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: Transforms 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
32 U_NAMESPACE_BEGIN
33
34 class UnicodeFilter;
35 class UnicodeSet;
36 class TransliteratorParser;
37 class NormalizationTransliterator;
38 class 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 * identity 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> properly. 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 < and > and <>.
259 *
260 * <p>Rule statements take one of the following forms:
261 *
262 * <dl>
263 * <dt><code>$alefmadda=\\u0622;</code></dt>
264 * <dd><strong>Variable definition.</strong> The name on the
265 * left is assigned the text on the right. In this example,
266 * after this statement, instances of the left hand name,
267 * "<code>$alefmadda</code>", 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 ("<code>$empty=;</code>").
274 * The right hand side may contain embedded <code>UnicodeSet</code>
275 * patterns, for example, "<code>$softvowel=[eiyEIY]</code>".</dd>
276 * <dt><code>ai>$alefmadda;</code></dt>
277 * <dd><strong>Forward translation rule.</strong> This rule
278 * states that the string on the left will be changed to the
279 * string on the right when performing forward
280 * transliteration.</dd>
281 * <dt><code>ai<$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<>$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 "<code>abc{def}</code>"
304 * indicates the characters "<code>def</code>" must be
305 * preceded by "<code>abc</code>" for a successful match.
306 * If there is a successful match, "<code>def</code>" will
307 * be replaced, but not "<code>abc</code>". The final '<code>}</code>'
308 * is optional, so "<code>abc{def</code>" is equivalent to
309 * "<code>abc{def}</code>". Another example is "<code>{123}456</code>"
310 * (or "<code>123}456</code>") in which the literal
311 * pattern "<code>123</code>" must be followed by "<code>456</code>".
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 preceding or following context by using the
321 * special character '@'. Examples:
322 *
323 * <pre>
324 * a {foo} z > | @ bar; # foo -> bar, move cursor before a
325 * {foo} xyz > bar @@|; # foo -> 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 "<code>$a=[a-z];$not_a=[^$a]</code>",
334 * or "<code>$range=a-z;$ll=[$range]</code>".
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>'*'; # One way to do this
341 * [aeiou]>'*'; # 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]) > $1 $1; # double lowercase letters
354 * ([:Lu:]) ([:Ll:]) > $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 * "<code>(</code>" and "<code>)</code>". Up to
359 * nine segments may be defined. Segments may not overlap. In the
360 * output string, "<code>$1</code>" through "<code>$9</code>"
361 * represent the input string segments, in left-to-right order of
362 * definition.
363 *
364 * <p><b>Anchors</b>
365 *
366 * <p>Patterns can be anchored to the beginning or the end of the text. This is done with the
367 * special characters '<code>^</code>' and '<code>$</code>'. For example:
368 *
369 * <pre>
370 * ^ a > 'BEG_A'; # match 'a' at start of text
371 * a > 'A'; # match other instances of 'a'
372 * z $ > 'END_Z'; # match 'z' at end of text
373 * z > 'Z'; # 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 character 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 * $x = [a-z$]; # match 'a' through 'z' OR anchor
384 * $x 1 > 2; # match '1' after a-z or at the start
385 * 3 $x > 4; # 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}>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>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>q</code></td>
405 * </tr>
406 * </table>
407 *
408 * <p>Applying these rules to the string "<code>adefabcdefz</code>"
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 "<code>def</code>"
446 * with "<code>xy</code>" 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 "<code>xyz</code>" 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 "<code>yz</code>"
455 * with "<code>q</code>".</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 "<code>'>'>o''clock</code>" changes the
475 * string "<code>></code>" to the string "<code>o'clock</code>".
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 * "a>x" is followed by the rule "ab>y",
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 */
490 class U_I18N_API Transliterator : public UObject {
491
492 private:
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
562 protected:
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
605 public:
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
687 * {@link #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 #transliterate(Replaceable&, UTransPosition&, const UnicodeString&, UErrorCode&) const }
745 * for details.
746 * @param text the buffer holding transliterated and
747 * untransliterated text
748 * @param index an array of three integers.
749 * @param status Output param to filled in with a success or an error.
750 * @see #transliterate(Replaceable&, UTransPosition&, const UnicodeString&, UErrorCode &) const
751 * @stable ICU 2.0
752 */
753 virtual void transliterate(Replaceable& text, UTransPosition& index,
754 UErrorCode& status) const;
755
756 /**
757 * Finishes any pending transliterations that were waiting for
758 * more characters. Clients should call this method as the last
759 * call after a sequence of one or more calls to
760 * <code>transliterate()</code>.
761 * @param text the buffer holding transliterated and
762 * untransliterated text.
