valid index, the format will be ignored. No error is thrown.
\r
865 * @param argumentName the name of the argument to change
\r
866 * @param newFormat the new format to use
\r
869 public void setFormatByArgumentName(String argumentName, Format newFormat) {
\r
870 for (int i = 0; i <= maxOffset; ++i) {
\r
871 if (argumentName.equals(argumentNames[i])) {
\r
872 formats[i] = newFormat;
\r
878 * Sets the format to use for the format element with the given
\r
879 * format element index within the previously set pattern string.
\r
880 * The format element index is the zero-based number of the format
\r
881 * element counting from the start of the pattern string.
\r
883 * Since the order of format elements in a pattern string often
\r
884 * changes during localization, it is generally better to use the
\r
885 * {@link #setFormatByArgumentIndex setFormatByArgumentIndex}
\r
886 * method, which accesses format elements based on the argument
\r
887 * index they specify.
\r
889 * @param formatElementIndex the index of a format element within the pattern
\r
890 * @param newFormat the format to use for the specified format element
\r
891 * @exception ArrayIndexOutOfBoundsException if formatElementIndex is equal to or
\r
892 * larger than the number of format elements in the pattern string
\r
895 public void setFormat(int formatElementIndex, Format newFormat) {
\r
896 formats[formatElementIndex] = newFormat;
\r
900 * Returns the formats used for the values passed into
\r
901 * <code>format</code> methods or returned from <code>parse</code>
\r
902 * methods. The indices of elements in the returned array
\r
903 * correspond to the argument indices used in the previously set
\r
905 * The order of formats in the returned array thus corresponds to
\r
906 * the order of elements in the <code>arguments</code> array passed
\r
907 * to the <code>format</code> methods or the result array returned
\r
908 * by the <code>parse</code> methods.
\r
910 * If an argument index is used for more than one format element
\r
911 * in the pattern string, then the format used for the last such
\r
912 * format element is returned in the array. If an argument index
\r
913 * is not used for any format element in the pattern string, then
\r
914 * null is returned in the array.
\r
916 * This method is only supported when exclusively numbers are used for
\r
917 * argument names. Otherwise an IllegalArgumentException is thrown.
\r
919 * @return the formats used for the arguments within the pattern
\r
920 * @throws IllegalArgumentException if this format uses named arguments
\r
923 public Format[] getFormatsByArgumentIndex() {
\r
924 if (!argumentNamesAreNumeric) {
\r
925 throw new IllegalArgumentException(
\r
926 "This method is not available in MessageFormat objects " +
\r
927 "that use alphanumeric argument names.");
\r
929 int maximumArgumentNumber = -1;
\r
930 for (int i = 0; i <= maxOffset; i++) {
\r
931 int argumentNumber = Integer.parseInt(argumentNames[i]);
\r
932 if (argumentNumber > maximumArgumentNumber) {
\r
933 maximumArgumentNumber = argumentNumber;
\r
936 Format[] resultArray = new Format[maximumArgumentNumber + 1];
\r
937 for (int i = 0; i <= maxOffset; i++) {
\r
938 resultArray[Integer.parseInt(argumentNames[i])] = formats[i];
\r
940 return resultArray;
\r
942 // TODO: provide method public Map getFormatsByArgumentName().
\r
943 // Where Map is: String argumentName --> Format format.
\r
946 * Returns the formats used for the format elements in the
\r
947 * previously set pattern string.
\r
948 * The order of formats in the returned array corresponds to
\r
949 * the order of format elements in the pattern string.
\r
951 * Since the order of format elements in a pattern string often
\r
952 * changes during localization, it's generally better to use the
\r
953 * {@link #getFormatsByArgumentIndex()}
\r
954 * method, which assumes an order of formats corresponding to the
\r
955 * order of elements in the <code>arguments</code> array passed to
\r
956 * the <code>format</code> methods or the result array returned by
\r
957 * the <code>parse</code> methods.
\r
959 * This method is only supported when exclusively numbers are used for
\r
960 * argument names. Otherwise an IllegalArgumentException is thrown.
\r
962 * @return the formats used for the format elements in the pattern
\r
963 * @throws IllegalArgumentException if this format uses named arguments
\r
966 public Format[] getFormats() {
\r
967 Format[] resultArray = new Format[maxOffset + 1];
\r
968 System.arraycopy(formats, 0, resultArray, 0, maxOffset + 1);
\r
969 return resultArray;
\r
973 * {@icu} Returns the format argument names. For more details, see
\r
974 * {@link #setFormatByArgumentName(String, Format)}.
\r
975 * @return List of names
\r
977 * @deprecated This API is ICU internal only.
\r
979 public Set<String> getFormatArgumentNames() {
\r
980 Set<String> result = new HashSet<String>();
\r
981 for (int i = 0; i <= maxOffset; ++i) {
\r
982 result.add(argumentNames[i]);
\r
988 * {@icu} Returns the formats according to their argument names. For more details, see
\r
989 * {@link #setFormatByArgumentName(String, Format)}.
\r
990 * @return format associated with the name, or null if there isn't one.
\r
992 * @deprecated This API is ICU internal only.
\r
994 public Format getFormatByArgumentName(String argumentName) {
\r
995 for (int i = 0; i <= maxOffset; ++i) {
\r
996 if (argumentName.equals(argumentNames[i])) {
\r
1004 * Formats an array of objects and appends the <code>MessageFormat</code>'s
\r
1005 * pattern, with format elements replaced by the formatted objects, to the
\r
1006 * provided <code>StringBuffer</code>.
\r
1008 * The text substituted for the individual format elements is derived from
\r
1009 * the current subformat of the format element and the
\r
1010 * <code>arguments</code> element at the format element's argument index
\r
1011 * as indicated by the first matching line of the following table. An
\r
1012 * argument is <i>unavailable</i> if <code>arguments</code> is
\r
1013 * <code>null</code> or has fewer than argumentIndex+1 elements. When
\r
1014 * an argument is unavailable no substitution is performed.
\r
1016 * <table border=1>
\r
1020 * <th>Formatted Text
\r
1023 * <td><i>unavailable</i>
\r
1024 * <td><code>"{" + argumentIndex + "}"</code>
\r
1027 * <td><code>null</code>
\r
1028 * <td><code>"null"</code>
\r
1030 * <td><code>instanceof ChoiceFormat</code>
\r
1032 * <td><code>subformat.format(argument).indexOf('{') >= 0 ?<br>
\r
1033 * (new MessageFormat(subformat.format(argument), getLocale())).format(argument) :
\r
1034 * subformat.format(argument)</code>
\r
1036 * <td><code>!= null</code>
\r
1038 * <td><code>subformat.format(argument)</code>
\r
1040 * <td><code>null</code>
\r
1041 * <td><code>instanceof Number</code>
\r
1042 * <td><code>NumberFormat.getInstance(getLocale()).format(argument)</code>
\r
1044 * <td><code>null</code>
\r
1045 * <td><code>instanceof Date</code>
\r
1046 * <td><code>DateFormat.getDateTimeInstance(DateFormat.SHORT,
\r
1047 * DateFormat.SHORT, getLocale()).format(argument)</code>
\r
1049 * <td><code>null</code>
\r
1050 * <td><code>instanceof String</code>
\r
1051 * <td><code>argument</code>
\r
1053 * <td><code>null</code>
\r
1055 * <td><code>argument.toString()</code>
\r
1058 * If <code>pos</code> is non-null, and refers to
\r
1059 * <code>Field.ARGUMENT</code>, the location of the first formatted
\r
1060 * string will be returned.
