1// © 2016 and later: Unicode, Inc. and others.
2// License & terms of use: http://www.unicode.org/copyright.html
3/*
4**********************************************************************
5* Copyright (C) 1999-2014, International Business Machines
6* Corporation and others. All Rights Reserved.
7**********************************************************************
8* Date Name Description
9* 11/17/99 aliu Creation.
10**********************************************************************
11*/
12#ifndef TRANSLIT_H
13#define TRANSLIT_H
14
15#include "unicode/utypes.h"
16
17#if U_SHOW_CPLUSPLUS_API
18
19/**
20 * \file
21 * \brief C++ API: Tranforms text from one format to another.
22 */
23
24#if !UCONFIG_NO_TRANSLITERATION
25
26#include "unicode/uobject.h"
27#include "unicode/unistr.h"
28#include "unicode/parseerr.h"
29#include "unicode/utrans.h" // UTransPosition, UTransDirection
30#include "unicode/strenum.h"
31
32U_NAMESPACE_BEGIN
33
34class UnicodeFilter;
35class UnicodeSet;
36class TransliteratorParser;
37class NormalizationTransliterator;
38class TransliteratorIDParser;
39
40/**
41 *
42 * <code>Transliterator</code> is an abstract class that
43 * transliterates text from one format to another. The most common
44 * kind of transliterator is a script, or alphabet, transliterator.
45 * For example, a Russian to Latin transliterator changes Russian text
46 * written in Cyrillic characters to phonetically equivalent Latin
47 * characters. It does not <em>translate</em> Russian to English!
48 * Transliteration, unlike translation, operates on characters, without
49 * reference to the meanings of words and sentences.
50 *
51 * <p>Although script conversion is its most common use, a
52 * transliterator can actually perform a more general class of tasks.
53 * In fact, <code>Transliterator</code> defines a very general API
54 * which specifies only that a segment of the input text is replaced
55 * by new text. The particulars of this conversion are determined
56 * entirely by subclasses of <code>Transliterator</code>.
57 *
58 * <p><b>Transliterators are stateless</b>
59 *
60 * <p><code>Transliterator</code> objects are <em>stateless</em>; they
61 * retain no information between calls to
62 * <code>transliterate()</code>. (However, this does <em>not</em>
63 * mean that threads may share transliterators without synchronizing
64 * them. Transliterators are not immutable, so they must be
65 * synchronized when shared between threads.) This might seem to
66 * limit the complexity of the transliteration operation. In
67 * practice, subclasses perform complex transliterations by delaying
68 * the replacement of text until it is known that no other
69 * replacements are possible. In other words, although the
70 * <code>Transliterator</code> objects are stateless, the source text
71 * itself embodies all the needed information, and delayed operation
72 * allows arbitrary complexity.
73 *
74 * <p><b>Batch transliteration</b>
75 *
76 * <p>The simplest way to perform transliteration is all at once, on a
77 * string of existing text. This is referred to as <em>batch</em>
78 * transliteration. For example, given a string <code>input</code>
79 * and a transliterator <code>t</code>, the call
80 *
81 * String result = t.transliterate(input);
82 *
83 * will transliterate it and return the result. Other methods allow
84 * the client to specify a substring to be transliterated and to use
85 * {@link Replaceable } objects instead of strings, in order to
86 * preserve out-of-band information (such as text styles).
87 *
88 * <p><b>Keyboard transliteration</b>
89 *
90 * <p>Somewhat more involved is <em>keyboard</em>, or incremental
91 * transliteration. This is the transliteration of text that is
92 * arriving from some source (typically the user's keyboard) one
93 * character at a time, or in some other piecemeal fashion.
94 *
95 * <p>In keyboard transliteration, a <code>Replaceable</code> buffer
96 * stores the text. As text is inserted, as much as possible is
97 * transliterated on the fly. This means a GUI that displays the
98 * contents of the buffer may show text being modified as each new
99 * character arrives.
100 *
101 * <p>Consider the simple rule-based Transliterator:
102 * <pre>
103 * th>{theta}
104 * t>{tau}
105 * </pre>
106 *
107 * When the user types 't', nothing will happen, since the
108 * transliterator is waiting to see if the next character is 'h'. To
109 * remedy this, we introduce the notion of a cursor, marked by a '|'
110 * in the output string:
111 * <pre>
112 * t>|{tau}
113 * {tau}h>{theta}
114 * </pre>
115 *
116 * Now when the user types 't', tau appears, and if the next character
117 * is 'h', the tau changes to a theta. This is accomplished by
118 * maintaining a cursor position (independent of the insertion point,
119 * and invisible in the GUI) across calls to
120 * <code>transliterate()</code>. Typically, the cursor will
121 * be coincident with the insertion point, but in a case like the one
122 * above, it will precede the insertion point.
123 *
124 * <p>Keyboard transliteration methods maintain a set of three indices
125 * that are updated with each call to
126 * <code>transliterate()</code>, including the cursor, start,
127 * and limit. Since these indices are changed by the method, they are
128 * passed in an <code>int[]</code> array. The <code>START</code> index
129 * marks the beginning of the substring that the transliterator will
130 * look at. It is advanced as text becomes committed (but it is not
131 * the committed index; that's the <code>CURSOR</code>). The
132 * <code>CURSOR</code> index, described above, marks the point at
133 * which the transliterator last stopped, either because it reached
134 * the end, or because it required more characters to disambiguate
135 * between possible inputs. The <code>CURSOR</code> can also be
136 * explicitly set by rules in a rule-based Transliterator.
137 * Any characters before the <code>CURSOR</code> index are frozen;
138 * future keyboard transliteration calls within this input sequence
139 * will not change them. New text is inserted at the
140 * <code>LIMIT</code> index, which marks the end of the substring that
141 * the transliterator looks at.
142 *
143 * <p>Because keyboard transliteration assumes that more characters
144 * are to arrive, it is conservative in its operation. It only
145 * transliterates when it can do so unambiguously. Otherwise it waits
146 * for more characters to arrive. When the client code knows that no
147 * more characters are forthcoming, perhaps because the user has
148 * performed some input termination operation, then it should call
149 * <code>finishTransliteration()</code> to complete any
150 * pending transliterations.
151 *
152 * <p><b>Inverses</b>
153 *
154 * <p>Pairs of transliterators may be inverses of one another. For
155 * example, if transliterator <b>A</b> transliterates characters by
156 * incrementing their Unicode value (so "abc" -> "def"), and
157 * transliterator <b>B</b> decrements character values, then <b>A</b>
158 * is an inverse of <b>B</b> and vice versa. If we compose <b>A</b>
159 * with <b>B</b> in a compound transliterator, the result is the
160 * indentity transliterator, that is, a transliterator that does not
161 * change its input text.
162 *
163 * The <code>Transliterator</code> method <code>getInverse()</code>
164 * returns a transliterator's inverse, if one exists, or
165 * <code>null</code> otherwise. However, the result of
166 * <code>getInverse()</code> usually will <em>not</em> be a true
167 * mathematical inverse. This is because true inverse transliterators
168 * are difficult to formulate. For example, consider two
169 * transliterators: <b>AB</b>, which transliterates the character 'A'
170 * to 'B', and <b>BA</b>, which transliterates 'B' to 'A'. It might
171 * seem that these are exact inverses, since
172 *
173 * \htmlonly<blockquote>\endhtmlonly"A" x <b>AB</b> -> "B"<br>
174 * "B" x <b>BA</b> -> "A"\htmlonly</blockquote>\endhtmlonly
175 *
176 * where 'x' represents transliteration. However,
177 *
178 * \htmlonly<blockquote>\endhtmlonly"ABCD" x <b>AB</b> -> "BBCD"<br>
179 * "BBCD" x <b>BA</b> -> "AACD"\htmlonly</blockquote>\endhtmlonly
180 *
181 * so <b>AB</b> composed with <b>BA</b> is not the
182 * identity. Nonetheless, <b>BA</b> may be usefully considered to be
183 * <b>AB</b>'s inverse, and it is on this basis that
184 * <b>AB</b><code>.getInverse()</code> could legitimately return
185 * <b>BA</b>.