763 * @param index the array of indices previously passed to {@link #transliterate }
764 * @stable ICU 2.0
765 */
766 virtual void finishTransliteration(Replaceable& text,
767 UTransPosition& index) const;
768
769 private:
770
771 /**
772 * This internal method does incremental transliteration. If the
773 * 'insertion' is non-null then we append it to 'text' before
774 * proceeding. This method calls through to the pure virtual
775 * framework method handleTransliterate() to do the actual
776 * work.
777 * @param text the buffer holding transliterated and
778 * untransliterated text
779 * @param index an array of three integers. See {@link
780 * #transliterate(Replaceable, int[], String)}.
781 * @param insertion text to be inserted and possibly
782 * transliterated into the translation buffer at
783 * <code>index.limit</code>.
784 * @param status Output param to filled in with a success or an error.
785 */
786 void _transliterate(Replaceable& text,
787 UTransPosition& index,
788 const UnicodeString* insertion,
789 UErrorCode &status) const;
790
791 protected:
792
793 /**
794 * Abstract method that concrete subclasses define to implement
795 * their transliteration algorithm. This method handles both
796 * incremental and non-incremental transliteration. Let
797 * <code>originalStart</code> refer to the value of
798 * <code>pos.start</code> upon entry.
799 *
800 * <ul>
801 * <li>If <code>incremental</code> is false, then this method
802 * should transliterate all characters between
803 * <code>pos.start</code> and <code>pos.limit</code>. Upon return
804 * <code>pos.start</code> must == <code> pos.limit</code>.</li>
805 *
806 * <li>If <code>incremental</code> is true, then this method
807 * should transliterate all characters between
808 * <code>pos.start</code> and <code>pos.limit</code> that can be
809 * unambiguously transliterated, regardless of future insertions
810 * of text at <code>pos.limit</code>. Upon return,
811 * <code>pos.start</code> should be in the range
812 * [<code>originalStart</code>, <code>pos.limit</code>).
813 * <code>pos.start</code> should be positioned such that
814 * characters [<code>originalStart</code>, <code>
815 * pos.start</code>) will not be changed in the future by this
816 * transliterator and characters [<code>pos.start</code>,
817 * <code>pos.limit</code>) are unchanged.</li>
818 * </ul>
819 *
820 * <p>Implementations of this method should also obey the
821 * following invariants:</p>
822 *
823 * <ul>
824 * <li> <code>pos.limit</code> and <code>pos.contextLimit</code>
825 * should be updated to reflect changes in length of the text
826 * between <code>pos.start</code> and <code>pos.limit</code>. The
827 * difference <code> pos.contextLimit - pos.limit</code> should
828 * not change.</li>
829 *
830 * <li><code>pos.contextStart</code> should not change.</li>
831 *
832 * <li>Upon return, neither <code>pos.start</code> nor
833 * <code>pos.limit</code> should be less than
834 * <code>originalStart</code>.</li>
835 *
836 * <li>Text before <code>originalStart</code> and text after
837 * <code>pos.limit</code> should not change.</li>
838 *
839 * <li>Text before <code>pos.contextStart</code> and text after
840 * <code> pos.contextLimit</code> should be ignored.</li>
841 * </ul>
842 *
843 * <p>Subclasses may safely assume that all characters in
844 * [<code>pos.start</code>, <code>pos.limit</code>) are filtered.
845 * In other words, the filter has already been applied by the time
846 * this method is called. See
847 * <code>filteredTransliterate()</code>.
848 *
849 * <p>This method is <b>not</b> for public consumption. Calling
850 * this method directly will transliterate
851 * [<code>pos.start</code>, <code>pos.limit</code>) without
852 * applying the filter. End user code should call <code>
853 * transliterate()</code> instead of this method. Subclass code
854 * and wrapping transliterators should call
855 * <code>filteredTransliterate()</code> instead of this method.<p>
856 *
857 * @param text the buffer holding transliterated and
858 * untransliterated text
859 *
860 * @param pos the indices indicating the start, limit, context
861 * start, and context limit of the text.
862 *
863 * @param incremental if true, assume more text may be inserted at
864 * <code>pos.limit</code> and act accordingly. Otherwise,
865 * transliterate all text between <code>pos.start</code> and
866 * <code>pos.limit</code> and move <code>pos.start</code> up to
867 * <code>pos.limit</code>.
868 *
869 * @see #transliterate
870 * @stable ICU 2.4
871 */
872 virtual void handleTransliterate(Replaceable& text,
873 UTransPosition& pos,
874 UBool incremental) const = 0;
875
876 public:
877 /**
878 * Transliterate a substring of text, as specified by index, taking filters
879 * into account. This method is for subclasses that need to delegate to
880 * another transliterator.