\r
1062 * This method is only supported when the format does not use named
\r
1063 * arguments, otherwise an IllegalArgumentException is thrown.
\r
1065 * @param arguments an array of objects to be formatted and substituted.
\r
1066 * @param result where text is appended.
\r
1067 * @param pos On input: an alignment field, if desired.
\r
1068 * On output: the offsets of the alignment field.
\r
1069 * @throws IllegalArgumentException if an argument in the
\r
1070 * <code>arguments</code> array is not of the type
\r
1071 * expected by the format element(s) that use it.
\r
1072 * @throws IllegalArgumentException if this format uses named arguments
\r
1075 public final StringBuffer format(Object[] arguments, StringBuffer result,
\r
1076 FieldPosition pos)
\r
1078 if (!argumentNamesAreNumeric) {
\r
1079 throw new IllegalArgumentException(
\r
1080 "This method is not available in MessageFormat objects " +
\r
1081 "that use alphanumeric argument names.");
\r
1083 return subformat(arguments, result, pos, null);
\r
1087 * Formats a map of objects and appends the <code>MessageFormat</code>'s
\r
1088 * pattern, with format elements replaced by the formatted objects, to the
\r
1089 * provided <code>StringBuffer</code>.
\r
1091 * The text substituted for the individual format elements is derived from
\r
1092 * the current subformat of the format element and the
\r
1093 * <code>arguments</code> value corresopnding to the format element's
\r
1096 * This API may be called on formats that do not use named arguments.
\r
1097 * In this case the the keys in <code>arguments</code> must be numeric
\r
1098 * strings (e.g. "0", "1", "2"...).
\r
1100 * An argument is <i>unavailable</i> if <code>arguments</code> is
\r
1101 * <code>null</code> or does not have a value corresponding to an argument
\r
1102 * name in the pattern. When an argument is unavailable no substitution
\r
1105 * @param arguments a map of objects to be formatted and substituted.
\r
1106 * @param result where text is appended.
\r
1107 * @param pos On input: an alignment field, if desired.
\r
1108 * On output: the offsets of the alignment field.
\r
1109 * @throws IllegalArgumentException if an argument in the
\r
1110 * <code>arguments</code> array is not of the type
\r
1111 * expected by the format element(s) that use it.
\r
1112 * @return the passed-in StringBuffer
\r
1115 public final StringBuffer format(Map<String, Object> arguments, StringBuffer result,
\r
1116 FieldPosition pos) {
\r
1117 return subformat(arguments, result, pos, null);
\r
1121 * Creates a MessageFormat with the given pattern and uses it
\r
1122 * to format the given arguments. This is equivalent to
\r
1124 * <code>(new {@link #MessageFormat(String) MessageFormat}(pattern)).{@link
\r
1125 * #format(java.lang.Object[], java.lang.StringBuffer, java.text.FieldPosition)
\r
1126 * format}(arguments, new StringBuffer(), null).toString()</code>
\r
1129 * @throws IllegalArgumentException if the pattern is invalid,
\r
1130 * or if an argument in the <code>arguments</code> array
\r
1131 * is not of the type expected by the format element(s)
\r
1133 * @throws IllegalArgumentException if this format uses named arguments
\r
1136 public static String format(String pattern, Object[] arguments) {
\r
1137 MessageFormat temp = new MessageFormat(pattern);
\r
1138 return temp.format(arguments);
\r
1142 * Creates a MessageFormat with the given pattern and uses it to
\r
1143 * format the given arguments. The pattern must identifyarguments
\r
1144 * by name instead of by number.
\r
1146 * @throws IllegalArgumentException if the pattern is invalid,
\r
1147 * or if an argument in the <code>arguments</code> map
\r
1148 * is not of the type expected by the format element(s)
\r
1150 * @see #format(Map, StringBuffer, FieldPosition)
\r
1151 * @see #format(String, Object[])
\r
1154 public static String format(String pattern, Map<String, Object> arguments) {
\r
1155 MessageFormat temp = new MessageFormat(pattern);
\r
1156 return temp.format(arguments);
\r
1160 * {@icu} Returns true if this MessageFormat uses named arguments,
\r
1161 * and false otherwise. See class description.
\r
1163 * @return true if named arguments are used.
\r
1166 public boolean usesNamedArguments() {
\r
1167 return !argumentNamesAreNumeric;
\r
1172 * Formats a map or array of objects and appends the <code>MessageFormat</code>'s
\r
1173 * pattern, with format elements replaced by the formatted objects, to the
\r
1174 * provided <code>StringBuffer</code>.
\r
1175 * This is equivalent to either of
\r
1177 * <code>{@link #format(java.lang.Object[], java.lang.StringBuffer,
\r
1178 * java.text.FieldPosition) format}((Object[]) arguments, result, pos)</code>
\r
1179 * <code>{@link #format(java.util.Map, java.lang.StringBuffer,
\r
1180 * java.text.FieldPosition) format}((Map) arguments, result, pos)</code>
\r
1182 * A map must be provided if this format uses named arguments, otherwise
\r
1183 * an IllegalArgumentException will be thrown.
\r
1184 * @param arguments a map or array of objects to be formatted
\r
1185 * @param result where text is appended
\r
1186 * @param pos On input: an alignment field, if desired
\r
1187 * On output: the offsets of the alignment field
\r
1188 * @throws IllegalArgumentException if an argument in
\r
1189 * <code>arguments</code> is not of the type
\r
1190 * expected by the format element(s) that use it
\r
1191 * @throws IllegalArgumentException if <code>arguments<code> is
\r
1192 * an array of Object and this format uses named arguments
\r
1195 @SuppressWarnings("unchecked")
\r
1196 public final StringBuffer format(Object arguments, StringBuffer result,
\r
1197 FieldPosition pos)
\r
1199 if ((arguments == null || arguments instanceof Map)) {
\r
1200 return subformat((Map<String, Object>)arguments, result, pos, null);
\r
1202 if (!argumentNamesAreNumeric) {
\r
1203 throw new IllegalArgumentException(
\r
1204 "This method is not available in MessageFormat objects " +
\r
1205 "that use alphanumeric argument names.");
\r
1207 return subformat((Object[]) arguments, result, pos, null);
\r
1212 * Formats an array of objects and inserts them into the
\r
1213 * <code>MessageFormat</code>'s pattern, producing an
\r
1214 * <code>AttributedCharacterIterator</code>.
\r
1215 * You can use the returned <code>AttributedCharacterIterator</code>
\r
1216 * to build the resulting String, as well as to determine information
\r
1217 * about the resulting String.
