]>
Commit | Line | Data |
---|---|---|
b75a7d8f A |
1 | /* |
2 | ******************************************************************************* | |
374ca955 | 3 | * Copyright (C) 1997-2004, International Business Machines Corporation and * |
b75a7d8f A |
4 | * others. All Rights Reserved. * |
5 | ******************************************************************************* | |
6 | * | |
7 | * File MSGFMT.CPP | |
8 | * | |
9 | * Modification History: | |
10 | * | |
11 | * Date Name Description | |
12 | * 02/19/97 aliu Converted from java. | |
13 | * 03/20/97 helena Finished first cut of implementation. | |
14 | * 04/10/97 aliu Made to work on AIX. Added stoi to replace wtoi. | |
15 | * 06/11/97 helena Fixed addPattern to take the pattern correctly. | |
16 | * 06/17/97 helena Fixed the getPattern to return the correct pattern. | |
17 | * 07/09/97 helena Made ParsePosition into a class. | |
18 | * 02/22/99 stephen Removed character literals for EBCDIC safety | |
19 | ******************************************************************************** | |
20 | */ | |
21 | ||
22 | #include "unicode/utypes.h" | |
23 | ||
24 | #if !UCONFIG_NO_FORMATTING | |
25 | ||
26 | #include "unicode/msgfmt.h" | |
27 | #include "unicode/decimfmt.h" | |
28 | #include "unicode/datefmt.h" | |
29 | #include "unicode/smpdtfmt.h" | |
30 | #include "unicode/choicfmt.h" | |
31 | #include "unicode/ustring.h" | |
32 | #include "unicode/ucnv_err.h" | |
33 | #include "unicode/uchar.h" | |
374ca955 | 34 | #include "unicode/rbnf.h" |
b75a7d8f A |
35 | #include "ustrfmt.h" |
36 | #include "cmemory.h" | |
374ca955 | 37 | #include "util.h" |
b75a7d8f A |
38 | #include "uassert.h" |
39 | ||
40 | // ***************************************************************************** | |
41 | // class MessageFormat | |
42 | // ***************************************************************************** | |
43 | ||
44 | #define COMMA ((UChar)0x002C) | |
45 | #define SINGLE_QUOTE ((UChar)0x0027) | |
46 | #define LEFT_CURLY_BRACE ((UChar)0x007B) | |
47 | #define RIGHT_CURLY_BRACE ((UChar)0x007D) | |
48 | ||
49 | //--------------------------------------- | |
50 | // static data | |
51 | ||
374ca955 A |
52 | static const UChar ID_EMPTY[] = { |
53 | 0 /* empty string, used for default so that null can mark end of list */ | |
54 | }; | |
55 | ||
b75a7d8f A |
56 | static const UChar ID_NUMBER[] = { |
57 | 0x6E, 0x75, 0x6D, 0x62, 0x65, 0x72, 0 /* "number" */ | |
58 | }; | |
59 | static const UChar ID_DATE[] = { | |
60 | 0x64, 0x61, 0x74, 0x65, 0 /* "date" */ | |
61 | }; | |
62 | static const UChar ID_TIME[] = { | |
63 | 0x74, 0x69, 0x6D, 0x65, 0 /* "time" */ | |
64 | }; | |
65 | static const UChar ID_CHOICE[] = { | |
66 | 0x63, 0x68, 0x6F, 0x69, 0x63, 0x65, 0 /* "choice" */ | |
67 | }; | |
374ca955 A |
68 | static const UChar ID_SPELLOUT[] = { |
69 | 0x73, 0x70, 0x65, 0x6c, 0x6c, 0x6f, 0x75, 0x74, 0 /* "spellout" */ | |
70 | }; | |
71 | static const UChar ID_ORDINAL[] = { | |
72 | 0x6f, 0x72, 0x64, 0x69, 0x6e, 0x61, 0x6c, 0 /* "ordinal" */ | |
73 | }; | |
74 | static const UChar ID_DURATION[] = { | |
75 | 0x64, 0x75, 0x72, 0x61, 0x74, 0x69, 0x6f, 0x6e, 0 /* "duration" */ | |
76 | }; | |
b75a7d8f A |
77 | |
78 | // MessageFormat Type List Number, Date, Time or Choice | |
79 | static const UChar * const TYPE_IDS[] = { | |
374ca955 | 80 | ID_EMPTY, |
b75a7d8f A |
81 | ID_NUMBER, |
82 | ID_DATE, | |
83 | ID_TIME, | |
374ca955 A |
84 | ID_CHOICE, |
85 | ID_SPELLOUT, | |
86 | ID_ORDINAL, | |
87 | ID_DURATION, | |
88 | NULL, | |
b75a7d8f A |
89 | }; |
90 | ||
91 | static const UChar ID_CURRENCY[] = { | |
92 | 0x63, 0x75, 0x72, 0x72, 0x65, 0x6E, 0x63, 0x79, 0 /* "currency" */ | |
93 | }; | |
94 | static const UChar ID_PERCENT[] = { | |
95 | 0x70, 0x65, 0x72, 0x63, 0x65, 0x6E, 0x74, 0 /* "percent" */ | |
96 | }; | |
97 | static const UChar ID_INTEGER[] = { | |
98 | 0x69, 0x6E, 0x74, 0x65, 0x67, 0x65, 0x72, 0 /* "integer" */ | |
99 | }; | |
100 | ||
101 | // NumberFormat modifier list, default, currency, percent or integer | |
102 | static const UChar * const NUMBER_STYLE_IDS[] = { | |
374ca955 | 103 | ID_EMPTY, |
b75a7d8f A |
104 | ID_CURRENCY, |
105 | ID_PERCENT, | |
106 | ID_INTEGER, | |
107 | NULL, | |
108 | }; | |
374ca955 | 109 | |
b75a7d8f A |
110 | static const UChar ID_SHORT[] = { |
111 | 0x73, 0x68, 0x6F, 0x72, 0x74, 0 /* "short" */ | |
112 | }; | |
113 | static const UChar ID_MEDIUM[] = { | |
114 | 0x6D, 0x65, 0x64, 0x69, 0x75, 0x6D, 0 /* "medium" */ | |
115 | }; | |
116 | static const UChar ID_LONG[] = { | |
117 | 0x6C, 0x6F, 0x6E, 0x67, 0 /* "long" */ | |
118 | }; | |
119 | static const UChar ID_FULL[] = { | |
120 | 0x66, 0x75, 0x6C, 0x6C, 0 /* "full" */ | |
121 | }; | |
122 | ||
123 | // DateFormat modifier list, default, short, medium, long or full | |
124 | static const UChar * const DATE_STYLE_IDS[] = { | |
374ca955 | 125 | ID_EMPTY, |
b75a7d8f A |
126 | ID_SHORT, |
127 | ID_MEDIUM, | |
128 | ID_LONG, | |
374ca955 A |
129 | ID_FULL, |
130 | NULL, | |
b75a7d8f A |
131 | }; |
132 | ||
133 | static const DateFormat::EStyle DATE_STYLES[] = { | |
134 | DateFormat::kDefault, | |
135 | DateFormat::kShort, | |
136 | DateFormat::kMedium, | |
137 | DateFormat::kLong, | |
138 | DateFormat::kFull, | |
139 | }; | |
140 | ||
b75a7d8f A |
141 | static const int32_t DEFAULT_INITIAL_CAPACITY = 10; |
142 | ||
143 | U_NAMESPACE_BEGIN | |
144 | ||
145 | // ------------------------------------- | |
374ca955 | 146 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(MessageFormat) |
b75a7d8f A |
147 | |
148 | //-------------------------------------------------------------------- | |
149 | ||
150 | /** | |
151 | * Convert a string to an unsigned decimal, ignoring rule whitespace. | |
152 | * @return a non-negative number if successful, or a negative number | |
153 | * upon failure. | |
154 | */ | |
155 | static int32_t stou(const UnicodeString& string) { | |
156 | int32_t n = 0; | |
157 | int32_t count = 0; | |
158 | UChar32 c; | |