186 *
187 * <p><b>IDs and display names</b>
188 *
189 * <p>A transliterator is designated by a short identifier string or
190 * <em>ID</em>. IDs follow the format <em>source-destination</em>,
191 * where <em>source</em> describes the entity being replaced, and
192 * <em>destination</em> describes the entity replacing
193 * <em>source</em>. The entities may be the names of scripts,
194 * particular sequences of characters, or whatever else it is that the
195 * transliterator converts to or from. For example, a transliterator
196 * from Russian to Latin might be named "Russian-Latin". A
197 * transliterator from keyboard escape sequences to Latin-1 characters
198 * might be named "KeyboardEscape-Latin1". By convention, system
199 * entity names are in English, with the initial letters of words
200 * capitalized; user entity names may follow any format so long as
201 * they do not contain dashes.
202 *
203 * <p>In addition to programmatic IDs, transliterator objects have
204 * display names for presentation in user interfaces, returned by
205 * {@link #getDisplayName }.
206 *
207 * <p><b>Factory methods and registration</b>
208 *
209 * <p>In general, client code should use the factory method
210 * {@link #createInstance } to obtain an instance of a
211 * transliterator given its ID. Valid IDs may be enumerated using
212 * <code>getAvailableIDs()</code>. Since transliterators are mutable,
213 * multiple calls to {@link #createInstance } with the same ID will
214 * return distinct objects.
215 *
216 * <p>In addition to the system transliterators registered at startup,
217 * user transliterators may be registered by calling
218 * <code>registerInstance()</code> at run time. A registered instance
219 * acts a template; future calls to {@link #createInstance } with the ID
220 * of the registered object return clones of that object. Thus any
221 * object passed to <tt>registerInstance()</tt> must implement
222 * <tt>clone()</tt> propertly. To register a transliterator subclass
223 * without instantiating it (until it is needed), users may call
224 * {@link #registerFactory }. In this case, the objects are
225 * instantiated by invoking the zero-argument public constructor of
226 * the class.
227 *
228 * <p><b>Subclassing</b>
229 *
230 * Subclasses must implement the abstract method
231 * <code>handleTransliterate()</code>. <p>Subclasses should override
232 * the <code>transliterate()</code> method taking a
233 * <code>Replaceable</code> and the <code>transliterate()</code>
234 * method taking a <code>String</code> and <code>StringBuffer</code>
235 * if the performance of these methods can be improved over the
236 * performance obtained by the default implementations in this class.
237 *
238 * <p><b>Rule syntax</b>
239 *
240 * <p>A set of rules determines how to perform translations.
241 * Rules within a rule set are separated by semicolons (';').
242 * To include a literal semicolon, prefix it with a backslash ('\').
243 * Unicode Pattern_White_Space is ignored.
244 * If the first non-blank character on a line is '#',
245 * the entire line is ignored as a comment.
246 *
247 * <p>Each set of rules consists of two groups, one forward, and one
248 * reverse. This is a convention that is not enforced; rules for one
249 * direction may be omitted, with the result that translations in
250 * that direction will not modify the source text. In addition,
251 * bidirectional forward-reverse rules may be specified for
252 * symmetrical transformations.
253 *
254 * <p>Note: Another description of the Transliterator rule syntax is available in
255 * <a href="https://www.unicode.org/reports/tr35/tr35-general.html#Transform_Rules_Syntax">section
256 * Transform Rules Syntax of UTS #35: Unicode LDML</a>.
257 * The rules are shown there using arrow symbols ← and → and ↔.
258 * ICU supports both those and the equivalent ASCII symbols &lt; and &gt; and &lt;&gt;.
259 *
260 * <p>Rule statements take one of the following forms:
261 *
262 * <dl>
263 * <dt><code>$alefmadda=\\u0622;</code></dt>
264 * <dd><strong>Variable definition.</strong> The name on the
265 * left is assigned the text on the right. In this example,
266 * after this statement, instances of the left hand name,
267 * &quot;<code>$alefmadda</code>&quot;, will be replaced by
268 * the Unicode character U+0622. Variable names must begin
269 * with a letter and consist only of letters, digits, and
270 * underscores. Case is significant. Duplicate names cause
271 * an exception to be thrown, that is, variables cannot be
272 * redefined. The right hand side may contain well-formed
273 * text of any length, including no text at all (&quot;<code>$empty=;</code>&quot;).
274 * The right hand side may contain embedded <code>UnicodeSet</code>
275 * patterns, for example, &quot;<code>$softvowel=[eiyEIY]</code>&quot;.</dd>
276 * <dt><code>ai&gt;$alefmadda;</code></dt>
277 * <dd><strong>Forward translation rule.</strong> This rule
278 * states that the string on the left will be changed to the
279 * string on the right when performing forward
280 * transliteration.</dd>
281 * <dt><code>ai&lt;$alefmadda;</code></dt>
282 * <dd><strong>Reverse translation rule.</strong> This rule
283 * states that the string on the right will be changed to
284 * the string on the left when performing reverse
285 * transliteration.</dd>
286 * </dl>
287 *
288 * <dl>
289 * <dt><code>ai&lt;&gt;$alefmadda;</code></dt>
290 * <dd><strong>Bidirectional translation rule.</strong> This
291 * rule states that the string on the right will be changed
292 * to the string on the left when performing forward
293 * transliteration, and vice versa when performing reverse
294 * transliteration.</dd>
295 * </dl>
296 *
297 * <p>Translation rules consist of a <em>match pattern</em> and an <em>output
298 * string</em>. The match pattern consists of literal characters,
299 * optionally preceded by context, and optionally followed by
300 * context. Context characters, like literal pattern characters,
301 * must be matched in the text being transliterated. However, unlike
302 * literal pattern characters, they are not replaced by the output
303 * text. For example, the pattern &quot;<code>abc{def}</code>&quot;
304 * indicates the characters &quot;<code>def</code>&quot; must be
305 * preceded by &quot;<code>abc</code>&quot; for a successful match.
306 * If there is a successful match, &quot;<code>def</code>&quot; will
307 * be replaced, but not &quot;<code>abc</code>&quot;. The final '<code>}</code>'
308 * is optional, so &quot;<code>abc{def</code>&quot; is equivalent to
309 * &quot;<code>abc{def}</code>&quot;. Another example is &quot;<code>{123}456</code>&quot;
310 * (or &quot;<code>123}456</code>&quot;) in which the literal
311 * pattern &quot;<code>123</code>&quot; must be followed by &quot;<code>456</code>&quot;.
312 *
313 * <p>The output string of a forward or reverse rule consists of
314 * characters to replace the literal pattern characters. If the
315 * output string contains the character '<code>|</code>', this is
316 * taken to indicate the location of the <em>cursor</em> after
317 * replacement. The cursor is the point in the text at which the
318 * next replacement, if any, will be applied. The cursor is usually
319 * placed within the replacement text; however, it can actually be
320 * placed into the precending or following context by using the
321 * special character '@'. Examples:
322 *
323 * <pre>
324 * a {foo} z &gt; | @ bar; # foo -&gt; bar, move cursor before a
325 * {foo} xyz &gt; bar @@|; #&nbsp;foo -&gt; bar, cursor between y and z
326 * </pre>
327 *
328 * <p><b>UnicodeSet</b>
329 *
330 * <p><code>UnicodeSet</code> patterns may appear anywhere that
331 * makes sense. They may appear in variable definitions.
332 * Contrariwise, <code>UnicodeSet</code> patterns may themselves
333 * contain variable references, such as &quot;<code>$a=[a-z];$not_a=[^$a]</code>&quot;,
334 * or &quot;<code>$range=a-z;$ll=[$range]</code>&quot;.
335 *
336 * <p><code>UnicodeSet</code> patterns may also be embedded directly
337 * into rule strings. Thus, the following two rules are equivalent:
338 *
339 * <pre>
340 * $vowel=[aeiou]; $vowel&gt;'*'; # One way to do this
341 * [aeiou]&gt;'*'; # Another way
342 * </pre>
343 *
344 * <p>See {@link UnicodeSet} for more documentation and examples.