\r
1219 * The text of the returned <code>AttributedCharacterIterator</code> is
\r
1220 * the same that would be returned by
\r
1222 * <code>{@link #format(java.lang.Object[], java.lang.StringBuffer,
\r
1223 * java.text.FieldPosition) format}(arguments, new StringBuffer(), null).toString()</code>
\r
1226 * In addition, the <code>AttributedCharacterIterator</code> contains at
\r
1227 * least attributes indicating where text was generated from an
\r
1228 * argument in the <code>arguments</code> array. The keys of these attributes are of
\r
1229 * type <code>MessageFormat.Field</code>, their values are
\r
1230 * <code>Integer</code> objects indicating the index in the <code>arguments</code>
\r
1231 * array of the argument from which the text was generated.
\r
1233 * The attributes/value from the underlying <code>Format</code>
\r
1234 * instances that <code>MessageFormat</code> uses will also be
\r
1235 * placed in the resulting <code>AttributedCharacterIterator</code>.
\r
1236 * This allows you to not only find where an argument is placed in the
\r
1237 * resulting String, but also which fields it contains in turn.
\r
1239 * @param arguments an array of objects to be formatted and substituted.
\r
1240 * @return AttributedCharacterIterator describing the formatted value.
\r
1241 * @exception NullPointerException if <code>arguments</code> is null.
\r
1242 * @exception IllegalArgumentException if an argument in the
\r
1243 * <code>arguments</code> array is not of the type
\r
1244 * expected by the format element(s) that use it.
\r
1247 @SuppressWarnings("unchecked")
\r
1248 public AttributedCharacterIterator formatToCharacterIterator(Object arguments) {
\r
1249 StringBuffer result = new StringBuffer();
\r
1250 ArrayList<AttributedCharacterIterator> iterators =
\r
1251 new ArrayList<AttributedCharacterIterator>();
\r
1253 if (arguments == null) {
\r
1254 throw new NullPointerException(
\r
1255 "formatToCharacterIterator must be passed non-null object");
\r
1257 if (arguments instanceof Map) {
\r
1258 subformat((Map<String, Object>)arguments, result, null, iterators);
\r
1260 subformat((Object[]) arguments, result, null, iterators);
\r
1262 if (iterators.size() == 0) {
\r
1263 return _createAttributedCharacterIterator("");
\r
1265 return _createAttributedCharacterIterator(
\r
1266 iterators.toArray(new AttributedCharacterIterator[iterators.size()]));
\r
1270 * Parses the string.
\r
1272 * <p>Caveats: The parse may fail in a number of circumstances.
\r
1275 * <li>If one of the arguments does not occur in the pattern.
\r
1276 * <li>If the format of an argument loses information, such as
\r
1277 * with a choice format where a large number formats to "many".
\r
1278 * <li>Does not yet handle recursion (where
\r
1279 * the substituted strings contain {n} references.)
\r
1280 * <li>Will not always find a match (or the correct match)
\r
1281 * if some part of the parse is ambiguous.
\r
1282 * For example, if the pattern "{1},{2}" is used with the
\r
1283 * string arguments {"a,b", "c"}, it will format as "a,b,c".
\r
1284 * When the result is parsed, it will return {"a", "b,c"}.
\r
1285 * <li>If a single argument is parsed more than once in the string,
\r
1286 * then the later parse wins.
\r
1288 * When the parse fails, use ParsePosition.getErrorIndex() to find out
\r
1289 * where in the string did the parsing failed. The returned error
\r
1290 * index is the starting offset of the sub-patterns that the string
\r
1291 * is comparing with. For example, if the parsing string "AAA {0} BBB"
\r
1292 * is comparing against the pattern "AAD {0} BBB", the error index is
\r
1293 * 0. When an error occurs, the call to this method will return null.
\r
1294 * If the source is null, return an empty array.
\r
1296 * This method is only supported with numbered arguments. If
\r
1297 * the format pattern used named argument an
\r
1298 * IllegalArgumentException is thrown.
\r
1300 * @throws IllegalArgumentException if this format uses named arguments
\r
1303 public Object[] parse(String source, ParsePosition pos) {
\r
1304 if (!argumentNamesAreNumeric) {
\r
1305 throw new IllegalArgumentException(
\r
1306 "This method is not available in MessageFormat objects " +
\r
1307 "that use named argument.");
\r
1309 Map<String, Object> objectMap = parseToMap(source, pos);
\r
1310 int maximumArgumentNumber = -1;
\r
1311 for (int i = 0; i <= maxOffset; i++) {
\r
1312 int argumentNumber = Integer.parseInt(argumentNames[i]);
\r
1313 if (argumentNumber > maximumArgumentNumber) {
\r
1314 maximumArgumentNumber = argumentNumber;
\r
1318 if (objectMap == null) {
\r
1322 Object[] resultArray = new Object[maximumArgumentNumber + 1];
\r
1323 for (String key : objectMap.keySet()) {
\r
1324 resultArray[Integer.parseInt(key)] = objectMap.get(key);
\r
1327 return resultArray;
\r
1331 * {@icu} Parses the string, returning the results in a Map.
\r
1332 * This is similar to the version that returns an array
\r
1333 * of Object. This supports both named and numbered
\r
1334 * arguments-- if numbered, the keys in the map are the
\r
1335 * corresponding Strings (e.g. "0", "1", "2"...).
\r
1337 * @param source the text to parse
\r
1338 * @param pos the position at which to start parsing. on return,
\r
1339 * contains the result of the parse.
\r
1340 * @return a Map containing key/value pairs for each parsed argument.