881 * @param text the text to be transliterated
882 * @param index the position indices
883 * @param incremental if true, then assume more characters may be inserted
884 * at index.limit, and postpone processing to accommodate future incoming
885 * characters
886 * @stable ICU 2.4
887 */
888 virtual void filteredTransliterate(Replaceable& text,
889 UTransPosition& index,
890 UBool incremental) const;
891
892 private:
893
894 /**
895 * Top-level transliteration method, handling filtering, incremental and
896 * non-incremental transliteration, and rollback. All transliteration
897 * public API methods eventually call this method with a rollback argument
898 * of true. Other entities may call this method but rollback should be
899 * false.
900 *
901 * <p>If this transliterator has a filter, break up the input text into runs
902 * of unfiltered characters. Pass each run to
903 * subclass.handleTransliterate().
904 *
905 * <p>In incremental mode, if rollback is true, perform a special
906 * incremental procedure in which several passes are made over the input
907 * text, adding one character at a time, and committing successful
908 * transliterations as they occur. Unsuccessful transliterations are rolled
909 * back and retried with additional characters to give correct results.
910 *
911 * @param text the text to be transliterated
912 * @param index the position indices
913 * @param incremental if true, then assume more characters may be inserted
914 * at index.limit, and postpone processing to accommodate future incoming
915 * characters
916 * @param rollback if true and if incremental is true, then perform special
917 * incremental processing, as described above, and undo partial
918 * transliterations where necessary. If incremental is false then this
919 * parameter is ignored.
920 */
921 virtual void filteredTransliterate(Replaceable& text,
922 UTransPosition& index,
923 UBool incremental,
924 UBool rollback) const;
925
926 public:
927
928 /**
929 * Returns the length of the longest context required by this transliterator.
930 * This is <em>preceding</em> context. The default implementation supplied
931 * by <code>Transliterator</code> returns zero; subclasses
932 * that use preceding context should override this method to return the
933 * correct value. For example, if a transliterator translates "ddd" (where
934 * d is any digit) to "555" when preceded by "(ddd)", then the preceding
935 * context length is 5, the length of "(ddd)".
936 *
937 * @return The maximum number of preceding context characters this
938 * transliterator needs to examine
939 * @stable ICU 2.0
940 */
941 int32_t getMaximumContextLength(void) const;
942
943 protected:
944
945 /**
946 * Method for subclasses to use to set the maximum context length.
947 * @param maxContextLength the new value to be set.
948 * @see #getMaximumContextLength
949 * @stable ICU 2.4
950 */
951 void setMaximumContextLength(int32_t maxContextLength);
952
953 public:
954
955 /**
956 * Returns a programmatic identifier for this transliterator.
957 * If this identifier is passed to <code>createInstance()</code>, it
958 * will return this object, if it has been registered.
959 * @return a programmatic identifier for this transliterator.
960 * @see #registerInstance
961 * @see #registerFactory
962 * @see #getAvailableIDs
963 * @stable ICU 2.0
964 */
965 virtual const UnicodeString& getID(void) const;
966
967 /**
968 * Returns a name for this transliterator that is appropriate for
969 * display to the user in the default locale. See {@link #getDisplayName }
970 * for details.
971 * @param ID the string identifier for this transliterator
972 * @param result Output param to receive the display name
973 * @return A reference to 'result'.
974 * @stable ICU 2.0
975 */
976 static UnicodeString& U_EXPORT2 getDisplayName(const UnicodeString& ID,
977 UnicodeString& result);
978
979 /**
980 * Returns a name for this transliterator that is appropriate for
981 * display to the user in the given locale. This name is taken
982 * from the locale resource data in the standard manner of the
983 * <code>java.text</code> package.
984 *
985 * <p>If no localized names exist in the system resource bundles,
986 * a name is synthesized using a localized
987 * <code>MessageFormat</code> pattern from the resource data. The
988 * arguments to this pattern are an integer followed by one or two
989 * strings. The integer is the number of strings, either 1 or 2.
990 * The strings are formed by splitting the ID for this
991 * transliterator at the first '-'. If there is no '-', then the
992 * entire ID forms the only string.
993 * @param ID the string identifier for this transliterator
994 * @param inLocale the Locale in which the display name should be
995 * localized.
996 * @param result Output param to receive the display name
997 * @return A reference to 'result'.
998 * @stable ICU 2.0
999 */
1000 static UnicodeString& U_EXPORT2 getDisplayName(const UnicodeString& ID,
1001 const Locale& inLocale,
1002 UnicodeString& result);
1003
1004 /**
1005 * Returns the filter used by this transliterator, or <tt>NULL</tt>
1006 * if this transliterator uses no filter.