345 *
346 * <p><b>Segments</b>
347 *
348 * <p>Segments of the input string can be matched and copied to the
349 * output string. This makes certain sets of rules simpler and more
350 * general, and makes reordering possible. For example:
351 *
352 * <pre>
353 * ([a-z]) &gt; $1 $1; # double lowercase letters
354 * ([:Lu:]) ([:Ll:]) &gt; $2 $1; # reverse order of Lu-Ll pairs
355 * </pre>
356 *
357 * <p>The segment of the input string to be copied is delimited by
358 * &quot;<code>(</code>&quot; and &quot;<code>)</code>&quot;. Up to
359 * nine segments may be defined. Segments may not overlap. In the
360 * output string, &quot;<code>$1</code>&quot; through &quot;<code>$9</code>&quot;
361 * represent the input string segments, in left-to-right order of
362 * definition.
363 *
364 * <p><b>Anchors</b>
365 *
366 * <p>Patterns can be anchored to the beginning or the end of the text. This is done with the
367 * special characters '<code>^</code>' and '<code>$</code>'. For example:
368 *
369 * <pre>
370 * ^ a&nbsp;&nbsp; &gt; 'BEG_A'; &nbsp;&nbsp;# match 'a' at start of text
371 * &nbsp; a&nbsp;&nbsp; &gt; 'A'; # match other instances of 'a'
372 * &nbsp; z $ &gt; 'END_Z'; &nbsp;&nbsp;# match 'z' at end of text
373 * &nbsp; z&nbsp;&nbsp; &gt; 'Z';&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; # match other instances of 'z'
374 * </pre>
375 *
376 * <p>It is also possible to match the beginning or the end of the text using a <code>UnicodeSet</code>.
377 * This is done by including a virtual anchor character '<code>$</code>' at the end of the
378 * set pattern. Although this is usually the match chafacter for the end anchor, the set will
379 * match either the beginning or the end of the text, depending on its placement. For
380 * example:
381 *
382 * <pre>
383 * $x = [a-z$]; &nbsp;&nbsp;# match 'a' through 'z' OR anchor
384 * $x 1&nbsp;&nbsp;&nbsp; &gt; 2;&nbsp;&nbsp; # match '1' after a-z or at the start
385 * &nbsp;&nbsp; 3 $x &gt; 4; &nbsp;&nbsp;# match '3' before a-z or at the end
386 * </pre>
387 *
388 * <p><b>Example</b>
389 *
390 * <p>The following example rules illustrate many of the features of
391 * the rule language.
392 *
393 * <table border="0" cellpadding="4">
394 * <tr>
395 * <td style="vertical-align: top;">Rule 1.</td>
396 * <td style="vertical-align: top; write-space: nowrap;"><code>abc{def}&gt;x|y</code></td>
397 * </tr>
398 * <tr>
399 * <td style="vertical-align: top;">Rule 2.</td>
400 * <td style="vertical-align: top; write-space: nowrap;"><code>xyz&gt;r</code></td>
401 * </tr>
402 * <tr>
403 * <td style="vertical-align: top;">Rule 3.</td>
404 * <td style="vertical-align: top; write-space: nowrap;"><code>yz&gt;q</code></td>
405 * </tr>
406 * </table>
407 *
408 * <p>Applying these rules to the string &quot;<code>adefabcdefz</code>&quot;
409 * yields the following results:
410 *
411 * <table border="0" cellpadding="4">
412 * <tr>
413 * <td style="vertical-align: top; write-space: nowrap;"><code>|adefabcdefz</code></td>
414 * <td style="vertical-align: top;">Initial state, no rules match. Advance
415 * cursor.</td>
416 * </tr>
417 * <tr>
418 * <td style="vertical-align: top; write-space: nowrap;"><code>a|defabcdefz</code></td>
419 * <td style="vertical-align: top;">Still no match. Rule 1 does not match
420 * because the preceding context is not present.</td>
421 * </tr>
422 * <tr>
423 * <td style="vertical-align: top; write-space: nowrap;"><code>ad|efabcdefz</code></td>
424 * <td style="vertical-align: top;">Still no match. Keep advancing until
425 * there is a match...</td>
426 * </tr>
427 * <tr>
428 * <td style="vertical-align: top; write-space: nowrap;"><code>ade|fabcdefz</code></td>
429 * <td style="vertical-align: top;">...</td>
430 * </tr>
431 * <tr>
432 * <td style="vertical-align: top; write-space: nowrap;"><code>adef|abcdefz</code></td>
433 * <td style="vertical-align: top;">...</td>
434 * </tr>
435 * <tr>
436 * <td style="vertical-align: top; write-space: nowrap;"><code>adefa|bcdefz</code></td>
437 * <td style="vertical-align: top;">...</td>
438 * </tr>
439 * <tr>
440 * <td style="vertical-align: top; write-space: nowrap;"><code>adefab|cdefz</code></td>
441 * <td style="vertical-align: top;">...</td>
442 * </tr>
443 * <tr>
444 * <td style="vertical-align: top; write-space: nowrap;"><code>adefabc|defz</code></td>
445 * <td style="vertical-align: top;">Rule 1 matches; replace &quot;<code>def</code>&quot;
446 * with &quot;<code>xy</code>&quot; and back up the cursor
447 * to before the '<code>y</code>'.</td>
448 * </tr>
449 * <tr>
450 * <td style="vertical-align: top; write-space: nowrap;"><code>adefabcx|yz</code></td>
451 * <td style="vertical-align: top;">Although &quot;<code>xyz</code>&quot; is
452 * present, rule 2 does not match because the cursor is
453 * before the '<code>y</code>', not before the '<code>x</code>'.
454 * Rule 3 does match. Replace &quot;<code>yz</code>&quot;
455 * with &quot;<code>q</code>&quot;.</td>
456 * </tr>
457 * <tr>
458 * <td style="vertical-align: top; write-space: nowrap;"><code>adefabcxq|</code></td>
459 * <td style="vertical-align: top;">The cursor is at the end;
460 * transliteration is complete.</td>
461 * </tr>
462 * </table>
463 *
464 * <p>The order of rules is significant. If multiple rules may match
465 * at some point, the first matching rule is applied.
466 *
467 * <p>Forward and reverse rules may have an empty output string.
468 * Otherwise, an empty left or right hand side of any statement is a
469 * syntax error.
470 *
471 * <p>Single quotes are used to quote any character other than a
472 * digit or letter. To specify a single quote itself, inside or
473 * outside of quotes, use two single quotes in a row. For example,
474 * the rule &quot;<code>'&gt;'&gt;o''clock</code>&quot; changes the
475 * string &quot;<code>&gt;</code>&quot; to the string &quot;<code>o'clock</code>&quot;.
476 *
477 * <p><b>Notes</b>
478 *
479 * <p>While a Transliterator is being built from rules, it checks that
480 * the rules are added in proper order. For example, if the rule
481 * &quot;a&gt;x&quot; is followed by the rule &quot;ab&gt;y&quot;,
482 * then the second rule will throw an exception. The reason is that
483 * the second rule can never be triggered, since the first rule
484 * always matches anything it matches. In other words, the first
485 * rule <em>masks</em> the second rule.
486 *
487 * @author Alan Liu
488 * @stable ICU 2.0
489 */
490class U_I18N_API Transliterator : public UObject {
491
492private:
493
494 /**
495 * Programmatic name, e.g., "Latin-Arabic".
496 */
497 UnicodeString ID;
498
499 /**
500 * This transliterator's filter. Any character for which
501 * <tt>filter.contains()</tt> returns <tt>false</tt> will not be
502 * altered by this transliterator. If <tt>filter</tt> is
503 * <tt>null</tt> then no filtering is applied.
504 */
505 UnicodeFilter* filter;
506
507 int32_t maximumContextLength;
508
509 public:
510
511 /**
512 * A context integer or pointer for a factory function, passed by
513 * value.
514 * @stable ICU 2.4
515 */
516 union Token {
517 /**
518 * This token, interpreted as a 32-bit integer.
519 * @stable ICU 2.4
520 */
521 int32_t integer;
522 /**
523 * This token, interpreted as a native pointer.
524 * @stable ICU 2.4
525 */
526 void* pointer;
527 };
528
529#ifndef U_HIDE_INTERNAL_API
530 /**
531 * Return a token containing an integer.
532 * @return a token containing an integer.