\r
1343 public Map<String, Object> parseToMap(String source, ParsePosition pos) {
\r
1344 if (source == null) {
\r
1345 Map<String, Object> empty = new HashMap<String, Object>();
\r
1349 Map<String, Object> resultMap = new HashMap<String, Object>();
\r
1351 int patternOffset = 0;
\r
1352 int sourceOffset = pos.getIndex();
\r
1353 ParsePosition tempStatus = new ParsePosition(0);
\r
1354 for (int i = 0; i <= maxOffset; ++i) {
\r
1355 // match up to format
\r
1356 int len = offsets[i] - patternOffset;
\r
1357 if (len == 0 || pattern.regionMatches(patternOffset,
\r
1358 source, sourceOffset, len)) {
\r
1359 sourceOffset += len;
\r
1360 patternOffset += len;
\r
1362 pos.setErrorIndex(sourceOffset);
\r
1363 return null; // leave index as is to signal error
\r
1367 if (formats[i] == null) { // string format
\r
1368 // if at end, use longest possible match
\r
1369 // otherwise uses first match to intervening string
\r
1370 // does NOT recursively try all possibilities
\r
1371 int tempLength = (i != maxOffset) ? offsets[i+1] : pattern.length();
\r
1374 if (patternOffset >= tempLength) {
\r
1375 next = source.length();
\r
1377 next = source.indexOf( pattern.substring(patternOffset,tempLength), sourceOffset);
\r
1381 pos.setErrorIndex(sourceOffset);
\r
1382 return null; // leave index as is to signal error
\r
1384 String strValue = source.substring(sourceOffset, next);
\r
1385 if (!strValue.equals("{" + argumentNames[i] + "}"))
\r
1386 resultMap.put(argumentNames[i], source.substring(sourceOffset, next));
\r
1387 // resultArray[Integer.parseInt(argumentNames[i])] =
\r
1388 // source.substring(sourceOffset, next);
\r
1389 sourceOffset = next;
\r
1392 tempStatus.setIndex(sourceOffset);
\r
1393 resultMap.put(argumentNames[i], formats[i].parseObject(source, tempStatus));
\r
1394 // resultArray[Integer.parseInt(argumentNames[i])] =
\r
1395 // formats[i].parseObject(source, tempStatus);
\r
1396 if (tempStatus.getIndex() == sourceOffset) {
\r
1397 pos.setErrorIndex(sourceOffset);
\r
1398 return null; // leave index as is to signal error
\r
1400 sourceOffset = tempStatus.getIndex(); // update
\r
1403 int len = pattern.length() - patternOffset;
\r
1404 if (len == 0 || pattern.regionMatches(patternOffset,
\r
1405 source, sourceOffset, len)) {
\r
1406 pos.setIndex(sourceOffset + len);
\r
1408 pos.setErrorIndex(sourceOffset);
\r
1409 return null; // leave index as is to signal error
\r
1415 * Parses text from the beginning of the given string to produce an object
\r
1417 * The method may not use the entire text of the given string.
\r
1419 * See the {@link #parse(String, ParsePosition)} method for more information
\r
1420 * on message parsing.
\r
1422 * @param source A <code>String</code> whose beginning should be parsed.
\r
1423 * @return An <code>Object</code> array parsed from the string.
\r
1424 * @exception ParseException if the beginning of the specified string cannot be parsed.
\r
1425 * @exception IllegalArgumentException if this format uses named arguments
\r
1428 public Object[] parse(String source) throws ParseException {
\r
1429 ParsePosition pos = new ParsePosition(0);
\r
1430 Object[] result = parse(source, pos);
\r
1431 if (pos.getIndex() == 0) // unchanged, returned object is null
\r
1432 throw new ParseException("MessageFormat parse error!",
\r
1433 pos.getErrorIndex());
\r
1439 * {@icu} Parses text from the beginning of the given string to produce a map from
\r
1440 * argument to values. The method may not use the entire text of the given string.
\r
1442 * <p>See the {@link #parse(String, ParsePosition)} method for more information on
\r
1443 * message parsing.
\r
1445 * @param source A <code>String</code> whose beginning should be parsed.
\r
1446 * @return A <code>Map</code> parsed from the string.
\r
1447 * @throws ParseException if the beginning of the specified string cannot
\r
1449 * @see #parseToMap(String, ParsePosition)
\r
1452 public Map<String, Object> parseToMap(String source) throws ParseException {
\r
1454 ParsePosition pos = new ParsePosition(0);
\r
1455 Map<String, Object> result = parseToMap(source, pos);
\r
1456 if (pos.getIndex() == 0) // unchanged, returned object is null
\r
1457 throw new ParseException("MessageFormat parse error!",
\r
1458 pos.getErrorIndex());
\r
1464 * Parses text from a string to produce an object array or Map.
\r
1466 * The method attempts to parse text starting at the index given by
\r
1467 * <code>pos</code>.
\r
1468 * If parsing succeeds, then the index of <code>pos</code> is updated
\r
1469 * to the index after the last character used (parsing does not necessarily
\r
1470 * use all characters up to the end of the string), and the parsed
\r
1471 * object array is returned. The updated <code>pos</code> can be used to
\r
1472 * indicate the starting point for the next call to this method.
\r
1473 * If an error occurs, then the index of <code>pos</code> is not
\r
1474 * changed, the error index of <code>pos</code> is set to the index of
\r
1475 * the character where the error occurred, and null is returned.
\r
1477 * See the {@link #parse(String, ParsePosition)} method for more information
\r
1478 * on message parsing.
\r
1480 * @param source A <code>String</code>, part of which should be parsed.
\r
1481 * @param pos A <code>ParsePosition</code> object with index and error
\r
1482 * index information as described above.
\r
1483 * @return An <code>Object</code> parsed from the string, either an
\r
1484 * array of Object, or a Map, depending on whether named
\r
1485 * arguments are used. This can be queried using <code>usesNamedArguments</code>.
\r
1486 * In case of error, returns null.
\r
1487 * @throws NullPointerException if <code>pos</code> is null.
\r
1490 public Object parseObject(String source, ParsePosition pos) {
\r
1491 if (argumentNamesAreNumeric) {
\r
1492 return parse(source, pos);
\r
1494 return parseToMap(source, pos);
\r
1499 * Overrides clone.
\r
1501 * @return a clone of this instance.
\r
1504 public Object clone() {
\r
1505 MessageFormat other = (MessageFormat) super.clone();
\r
1507 // clone arrays. Can't do with utility because of bug in Cloneable
\r
1508 other.formats = formats.clone(); // shallow clone
\r
1509 for (int i = 0; i < formats.length; ++i) {
\r
1510 if (formats[i] != null)
\r
1511 other.formats[i] = (Format) formats[i].clone();
\r
1513 // for primitives or immutables, shallow clone is enough
\r
1514 other.offsets = offsets.clone();
\r
1515 other.argumentNames = argumentNames.clone();
\r
1516 other.argumentNamesAreNumeric = argumentNamesAreNumeric;
\r
1522 * Overrides equals.
\r
1525 public boolean equals(Object obj) {
\r
1526 if (this == obj) // quick check
\r
1528 if (obj == null || getClass() != obj.getClass())
\r
1530 MessageFormat other = (MessageFormat) obj;
\r
1531 return (maxOffset == other.maxOffset
\r
1532 && pattern.equals(other.pattern)
\r
1533 && Utility.objectEquals(ulocale, other.ulocale) // does null check
\r
1534 && Utility.arrayEquals(offsets, other.offsets)
\r
1535 && Utility.arrayEquals(argumentNames, other.argumentNames)
\r
1536 && Utility.arrayEquals(formats, other.formats)
\r
1537 && (argumentNamesAreNumeric == other.argumentNamesAreNumeric));
\r
1541 * Overrides hashCode.
\r
1544 public int hashCode() {
\r
1545 return pattern.hashCode(); // enough for reasonable distribution
\r
1549 * Defines constants that are used as attribute keys in the
\r
1550 * <code>AttributedCharacterIterator</code> returned
\r
1551 * from <code>MessageFormat.formatToCharacterIterator</code>.
\r
1555 public static class Field extends Format.Field {
\r
1557 private static final long serialVersionUID = 7510380454602616157L;
\r
1560 * Create a <code>Field</code> with the specified name.
\r
1562 * @param name The name of the attribute
\r
1566 protected Field(String name) {
\r
1571 * Resolves instances being deserialized to the predefined constants.