1007 * @return the filter used by this transliterator, or <tt>NULL</tt>
1008 * if this transliterator uses no filter.
1009 * @stable ICU 2.0
1010 */
1011 const UnicodeFilter* getFilter(void) const;
1012
1013 /**
1014 * Returns the filter used by this transliterator, or <tt>NULL</tt> if this
1015 * transliterator uses no filter. The caller must eventually delete the
1016 * result. After this call, this transliterator's filter is set to
1017 * <tt>NULL</tt>.
1018 * @return the filter used by this transliterator, or <tt>NULL</tt> if this
1019 * transliterator uses no filter.
1020 * @stable ICU 2.4
1021 */
1022 UnicodeFilter* orphanFilter(void);
1023
1024 /**
1025 * Changes the filter used by this transliterator. If the filter
1026 * is set to <tt>null</tt> then no filtering will occur.
1027 *
1028 * <p>Callers must take care if a transliterator is in use by
1029 * multiple threads. The filter should not be changed by one
1030 * thread while another thread may be transliterating.
1031 * @param adoptedFilter the new filter to be adopted.
1032 * @stable ICU 2.0
1033 */
1034 void adoptFilter(UnicodeFilter* adoptedFilter);
1035
1036 /**
1037 * Returns this transliterator's inverse. See the class
1038 * documentation for details. This implementation simply inverts
1039 * the two entities in the ID and attempts to retrieve the
1040 * resulting transliterator. That is, if <code>getID()</code>
1041 * returns "A-B", then this method will return the result of
1042 * <code>createInstance("B-A")</code>, or <code>null</code> if that
1043 * call fails.
1044 *
1045 * <p>Subclasses with knowledge of their inverse may wish to
1046 * override this method.
1047 *
1048 * @param status Output param to filled in with a success or an error.
1049 * @return a transliterator that is an inverse, not necessarily
1050 * exact, of this transliterator, or <code>null</code> if no such
1051 * transliterator is registered.
1052 * @see #registerInstance
1053 * @stable ICU 2.0
1054 */
1055 Transliterator* createInverse(UErrorCode& status) const;
1056
1057 /**
1058 * Returns a <code>Transliterator</code> object given its ID.
1059 * The ID must be either a system transliterator ID or a ID registered
1060 * using <code>registerInstance()</code>.
1061 *
1062 * @param ID a valid ID, as enumerated by <code>getAvailableIDs()</code>
1063 * @param dir either FORWARD or REVERSE.
1064 * @param parseError Struct to receive information on position
1065 * of error if an error is encountered
1066 * @param status Output param to filled in with a success or an error.
1067 * @return A <code>Transliterator</code> object with the given ID
1068 * @see #registerInstance
1069 * @see #getAvailableIDs
1070 * @see #getID
1071 * @stable ICU 2.0
1072 */
1073 static Transliterator* U_EXPORT2 createInstance(const UnicodeString& ID,
1074 UTransDirection dir,
1075 UParseError& parseError,
1076 UErrorCode& status);
1077
1078 /**
1079 * Returns a <code>Transliterator</code> object given its ID.
1080 * The ID must be either a system transliterator ID or a ID registered
1081 * using <code>registerInstance()</code>.
1082 * @param ID a valid ID, as enumerated by <code>getAvailableIDs()</code>
1083 * @param dir either FORWARD or REVERSE.
1084 * @param status Output param to filled in with a success or an error.
1085 * @return A <code>Transliterator</code> object with the given ID
1086 * @stable ICU 2.0
1087 */
1088 static Transliterator* U_EXPORT2 createInstance(const UnicodeString& ID,
1089 UTransDirection dir,
1090 UErrorCode& status);
1091
1092 /**
1093 * Returns a <code>Transliterator</code> object constructed from
1094 * the given rule string. This will be a rule-based Transliterator,
1095 * if the rule string contains only rules, or a
1096 * compound Transliterator, if it contains ID blocks, or a
1097 * null Transliterator, if it contains ID blocks which parse as
1098 * empty for the given direction.
1099 *
1100 * @param ID the id for the transliterator.
1101 * @param rules rules, separated by ';'
1102 * @param dir either FORWARD or REVERSE.
1103 * @param parseError Struct to receive information on position
1104 * of error if an error is encountered
1105 * @param status Output param set to success/failure code.
1106 * @return a newly created Transliterator
1107 * @stable ICU 2.0
1108 */
1109 static Transliterator* U_EXPORT2 createFromRules(const UnicodeString& ID,
1110 const UnicodeString& rules,
1111 UTransDirection dir,
1112 UParseError& parseError,
1113 UErrorCode& status);
1114
1115 /**
1116 * Create a rule string that can be passed to createFromRules()
1117 * to recreate this transliterator.