533 * @internal
534 */
535 inline static Token integerToken(int32_t);
536
537 /**
538 * Return a token containing a pointer.
539 * @return a token containing a pointer.
540 * @internal
541 */
542 inline static Token pointerToken(void*);
543#endif /* U_HIDE_INTERNAL_API */
544
545 /**
546 * A function that creates and returns a Transliterator. When
547 * invoked, it will be passed the ID string that is being
548 * instantiated, together with the context pointer that was passed
549 * in when the factory function was first registered. Many
550 * factory functions will ignore both parameters, however,
551 * functions that are registered to more than one ID may use the
552 * ID or the context parameter to parameterize the transliterator
553 * they create.
554 * @param ID the string identifier for this transliterator
555 * @param context a context pointer that will be stored and
556 * later passed to the factory function when an ID matching
557 * the registration ID is being instantiated with this factory.
558 * @stable ICU 2.4
559 */
560 typedef Transliterator* (U_EXPORT2 *Factory)(const UnicodeString& ID, Token context);
561
562protected:
563
564 /**
565 * Default constructor.
566 * @param ID the string identifier for this transliterator
567 * @param adoptedFilter the filter. Any character for which
568 * <tt>filter.contains()</tt> returns <tt>false</tt> will not be
569 * altered by this transliterator. If <tt>filter</tt> is
570 * <tt>null</tt> then no filtering is applied.
571 * @stable ICU 2.4
572 */
573 Transliterator(const UnicodeString& ID, UnicodeFilter* adoptedFilter);
574
575 /**
576 * Copy constructor.
577 * @stable ICU 2.4
578 */
579 Transliterator(const Transliterator&);
580
581 /**
582 * Assignment operator.
583 * @stable ICU 2.4
584 */
585 Transliterator& operator=(const Transliterator&);
586
587 /**
588 * Create a transliterator from a basic ID. This is an ID
589 * containing only the forward direction source, target, and
590 * variant.
591 * @param id a basic ID of the form S-T or S-T/V.
592 * @param canon canonical ID to assign to the object, or
593 * NULL to leave the ID unchanged
594 * @return a newly created Transliterator or null if the ID is
595 * invalid.
596 * @stable ICU 2.4
597 */
598 static Transliterator* createBasicInstance(const UnicodeString& id,
599 const UnicodeString* canon);
600
601 friend class TransliteratorParser; // for parseID()
602 friend class TransliteratorIDParser; // for createBasicInstance()
603 friend class TransliteratorAlias; // for setID()
604
605public:
606
607 /**
608 * Destructor.
609 * @stable ICU 2.0
610 */
611 virtual ~Transliterator();
612
613 /**
614 * Implements Cloneable.
615 * All subclasses are encouraged to implement this method if it is
616 * possible and reasonable to do so. Subclasses that are to be
617 * registered with the system using <tt>registerInstance()</tt>
618 * are required to implement this method. If a subclass does not
619 * implement clone() properly and is registered with the system
620 * using registerInstance(), then the default clone() implementation
621 * will return null, and calls to createInstance() will fail.
622 *
623 * @return a copy of the object.
624 * @see #registerInstance
625 * @stable ICU 2.0
626 */
627 virtual Transliterator* clone() const;
628
629 /**
630 * Transliterates a segment of a string, with optional filtering.
631 *
632 * @param text the string to be transliterated
633 * @param start the beginning index, inclusive; <code>0 <= start
634 * <= limit</code>.
635 * @param limit the ending index, exclusive; <code>start <= limit
636 * <= text.length()</code>.
637 * @return The new limit index. The text previously occupying <code>[start,
638 * limit)</code> has been transliterated, possibly to a string of a different
639 * length, at <code>[start, </code><em>new-limit</em><code>)</code>, where
640 * <em>new-limit</em> is the return value. If the input offsets are out of bounds,
641 * the returned value is -1 and the input string remains unchanged.
642 * @stable ICU 2.0
643 */
644 virtual int32_t transliterate(Replaceable& text,
645 int32_t start, int32_t limit) const;
646
647 /**
648 * Transliterates an entire string in place. Convenience method.
649 * @param text the string to be transliterated
650 * @stable ICU 2.0
651 */
652 virtual void transliterate(Replaceable& text) const;
653
654 /**
655 * Transliterates the portion of the text buffer that can be
656 * transliterated unambiguosly after new text has been inserted,
657 * typically as a result of a keyboard event. The new text in
658 * <code>insertion</code> will be inserted into <code>text</code>
659 * at <code>index.limit</code>, advancing
660 * <code>index.limit</code> by <code>insertion.length()</code>.
661 * Then the transliterator will try to transliterate characters of
662 * <code>text</code> between <code>index.cursor</code> and
663 * <code>index.limit</code>. Characters before
664 * <code>index.cursor</code> will not be changed.
665 *
666 * <p>Upon return, values in <code>index</code> will be updated.
667 * <code>index.start</code> will be advanced to the first
668 * character that future calls to this method will read.
669 * <code>index.cursor</code> and <code>index.limit</code> will
670 * be adjusted to delimit the range of text that future calls to
671 * this method may change.
672 *
673 * <p>Typical usage of this method begins with an initial call
674 * with <code>index.start</code> and <code>index.limit</code>
675 * set to indicate the portion of <code>text</code> to be
676 * transliterated, and <code>index.cursor == index.start</code>.
677 * Thereafter, <code>index</code> can be used without
678 * modification in future calls, provided that all changes to
679 * <code>text</code> are made via this method.
680 *
681 * <p>This method assumes that future calls may be made that will
682 * insert new text into the buffer. As a result, it only performs
683 * unambiguous transliterations. After the last call to this
684 * method, there may be untransliterated text that is waiting for
685 * more input to resolve an ambiguity. In order to perform these
686 * pending transliterations, clients should call {@link
687 * #finishTransliteration } after the last call to this
688 * method has been made.
689 *
690 * @param text the buffer holding transliterated and untransliterated text
691 * @param index an array of three integers.
692 *
693 * <ul><li><code>index.start</code>: the beginning index,
694 * inclusive; <code>0 <= index.start <= index.limit</code>.
695 *
696 * <li><code>index.limit</code>: the ending index, exclusive;
697 * <code>index.start <= index.limit <= text.length()</code>.
698 * <code>insertion</code> is inserted at
699 * <code>index.limit</code>.
700 *
701 * <li><code>index.cursor</code>: the next character to be
702 * considered for transliteration; <code>index.start <=
703 * index.cursor <= index.limit</code>. Characters before
704 * <code>index.cursor</code> will not be changed by future calls
705 * to this method.</ul>
706 *
707 * @param insertion text to be inserted and possibly
708 * transliterated into the translation buffer at
709 * <code>index.limit</code>. If <code>null</code> then no text
710 * is inserted.
711 * @param status Output param to filled in with a success or an error.
712 * @see #handleTransliterate
713 * @exception IllegalArgumentException if <code>index</code>
714 * is invalid
715 * @see UTransPosition
716 * @stable ICU 2.0
717 */
718 virtual void transliterate(Replaceable& text, UTransPosition& index,
719 const UnicodeString& insertion,
720 UErrorCode& status) const;
721
722 /**
723 * Transliterates the portion of the text buffer that can be
724 * transliterated unambiguosly after a new character has been
725 * inserted, typically as a result of a keyboard event. This is a
726 * convenience method.
727 * @param text the buffer holding transliterated and
728 * untransliterated text
729 * @param index an array of three integers.
730 * @param insertion text to be inserted and possibly
731 * transliterated into the translation buffer at
732 * <code>index.limit</code>.
733 * @param status Output param to filled in with a success or an error.
734 * @see #transliterate(Replaceable&, UTransPosition&, const UnicodeString&, UErrorCode&) const
735 * @stable ICU 2.0
736 */
737 virtual void transliterate(Replaceable& text, UTransPosition& index,
738 UChar32 insertion,
739 UErrorCode& status) const;
740
741 /**
742 * Transliterates the portion of the text buffer that can be
743 * transliterated unambiguosly. This is a convenience method; see
744 * {@link
745 * #transliterate(Replaceable&, UTransPosition&, const UnicodeString&, UErrorCode&) const }
746 * for details.