\r
1573 * @return resolved MessageFormat.Field constant
\r
1574 * @throws InvalidObjectException if the constant could not be resolved.
\r
1578 protected Object readResolve() throws InvalidObjectException {
\r
1579 if (this.getClass() != MessageFormat.Field.class) {
\r
1580 throw new InvalidObjectException(
\r
1581 "A subclass of MessageFormat.Field must implement readResolve.");
\r
1583 if (this.getName().equals(ARGUMENT.getName())) {
\r
1586 throw new InvalidObjectException("Unknown attribute name.");
\r
1591 * Constant identifying a portion of a message that was generated
\r
1592 * from an argument passed into <code>formatToCharacterIterator</code>.
\r
1593 * The value associated with the key will be an <code>Integer</code>
\r
1594 * indicating the index in the <code>arguments</code> array of the
\r
1595 * argument from which the text was generated.
\r
1599 public static final Field ARGUMENT = new Field("message argument field");
\r
1603 // ===========================privates============================
\r
1606 * The locale to use for formatting numbers and dates.
\r
1607 * This is no longer used, and here only for serialization compatibility.
\r
1610 private Locale locale;
\r
1613 * The locale to use for formatting numbers and dates.
\r
1616 private ULocale ulocale;
\r
1619 * The string that the formatted values are to be plugged into. In other words, this
\r
1620 * is the pattern supplied on construction with all of the {} expressions taken out.
\r
1623 private String pattern = "";
\r
1625 /** The initially expected number of subformats in the format */
\r
1626 private static final int INITIAL_FORMATS = 10;
\r
1629 * An array of formatters, which are used to format the arguments.
\r
1632 private Format[] formats = new Format[INITIAL_FORMATS];
\r
1635 * The positions where the results of formatting each argument are to be
\r
1636 * inserted into the pattern.
\r
1640 private int[] offsets = new int[INITIAL_FORMATS];
\r
1643 * The argument numbers corresponding to each formatter. (The formatters are stored
\r
1644 * in the order they occur in the pattern, not in the order in which the arguments
\r
1648 // retained for backwards compatibility
\r
1649 private int[] argumentNumbers = new int[INITIAL_FORMATS];
\r
1652 * The argument names corresponding to each formatter. (The formatters are
\r
1653 * stored in the order they occur in the pattern, not in the order in which
\r
1654 * the arguments are specified.)
\r
1658 private String[] argumentNames = new String[INITIAL_FORMATS];
\r
1661 * Is true iff all argument names are non-negative numbers.
\r
1665 private boolean argumentNamesAreNumeric = true;
\r
1668 * One less than the number of entries in <code>offsets</code>. Can also be thought of
\r
1669 * as the index of the highest-numbered element in <code>offsets</code> that is being used.
\r
1670 * All of these arrays should have the same number of elements being used as <code>offsets</code>
\r
1671 * does, and so this variable suffices to tell us how many entries are in all of them.
\r
1674 private int maxOffset = -1;
\r
1677 * Internal routine used by format. If <code>characterIterators</code> is
\r
1678 * non-null, AttributedCharacterIterator will be created from the
\r
1679 * subformats as necessary. If <code>characterIterators</code> is null
\r
1680 * and <code>fp</code> is non-null and identifies
\r
1681 * <code>Field.MESSAGE_ARGUMENT</code>, the location of
\r
1682 * the first replaced argument will be set in it.
\r
1684 * @exception IllegalArgumentException if an argument in the
\r
1685 * <code>arguments</code> array is not of the type
\r
1686 * expected by the format element(s) that use it.
\r
1688 private StringBuffer subformat(Object[] arguments, StringBuffer result,
\r
1689 FieldPosition fp, List<AttributedCharacterIterator> characterIterators) {
\r
1690 return subformat(arrayToMap(arguments), result, fp, characterIterators);
\r
1694 * Internal routine used by format.
\r
1696 * @throws IllegalArgumentException if an argument in the
\r
1697 * <code>arguments</code> map is not of the type
\r
1698 * expected by the format element(s) that use it.
\r
1700 private StringBuffer subformat(Map<String, Object> arguments, StringBuffer result,
\r
1701 FieldPosition fp, List<AttributedCharacterIterator> characterIterators) {
\r
1702 // note: this implementation assumes a fast substring & index.
\r
1703 // if this is not true, would be better to append chars one by one.
\r
1704 int lastOffset = 0;
\r
1705 int last = result.length();
\r
1707 for (int i = 0; i <= maxOffset; ++i) {
\r
1708 result.append(pattern.substring(lastOffset, offsets[i]));
\r
1709 lastOffset = offsets[i];
\r
1710 String argumentName = argumentNames[i];
\r
1711 if (arguments == null || !arguments.containsKey(argumentName)) {
\r
1712 result.append("{" + argumentName + "}");
\r
1715 // int argRecursion = ((recursionProtection >> (argumentNumber*2)) & 0x3);
\r
1716 //Eclipse stated the following is "dead code"
\r
1717 /*if (false) { // if (argRecursion == 3){
\r
1718 // prevent loop!!!
\r
1719 result.append('\uFFFD');
\r
1721 Object obj = arguments.get(argumentName);
\r
1722 String arg = null;
\r
1723 Format subFormatter = null;
\r
1724 if (obj == null) {
\r
1726 } else if (formats[i] != null) {
\r
1727 subFormatter = formats[i];
\r
1728 if (subFormatter instanceof ChoiceFormat
\r
1729 || subFormatter instanceof PluralFormat
\r
1730 || subFormatter instanceof SelectFormat) {
\r
1731 arg = formats[i].format(obj);
\r
1732 // TODO: This should be made more robust.
\r
1733 // Does this work with '{' in quotes?
\r
1734 if (arg.indexOf('{') >= 0) {
\r
1735 subFormatter = new MessageFormat(arg, ulocale);
\r
1740 } else if (obj instanceof Number) {
\r
1741 // format number if can
\r
1742 subFormatter = NumberFormat.getInstance(ulocale);
\r
1743 } else if (obj instanceof Date) {
\r
1744 // format a Date if can
\r
1745 subFormatter = DateFormat.getDateTimeInstance(
\r
1746 DateFormat.SHORT, DateFormat.SHORT, ulocale);//fix
\r
1747 } else if (obj instanceof String) {
\r
1748 arg = (String) obj;
\r
1751 arg = obj.toString();
\r
1752 if (arg == null) arg = "null";
\r
1755 // At this point we are in two states, either subFormatter
\r
1756 // is non-null indicating we should format obj using it,
\r
1757 // or arg is non-null and we should use it as the value.
\r
1759 if (characterIterators != null) {
\r
1760 // If characterIterators is non-null, it indicates we need
\r
1761 // to get the CharacterIterator from the child formatter.