1118 * @param result the string to receive the rules. Previous
1119 * contents will be deleted.
1120 * @param escapeUnprintable if true then convert unprintable
1121 * character to their hex escape representations, \\uxxxx or
1122 * \\Uxxxxxxxx. Unprintable characters are those other than
1123 * U+000A, U+0020..U+007E.
1124 * @stable ICU 2.0
1125 */
1126 virtual UnicodeString& toRules(UnicodeString& result,
1127 UBool escapeUnprintable) const;
1128
1129 /**
1130 * Return the number of elements that make up this transliterator.
1131 * For example, if the transliterator "NFD;Jamo-Latin;Latin-Greek"
1132 * were created, the return value of this method would be 3.
1133 *
1134 * <p>If this transliterator is not composed of other
1135 * transliterators, then this method returns 1.
1136 * @return the number of transliterators that compose this
1137 * transliterator, or 1 if this transliterator is not composed of
1138 * multiple transliterators
1139 * @stable ICU 3.0
1140 */
1141 int32_t countElements() const;
1142
1143 /**
1144 * Return an element that makes up this transliterator. For
1145 * example, if the transliterator "NFD;Jamo-Latin;Latin-Greek"
1146 * were created, the return value of this method would be one
1147 * of the three transliterator objects that make up that
1148 * transliterator: [NFD, Jamo-Latin, Latin-Greek].
1149 *
1150 * <p>If this transliterator is not composed of other
1151 * transliterators, then this method will return a reference to
1152 * this transliterator when given the index 0.
1153 * @param index a value from 0..countElements()-1 indicating the
1154 * transliterator to return
1155 * @param ec input-output error code
1156 * @return one of the transliterators that makes up this
1157 * transliterator, if this transliterator is made up of multiple
1158 * transliterators, otherwise a reference to this object if given
1159 * an index of 0
1160 * @stable ICU 3.0
1161 */
1162 const Transliterator& getElement(int32_t index, UErrorCode& ec) const;
1163
1164 /**
1165 * Returns the set of all characters that may be modified in the
1166 * input text by this Transliterator. This incorporates this
1167 * object's current filter; if the filter is changed, the return
1168 * value of this function will change. The default implementation
1169 * returns an empty set. Some subclasses may override
1170 * {@link #handleGetSourceSet } to return a more precise result. The
1171 * return result is approximate in any case and is intended for
1172 * use by tests, tools, or utilities.
1173 * @param result receives result set; previous contents lost
1174 * @return a reference to result
1175 * @see #getTargetSet
1176 * @see #handleGetSourceSet
1177 * @stable ICU 2.4
1178 */
1179 UnicodeSet& getSourceSet(UnicodeSet& result) const;
1180
1181 /**
1182 * Framework method that returns the set of all characters that
1183 * may be modified in the input text by this Transliterator,
1184 * ignoring the effect of this object's filter. The base class
1185 * implementation returns the empty set. Subclasses that wish to
1186 * implement this should override this method.
1187 * @return the set of characters that this transliterator may
1188 * modify. The set may be modified, so subclasses should return a
1189 * newly-created object.
1190 * @param result receives result set; previous contents lost
1191 * @see #getSourceSet
1192 * @see #getTargetSet
1193 * @stable ICU 2.4
1194 */
1195 virtual void handleGetSourceSet(UnicodeSet& result) const;
1196
1197 /**
1198 * Returns the set of all characters that may be generated as
1199 * replacement text by this transliterator. The default
1200 * implementation returns the empty set. Some subclasses may
1201 * override this method to return a more precise result. The
1202 * return result is approximate in any case and is intended for
1203 * use by tests, tools, or utilities requiring such
1204 * meta-information.
1205 * @param result receives result set; previous contents lost
1206 * @return a reference to result
1207 * @see #getTargetSet
1208 * @stable ICU 2.4
1209 */
1210 virtual UnicodeSet& getTargetSet(UnicodeSet& result) const;
1211
1212 public:
1213
1214 /**
1215 * Registers a factory function that creates transliterators of
1216 * a given ID.
1217 *
1218 * Because ICU may choose to cache Transliterators internally, this must
1219 * be called at application startup, prior to any calls to
1220 * Transliterator::createXXX to avoid undefined behavior.
1221 *
1222 * @param id the ID being registered
1223 * @param factory a function pointer that will be copied and
1224 * called later when the given ID is passed to createInstance()
1225 * @param context a context pointer that will be stored and
1226 * later passed to the factory function when an ID matching
1227 * the registration ID is being instantiated with this factory.