159 | for (int32_t i=0; i<string.length(); i+=U16_LENGTH(c)) { | |
160 | c = string.char32At(i); | |
161 | if (uprv_isRuleWhiteSpace(c)) { | |
162 | continue; | |
163 | } | |
164 | int32_t d = u_digit(c, 10); | |
165 | if (d < 0 || ++count > 10) { | |
166 | return -1; | |
167 | } | |
168 | n = 10*n + d; | |
169 | } | |
170 | return n; | |
171 | } | |
172 | ||
173 | /** | |
174 | * Convert an integer value to a string and append the result to | |
175 | * the given UnicodeString. | |
176 | */ | |
177 | static UnicodeString& itos(int32_t i, UnicodeString& appendTo) { | |
178 | UChar temp[16]; | |
179 | uprv_itou(temp,16,i,10,0); // 10 == radix | |
180 | appendTo.append(temp); | |
181 | return appendTo; | |
182 | } | |
183 | ||
184 | // ------------------------------------- | |
185 | // Creates a MessageFormat instance based on the pattern. | |
186 | ||
187 | MessageFormat::MessageFormat(const UnicodeString& pattern, | |
188 | UErrorCode& success) | |
189 | : fLocale(Locale::getDefault()), // Uses the default locale | |
190 | formatAliases(NULL), | |
191 | formatAliasesCapacity(0), | |
192 | subformats(NULL), | |
193 | subformatCount(0), | |
194 | subformatCapacity(0), | |
195 | argTypes(NULL), | |
196 | argTypeCount(0), | |
197 | argTypeCapacity(0), | |
198 | defaultNumberFormat(NULL), | |
199 | defaultDateFormat(NULL) | |
200 | { | |
201 | if (!allocateSubformats(DEFAULT_INITIAL_CAPACITY) || | |
202 | !allocateArgTypes(DEFAULT_INITIAL_CAPACITY)) { | |
203 | success = U_MEMORY_ALLOCATION_ERROR; | |
204 | return; | |
205 | } | |
206 | applyPattern(pattern, success); | |
374ca955 | 207 | setLocaleIDs(fLocale.getName(), fLocale.getName()); |
b75a7d8f A |
208 | } |
209 | ||
210 | MessageFormat::MessageFormat(const UnicodeString& pattern, | |
211 | const Locale& newLocale, | |
212 | UErrorCode& success) | |
213 | : fLocale(newLocale), | |
214 | formatAliases(NULL), | |
215 | formatAliasesCapacity(0), | |
216 | subformats(NULL), | |
217 | subformatCount(0), | |
218 | subformatCapacity(0), | |
219 | argTypes(NULL), | |
220 | argTypeCount(0), | |
221 | argTypeCapacity(0), | |
222 | defaultNumberFormat(NULL), | |
223 | defaultDateFormat(NULL) | |
224 | { | |
225 | if (!allocateSubformats(DEFAULT_INITIAL_CAPACITY) || | |
226 | !allocateArgTypes(DEFAULT_INITIAL_CAPACITY)) { | |
227 | success = U_MEMORY_ALLOCATION_ERROR; | |
228 | return; | |
229 | } | |
230 | applyPattern(pattern, success); | |
374ca955 | 231 | setLocaleIDs(fLocale.getName(), fLocale.getName()); |
b75a7d8f A |
232 | } |
233 | ||
234 | MessageFormat::MessageFormat(const UnicodeString& pattern, | |
235 | const Locale& newLocale, | |
236 | UParseError& parseError, | |
237 | UErrorCode& success) | |
238 | : fLocale(newLocale), | |
239 | formatAliases(NULL), | |
240 | formatAliasesCapacity(0), | |
241 | subformats(NULL), | |
242 | subformatCount(0), | |
243 | subformatCapacity(0), | |
244 | argTypes(NULL), | |
245 | argTypeCount(0), | |
246 | argTypeCapacity(0), | |
247 | defaultNumberFormat(NULL), | |
248 | defaultDateFormat(NULL) | |
249 | { | |
250 | if (!allocateSubformats(DEFAULT_INITIAL_CAPACITY) || | |
251 | !allocateArgTypes(DEFAULT_INITIAL_CAPACITY)) { | |
252 | success = U_MEMORY_ALLOCATION_ERROR; | |
253 | return; | |
254 | } | |
255 | applyPattern(pattern, parseError, success); | |
374ca955 | 256 | setLocaleIDs(fLocale.getName(), fLocale.getName()); |
b75a7d8f A |
257 | } |
258 | ||
259 | MessageFormat::MessageFormat(const MessageFormat& that) | |
260 | : Format(that), | |
261 | formatAliases(NULL), | |
262 | formatAliasesCapacity(0), | |
263 | subformats(NULL), | |
264 | subformatCount(0), | |
265 | subformatCapacity(0), | |
266 | argTypes(NULL), | |
267 | argTypeCount(0), | |
268 | argTypeCapacity(0), | |
269 | defaultNumberFormat(NULL), | |
270 | defaultDateFormat(NULL) | |
271 | { | |
272 | *this = that; | |
273 | } | |
274 | ||
275 | MessageFormat::~MessageFormat() | |
276 | { | |
277 | int32_t idx; | |
278 | for (idx = 0; idx < subformatCount; idx++) { | |
279 | delete subformats[idx].format; | |
280 | } | |
281 | uprv_free(subformats); | |
282 | subformats = NULL; | |
283 | subformatCount = subformatCapacity = 0; | |
284 | ||
285 | uprv_free(argTypes); | |
286 | argTypes = NULL; | |
287 | argTypeCount = argTypeCapacity = 0; | |
288 | ||
289 | uprv_free(formatAliases); | |
290 | ||
291 | delete defaultNumberFormat; | |
292 | delete defaultDateFormat; | |
293 | } | |
294 | ||
295 | //-------------------------------------------------------------------- | |
296 | // Variable-size array management | |
297 | ||
298 | /** | |
299 | * Allocate subformats[] to at least the given capacity and return | |
300 | * TRUE if successful. If not, leave subformats[] unchanged. | |
301 | * | |
302 | * If subformats is NULL, allocate it. If it is not NULL, enlarge it | |
303 | * if necessary to be at least as large as specified. | |
304 | */ | |
305 | UBool MessageFormat::allocateSubformats(int32_t capacity) { | |
306 | if (subformats == NULL) { | |
307 | subformats = (Subformat*) uprv_malloc(sizeof(*subformats) * capacity); | |
308 | subformatCapacity = capacity; | |
309 | subformatCount = 0; | |
310 | if (subformats == NULL) { | |
311 | subformatCapacity = 0; | |
312 | return FALSE; | |
313 | } | |
314 | } else if (subformatCapacity < capacity) { | |
315 | if (capacity < 2*subformatCapacity) { | |
316 | capacity = 2*subformatCapacity; | |
317 | } | |
318 | Subformat* a = (Subformat*) | |
319 | uprv_realloc(subformats, sizeof(*subformats) * capacity); | |
320 | if (a == NULL) { | |
321 | return FALSE; // request failed | |
322 | } | |
323 | subformats = a; | |
324 | subformatCapacity = capacity; | |
325 | } | |
326 | return TRUE; | |
327 | } | |
328 | ||
329 | /** | |
330 | * Allocate argTypes[] to at least the given capacity and return | |
331 | * TRUE if successful. If not, leave argTypes[] unchanged. | |
332 | * | |
333 | * If argTypes is NULL, allocate it. If it is not NULL, enlarge it | |
334 | * if necessary to be at least as large as specified. | |
335 | */ | |