747 * @param text the buffer holding transliterated and
748 * untransliterated text
749 * @param index an array of three integers.
750 * @param status Output param to filled in with a success or an error.
751 * @see #transliterate(Replaceable&, UTransPosition&, const UnicodeString&, UErrorCode &) const
752 * @stable ICU 2.0
753 */
754 virtual void transliterate(Replaceable& text, UTransPosition& index,
755 UErrorCode& status) const;
756
757 /**
758 * Finishes any pending transliterations that were waiting for
759 * more characters. Clients should call this method as the last
760 * call after a sequence of one or more calls to
761 * <code>transliterate()</code>.
762 * @param text the buffer holding transliterated and
763 * untransliterated text.
764 * @param index the array of indices previously passed to {@link
765 * #transliterate }
766 * @stable ICU 2.0
767 */
768 virtual void finishTransliteration(Replaceable& text,
769 UTransPosition& index) const;
770
771private:
772
773 /**
774 * This internal method does incremental transliteration. If the
775 * 'insertion' is non-null then we append it to 'text' before
776 * proceeding. This method calls through to the pure virtual
777 * framework method handleTransliterate() to do the actual
778 * work.
779 * @param text the buffer holding transliterated and
780 * untransliterated text
781 * @param index an array of three integers. See {@link
782 * #transliterate(Replaceable, int[], String)}.
783 * @param insertion text to be inserted and possibly
784 * transliterated into the translation buffer at
785 * <code>index.limit</code>.
786 * @param status Output param to filled in with a success or an error.
787 */
788 void _transliterate(Replaceable& text,
789 UTransPosition& index,
790 const UnicodeString* insertion,
791 UErrorCode &status) const;
792
793protected:
794
795 /**
796 * Abstract method that concrete subclasses define to implement
797 * their transliteration algorithm. This method handles both
798 * incremental and non-incremental transliteration. Let
799 * <code>originalStart</code> refer to the value of
800 * <code>pos.start</code> upon entry.
801 *
802 * <ul>
803 * <li>If <code>incremental</code> is false, then this method
804 * should transliterate all characters between
805 * <code>pos.start</code> and <code>pos.limit</code>. Upon return
806 * <code>pos.start</code> must == <code> pos.limit</code>.</li>
807 *
808 * <li>If <code>incremental</code> is true, then this method
809 * should transliterate all characters between
810 * <code>pos.start</code> and <code>pos.limit</code> that can be
811 * unambiguously transliterated, regardless of future insertions
812 * of text at <code>pos.limit</code>. Upon return,
813 * <code>pos.start</code> should be in the range
814 * [<code>originalStart</code>, <code>pos.limit</code>).
815 * <code>pos.start</code> should be positioned such that
816 * characters [<code>originalStart</code>, <code>
817 * pos.start</code>) will not be changed in the future by this
818 * transliterator and characters [<code>pos.start</code>,
819 * <code>pos.limit</code>) are unchanged.</li>
820 * </ul>
821 *
822 * <p>Implementations of this method should also obey the
823 * following invariants:</p>
824 *
825 * <ul>
826 * <li> <code>pos.limit</code> and <code>pos.contextLimit</code>
827 * should be updated to reflect changes in length of the text
828 * between <code>pos.start</code> and <code>pos.limit</code>. The
829 * difference <code> pos.contextLimit - pos.limit</code> should
830 * not change.</li>
831 *
832 * <li><code>pos.contextStart</code> should not change.</li>
833 *
834 * <li>Upon return, neither <code>pos.start</code> nor
835 * <code>pos.limit</code> should be less than
836 * <code>originalStart</code>.</li>
837 *
838 * <li>Text before <code>originalStart</code> and text after
839 * <code>pos.limit</code> should not change.</li>
840 *
841 * <li>Text before <code>pos.contextStart</code> and text after
842 * <code> pos.contextLimit</code> should be ignored.</li>
843 * </ul>
844 *
845 * <p>Subclasses may safely assume that all characters in
846 * [<code>pos.start</code>, <code>pos.limit</code>) are filtered.
847 * In other words, the filter has already been applied by the time
848 * this method is called. See
849 * <code>filteredTransliterate()</code>.
850 *
851 * <p>This method is <b>not</b> for public consumption. Calling
852 * this method directly will transliterate
853 * [<code>pos.start</code>, <code>pos.limit</code>) without
854 * applying the filter. End user code should call <code>
855 * transliterate()</code> instead of this method. Subclass code
856 * and wrapping transliterators should call
857 * <code>filteredTransliterate()</code> instead of this method.<p>
858 *
859 * @param text the buffer holding transliterated and
860 * untransliterated text
861 *
862 * @param pos the indices indicating the start, limit, context
863 * start, and context limit of the text.
864 *
865 * @param incremental if true, assume more text may be inserted at
866 * <code>pos.limit</code> and act accordingly. Otherwise,
867 * transliterate all text between <code>pos.start</code> and
868 * <code>pos.limit</code> and move <code>pos.start</code> up to
869 * <code>pos.limit</code>.
870 *
871 * @see #transliterate
872 * @stable ICU 2.4
873 */
874 virtual void handleTransliterate(Replaceable& text,
875 UTransPosition& pos,
876 UBool incremental) const = 0;
877
878public:
879 /**
880 * Transliterate a substring of text, as specified by index, taking filters
881 * into account. This method is for subclasses that need to delegate to
882 * another transliterator.
883 * @param text the text to be transliterated
884 * @param index the position indices
885 * @param incremental if TRUE, then assume more characters may be inserted
886 * at index.limit, and postpone processing to accomodate future incoming
887 * characters
888 * @stable ICU 2.4
889 */
890 virtual void filteredTransliterate(Replaceable& text,
891 UTransPosition& index,
892 UBool incremental) const;
893
894private:
895
896 /**
897 * Top-level transliteration method, handling filtering, incremental and
898 * non-incremental transliteration, and rollback. All transliteration
899 * public API methods eventually call this method with a rollback argument
900 * of TRUE. Other entities may call this method but rollback should be
901 * FALSE.
902 *
903 * <p>If this transliterator has a filter, break up the input text into runs
904 * of unfiltered characters. Pass each run to
905 * subclass.handleTransliterate().
906 *
907 * <p>In incremental mode, if rollback is TRUE, perform a special
908 * incremental procedure in which several passes are made over the input
909 * text, adding one character at a time, and committing successful
910 * transliterations as they occur. Unsuccessful transliterations are rolled
911 * back and retried with additional characters to give correct results.
912 *
913 * @param text the text to be transliterated
914 * @param index the position indices
915 * @param incremental if TRUE, then assume more characters may be inserted
916 * at index.limit, and postpone processing to accomodate future incoming
917 * characters
918 * @param rollback if TRUE and if incremental is TRUE, then perform special
919 * incremental processing, as described above, and undo partial
920 * transliterations where necessary. If incremental is FALSE then this
921 * parameter is ignored.
922 */
923 virtual void filteredTransliterate(Replaceable& text,
924 UTransPosition& index,
925 UBool incremental,
926 UBool rollback) const;
927
928public:
929
930 /**
931 * Returns the length of the longest context required by this transliterator.
932 * This is <em>preceding</em> context. The default implementation supplied
933 * by <code>Transliterator</code> returns zero; subclasses
934 * that use preceding context should override this method to return the
935 * correct value. For example, if a transliterator translates "ddd" (where
936 * d is any digit) to "555" when preceded by "(ddd)", then the preceding
937 * context length is 5, the length of "(ddd)".
938 *
939 * @return The maximum number of preceding context characters this
940 * transliterator needs to examine
941 * @stable ICU 2.0
942 */
943 int32_t getMaximumContextLength(void) const;
944
945protected:
946
947 /**
948 * Method for subclasses to use to set the maximum context length.
949 * @param maxContextLength the new value to be set.
950 * @see #getMaximumContextLength
951 * @stable ICU 2.4
952 */
953 void setMaximumContextLength(int32_t maxContextLength);
954
955public:
956
957 /**
958 * Returns a programmatic identifier for this transliterator.
959 * If this identifier is passed to <code>createInstance()</code>, it
960 * will return this object, if it has been registered.