1228 * @stable ICU 2.0
1229 */
1230 static void U_EXPORT2 registerFactory(const UnicodeString& id,
1231 Factory factory,
1232 Token context);
1233
1234 /**
1235 * Registers an instance <tt>obj</tt> of a subclass of
1236 * <code>Transliterator</code> with the system. When
1237 * <tt>createInstance()</tt> is called with an ID string that is
1238 * equal to <tt>obj->getID()</tt>, then <tt>obj->clone()</tt> is
1239 * returned.
1240 *
1241 * After this call the Transliterator class owns the adoptedObj
1242 * and will delete it.
1243 *
1244 * Because ICU may choose to cache Transliterators internally, this must
1245 * be called at application startup, prior to any calls to
1246 * Transliterator::createXXX to avoid undefined behavior.
1247 *
1248 * @param adoptedObj an instance of subclass of
1249 * <code>Transliterator</code> that defines <tt>clone()</tt>
1250 * @see #createInstance
1251 * @see #registerFactory
1252 * @see #unregister
1253 * @stable ICU 2.0
1254 */
1255 static void U_EXPORT2 registerInstance(Transliterator* adoptedObj);
1256
1257 /**
1258 * Registers an ID string as an alias of another ID string.
1259 * That is, after calling this function, <tt>createInstance(aliasID)</tt>
1260 * will return the same thing as <tt>createInstance(realID)</tt>.
1261 * This is generally used to create shorter, more mnemonic aliases
1262 * for long compound IDs.
1263 *
1264 * @param aliasID The new ID being registered.
1265 * @param realID The ID that the new ID is to be an alias for.
1266 * This can be a compound ID and can include filters and should
1267 * refer to transliterators that have already been registered with
1268 * the framework, although this isn't checked.
1269 * @stable ICU 3.6
1270 */
1271 static void U_EXPORT2 registerAlias(const UnicodeString& aliasID,
1272 const UnicodeString& realID);
1273
1274 protected:
1275
1276 #ifndef U_HIDE_INTERNAL_API
1277 /**
1278 * @param id the ID being registered
1279 * @param factory a function pointer that will be copied and
1280 * called later when the given ID is passed to createInstance()
1281 * @param context a context pointer that will be stored and
1282 * later passed to the factory function when an ID matching
1283 * the registration ID is being instantiated with this factory.
1284 * @internal
1285 */
1286 static void _registerFactory(const UnicodeString& id,
1287 Factory factory,
1288 Token context);
1289
1290 /**
1291 * @internal
1292 */
1293 static void _registerInstance(Transliterator* adoptedObj);
1294
1295 /**
1296 * @internal
1297 */
1298 static void _registerAlias(const UnicodeString& aliasID, const UnicodeString& realID);
1299
1300 /**
1301 * Register two targets as being inverses of one another. For
1302 * example, calling registerSpecialInverse("NFC", "NFD", true) causes
1303 * Transliterator to form the following inverse relationships:
1304 *
1305 * <pre>NFC => NFD
1306 * Any-NFC => Any-NFD
1307 * NFD => NFC
1308 * Any-NFD => Any-NFC</pre>
1309 *
1310 * (Without the special inverse registration, the inverse of NFC
1311 * would be NFC-Any.) Note that NFD is shorthand for Any-NFD, but
1312 * that the presence or absence of "Any-" is preserved.
1313 *
1314 * <p>The relationship is symmetrical; registering (a, b) is
1315 * equivalent to registering (b, a).
1316 *
1317 * <p>The relevant IDs must still be registered separately as
1318 * factories or classes.
1319 *
1320 * <p>Only the targets are specified. Special inverses always
1321 * have the form Any-Target1 <=> Any-Target2. The target should
1322 * have canonical casing (the casing desired to be produced when
1323 * an inverse is formed) and should contain no whitespace or other
1324 * extraneous characters.
1325 *
1326 * @param target the target against which to register the inverse
1327 * @param inverseTarget the inverse of target, that is
1328 * Any-target.getInverse() => Any-inverseTarget
1329 * @param bidirectional if true, register the reverse relation
1330 * as well, that is, Any-inverseTarget.getInverse() => Any-target
1331 * @internal
1332 */
1333 static void _registerSpecialInverse(const UnicodeString& target,
1334 const UnicodeString& inverseTarget,
1335 UBool bidirectional);
1336 #endif /* U_HIDE_INTERNAL_API */
1337
1338 public:
1339
1340 /**
1341 * Unregisters a transliterator or class. This may be either
1342 * a system transliterator or a user transliterator or class.
1343 * Any attempt to construct an unregistered transliterator based
1344 * on its ID will fail.