336 | UBool MessageFormat::allocateArgTypes(int32_t capacity) { | |
337 | if (argTypes == NULL) { | |
338 | argTypes = (Formattable::Type*) uprv_malloc(sizeof(*argTypes) * capacity); | |
339 | argTypeCount = 0; | |
340 | argTypeCapacity = capacity; | |
341 | if (argTypes == NULL) { | |
342 | argTypeCapacity = 0; | |
343 | return FALSE; | |
344 | } | |
345 | for (int32_t i=0; i<capacity; ++i) { | |
346 | argTypes[i] = Formattable::kString; | |
347 | } | |
348 | } else if (argTypeCapacity < capacity) { | |
349 | if (capacity < 2*argTypeCapacity) { | |
350 | capacity = 2*argTypeCapacity; | |
351 | } | |
352 | Formattable::Type* a = (Formattable::Type*) | |
353 | uprv_realloc(argTypes, sizeof(*argTypes) * capacity); | |
354 | if (a == NULL) { | |
355 | return FALSE; // request failed | |
356 | } | |
357 | for (int32_t i=argTypeCapacity; i<capacity; ++i) { | |
358 | a[i] = Formattable::kString; | |
359 | } | |
360 | argTypes = a; | |
361 | argTypeCapacity = capacity; | |
362 | } | |
363 | return TRUE; | |
364 | } | |
365 | ||
366 | // ------------------------------------- | |
367 | // assignment operator | |
368 | ||
369 | const MessageFormat& | |
370 | MessageFormat::operator=(const MessageFormat& that) | |
371 | { | |
372 | // Reallocate the arrays BEFORE changing this object | |
373 | if (this != &that && | |
374 | allocateSubformats(that.subformatCount) && | |
375 | allocateArgTypes(that.argTypeCount)) { | |
376 | ||
377 | // Calls the super class for assignment first. | |
378 | Format::operator=(that); | |
379 | ||
380 | fPattern = that.fPattern; | |
381 | setLocale(that.fLocale); | |
382 | ||
383 | int32_t j; | |
384 | for (j=0; j<subformatCount; ++j) { | |
385 | delete subformats[j].format; | |
386 | } | |
387 | subformatCount = 0; | |
388 | ||
389 | for (j=0; j<that.subformatCount; ++j) { | |
390 | // Subformat::operator= does NOT delete this.format | |
391 | subformats[j] = that.subformats[j]; | |
392 | } | |
393 | subformatCount = that.subformatCount; | |
394 | ||
395 | for (j=0; j<that.argTypeCount; ++j) { | |
396 | argTypes[j] = that.argTypes[j]; | |
397 | } | |
398 | argTypeCount = that.argTypeCount; | |
399 | } | |
400 | return *this; | |
401 | } | |
402 | ||
403 | UBool | |
404 | MessageFormat::operator==(const Format& rhs) const | |
405 | { | |
406 | if (this == &rhs) return TRUE; | |
407 | ||
408 | MessageFormat& that = (MessageFormat&)rhs; | |
409 | ||
410 | // Check class ID before checking MessageFormat members | |
411 | if (!Format::operator==(rhs) || | |
b75a7d8f A |
412 | fPattern != that.fPattern || |
413 | fLocale != that.fLocale) { | |
414 | return FALSE; | |
415 | } | |
416 | ||
417 | int32_t j; | |
418 | for (j=0; j<subformatCount; ++j) { | |
419 | if (subformats[j] != that.subformats[j]) { | |
420 | return FALSE; | |
421 | } | |
422 | } | |
423 | ||
424 | return TRUE; | |
425 | } | |
426 | ||
427 | // ------------------------------------- | |
428 | // Creates a copy of this MessageFormat, the caller owns the copy. | |
429 | ||
430 | Format* | |
431 | MessageFormat::clone() const | |
432 | { | |
433 | return new MessageFormat(*this); | |
434 | } | |
435 | ||
436 | // ------------------------------------- | |
437 | // Sets the locale of this MessageFormat object to theLocale. | |
438 | ||
439 | void | |
440 | MessageFormat::setLocale(const Locale& theLocale) | |
441 | { | |
442 | if (fLocale != theLocale) { | |
443 | delete defaultNumberFormat; | |
444 | defaultNumberFormat = NULL; | |
445 | delete defaultDateFormat; | |
446 | defaultDateFormat = NULL; | |
447 | } | |
448 | fLocale = theLocale; | |
374ca955 | 449 | setLocaleIDs(fLocale.getName(), fLocale.getName()); |
b75a7d8f A |
450 | } |
451 | ||
452 | // ------------------------------------- | |
453 | // Gets the locale of this MessageFormat object. | |
454 | ||
455 | const Locale& | |
456 | MessageFormat::getLocale() const | |
457 | { | |
458 | return fLocale; | |
459 | } | |
460 | ||
461 | ||
462 | ||
463 | ||
464 | void | |
465 | MessageFormat::applyPattern(const UnicodeString& newPattern, | |
466 | UErrorCode& status) | |
467 | { | |
468 | UParseError parseError; | |
469 | applyPattern(newPattern,parseError,status); | |
470 | } | |
471 | ||
472 | ||
473 | // ------------------------------------- | |
474 | // Applies the new pattern and returns an error if the pattern | |
475 | // is not correct. | |
476 | void | |
477 | MessageFormat::applyPattern(const UnicodeString& pattern, | |
478 | UParseError& parseError, | |
479 | UErrorCode& ec) | |
480 | { | |
481 | if(U_FAILURE(ec)) { | |
482 | return; | |
483 | } | |
484 | // The pattern is broken up into segments. Each time a subformat | |
485 | // is encountered, 4 segments are recorded. For example, consider | |
486 | // the pattern: | |
487 | // "There {0,choice,0.0#are no files|1.0#is one file|1.0<are {0, number} files} on disk {1}." | |
488 | // The first set of segments is: | |
489 | // segments[0] = "There " | |
490 | // segments[1] = "0" | |
491 | // segments[2] = "choice" | |
492 | // segments[3] = "0.0#are no files|1.0#is one file|1.0<are {0, number} files" | |
493 | ||
494 | // During parsing, the plain text is accumulated into segments[0]. | |
495 | // Segments 1..3 are used to parse each subpattern. Each time a | |
496 | // subpattern is parsed, it creates a format object that is stored | |
497 | // in the subformats array, together with an offset and argument | |
498 | // number. The offset into the plain text stored in | |
499 | // segments[0]. | |
500 | ||
501 | // Quotes in segment 0 are handled normally. They are removed. | |
502 | // Quotes may not occur in segments 1 or 2. | |
503 | // Quotes in segment 3 are parsed and _copied_. This makes | |
504 | // subformat patterns work, e.g., {1,number,'#'.##} passes | |
505 | // the pattern "'#'.##" to DecimalFormat. | |
506 | ||
507 | UnicodeString segments[4]; | |
508 | int32_t part = 0; // segment we are in, 0..3 | |
509 | // Record the highest argument number in the pattern. (In the | |