961 * @return a programmatic identifier for this transliterator.
962 * @see #registerInstance
963 * @see #registerFactory
964 * @see #getAvailableIDs
965 * @stable ICU 2.0
966 */
967 virtual const UnicodeString& getID(void) const;
968
969 /**
970 * Returns a name for this transliterator that is appropriate for
971 * display to the user in the default locale. See {@link
972 * #getDisplayName } for details.
973 * @param ID the string identifier for this transliterator
974 * @param result Output param to receive the display name
975 * @return A reference to 'result'.
976 * @stable ICU 2.0
977 */
978 static UnicodeString& U_EXPORT2 getDisplayName(const UnicodeString& ID,
979 UnicodeString& result);
980
981 /**
982 * Returns a name for this transliterator that is appropriate for
983 * display to the user in the given locale. This name is taken
984 * from the locale resource data in the standard manner of the
985 * <code>java.text</code> package.
986 *
987 * <p>If no localized names exist in the system resource bundles,
988 * a name is synthesized using a localized
989 * <code>MessageFormat</code> pattern from the resource data. The
990 * arguments to this pattern are an integer followed by one or two
991 * strings. The integer is the number of strings, either 1 or 2.
992 * The strings are formed by splitting the ID for this
993 * transliterator at the first '-'. If there is no '-', then the
994 * entire ID forms the only string.
995 * @param ID the string identifier for this transliterator
996 * @param inLocale the Locale in which the display name should be
997 * localized.
998 * @param result Output param to receive the display name
999 * @return A reference to 'result'.
1000 * @stable ICU 2.0
1001 */
1002 static UnicodeString& U_EXPORT2 getDisplayName(const UnicodeString& ID,
1003 const Locale& inLocale,
1004 UnicodeString& result);
1005
1006 /**
1007 * Returns the filter used by this transliterator, or <tt>NULL</tt>
1008 * if this transliterator uses no filter.
1009 * @return the filter used by this transliterator, or <tt>NULL</tt>
1010 * if this transliterator uses no filter.
1011 * @stable ICU 2.0
1012 */
1013 const UnicodeFilter* getFilter(void) const;
1014
1015 /**
1016 * Returns the filter used by this transliterator, or <tt>NULL</tt> if this
1017 * transliterator uses no filter. The caller must eventually delete the
1018 * result. After this call, this transliterator's filter is set to
1019 * <tt>NULL</tt>.
1020 * @return the filter used by this transliterator, or <tt>NULL</tt> if this
1021 * transliterator uses no filter.
1022 * @stable ICU 2.4
1023 */
1024 UnicodeFilter* orphanFilter(void);
1025
1026 /**
1027 * Changes the filter used by this transliterator. If the filter
1028 * is set to <tt>null</tt> then no filtering will occur.
1029 *
1030 * <p>Callers must take care if a transliterator is in use by
1031 * multiple threads. The filter should not be changed by one
1032 * thread while another thread may be transliterating.
1033 * @param adoptedFilter the new filter to be adopted.
1034 * @stable ICU 2.0
1035 */
1036 void adoptFilter(UnicodeFilter* adoptedFilter);
1037
1038 /**
1039 * Returns this transliterator's inverse. See the class
1040 * documentation for details. This implementation simply inverts
1041 * the two entities in the ID and attempts to retrieve the
1042 * resulting transliterator. That is, if <code>getID()</code>
1043 * returns "A-B", then this method will return the result of
1044 * <code>createInstance("B-A")</code>, or <code>null</code> if that
1045 * call fails.
1046 *
1047 * <p>Subclasses with knowledge of their inverse may wish to
1048 * override this method.
1049 *
1050 * @param status Output param to filled in with a success or an error.
1051 * @return a transliterator that is an inverse, not necessarily
1052 * exact, of this transliterator, or <code>null</code> if no such
1053 * transliterator is registered.
1054 * @see #registerInstance
1055 * @stable ICU 2.0
1056 */
1057 Transliterator* createInverse(UErrorCode& status) const;
1058
1059 /**
1060 * Returns a <code>Transliterator</code> object given its ID.
1061 * The ID must be either a system transliterator ID or a ID registered
1062 * using <code>registerInstance()</code>.
1063 *
1064 * @param ID a valid ID, as enumerated by <code>getAvailableIDs()</code>
1065 * @param dir either FORWARD or REVERSE.
1066 * @param parseError Struct to recieve information on position
1067 * of error if an error is encountered
1068 * @param status Output param to filled in with a success or an error.
1069 * @return A <code>Transliterator</code> object with the given ID
1070 * @see #registerInstance
1071 * @see #getAvailableIDs
1072 * @see #getID
1073 * @stable ICU 2.0
1074 */
1075 static Transliterator* U_EXPORT2 createInstance(const UnicodeString& ID,
1076 UTransDirection dir,
1077 UParseError& parseError,
1078 UErrorCode& status);
1079
1080 /**
1081 * Returns a <code>Transliterator</code> object given its ID.
1082 * The ID must be either a system transliterator ID or a ID registered
1083 * using <code>registerInstance()</code>.
1084 * @param ID a valid ID, as enumerated by <code>getAvailableIDs()</code>
1085 * @param dir either FORWARD or REVERSE.
1086 * @param status Output param to filled in with a success or an error.
1087 * @return A <code>Transliterator</code> object with the given ID
1088 * @stable ICU 2.0
1089 */
1090 static Transliterator* U_EXPORT2 createInstance(const UnicodeString& ID,
1091 UTransDirection dir,
1092 UErrorCode& status);
1093
1094 /**
1095 * Returns a <code>Transliterator</code> object constructed from
1096 * the given rule string. This will be a rule-based Transliterator,
1097 * if the rule string contains only rules, or a
1098 * compound Transliterator, if it contains ID blocks, or a
1099 * null Transliterator, if it contains ID blocks which parse as
1100 * empty for the given direction.
1101 *
1102 * @param ID the id for the transliterator.
1103 * @param rules rules, separated by ';'
1104 * @param dir either FORWARD or REVERSE.
1105 * @param parseError Struct to receive information on position
1106 * of error if an error is encountered
1107 * @param status Output param set to success/failure code.
1108 * @return a newly created Transliterator
1109 * @stable ICU 2.0
1110 */
1111 static Transliterator* U_EXPORT2 createFromRules(const UnicodeString& ID,
1112 const UnicodeString& rules,
1113 UTransDirection dir,
1114 UParseError& parseError,
1115 UErrorCode& status);
1116
1117 /**
1118 * Create a rule string that can be passed to createFromRules()
1119 * to recreate this transliterator.
1120 * @param result the string to receive the rules. Previous
1121 * contents will be deleted.
1122 * @param escapeUnprintable if TRUE then convert unprintable
1123 * character to their hex escape representations, \\uxxxx or
1124 * \\Uxxxxxxxx. Unprintable characters are those other than
1125 * U+000A, U+0020..U+007E.
1126 * @stable ICU 2.0
1127 */
1128 virtual UnicodeString& toRules(UnicodeString& result,
1129 UBool escapeUnprintable) const;
1130
1131 /**
1132 * Return the number of elements that make up this transliterator.
1133 * For example, if the transliterator "NFD;Jamo-Latin;Latin-Greek"
1134 * were created, the return value of this method would be 3.
1135 *
1136 * <p>If this transliterator is not composed of other
1137 * transliterators, then this method returns 1.
1138 * @return the number of transliterators that compose this
1139 * transliterator, or 1 if this transliterator is not composed of
1140 * multiple transliterators
1141 * @stable ICU 3.0
1142 */
1143 int32_t countElements() const;
1144
1145 /**
1146 * Return an element that makes up this transliterator. For
1147 * example, if the transliterator "NFD;Jamo-Latin;Latin-Greek"
1148 * were created, the return value of this method would be one
1149 * of the three transliterator objects that make up that
1150 * transliterator: [NFD, Jamo-Latin, Latin-Greek].
1151 *
1152 * <p>If this transliterator is not composed of other
1153 * transliterators, then this method will return a reference to
1154 * this transliterator when given the index 0.