\r
1762 if (last != result.length()) {
\r
1763 characterIterators.add(
\r
1764 _createAttributedCharacterIterator(result.substring
\r
1766 last = result.length();
\r
1768 if (subFormatter != null) {
\r
1769 AttributedCharacterIterator subIterator =
\r
1770 subFormatter.formatToCharacterIterator(obj);
\r
1772 append(result, subIterator);
\r
1773 if (last != result.length()) {
\r
1774 characterIterators.add(
\r
1775 _createAttributedCharacterIterator(
\r
1776 subIterator, Field.ARGUMENT,
\r
1777 argumentNamesAreNumeric ?
\r
1778 (Object)new Integer(argumentName) :
\r
1779 (Object)argumentName));
\r
1780 last = result.length();
\r
1784 if (arg != null && arg.length() > 0) {
\r
1785 result.append(arg);
\r
1786 characterIterators.add(
\r
1787 _createAttributedCharacterIterator(
\r
1788 arg, Field.ARGUMENT,
\r
1789 argumentNamesAreNumeric ?
\r
1790 (Object)new Integer(argumentName) :
\r
1791 (Object)argumentName));
\r
1792 last = result.length();
\r
1795 if (subFormatter != null) {
\r
1796 arg = subFormatter.format(obj);
\r
1798 last = result.length();
\r
1799 result.append(arg);
\r
1800 if (i == 0 && fp != null && Field.ARGUMENT.equals(
\r
1801 fp.getFieldAttribute())) {
\r
1802 fp.setBeginIndex(last);
\r
1803 fp.setEndIndex(result.length());
\r
1805 last = result.length();
\r
1809 result.append(pattern.substring(lastOffset, pattern.length()));
\r
1810 if (characterIterators != null && last != result.length()) {
\r
1811 characterIterators.add(_createAttributedCharacterIterator(
\r
1812 result.substring(last)));
\r
1818 * Convenience method to append all the characters in
\r
1819 * <code>iterator</code> to the StringBuffer <code>result</code>.
\r
1821 private void append(StringBuffer result, CharacterIterator iterator) {
\r
1822 if (iterator.first() != CharacterIterator.DONE) {
\r
1825 result.append(iterator.first());
\r
1826 while ((aChar = iterator.next()) != CharacterIterator.DONE) {
\r
1827 result.append(aChar);
\r
1832 private static final String[] typeList =
\r
1833 {"", "number", "date", "time", "choice", "spellout", "ordinal",
\r
1834 "duration", "plural", "select" };
\r
1835 private static final int
\r
1841 TYPE_SPELLOUT = 5,
\r
1843 TYPE_DURATION = 7,
\r
1847 private static final String[] modifierList =
\r
1848 {"", "currency", "percent", "integer"};
\r
1850 private static final int
\r
1851 MODIFIER_EMPTY = 0,
\r
1852 MODIFIER_CURRENCY = 1,
\r
1853 MODIFIER_PERCENT = 2,
\r
1854 MODIFIER_INTEGER = 3;
\r
1856 private static final String[] dateModifierList =
\r
1857 {"", "short", "medium", "long", "full"};
\r
1859 private static final int
\r
1860 DATE_MODIFIER_EMPTY = 0,
\r
1861 DATE_MODIFIER_SHORT = 1,
\r
1862 DATE_MODIFIER_MEDIUM = 2,
\r
1863 DATE_MODIFIER_LONG = 3,
\r
1864 DATE_MODIFIER_FULL = 4;
\r
1866 private void makeFormat(int position, int offsetNumber,
\r
1867 StringBuilder[] segments)
\r
1869 // get the argument number
\r
1870 // int argumentNumber;
\r
1872 // argumentNumber = Integer.parseInt(segments[1].toString()); // always unlocalized!
\r
1873 // } catch (NumberFormatException e) {
\r
1874 // throw new IllegalArgumentException("can't parse argument number "
\r
1875 // + segments[1]);
\r
1877 // if (argumentNumber < 0) {
\r
1878 // throw new IllegalArgumentException("negative argument number "
\r
1879 // + argumentNumber);
\r
1882 // resize format information arrays if necessary
\r
1883 if (offsetNumber >= formats.length) {
\r
1884 int newLength = formats.length * 2;
\r
1885 Format[] newFormats = new Format[newLength];
\r
1886 int[] newOffsets = new int[newLength];
\r
1887 String[] newArgumentNames = new String[newLength];
\r
1888 System.arraycopy(formats, 0, newFormats, 0, maxOffset + 1);
\r
1889 System.arraycopy(offsets, 0, newOffsets, 0, maxOffset + 1);
\r
1890 System.arraycopy(argumentNames, 0, newArgumentNames, 0,
\r
1892 formats = newFormats;
\r
1893 offsets = newOffsets;
\r
1894 argumentNames = newArgumentNames;
\r
1896 int oldMaxOffset = maxOffset;
\r
1897 maxOffset = offsetNumber;
\r
1898 offsets[offsetNumber] = segments[0].length();
\r
1899 argumentNames[offsetNumber] = segments[1].toString();
\r
1900 // All argument names numeric ?
\r
1901 int argumentNumber;
\r
1903 // always unlocalized!
\r
1904 argumentNumber = Integer.parseInt(segments[1].toString());
\r
1905 } catch (NumberFormatException e) {
\r
1906 argumentNumber = -1;
\r
1908 if (offsetNumber == 0) {
\r
1909 // First argument determines whether all argument identifiers have
\r
1910 // to be numbers or (IDStartChars IDContChars*) strings.