510 | // subpattern {3,number} the argument number is 3.) | |
511 | int32_t formatNumber = 0; | |
512 | UBool inQuote = FALSE; | |
513 | int32_t braceStack = 0; | |
514 | // Clear error struct | |
515 | parseError.offset = -1; | |
516 | parseError.preContext[0] = parseError.postContext[0] = (UChar)0; | |
517 | int32_t patLen = pattern.length(); | |
518 | int32_t i; | |
519 | ||
520 | for (i=0; i<subformatCount; ++i) { | |
521 | delete subformats[i].format; | |
522 | } | |
523 | subformatCount = 0; | |
524 | argTypeCount = 0; | |
525 | ||
526 | for (i=0; i<patLen; ++i) { | |
527 | UChar ch = pattern[i]; | |
528 | if (part == 0) { | |
529 | // In segment 0, recognize and remove quotes | |
530 | if (ch == SINGLE_QUOTE) { | |
531 | if (i+1 < patLen && pattern[i+1] == SINGLE_QUOTE) { | |
532 | segments[0] += ch; | |
533 | ++i; | |
534 | } else { | |
535 | inQuote = !inQuote; | |
536 | } | |
537 | } else if (ch == LEFT_CURLY_BRACE && !inQuote) { | |
538 | // The only way we get from segment 0 to 1 is via an | |
539 | // unquoted '{'. | |
540 | part = 1; | |
541 | } else { | |
542 | segments[0] += ch; | |
543 | } | |
544 | } else if (inQuote) { | |
545 | // In segments 1..3, recognize quoted matter, and copy it | |
546 | // into the segment, together with the quotes. This takes | |
547 | // care of '' as well. | |
548 | segments[part] += ch; | |
549 | if (ch == SINGLE_QUOTE) { | |
550 | inQuote = FALSE; | |
551 | } | |
552 | } else { | |
553 | // We have an unquoted character in segment 1..3 | |
554 | switch (ch) { | |
555 | case COMMA: | |
556 | // Commas bump us to the next segment, except for segment 3, | |
557 | // which can contain commas. See example above. | |
558 | if (part < 3) | |
559 | part += 1; | |
560 | else | |
561 | segments[3] += ch; | |
562 | break; | |
563 | case LEFT_CURLY_BRACE: | |
564 | // Handle '{' within segment 3. The initial '{' | |
565 | // before segment 1 is handled above. | |
566 | if (part != 3) { | |
567 | ec = U_PATTERN_SYNTAX_ERROR; | |
568 | goto SYNTAX_ERROR; | |
569 | } | |
570 | ++braceStack; | |
571 | segments[part] += ch; | |
572 | break; | |
573 | case RIGHT_CURLY_BRACE: | |
574 | if (braceStack == 0) { | |
575 | makeFormat(formatNumber, segments, parseError,ec); | |
576 | if (U_FAILURE(ec)){ | |
577 | goto SYNTAX_ERROR; | |
578 | } | |
579 | formatNumber++; | |
580 | segments[1].remove(); | |
581 | segments[2].remove(); | |
582 | segments[3].remove(); | |
583 | part = 0; | |
584 | } else { | |
585 | --braceStack; | |
586 | segments[part] += ch; | |
587 | } | |
588 | break; | |
589 | case SINGLE_QUOTE: | |
590 | inQuote = TRUE; | |
591 | // fall through (copy quote chars in segments 1..3) | |
592 | default: | |
593 | segments[part] += ch; | |
594 | break; | |
595 | } | |
596 | } | |
597 | } | |
598 | if (braceStack != 0 || part != 0) { | |
599 | // Unmatched braces in the pattern | |
600 | ec = U_UNMATCHED_BRACES; | |
601 | goto SYNTAX_ERROR; | |
602 | } | |
603 | fPattern = segments[0]; | |
604 | return; | |
605 | ||
606 | SYNTAX_ERROR: | |
607 | syntaxError(pattern, i, parseError); | |
608 | for (i=0; i<subformatCount; ++i) { | |
609 | delete subformats[i].format; | |
610 | } | |
611 | argTypeCount = subformatCount = 0; | |
612 | } | |
613 | // ------------------------------------- | |
614 | // Converts this MessageFormat instance to a pattern. | |
615 | ||
616 | UnicodeString& | |
617 | MessageFormat::toPattern(UnicodeString& appendTo) const { | |
618 | // later, make this more extensible | |
619 | int32_t lastOffset = 0; | |
620 | int32_t i; | |
621 | for (i=0; i<subformatCount; ++i) { | |
622 | copyAndFixQuotes(fPattern, lastOffset, subformats[i].offset, appendTo); | |
623 | lastOffset = subformats[i].offset; | |
624 | appendTo += LEFT_CURLY_BRACE; | |
625 | itos(subformats[i].arg, appendTo); | |
626 | Format* fmt = subformats[i].format; | |
627 | if (fmt == NULL) { | |
628 | // do nothing, string format | |
629 | } | |
630 | else if (fmt->getDynamicClassID() == DecimalFormat::getStaticClassID()) { | |
631 | ||
632 | UErrorCode ec = U_ZERO_ERROR; | |
633 | NumberFormat& formatAlias = *(NumberFormat*)fmt; | |
634 | NumberFormat *defaultTemplate = NumberFormat::createInstance(fLocale, ec); | |
635 | NumberFormat *currencyTemplate = NumberFormat::createCurrencyInstance(fLocale, ec); | |
636 | NumberFormat *percentTemplate = NumberFormat::createPercentInstance(fLocale, ec); | |
637 | NumberFormat *integerTemplate = createIntegerFormat(fLocale, ec); | |
638 | ||
639 | appendTo += COMMA; | |
640 | appendTo += ID_NUMBER; | |
641 | if (formatAlias != *defaultTemplate) { | |
642 | appendTo += COMMA; | |
643 | if (formatAlias == *currencyTemplate) { | |
644 | appendTo += ID_CURRENCY; | |
645 | } | |
646 | else if (formatAlias == *percentTemplate) { | |
647 | appendTo += ID_PERCENT; | |
648 | } | |
649 | else if (formatAlias == *integerTemplate) { | |
650 | appendTo += ID_INTEGER; | |
651 | } | |
652 | else { | |
653 | UnicodeString buffer; | |
654 | appendTo += ((DecimalFormat*)fmt)->toPattern(buffer); | |
655 | } | |
656 | } | |
657 | ||
658 | delete defaultTemplate; | |
659 | delete currencyTemplate; | |
660 | delete percentTemplate; | |
661 | delete integerTemplate; | |
662 | } | |
663 | else if (fmt->getDynamicClassID() == SimpleDateFormat::getStaticClassID()) { | |
664 | DateFormat& formatAlias = *(DateFormat*)fmt; | |
665 | DateFormat *defaultDateTemplate = DateFormat::createDateInstance(DateFormat::kDefault, fLocale); | |
666 | DateFormat *shortDateTemplate = DateFormat::createDateInstance(DateFormat::kShort, fLocale); | |
667 | DateFormat *longDateTemplate = DateFormat::createDateInstance(DateFormat::kLong, fLocale); | |
668 | DateFormat *fullDateTemplate = DateFormat::createDateInstance(DateFormat::kFull, fLocale); | |
669 | DateFormat *defaultTimeTemplate = DateFormat::createTimeInstance(DateFormat::kDefault, fLocale); | |
670 | DateFormat *shortTimeTemplate = DateFormat::createTimeInstance(DateFormat::kShort, fLocale); | |