1155 * @param index a value from 0..countElements()-1 indicating the
1156 * transliterator to return
1157 * @param ec input-output error code
1158 * @return one of the transliterators that makes up this
1159 * transliterator, if this transliterator is made up of multiple
1160 * transliterators, otherwise a reference to this object if given
1161 * an index of 0
1162 * @stable ICU 3.0
1163 */
1164 const Transliterator& getElement(int32_t index, UErrorCode& ec) const;
1165
1166 /**
1167 * Returns the set of all characters that may be modified in the
1168 * input text by this Transliterator. This incorporates this
1169 * object's current filter; if the filter is changed, the return
1170 * value of this function will change. The default implementation
1171 * returns an empty set. Some subclasses may override {@link
1172 * #handleGetSourceSet } to return a more precise result. The
1173 * return result is approximate in any case and is intended for
1174 * use by tests, tools, or utilities.
1175 * @param result receives result set; previous contents lost
1176 * @return a reference to result
1177 * @see #getTargetSet
1178 * @see #handleGetSourceSet
1179 * @stable ICU 2.4
1180 */
1181 UnicodeSet& getSourceSet(UnicodeSet& result) const;
1182
1183 /**
1184 * Framework method that returns the set of all characters that
1185 * may be modified in the input text by this Transliterator,
1186 * ignoring the effect of this object's filter. The base class
1187 * implementation returns the empty set. Subclasses that wish to
1188 * implement this should override this method.
1189 * @return the set of characters that this transliterator may
1190 * modify. The set may be modified, so subclasses should return a
1191 * newly-created object.
1192 * @param result receives result set; previous contents lost
1193 * @see #getSourceSet
1194 * @see #getTargetSet
1195 * @stable ICU 2.4
1196 */
1197 virtual void handleGetSourceSet(UnicodeSet& result) const;
1198
1199 /**
1200 * Returns the set of all characters that may be generated as
1201 * replacement text by this transliterator. The default
1202 * implementation returns the empty set. Some subclasses may
1203 * override this method to return a more precise result. The
1204 * return result is approximate in any case and is intended for
1205 * use by tests, tools, or utilities requiring such
1206 * meta-information.
1207 * @param result receives result set; previous contents lost
1208 * @return a reference to result
1209 * @see #getTargetSet
1210 * @stable ICU 2.4
1211 */
1212 virtual UnicodeSet& getTargetSet(UnicodeSet& result) const;
1213
1214public:
1215
1216 /**
1217 * Registers a factory function that creates transliterators of
1218 * a given ID.
1219 *
1220 * Because ICU may choose to cache Transliterators internally, this must
1221 * be called at application startup, prior to any calls to
1222 * Transliterator::createXXX to avoid undefined behavior.
1223 *
1224 * @param id the ID being registered
1225 * @param factory a function pointer that will be copied and
1226 * called later when the given ID is passed to createInstance()
1227 * @param context a context pointer that will be stored and
1228 * later passed to the factory function when an ID matching
1229 * the registration ID is being instantiated with this factory.
1230 * @stable ICU 2.0
1231 */
1232 static void U_EXPORT2 registerFactory(const UnicodeString& id,
1233 Factory factory,
1234 Token context);
1235
1236 /**
1237 * Registers an instance <tt>obj</tt> of a subclass of
1238 * <code>Transliterator</code> with the system. When
1239 * <tt>createInstance()</tt> is called with an ID string that is
1240 * equal to <tt>obj->getID()</tt>, then <tt>obj->clone()</tt> is
1241 * returned.
1242 *
1243 * After this call the Transliterator class owns the adoptedObj
1244 * and will delete it.
1245 *
1246 * Because ICU may choose to cache Transliterators internally, this must
1247 * be called at application startup, prior to any calls to
1248 * Transliterator::createXXX to avoid undefined behavior.
1249 *
1250 * @param adoptedObj an instance of subclass of
1251 * <code>Transliterator</code> that defines <tt>clone()</tt>
1252 * @see #createInstance
1253 * @see #registerFactory
1254 * @see #unregister
1255 * @stable ICU 2.0
1256 */
1257 static void U_EXPORT2 registerInstance(Transliterator* adoptedObj);
1258
1259 /**
1260 * Registers an ID string as an alias of another ID string.
1261 * That is, after calling this function, <tt>createInstance(aliasID)</tt>
1262 * will return the same thing as <tt>createInstance(realID)</tt>.
1263 * This is generally used to create shorter, more mnemonic aliases
1264 * for long compound IDs.
1265 *
1266 * @param aliasID The new ID being registered.
1267 * @param realID The ID that the new ID is to be an alias for.
1268 * This can be a compound ID and can include filters and should
1269 * refer to transliterators that have already been registered with
1270 * the framework, although this isn't checked.
1271 * @stable ICU 3.6
1272 */
1273 static void U_EXPORT2 registerAlias(const UnicodeString& aliasID,
1274 const UnicodeString& realID);
1275
1276protected:
1277
1278#ifndef U_HIDE_INTERNAL_API
1279 /**
1280 * @param id the ID being registered
1281 * @param factory a function pointer that will be copied and
1282 * called later when the given ID is passed to createInstance()
1283 * @param context a context pointer that will be stored and
1284 * later passed to the factory function when an ID matching
1285 * the registration ID is being instantiated with this factory.
1286 * @internal
1287 */
1288 static void _registerFactory(const UnicodeString& id,
1289 Factory factory,
1290 Token context);
1291
1292 /**
1293 * @internal
1294 */
1295 static void _registerInstance(Transliterator* adoptedObj);
1296
1297 /**
1298 * @internal
1299 */
1300 static void _registerAlias(const UnicodeString& aliasID, const UnicodeString& realID);
1301
1302 /**
1303 * Register two targets as being inverses of one another. For
1304 * example, calling registerSpecialInverse("NFC", "NFD", true) causes
1305 * Transliterator to form the following inverse relationships:
1306 *
1307 * <pre>NFC => NFD
1308 * Any-NFC => Any-NFD
1309 * NFD => NFC
1310 * Any-NFD => Any-NFC</pre>
1311 *
1312 * (Without the special inverse registration, the inverse of NFC
1313 * would be NFC-Any.) Note that NFD is shorthand for Any-NFD, but
1314 * that the presence or absence of "Any-" is preserved.
1315 *
1316 * <p>The relationship is symmetrical; registering (a, b) is
1317 * equivalent to registering (b, a).
1318 *
1319 * <p>The relevant IDs must still be registered separately as
1320 * factories or classes.
1321 *
1322 * <p>Only the targets are specified. Special inverses always
1323 * have the form Any-Target1 <=> Any-Target2. The target should
1324 * have canonical casing (the casing desired to be produced when
1325 * an inverse is formed) and should contain no whitespace or other
1326 * extraneous characters.
1327 *
1328 * @param target the target against which to register the inverse
1329 * @param inverseTarget the inverse of target, that is
1330 * Any-target.getInverse() => Any-inverseTarget
1331 * @param bidirectional if true, register the reverse relation
1332 * as well, that is, Any-inverseTarget.getInverse() => Any-target
1333 * @internal
1334 */
1335 static void _registerSpecialInverse(const UnicodeString& target,
1336 const UnicodeString& inverseTarget,
1337 UBool bidirectional);
1338#endif /* U_HIDE_INTERNAL_API */
1339
1340public:
1341
1342 /**
1343 * Unregisters a transliterator or class. This may be either
1344 * a system transliterator or a user transliterator or class.
1345 * Any attempt to construct an unregistered transliterator based
1346 * on its ID will fail.
1347 *
1348 * Because ICU may choose to cache Transliterators internally, this should
1349 * be called during application shutdown, after all calls to
1350 * Transliterator::createXXX to avoid undefined behavior.
1351 *
1352 * @param ID the ID of the transliterator or class
1353 * @return the <code>Object</code> that was registered with
1354 * <code>ID</code>, or <code>null</code> if none was
1355 * @see #registerInstance
1356 * @see #registerFactory
1357 * @stable ICU 2.0
1358 */
1359 static void U_EXPORT2 unregister(const UnicodeString& ID);
1360
1361public:
1362
1363 /**
1364 * Return a StringEnumeration over the IDs available at the time of the
1365 * call, including user-registered IDs.