\r
1911 argumentNamesAreNumeric = argumentNumber >= 0;
\r
1914 if (argumentNamesAreNumeric && argumentNumber < 0 ||
\r
1915 !argumentNamesAreNumeric &&
\r
1916 !isAlphaIdentifier(argumentNames[offsetNumber])) {
\r
1917 throw new IllegalArgumentException(
\r
1918 "All argument identifiers have to be either non-negative " +
\r
1919 "numbers or strings following the pattern " +
\r
1920 "([:ID_Start:] [:ID_Continue:]*).\n" +
\r
1921 "For more details on these unicode sets, visit " +
\r
1922 "http://demo.icu-project.org/icu-bin/ubrowse");
\r
1925 // now get the format
\r
1926 Format newFormat = null;
\r
1927 int subformatType = findKeyword(segments[2].toString(), typeList);
\r
1928 switch (subformatType){
\r
1932 switch (findKeyword(segments[3].toString(), modifierList)) {
\r
1933 case MODIFIER_EMPTY:
\r
1934 newFormat = NumberFormat.getInstance(ulocale);
\r
1936 case MODIFIER_CURRENCY:
\r
1937 newFormat = NumberFormat.getCurrencyInstance(ulocale);
\r
1939 case MODIFIER_PERCENT:
\r
1940 newFormat = NumberFormat.getPercentInstance(ulocale);
\r
1942 case MODIFIER_INTEGER:
\r
1943 newFormat = NumberFormat.getIntegerInstance(ulocale);
\r
1945 default: // pattern
\r
1946 newFormat = new DecimalFormat(segments[3].toString(),
\r
1947 new DecimalFormatSymbols(ulocale));
\r
1952 switch (findKeyword(segments[3].toString(), dateModifierList)) {
\r
1953 case DATE_MODIFIER_EMPTY:
\r
1954 newFormat = DateFormat.getDateInstance(DateFormat.DEFAULT, ulocale);
\r
1956 case DATE_MODIFIER_SHORT:
\r
1957 newFormat = DateFormat.getDateInstance(DateFormat.SHORT, ulocale);
\r
1959 case DATE_MODIFIER_MEDIUM:
\r
1960 newFormat = DateFormat.getDateInstance(DateFormat.DEFAULT, ulocale);
\r
1962 case DATE_MODIFIER_LONG:
\r
1963 newFormat = DateFormat.getDateInstance(DateFormat.LONG, ulocale);
\r
1965 case DATE_MODIFIER_FULL:
\r
1966 newFormat = DateFormat.getDateInstance(DateFormat.FULL, ulocale);
\r
1969 newFormat = new SimpleDateFormat(segments[3].toString(), ulocale);
\r
1974 switch (findKeyword(segments[3].toString(), dateModifierList)) {
\r
1975 case DATE_MODIFIER_EMPTY:
\r
1976 newFormat = DateFormat.getTimeInstance(DateFormat.DEFAULT, ulocale);
\r
1978 case DATE_MODIFIER_SHORT:
\r
1979 newFormat = DateFormat.getTimeInstance(DateFormat.SHORT, ulocale);
\r
1981 case DATE_MODIFIER_MEDIUM:
\r
1982 newFormat = DateFormat.getTimeInstance(DateFormat.DEFAULT, ulocale);
\r
1984 case DATE_MODIFIER_LONG:
\r
1985 newFormat = DateFormat.getTimeInstance(DateFormat.LONG, ulocale);
\r
1987 case DATE_MODIFIER_FULL:
\r
1988 newFormat = DateFormat.getTimeInstance(DateFormat.FULL, ulocale);
\r
1991 newFormat = new SimpleDateFormat(segments[3].toString(), ulocale);
\r
1997 newFormat = new ChoiceFormat(segments[3].toString());
\r
1998 } catch (Exception e) {
\r
1999 maxOffset = oldMaxOffset;
\r
2000 throw new IllegalArgumentException("Choice Pattern incorrect");
\r
2003 case TYPE_SPELLOUT:
\r
2005 RuleBasedNumberFormat rbnf = new RuleBasedNumberFormat(ulocale,
\r
2006 RuleBasedNumberFormat.SPELLOUT);
\r
2007 String ruleset = segments[3].toString().trim();
\r
2008 if (ruleset.length() != 0) {
\r
2010 rbnf.setDefaultRuleSet(ruleset);
\r
2012 catch (Exception e) {
\r
2013 // warn invalid ruleset
\r
2019 case TYPE_ORDINAL:
\r
2021 RuleBasedNumberFormat rbnf = new RuleBasedNumberFormat(ulocale,
\r
2022 RuleBasedNumberFormat.ORDINAL);
\r
2023 String ruleset = segments[3].toString().trim();
\r
2024 if (ruleset.length() != 0) {
\r
2026 rbnf.setDefaultRuleSet(ruleset);
\r
2028 catch (Exception e) {
\r
2029 // warn invalid ruleset
\r
2035 case TYPE_DURATION:
\r
2037 RuleBasedNumberFormat rbnf = new RuleBasedNumberFormat(ulocale,
\r
2038 RuleBasedNumberFormat.DURATION);
\r
2039 String ruleset = segments[3].toString().trim();
\r
2040 if (ruleset.length() != 0) {
\r
2042 rbnf.setDefaultRuleSet(ruleset);
\r
2044 catch (Exception e) {
\r
2045 // warn invalid ruleset
\r
2054 // PluralFormat and SelectFormat does not handle quotes.
\r
2056 // TODO: Should PluralFormat and SelectFormat handle quotes?
\r
2057 StringBuilder unquotedPattern = new StringBuilder();
\r
2058 String quotedPattern = segments[3].toString();
\r
2059 boolean inQuote = false;
\r
2060 for (int i = 0; i < quotedPattern.length(); ++i) {
\r
2061 char ch = quotedPattern.charAt(i);
\r
2063 if (i+1 < quotedPattern.length() &&
\r
2064 quotedPattern.charAt(i+1) == '\'') {
\r
2065 unquotedPattern.append(ch);
\r
2068 inQuote = !inQuote;
\r
2071 unquotedPattern.append(ch);
\r
2075 if( subformatType == TYPE_PLURAL){
\r
2076 PluralFormat pls = new PluralFormat(ulocale,
\r
2077 unquotedPattern.toString());
\r
2080 newFormat = new SelectFormat( unquotedPattern.toString());
\r
2085 maxOffset = oldMaxOffset;
\r
2086 throw new IllegalArgumentException("unknown format type at ");
\r
2088 formats[offsetNumber] = newFormat;
\r
2089 segments[1].setLength(0); // throw away other segments
\r
2090 segments[2].setLength(0);
\r
2091 segments[3].setLength(0);
\r
2094 private static final int findKeyword(String s, String[] list) {
\r
2095 s = s.trim().toLowerCase();
\r
2096 for (int i = 0; i < list.length; ++i) {
\r
2097 if (s.equals(list[i]))
\r
2103 private static final void copyAndFixQuotes(String source, int start, int end,
\r
2104 StringBuilder target) {
\r
2105 // added 'gotLB' logic from ICU4C - questionable [alan]
\r
2106 boolean gotLB = false;
\r
2107 for (int i = start; i < end; ++i) {
\r
2108 char ch = source.charAt(i);
\r
2110 target.append("'{'");
\r
2112 } else if (ch == '}') {
\r
2114 target.append(ch);
\r
2117 target.append("'}'");
\r
2119 } else if (ch == '\'') {
\r
2120 target.append("''");
\r
2122 target.append(ch);
\r
2128 * After reading an object from the input stream, do a simple verification
\r
2129 * to maintain class invariants.
\r
2130 * @throws InvalidObjectException if the objects read from the stream is invalid.
\r
2132 private void readObject(ObjectInputStream in) throws IOException, ClassNotFoundException {
\r
2133 in.defaultReadObject();
\r
2134 if (argumentNames == null) { // name mod, rev
\r
2135 argumentNamesAreNumeric = true;
\r
2136 argumentNames = new String[argumentNumbers.length];
\r
2137 for (int i = 0; i < argumentNumbers.length; ++i) {
\r
2138 argumentNames[i] = String.valueOf(argumentNumbers[i]);
\r
2141 boolean isValid = maxOffset >= -1
\r
2142 && formats.length > maxOffset
\r
2143 && offsets.length > maxOffset
\r
2144 && argumentNames.length > maxOffset;
\r
2146 int lastOffset = pattern.length() + 1;
\r
2147 for (int i = maxOffset; i >= 0; --i) {
\r
2148 if ((offsets[i] < 0) || (offsets[i] > lastOffset)) {
\r
2152 lastOffset = offsets[i];
\r
2157 throw new InvalidObjectException(
\r
2158 "Could not reconstruct MessageFormat from corrupt stream.");
\r
2160 if (ulocale == null) {
\r
2161 ulocale = ULocale.forLocale(locale);
\r
2166 * This is a helper method for converting an object array into a map. The
\r
2167 * key set of the map is [0, ..., array.length]. The value associated with
\r
2168 * each key is the ith entry of the passed object array.