1345 *
1346 * Because ICU may choose to cache Transliterators internally, this should
1347 * be called during application shutdown, after all calls to
1348 * Transliterator::createXXX to avoid undefined behavior.
1349 *
1350 * @param ID the ID of the transliterator or class
1351 * @return the <code>Object</code> that was registered with
1352 * <code>ID</code>, or <code>null</code> if none was
1353 * @see #registerInstance
1354 * @see #registerFactory
1355 * @stable ICU 2.0
1356 */
1357 static void U_EXPORT2 unregister(const UnicodeString& ID);
1358
1359 public:
1360
1361 /**
1362 * Return a StringEnumeration over the IDs available at the time of the
1363 * call, including user-registered IDs.
1364 * @param ec input-output error code
1365 * @return a newly-created StringEnumeration over the transliterators
1366 * available at the time of the call. The caller should delete this object
1367 * when done using it.
1368 * @stable ICU 3.0
1369 */
1370 static StringEnumeration* U_EXPORT2 getAvailableIDs(UErrorCode& ec);
1371
1372 /**
1373 * Return the number of registered source specifiers.
1374 * @return the number of registered source specifiers.
1375 * @stable ICU 2.0
1376 */
1377 static int32_t U_EXPORT2 countAvailableSources(void);
1378
1379 /**
1380 * Return a registered source specifier.
1381 * @param index which specifier to return, from 0 to n-1, where
1382 * n = countAvailableSources()
1383 * @param result fill-in parameter to receive the source specifier.
1384 * If index is out of range, result will be empty.
1385 * @return reference to result
1386 * @stable ICU 2.0
1387 */
1388 static UnicodeString& U_EXPORT2 getAvailableSource(int32_t index,
1389 UnicodeString& result);
1390
1391 /**
1392 * Return the number of registered target specifiers for a given
1393 * source specifier.
1394 * @param source the given source specifier.
1395 * @return the number of registered target specifiers for a given
1396 * source specifier.
1397 * @stable ICU 2.0
1398 */
1399 static int32_t U_EXPORT2 countAvailableTargets(const UnicodeString& source);
1400
1401 /**
1402 * Return a registered target specifier for a given source.
1403 * @param index which specifier to return, from 0 to n-1, where
1404 * n = countAvailableTargets(source)
1405 * @param source the source specifier
1406 * @param result fill-in parameter to receive the target specifier.
1407 * If source is invalid or if index is out of range, result will
1408 * be empty.
1409 * @return reference to result
1410 * @stable ICU 2.0
1411 */
1412 static UnicodeString& U_EXPORT2 getAvailableTarget(int32_t index,
1413 const UnicodeString& source,
1414 UnicodeString& result);
1415
1416 /**
1417 * Return the number of registered variant specifiers for a given
1418 * source-target pair.
1419 * @param source the source specifiers.
1420 * @param target the target specifiers.
1421 * @stable ICU 2.0
1422 */
1423 static int32_t U_EXPORT2 countAvailableVariants(const UnicodeString& source,
1424 const UnicodeString& target);
1425
1426 /**
1427 * Return a registered variant specifier for a given source-target
1428 * pair.
1429 * @param index which specifier to return, from 0 to n-1, where
1430 * n = countAvailableVariants(source, target)
1431 * @param source the source specifier
1432 * @param target the target specifier
1433 * @param result fill-in parameter to receive the variant
1434 * specifier. If source is invalid or if target is invalid or if
1435 * index is out of range, result will be empty.
1436 * @return reference to result
1437 * @stable ICU 2.0
1438 */
1439 static UnicodeString& U_EXPORT2 getAvailableVariant(int32_t index,
1440 const UnicodeString& source,
1441 const UnicodeString& target,
1442 UnicodeString& result);
1443
1444 protected:
1445
1446 #ifndef U_HIDE_INTERNAL_API
1447 /**
1448 * Non-mutexed internal method
1449 * @internal
1450 */
1451 static int32_t _countAvailableSources(void);
1452
1453 /**
1454 * Non-mutexed internal method
1455 * @internal
1456 */
1457 static UnicodeString& _getAvailableSource(int32_t index,
1458 UnicodeString& result);
1459
1460 /**
1461 * Non-mutexed internal method
1462 * @internal
1463 */
1464 static int32_t _countAvailableTargets(const UnicodeString& source);
1465
1466 /**
1467 * Non-mutexed internal method
1468 * @internal
1469 */
1470 static UnicodeString& _getAvailableTarget(int32_t index,
1471 const UnicodeString& source,
1472 UnicodeString& result);
1473
1474 /**
1475 * Non-mutexed internal method
1476 * @internal
1477 */
1478 static int32_t _countAvailableVariants(const UnicodeString& source,
1479 const UnicodeString& target);
1480
1481 /**
1482 * Non-mutexed internal method
1483 * @internal
1484 */
1485 static UnicodeString& _getAvailableVariant(int32_t index,
1486 const UnicodeString& source,
1487 const UnicodeString& target,
1488 UnicodeString& result);
1489 #endif /* U_HIDE_INTERNAL_API */
1490
1491 protected:
1492
1493 /**
1494 * Set the ID of this transliterators. Subclasses shouldn't do
1495 * this, unless the underlying script behavior has changed.