1366 * @param ec input-output error code
1367 * @return a newly-created StringEnumeration over the transliterators
1368 * available at the time of the call. The caller should delete this object
1369 * when done using it.
1370 * @stable ICU 3.0
1371 */
1372 static StringEnumeration* U_EXPORT2 getAvailableIDs(UErrorCode& ec);
1373
1374 /**
1375 * Return the number of registered source specifiers.
1376 * @return the number of registered source specifiers.
1377 * @stable ICU 2.0
1378 */
1379 static int32_t U_EXPORT2 countAvailableSources(void);
1380
1381 /**
1382 * Return a registered source specifier.
1383 * @param index which specifier to return, from 0 to n-1, where
1384 * n = countAvailableSources()
1385 * @param result fill-in paramter to receive the source specifier.
1386 * If index is out of range, result will be empty.
1387 * @return reference to result
1388 * @stable ICU 2.0
1389 */
1390 static UnicodeString& U_EXPORT2 getAvailableSource(int32_t index,
1391 UnicodeString& result);
1392
1393 /**
1394 * Return the number of registered target specifiers for a given
1395 * source specifier.
1396 * @param source the given source specifier.
1397 * @return the number of registered target specifiers for a given
1398 * source specifier.
1399 * @stable ICU 2.0
1400 */
1401 static int32_t U_EXPORT2 countAvailableTargets(const UnicodeString& source);
1402
1403 /**
1404 * Return a registered target specifier for a given source.
1405 * @param index which specifier to return, from 0 to n-1, where
1406 * n = countAvailableTargets(source)
1407 * @param source the source specifier
1408 * @param result fill-in paramter to receive the target specifier.
1409 * If source is invalid or if index is out of range, result will
1410 * be empty.
1411 * @return reference to result
1412 * @stable ICU 2.0
1413 */
1414 static UnicodeString& U_EXPORT2 getAvailableTarget(int32_t index,
1415 const UnicodeString& source,
1416 UnicodeString& result);
1417
1418 /**
1419 * Return the number of registered variant specifiers for a given
1420 * source-target pair.
1421 * @param source the source specifiers.
1422 * @param target the target specifiers.
1423 * @stable ICU 2.0
1424 */
1425 static int32_t U_EXPORT2 countAvailableVariants(const UnicodeString& source,
1426 const UnicodeString& target);
1427
1428 /**
1429 * Return a registered variant specifier for a given source-target
1430 * pair.
1431 * @param index which specifier to return, from 0 to n-1, where
1432 * n = countAvailableVariants(source, target)
1433 * @param source the source specifier
1434 * @param target the target specifier
1435 * @param result fill-in paramter to receive the variant
1436 * specifier. If source is invalid or if target is invalid or if
1437 * index is out of range, result will be empty.
1438 * @return reference to result
1439 * @stable ICU 2.0
1440 */
1441 static UnicodeString& U_EXPORT2 getAvailableVariant(int32_t index,
1442 const UnicodeString& source,
1443 const UnicodeString& target,
1444 UnicodeString& result);
1445
1446protected:
1447
1448#ifndef U_HIDE_INTERNAL_API
1449 /**
1450 * Non-mutexed internal method
1451 * @internal
1452 */
1453 static int32_t _countAvailableSources(void);
1454
1455 /**
1456 * Non-mutexed internal method
1457 * @internal
1458 */
1459 static UnicodeString& _getAvailableSource(int32_t index,
1460 UnicodeString& result);
1461
1462 /**
1463 * Non-mutexed internal method
1464 * @internal
1465 */
1466 static int32_t _countAvailableTargets(const UnicodeString& source);
1467
1468 /**
1469 * Non-mutexed internal method
1470 * @internal
1471 */
1472 static UnicodeString& _getAvailableTarget(int32_t index,
1473 const UnicodeString& source,
1474 UnicodeString& result);
1475
1476 /**
1477 * Non-mutexed internal method
1478 * @internal
1479 */
1480 static int32_t _countAvailableVariants(const UnicodeString& source,
1481 const UnicodeString& target);
1482
1483 /**
1484 * Non-mutexed internal method
1485 * @internal
1486 */
1487 static UnicodeString& _getAvailableVariant(int32_t index,
1488 const UnicodeString& source,
1489 const UnicodeString& target,
1490 UnicodeString& result);
1491#endif /* U_HIDE_INTERNAL_API */
1492
1493protected:
1494
1495 /**
1496 * Set the ID of this transliterators. Subclasses shouldn't do
1497 * this, unless the underlying script behavior has changed.
1498 * @param id the new id t to be set.
1499 * @stable ICU 2.4
1500 */
1501 void setID(const UnicodeString& id);
1502
1503public:
1504
1505 /**
1506 * Return the class ID for this class. This is useful only for
1507 * comparing to a return value from getDynamicClassID().
1508 * Note that Transliterator is an abstract base class, and therefor
1509 * no fully constructed object will have a dynamic
1510 * UCLassID that equals the UClassID returned from
1511 * TRansliterator::getStaticClassID().
1512 * @return The class ID for class Transliterator.
1513 * @stable ICU 2.0
1514 */
1515 static UClassID U_EXPORT2 getStaticClassID(void);
1516
1517 /**
1518 * Returns a unique class ID <b>polymorphically</b>. This method
1519 * is to implement a simple version of RTTI, since not all C++
1520 * compilers support genuine RTTI. Polymorphic operator==() and
1521 * clone() methods call this method.
1522 *
1523 * <p>Concrete subclasses of Transliterator must use the
1524 * UOBJECT_DEFINE_RTTI_IMPLEMENTATION macro from
1525 * uobject.h to provide the RTTI functions.
1526 *
1527 * @return The class ID for this object. All objects of a given
1528 * class have the same class ID. Objects of other classes have
1529 * different class IDs.
1530 * @stable ICU 2.0
1531 */
1532 virtual UClassID getDynamicClassID(void) const = 0;
1533
1534private:
1535 static UBool initializeRegistry(UErrorCode &status);
1536
1537public:
1538#ifndef U_HIDE_OBSOLETE_API
1539 /**
1540 * Return the number of IDs currently registered with the system.
1541 * To retrieve the actual IDs, call getAvailableID(i) with
1542 * i from 0 to countAvailableIDs() - 1.
1543 * @return the number of IDs currently registered with the system.
1544 * @obsolete ICU 3.4 use getAvailableIDs() instead
1545 */
1546 static int32_t U_EXPORT2 countAvailableIDs(void);
1547
1548 /**
1549 * Return the index-th available ID. index must be between 0
1550 * and countAvailableIDs() - 1, inclusive. If index is out of
1551 * range, the result of getAvailableID(0) is returned.
1552 * @param index the given ID index.
1553 * @return the index-th available ID. index must be between 0
1554 * and countAvailableIDs() - 1, inclusive. If index is out of
1555 * range, the result of getAvailableID(0) is returned.
1556 * @obsolete ICU 3.4 use getAvailableIDs() instead; this function
1557 * is not thread safe, since it returns a reference to storage that
1558 * may become invalid if another thread calls unregister
1559 */
1560 static const UnicodeString& U_EXPORT2 getAvailableID(int32_t index);
1561#endif /* U_HIDE_OBSOLETE_API */
1562};
1563
1564inline int32_t Transliterator::getMaximumContextLength(void) const {
1565 return maximumContextLength;
1566}
1567
1568inline void Transliterator::setID(const UnicodeString& id) {
1569 ID = id;
1570 // NUL-terminate the ID string, which is a non-aliased copy.
1571 ID.append((char16_t)0);
1572 ID.truncate(ID.length()-1);
1573}
1574
1575#ifndef U_HIDE_INTERNAL_API
1576inline Transliterator::Token Transliterator::integerToken(int32_t i) {
1577 Token t;
1578 t.integer = i;
1579 return t;
1580}
1581
1582inline Transliterator::Token Transliterator::pointerToken(void* p) {
1583 Token t;
1584 t.pointer = p;
1585 return t;
1586}
1587#endif /* U_HIDE_INTERNAL_API */
1588
1589U_NAMESPACE_END
1590
1591#endif /* #if !UCONFIG_NO_TRANSLITERATION */
1592
1593#endif /* U_SHOW_CPLUSPLUS_API */
1594
1595#endif
1596