671 | DateFormat *longTimeTemplate = DateFormat::createTimeInstance(DateFormat::kLong, fLocale); | |
672 | DateFormat *fullTimeTemplate = DateFormat::createTimeInstance(DateFormat::kFull, fLocale); | |
673 | ||
674 | ||
675 | appendTo += COMMA; | |
676 | if (formatAlias == *defaultDateTemplate) { | |
677 | appendTo += ID_DATE; | |
678 | } | |
679 | else if (formatAlias == *shortDateTemplate) { | |
680 | appendTo += ID_DATE; | |
681 | appendTo += COMMA; | |
682 | appendTo += ID_SHORT; | |
683 | } | |
684 | else if (formatAlias == *defaultDateTemplate) { | |
685 | appendTo += ID_DATE; | |
686 | appendTo += COMMA; | |
687 | appendTo += ID_MEDIUM; | |
688 | } | |
689 | else if (formatAlias == *longDateTemplate) { | |
690 | appendTo += ID_DATE; | |
691 | appendTo += COMMA; | |
692 | appendTo += ID_LONG; | |
693 | } | |
694 | else if (formatAlias == *fullDateTemplate) { | |
695 | appendTo += ID_DATE; | |
696 | appendTo += COMMA; | |
697 | appendTo += ID_FULL; | |
698 | } | |
699 | else if (formatAlias == *defaultTimeTemplate) { | |
700 | appendTo += ID_TIME; | |
701 | } | |
702 | else if (formatAlias == *shortTimeTemplate) { | |
703 | appendTo += ID_TIME; | |
704 | appendTo += COMMA; | |
705 | appendTo += ID_SHORT; | |
706 | } | |
707 | else if (formatAlias == *defaultTimeTemplate) { | |
708 | appendTo += ID_TIME; | |
709 | appendTo += COMMA; | |
710 | appendTo += ID_MEDIUM; | |
711 | } | |
712 | else if (formatAlias == *longTimeTemplate) { | |
713 | appendTo += ID_TIME; | |
714 | appendTo += COMMA; | |
715 | appendTo += ID_LONG; | |
716 | } | |
717 | else if (formatAlias == *fullTimeTemplate) { | |
718 | appendTo += ID_TIME; | |
719 | appendTo += COMMA; | |
720 | appendTo += ID_FULL; | |
721 | } | |
722 | else { | |
723 | UnicodeString buffer; | |
724 | appendTo += ID_DATE; | |
725 | appendTo += COMMA; | |
726 | appendTo += ((SimpleDateFormat*)fmt)->toPattern(buffer); | |
727 | } | |
728 | ||
729 | delete defaultDateTemplate; | |
730 | delete shortDateTemplate; | |
731 | delete longDateTemplate; | |
732 | delete fullDateTemplate; | |
733 | delete defaultTimeTemplate; | |
734 | delete shortTimeTemplate; | |
735 | delete longTimeTemplate; | |
736 | delete fullTimeTemplate; | |
737 | // {sfb} there should be a more efficient way to do this! | |
738 | } | |
739 | else if (fmt->getDynamicClassID() == ChoiceFormat::getStaticClassID()) { | |
740 | UnicodeString buffer; | |
741 | appendTo += COMMA; | |
742 | appendTo += ID_CHOICE; | |
743 | appendTo += COMMA; | |
744 | appendTo += ((ChoiceFormat*)fmt)->toPattern(buffer); | |
745 | } | |
746 | else { | |
747 | //appendTo += ", unknown"; | |
748 | } | |
749 | appendTo += RIGHT_CURLY_BRACE; | |
750 | } | |
751 | copyAndFixQuotes(fPattern, lastOffset, fPattern.length(), appendTo); | |
752 | return appendTo; | |
753 | } | |
754 | ||
755 | // ------------------------------------- | |
756 | // Adopts the new formats array and updates the array count. | |
757 | // This MessageFormat instance owns the new formats. | |
758 | ||
759 | void | |
760 | MessageFormat::adoptFormats(Format** newFormats, | |
761 | int32_t count) { | |
762 | if (newFormats == NULL || count < 0) { | |
763 | return; | |
764 | } | |
765 | ||
766 | int32_t i; | |
767 | if (allocateSubformats(count)) { | |
768 | for (i=0; i<subformatCount; ++i) { | |
769 | delete subformats[i].format; | |
770 | } | |
771 | for (i=0; i<count; ++i) { | |
772 | subformats[i].format = newFormats[i]; | |
773 | } | |
774 | subformatCount = count; | |
775 | } else { | |
776 | // An adopt method must always take ownership. Delete | |
777 | // the incoming format objects and return unchanged. | |
778 | for (i=0; i<count; ++i) { | |
779 | delete newFormats[i]; | |
780 | } | |
781 | } | |
782 | ||
783 | // TODO: What about the .offset and .arg fields? | |
784 | } | |
785 | ||
786 | // ------------------------------------- | |
787 | // Sets the new formats array and updates the array count. | |
788 | // This MessageFormat instance maks a copy of the new formats. | |
789 | ||
790 | void | |
791 | MessageFormat::setFormats(const Format** newFormats, | |
792 | int32_t count) { | |
793 | if (newFormats == NULL || count < 0) { | |
794 | return; | |
795 | } | |
796 | ||
797 | if (allocateSubformats(count)) { | |
798 | int32_t i; | |
799 | for (i=0; i<subformatCount; ++i) { | |
800 | delete subformats[i].format; | |
801 | } | |
802 | subformatCount = 0; | |
803 | ||
804 | for (i=0; i<count; ++i) { | |
805 | subformats[i].format = newFormats[i] ? newFormats[i]->clone() : NULL; | |
806 | } | |
807 | subformatCount = count; | |
808 | } | |
809 | ||
810 | // TODO: What about the .offset and .arg fields? | |
811 | } | |
812 | ||
813 | // ------------------------------------- | |
814 | // Adopt a single format. | |
815 | // Do nothing is the format number is not less than the array count. | |
816 | ||
817 | void | |
818 | MessageFormat::adoptFormat(int32_t n, Format *newFormat) { | |
819 | if (n < 0 || n >= subformatCount) { | |
820 | delete newFormat; | |
821 | } else { | |
822 | delete subformats[n].format; | |
823 | subformats[n].format = newFormat; | |
824 | } | |
825 | } | |
826 | ||
827 | // ------------------------------------- | |
828 | // Set a single format. | |
829 | // Do nothing is the variable is not less than the array count. | |
830 | ||
831 | void | |
832 | MessageFormat::setFormat(int32_t n, const Format& newFormat) { | |
833 | if (n >= 0 && n < subformatCount) { | |
834 | delete subformats[n].format; | |
835 | if (&newFormat == NULL) { | |
836 | // This should never happen -- but we'll be nice if it does | |
837 | subformats[n].format = NULL; | |
838 | } else { | |
839 | subformats[n].format = newFormat.clone(); | |
840 | } | |
841 | } | |
842 | } | |
843 | ||
844 | // ------------------------------------- | |
845 | // Gets the format array. | |
846 | ||
847 | const Format** | |
848 | MessageFormat::getFormats(int32_t& cnt) const | |
849 | { | |
850 | // This old API returns an array (which we hold) of Format* | |
851 | // pointers. The array is valid up to the next call to any | |