1496 * @param id the new id t to be set.
1497 * @stable ICU 2.4
1498 */
1499 void setID(const UnicodeString& id);
1500
1501 public:
1502
1503 /**
1504 * Return the class ID for this class. This is useful only for
1505 * comparing to a return value from getDynamicClassID().
1506 * Note that Transliterator is an abstract base class, and therefor
1507 * no fully constructed object will have a dynamic
1508 * UCLassID that equals the UClassID returned from
1509 * TRansliterator::getStaticClassID().
1510 * @return The class ID for class Transliterator.
1511 * @stable ICU 2.0
1512 */
1513 static UClassID U_EXPORT2 getStaticClassID(void);
1514
1515 /**
1516 * Returns a unique class ID <b>polymorphically</b>. This method
1517 * is to implement a simple version of RTTI, since not all C++
1518 * compilers support genuine RTTI. Polymorphic operator==() and
1519 * clone() methods call this method.
1520 *
1521 * <p>Concrete subclasses of Transliterator must use the
1522 * UOBJECT_DEFINE_RTTI_IMPLEMENTATION macro from
1523 * uobject.h to provide the RTTI functions.
1524 *
1525 * @return The class ID for this object. All objects of a given
1526 * class have the same class ID. Objects of other classes have
1527 * different class IDs.
1528 * @stable ICU 2.0
1529 */
1530 virtual UClassID getDynamicClassID(void) const override = 0;
1531
1532 private:
1533 static UBool initializeRegistry(UErrorCode &status);
1534
1535 public:
1536 #ifndef U_HIDE_OBSOLETE_API
1537 /**
1538 * Return the number of IDs currently registered with the system.
1539 * To retrieve the actual IDs, call getAvailableID(i) with
1540 * i from 0 to countAvailableIDs() - 1.
1541 * @return the number of IDs currently registered with the system.
1542 * @obsolete ICU 3.4 use getAvailableIDs() instead
1543 */
1544 static int32_t U_EXPORT2 countAvailableIDs(void);
1545
1546 /**
1547 * Return the index-th available ID. index must be between 0
1548 * and countAvailableIDs() - 1, inclusive. If index is out of
1549 * range, the result of getAvailableID(0) is returned.
1550 * @param index the given ID index.
1551 * @return the index-th available ID. index must be between 0
1552 * and countAvailableIDs() - 1, inclusive. If index is out of
1553 * range, the result of getAvailableID(0) is returned.
1554 * @obsolete ICU 3.4 use getAvailableIDs() instead; this function
1555 * is not thread safe, since it returns a reference to storage that
1556 * may become invalid if another thread calls unregister
1557 */
1558 static const UnicodeString& U_EXPORT2 getAvailableID(int32_t index);
1559 #endif /* U_HIDE_OBSOLETE_API */
1560 };
1561
getMaximumContextLength(void)1562 inline int32_t Transliterator::getMaximumContextLength(void) const {
1563 return maximumContextLength;
1564 }
1565
setID(const UnicodeString & id)1566 inline void Transliterator::setID(const UnicodeString& id) {
1567 ID = id;
1568 // NUL-terminate the ID string, which is a non-aliased copy.
1569 ID.append((char16_t)0);
1570 ID.truncate(ID.length()-1);
1571 }
1572
1573 #ifndef U_HIDE_INTERNAL_API
integerToken(int32_t i)1574 inline Transliterator::Token Transliterator::integerToken(int32_t i) {
1575 Token t;
1576 t.integer = i;
1577 return t;
1578 }
1579
pointerToken(void * p)1580 inline Transliterator::Token Transliterator::pointerToken(void* p) {
1581 Token t;
1582 t.pointer = p;
1583 return t;
1584 }
1585 #endif /* U_HIDE_INTERNAL_API */
1586
1587 U_NAMESPACE_END
1588
1589 #endif /* #if !UCONFIG_NO_TRANSLITERATION */
1590
1591 #endif /* U_SHOW_CPLUSPLUS_API */
1592
1593 #endif
1594