\r
2170 * @throws InvalidObjectException
\r
2171 * if the objects read from the stream is invalid.
\r
2173 private Map<String, Object> arrayToMap(Object[] array) {
\r
2174 Map<String, Object> map = new HashMap<String, Object>();
\r
2175 if (array != null) {
\r
2176 for (int i = 0; i < array.length; ++i) {
\r
2177 map.put(Integer.toString(i), array[i]);
\r
2183 private boolean isAlphaIdentifier(String argument) {
\r
2184 if (argument.length() == 0) {
\r
2187 for (int i = 0; i < argument.length(); ++i ) {
\r
2188 if (i == 0 && !UCharacter.isUnicodeIdentifierStart(argument.charAt(i)) ||
\r
2189 i > 0 && !UCharacter.isUnicodeIdentifierPart(argument.charAt(i))){
\r
2196 private static final char SINGLE_QUOTE = '\'';
\r
2197 private static final char CURLY_BRACE_LEFT = '{';
\r
2198 private static final char CURLY_BRACE_RIGHT = '}';
\r
2200 private static final int STATE_INITIAL = 0;
\r
2201 private static final int STATE_SINGLE_QUOTE = 1;
\r
2202 private static final int STATE_IN_QUOTE = 2;
\r
2203 private static final int STATE_MSG_ELEMENT = 3;
\r
2206 * {@icu} Converts an 'apostrophe-friendly' pattern into a standard
\r
2207 * pattern. Standard patterns treat all apostrophes as
\r
2208 * quotes, which is problematic in some languages, e.g.
\r
2209 * French, where apostrophe is commonly used. This utility
\r
2210 * assumes that only an unpaired apostrophe immediately before
\r
2211 * a brace is a true quote. Other unpaired apostrophes are paired,
\r
2212 * and the resulting standard pattern string is returned.
\r
2214 * <p><b>Note</b> it is not guaranteed that the returned pattern
\r
2215 * is indeed a valid pattern. The only effect is to convert
\r
2216 * between patterns having different quoting semantics.
\r
2218 * @param pattern the 'apostrophe-friendly' patttern to convert
\r
2219 * @return the standard equivalent of the original pattern
\r
2222 public static String autoQuoteApostrophe(String pattern) {
\r
2223 StringBuilder buf = new StringBuilder(pattern.length() * 2);
\r
2224 int state = STATE_INITIAL;
\r
2225 int braceCount = 0;
\r
2226 for (int i = 0, j = pattern.length(); i < j; ++i) {
\r
2227 char c = pattern.charAt(i);
\r
2229 case STATE_INITIAL:
\r
2231 case SINGLE_QUOTE:
\r
2232 state = STATE_SINGLE_QUOTE;
\r
2234 case CURLY_BRACE_LEFT:
\r
2235 state = STATE_MSG_ELEMENT;
\r
2240 case STATE_SINGLE_QUOTE:
\r
2242 case SINGLE_QUOTE:
\r
2243 state = STATE_INITIAL;
\r
2245 case CURLY_BRACE_LEFT:
\r
2246 case CURLY_BRACE_RIGHT:
\r
2247 state = STATE_IN_QUOTE;
\r
2250 buf.append(SINGLE_QUOTE);
\r
2251 state = STATE_INITIAL;
\r
2255 case STATE_IN_QUOTE:
\r
2257 case SINGLE_QUOTE:
\r
2258 state = STATE_INITIAL;
\r
2262 case STATE_MSG_ELEMENT:
\r
2264 case CURLY_BRACE_LEFT:
\r
2267 case CURLY_BRACE_RIGHT:
\r
2268 if (--braceCount == 0) {
\r
2269 state = STATE_INITIAL;
\r
2275 default: // Never happens.
\r
2282 if (state == STATE_SINGLE_QUOTE || state == STATE_IN_QUOTE) {
\r
2283 buf.append(SINGLE_QUOTE);
\r
2285 return new String(buf);
\r
2289 // private methods for AttributedCharacterIterator support
\r
2291 // Note: The equivalent methods are defined as package local methods in
\r
2292 // java.text.Format. ICU cannot access these methods, so we have
\r
2293 // these methods locally, with "_" prefix for avoiding name collision.
\r
2294 // (The collision itself is not a problem, but Eclipse displays warnings
\r
2295 // by the default warning level.) We may move these utility methods
\r
2296 // up to com.ibm.icu.text.UFormat later. Yoshito
\r
2298 private static AttributedCharacterIterator _createAttributedCharacterIterator(String text) {
\r
2299 AttributedString as = new AttributedString(text);
\r
2300 return as.getIterator();
\r
2303 private static AttributedCharacterIterator _createAttributedCharacterIterator(
\r
2304 AttributedCharacterIterator[] iterators) {
\r
2305 if (iterators == null || iterators.length == 0) {
\r
2306 return _createAttributedCharacterIterator("");
\r
2308 // Create a single AttributedString
\r
2309 StringBuilder sb = new StringBuilder();
\r
2310 for (int i = 0; i < iterators.length; i++) {
\r
2311 int index = iterators[i].getBeginIndex();
\r
2312 int end = iterators[i].getEndIndex();
\r
2313 while (index < end) {
\r
2314 sb.append(iterators[i].setIndex(index++));
\r
2317 AttributedString as = new AttributedString(sb.toString());
\r
2321 for (int i = 0; i < iterators.length; i++) {
\r
2322 iterators[i].first();
\r
2323 int start = iterators[i].getBeginIndex();
\r
2325 Map<Attribute, Object> map = iterators[i].getAttributes();
\r
2326 int len = iterators[i].getRunLimit() - start; // run length
\r
2327 if (map.size() > 0) {
\r
2328 for (Map.Entry<Attribute, Object> entry : map.entrySet()) {
\r
2329 as.addAttribute(entry.getKey(), entry.getValue(),
\r
2330 offset, offset + len);
\r
2335 iterators[i].setIndex(start);
\r
2336 if (iterators[i].current() == CharacterIterator.DONE) {
\r
2342 return as.getIterator();
\r
2345 private static AttributedCharacterIterator _createAttributedCharacterIterator(
\r
2346 AttributedCharacterIterator iterator,
\r
2347 AttributedCharacterIterator.Attribute key, Object value) {
\r
2348 AttributedString as = new AttributedString(iterator);
\r
2349 as.addAttribute(key, value);
\r
2350 return as.getIterator();
\r
2353 private static AttributedCharacterIterator _createAttributedCharacterIterator(String text,
\r
2354 AttributedCharacterIterator.Attribute key, Object value) {
\r
2355 AttributedString as = new AttributedString(text);
\r
2356 as.addAttribute(key, value);
\r
2357 return as.getIterator();
\r
projects / Dictionary.git / blob