852 | // method on this object. We construct and resize an array | |
853 | // on demand that contains aliases to the subformats[i].format | |
854 | // pointers. | |
855 | MessageFormat* t = (MessageFormat*) this; | |
856 | cnt = 0; | |
857 | if (formatAliases == NULL) { | |
858 | t->formatAliasesCapacity = (subformatCount<10) ? 10 : subformatCount; | |
859 | Format** a = (Format**) | |
860 | uprv_malloc(sizeof(Format*) * formatAliasesCapacity); | |
861 | if (a == NULL) { | |
862 | return NULL; | |
863 | } | |
864 | t->formatAliases = a; | |
865 | } else if (subformatCount > formatAliasesCapacity) { | |
866 | Format** a = (Format**) | |
867 | uprv_realloc(formatAliases, sizeof(Format*) * subformatCount); | |
868 | if (a == NULL) { | |
869 | return NULL; | |
870 | } | |
871 | t->formatAliases = a; | |
872 | t->formatAliasesCapacity = subformatCount; | |
873 | } | |
874 | for (int32_t i=0; i<subformatCount; ++i) { | |
875 | t->formatAliases[i] = subformats[i].format; | |
876 | } | |
877 | cnt = subformatCount; | |
878 | return (const Format**)formatAliases; | |
879 | } | |
880 | ||
881 | // ------------------------------------- | |
882 | // Formats the source Formattable array and copy into the result buffer. | |
883 | // Ignore the FieldPosition result for error checking. | |
884 | ||
885 | UnicodeString& | |
886 | MessageFormat::format(const Formattable* source, | |
887 | int32_t cnt, | |
888 | UnicodeString& appendTo, | |
889 | FieldPosition& ignore, | |
890 | UErrorCode& success) const | |
891 | { | |
892 | if (U_FAILURE(success)) | |
893 | return appendTo; | |
894 | ||
895 | return format(source, cnt, appendTo, ignore, 0, success); | |
896 | } | |
897 | ||
898 | // ------------------------------------- | |
899 | // Internally creates a MessageFormat instance based on the | |
900 | // pattern and formats the arguments Formattable array and | |
901 | // copy into the appendTo buffer. | |
902 | ||
903 | UnicodeString& | |
904 | MessageFormat::format( const UnicodeString& pattern, | |
905 | const Formattable* arguments, | |
906 | int32_t cnt, | |
907 | UnicodeString& appendTo, | |
908 | UErrorCode& success) | |
909 | { | |
910 | MessageFormat temp(pattern, success); | |
911 | FieldPosition ignore(0); | |
912 | temp.format(arguments, cnt, appendTo, ignore, success); | |
913 | return appendTo; | |
914 | } | |
915 | ||
916 | // ------------------------------------- | |
917 | // Formats the source Formattable object and copy into the | |
918 | // appendTo buffer. The Formattable object must be an array | |
919 | // of Formattable instances, returns error otherwise. | |
920 | ||
921 | UnicodeString& | |
922 | MessageFormat::format(const Formattable& source, | |
923 | UnicodeString& appendTo, | |
924 | FieldPosition& ignore, | |
925 | UErrorCode& success) const | |
926 | { | |
927 | int32_t cnt; | |
928 | ||
929 | if (U_FAILURE(success)) | |
930 | return appendTo; | |
931 | if (source.getType() != Formattable::kArray) { | |
932 | success = U_ILLEGAL_ARGUMENT_ERROR; | |
933 | return appendTo; | |
934 | } | |
935 | const Formattable* tmpPtr = source.getArray(cnt); | |
936 | ||
937 | return format(tmpPtr, cnt, appendTo, ignore, 0, success); | |
938 | } | |
939 | ||
940 | // ------------------------------------- | |
941 | // Formats the arguments Formattable array and copy into the appendTo buffer. | |
942 | // Ignore the FieldPosition result for error checking. | |
943 | ||
944 | UnicodeString& | |
945 | MessageFormat::format(const Formattable* arguments, | |
946 | int32_t cnt, | |
947 | UnicodeString& appendTo, | |
948 | FieldPosition& status, | |
949 | int32_t recursionProtection, | |
950 | UErrorCode& success) const | |
951 | { | |
952 | // Allow NULL array only if cnt == 0 | |
953 | if (cnt < 0 || (cnt && arguments == NULL)) { | |
954 | success = U_ILLEGAL_ARGUMENT_ERROR; | |
955 | return appendTo; | |
956 | } | |
957 | ||
958 | int32_t lastOffset = 0; | |
959 | for (int32_t i=0; i<subformatCount; ++i) { | |
960 | // Append the prefix of current format element. | |
961 | appendTo.append(fPattern, lastOffset, subformats[i].offset - lastOffset); | |
962 | lastOffset = subformats[i].offset; | |
963 | int32_t argumentNumber = subformats[i].arg; | |
964 | // Checks the scope of the argument number. | |
965 | if (argumentNumber >= cnt) { | |
966 | appendTo += LEFT_CURLY_BRACE; | |
967 | itos(argumentNumber, appendTo); | |
968 | appendTo += RIGHT_CURLY_BRACE; | |
969 | continue; | |
970 | } | |
971 | ||
972 | const Formattable *obj = arguments + argumentNumber; | |
973 | Formattable::Type type = obj->getType(); | |
974 | ||
975 | // Recursively calling the format process only if the current | |
976 | // format argument refers to a ChoiceFormat object. | |
977 | Format* fmt = subformats[i].format; | |
978 | if (fmt != NULL) { | |
979 | UnicodeString arg; | |
980 | fmt->format(*obj, arg, success); | |
981 | ||
982 | // Needs to reprocess the ChoiceFormat option by using the | |
983 | // MessageFormat pattern application. | |
984 | if (fmt->getDynamicClassID() == ChoiceFormat::getStaticClassID() && | |
985 | arg.indexOf(LEFT_CURLY_BRACE) >= 0) { | |
986 | MessageFormat temp(arg, fLocale, success); | |
987 | // TODO: Implement recursion protection | |
988 | temp.format(arguments, cnt, appendTo, status, recursionProtection, success); | |
989 | if (U_FAILURE(success)) { | |
990 | return appendTo; | |
991 | } | |
992 | } | |
993 | else { | |
994 | appendTo += arg; | |
995 | } | |
996 | } | |
997 | // If the obj data type is a number, use a NumberFormat instance. | |
374ca955 A |
998 | else if ((type == Formattable::kDouble) || |
999 | (type == Formattable::kLong) || | |
1000 | (type == Formattable::kInt64)) { | |
1001 | ||
b75a7d8f A |
1002 | const NumberFormat* nf = getDefaultNumberFormat(success); |
1003 | if (nf == NULL) { | |
1004 | return appendTo; | |
1005 | } | |
1006 | if (type == Formattable::kDouble) { | |
1007 | nf->format(obj->getDouble(), appendTo); | |
374ca955 | 1008 | } else if (type == Formattable::kLong) { |
b75a7d8f | 1009 | nf->format(obj->getLong(), appendTo); |
374ca955 A |
1010 | } else { |
1011 | nf->format(obj->getInt64(), appendTo); | |
b75a7d8f A |
1012 | } |
1013 | } | |
1014 | // If the obj data type is a Date instance, use a DateFormat instance. | |
1015 | else if (type == Formattable::kDate) { | |
1016 | const DateFormat* df = getDefaultDateFormat(success); | |
1017 | if (df == NULL) { | |
1018 | return appendTo; | |
1019 | } | |
1020 | df->format(obj->getDate(), appendTo); | |
1021 | } | |
1022 | else if (type == Formattable::kString) { | |
1023 | appendTo += obj->getString(); | |
1024 | } | |
1025 | else { | |
1026 | success = U_ILLEGAL_ARGUMENT_ERROR; | |
1027 | return appendTo; | |
1028 | } | |
1029 | } | |
1030 | // Appends the rest of the pattern characters after the real last offset. | |
1031 | appendTo.append(fPattern, lastOffset, 0x7fffffff); | |
1032 | return appendTo; | |
1033 | } | |
1034 | ||
1035 | ||
1036 | // ------------------------------------- | |
1037 | // Parses the source pattern and returns the Formattable objects array, | |
1038 | // the array count and the ending parse position. The caller of this method | |
1039 | // owns the array. | |
1040 | ||
1041 | Formattable* | |
1042 | MessageFormat::parse(const UnicodeString& source, | |
1043 | ParsePosition& pos, | |
1044 | int32_t& count) const | |
1045 | { | |
1046 | // Allocate at least one element. Allocating an array of length | |
1047 | // zero causes problems on some platforms (e.g. Win32). | |
1048 | Formattable *resultArray = new Formattable[argTypeCount ? argTypeCount : 1]; | |
1049 | int32_t patternOffset = 0; | |
1050 | int32_t sourceOffset = pos.getIndex(); | |
1051 | ParsePosition tempPos(0); | |
1052 | count = 0; // {sfb} reset to zero | |
1053 | int32_t len; | |
1054 | for (int32_t i = 0; i < subformatCount; ++i) { | |
1055 | // match up to format | |
1056 | len = subformats[i].offset - patternOffset; | |
1057 | if (len == 0 || | |
1058 | fPattern.compare(patternOffset, len, source, sourceOffset, len) == 0) { | |
1059 | sourceOffset += len; | |
1060 | patternOffset += len; | |
1061 | } | |
1062 | else { | |
1063 | goto PARSE_ERROR; | |
1064 | } | |
1065 | ||
1066 | // now use format | |
1067 | Format* fmt = subformats[i].format; | |
1068 | int32_t arg = subformats[i].arg; | |
1069 | if (fmt == NULL) { // string format | |
1070 | // if at end, use longest possible match | |
1071 | // otherwise uses first match to intervening string | |
1072 | // does NOT recursively try all possibilities | |
1073 | int32_t tempLength = (i+1<subformatCount) ? | |
1074 | subformats[i+1].offset : fPattern.length(); | |
1075 | ||
1076 | int32_t next; | |
1077 | if (patternOffset >= tempLength) { | |
1078 | next = source.length(); | |
1079 | } | |
1080 | else { | |
1081 | UnicodeString buffer; | |
1082 | fPattern.extract(patternOffset,tempLength - patternOffset, buffer); | |
1083 | next = source.indexOf(buffer, sourceOffset); | |
1084 | } | |
1085 | ||
1086 | if (next < 0) { | |
1087 | goto PARSE_ERROR; | |
1088 | } | |
1089 | else { | |
1090 | UnicodeString buffer; | |
1091 | source.extract(sourceOffset,next - sourceOffset, buffer); | |
1092 | UnicodeString strValue = buffer; | |
1093 | UnicodeString temp(LEFT_CURLY_BRACE); | |
1094 | // {sfb} check this later | |
1095 | itos(arg, temp); | |
1096 | temp += RIGHT_CURLY_BRACE; | |
1097 | if (strValue != temp) { | |
1098 | source.extract(sourceOffset,next - sourceOffset, buffer); | |
1099 | resultArray[arg].setString(buffer); | |
1100 | // {sfb} not sure about this | |
1101 | if ((arg + 1) > count) { | |
1102 | count = arg + 1; | |
1103 | } | |
1104 | } | |
1105 | sourceOffset = next; | |
1106 | } | |
1107 | } | |
1108 | else { | |
1109 | tempPos.setIndex(sourceOffset); | |
1110 | fmt->parseObject(source, resultArray[arg], tempPos); | |
1111 | if (tempPos.getIndex() == sourceOffset) { | |
1112 | goto PARSE_ERROR; | |
1113 | } | |
1114 | ||
1115 | if ((arg + 1) > count) { | |
1116 | count = arg + 1; | |
1117 | } | |
1118 | sourceOffset = tempPos.getIndex(); // update | |
1119 | } | |
1120 | } | |
1121 | len = fPattern.length() - patternOffset; | |
1122 | if (len == 0 || | |
1123 | fPattern.compare(patternOffset, len, source, sourceOffset, len) == 0) { | |
1124 | pos.setIndex(sourceOffset + len); | |
1125 | return resultArray; | |
1126 | } | |
1127 | // else fall through... | |
1128 | ||
1129 | PARSE_ERROR: | |
1130 | pos.setErrorIndex(sourceOffset); | |
1131 | delete [] resultArray; | |
1132 | count = 0; | |
1133 | return NULL; // leave index as is to signal error | |
1134 | } | |
1135 | ||
1136 | // ------------------------------------- | |
1137 | // Parses the source string and returns the array of | |
1138 | // Formattable objects and the array count. The caller | |
1139 | // owns the returned array. | |
1140 | ||
1141 | Formattable* | |
1142 | MessageFormat::parse(const UnicodeString& source, | |
1143 | int32_t& cnt, | |
1144 | UErrorCode& success) const | |
1145 | { | |
1146 | ParsePosition status(0); | |
1147 | // Calls the actual implementation method and starts | |
1148 | // from zero offset of the source text. | |
1149 | Formattable* result = parse(source, status, cnt); | |
1150 | if (status.getIndex() == 0) { | |
1151 | success = U_MESSAGE_PARSE_ERROR; | |
1152 | delete[] result; | |
1153 | return NULL; | |
1154 | } | |
1155 | return result; | |
1156 | } | |
1157 | ||
1158 | // ------------------------------------- | |
1159 | // Parses the source text and copy into the result buffer. | |
1160 | ||
1161 | void | |
1162 | MessageFormat::parseObject( const UnicodeString& source, | |
1163 | Formattable& result, | |
1164 | ParsePosition& status) const | |
1165 | { | |
1166 | int32_t cnt = 0; | |
1167 | Formattable* tmpResult = parse(source, status, cnt); | |
1168 | if (tmpResult != NULL) | |
1169 | result.adoptArray(tmpResult, cnt); | |
1170 | } | |
1171 | ||
1172 | // ------------------------------------- | |
374ca955 A |
1173 | |
1174 | static Format* makeRBNF(URBNFRuleSetTag tag, const Locale& locale, const UnicodeString& defaultRuleSet, UErrorCode& ec) { | |
1175 | RuleBasedNumberFormat* fmt = new RuleBasedNumberFormat(tag, locale, ec); | |
1176 | if (U_SUCCESS(ec) && defaultRuleSet.length() > 0) { | |
1177 | fmt->setDefaultRuleSet(defaultRuleSet, ec); | |
1178 | if (U_FAILURE(ec)) { // ignore unrecognized default rule set | |
1179 | ec = U_ZERO_ERROR; | |
1180 | } | |
1181 | } | |
1182 | return fmt; | |
1183 | } | |
b75a7d8f A |
1184 | |
1185 | /** | |
1186 | * Reads the segments[] array (see applyPattern()) and parses the | |
1187 | * segments[1..3] into a Format* object. Stores the format object in | |
1188 | * the subformats[] array. Updates the argTypes[] array type | |
1189 | * information for the corresponding argument. | |
1190 | * | |
1191 | * @param formatNumber index into subformats[] for this format | |
1192 | * @param segments array of strings with the parsed pattern segments | |
1193 | * @param parseError parse error data (output param) | |
1194 | * @param ec error code | |
1195 | */ | |
1196 | void | |
1197 | MessageFormat::makeFormat(int32_t formatNumber, | |
1198 | UnicodeString* segments, | |
1199 | UParseError& parseError, | |
1200 | UErrorCode& ec) { | |
1201 | if (U_FAILURE(ec)) { | |
1202 | return; | |
1203 | } | |
1204 | ||
1205 | // Parse the argument number | |
1206 | int32_t argumentNumber = stou(segments[1]); // always unlocalized! | |
1207 | if (argumentNumber < 0) { | |
1208 | ec = U_INVALID_FORMAT_ERROR; | |
1209 | return; | |
1210 | } | |
1211 | ||
1212 | // Parse the format, recording the argument type and creating a | |
1213 | // new Format object (except for string arguments). | |
1214 | Formattable::Type argType; | |
1215 | Format *fmt = NULL; | |
1216 | int32_t typeID, styleID; | |
1217 | DateFormat::EStyle style; | |
1218 | ||
1219 | switch (typeID = findKeyword(segments[2], TYPE_IDS)) { | |
1220 | ||
1221 | case 0: // string | |
1222 | argType = Formattable::kString; | |
1223 | break; | |
1224 | ||
1225 | case 1: // number | |
1226 | argType = Formattable::kDouble; | |
1227 | ||
1228 | switch (findKeyword(segments[3], NUMBER_STYLE_IDS)) { | |
1229 | case 0: // default | |
1230 | fmt = NumberFormat::createInstance(fLocale, ec); | |
1231 | break; | |
1232 | case 1: // currency | |
1233 | fmt = NumberFormat::createCurrencyInstance(fLocale, ec); | |
1234 | break; | |
1235 | case 2: // percent | |
1236 | fmt = NumberFormat::createPercentInstance(fLocale, ec); | |
1237 | break; | |
1238 | case 3: // integer | |
1239 | argType = Formattable::kLong; | |
1240 | fmt = createIntegerFormat(fLocale, ec); | |
1241 | break; | |
1242 | default: // pattern | |
1243 | fmt = NumberFormat::createInstance(fLocale, ec); | |
1244 | if (fmt && | |
1245 | fmt->getDynamicClassID() == DecimalFormat::getStaticClassID()) { | |
1246 | ((DecimalFormat*)fmt)->applyPattern(segments[3],parseError,ec); | |
1247 | } | |
1248 | break; | |
1249 | } | |
1250 | break; | |
1251 | ||
1252 | case 2: // date | |
1253 | case 3: // time | |
1254 | argType = Formattable::kDate; | |
1255 | styleID = findKeyword(segments[3], DATE_STYLE_IDS); | |
1256 | style = (styleID >= 0) ? DATE_STYLES[styleID] : DateFormat::kDefault; | |
1257 | ||
1258 | if (typeID == 2) { | |
1259 | fmt = DateFormat::createDateInstance(style, fLocale); | |
1260 | } else { | |
1261 | fmt = DateFormat::createTimeInstance(style, fLocale); | |
1262 | } | |
1263 | ||
1264 | if (styleID < 0 && | |
1265 | fmt != NULL && | |
1266 | fmt->getDynamicClassID() == SimpleDateFormat::getStaticClassID()) { | |
1267 | ((SimpleDateFormat*)fmt)->applyPattern(segments[3]); | |
1268 | } | |
1269 | break; | |
1270 | ||
1271 | case 4: // choice | |
1272 | argType = Formattable::kDouble; | |
1273 | ||
1274 | fmt = new ChoiceFormat(segments[3], parseError, ec); | |
1275 | break; | |
1276 | ||
374ca955 A |
1277 | case 5: // spellout |
1278 | argType = Formattable::kDouble; | |
1279 | fmt = makeRBNF(URBNF_SPELLOUT, fLocale, segments[3], ec); | |
1280 | break; | |
1281 | case 6: // ordinal | |
1282 | argType = Formattable::kDouble; | |
1283 | fmt = makeRBNF(URBNF_ORDINAL, fLocale, segments[3], ec); | |
1284 | break; | |
1285 | case 7: // duration | |
1286 | argType = Formattable::kDouble; | |
1287 | fmt = makeRBNF(URBNF_DURATION, fLocale, segments[3], ec); | |
1288 | break; | |
b75a7d8f A |
1289 | default: |
1290 | argType = Formattable::kString; | |
1291 | ec = U_ILLEGAL_ARGUMENT_ERROR; | |
1292 | break; | |
1293 | } | |
1294 | ||
1295 | if (fmt==NULL && argType!=Formattable::kString && U_SUCCESS(ec)) { | |
1296 | ec = U_MEMORY_ALLOCATION_ERROR; | |
1297 | } | |
1298 | ||
1299 | if (!allocateSubformats(formatNumber+1) || | |
1300 | !allocateArgTypes(argumentNumber+1)) { | |
1301 | ec = U_MEMORY_ALLOCATION_ERROR; | |
1302 | } | |
1303 | ||
1304 | if (U_FAILURE(ec)) { | |
1305 | delete fmt; | |
1306 | return; | |
1307 | } | |
1308 | ||
1309 | // Parse succeeded; record results in our arrays | |
1310 | subformats[formatNumber].format = fmt; | |
1311 | subformats[formatNumber].offset = segments[0].length(); | |
1312 | subformats[formatNumber].arg = argumentNumber; | |
1313 | subformatCount = formatNumber+1; | |
1314 | ||
1315 | // Careful here: argumentNumber may in general arrive out of | |
1316 | // sequence, e.g., "There was {2} on {0,date} (see {1,number})." | |
1317 | argTypes[argumentNumber] = argType; | |
1318 | if (argumentNumber+1 > argTypeCount) { | |
1319 | argTypeCount = argumentNumber+1; | |
1320 | } | |
1321 | } | |
1322 | ||
1323 | // ------------------------------------- | |
1324 | // Finds the string, s, in the string array, list. | |
1325 | int32_t MessageFormat::findKeyword(const UnicodeString& s, | |
1326 | const UChar * const *list) | |
1327 | { | |
1328 | if (s.length() == 0) | |
374ca955 | 1329 | return 0; // default |
b75a7d8f A |
1330 | |
1331 | UnicodeString buffer = s; | |
1332 | // Trims the space characters and turns all characters | |
1333 | // in s to lower case. | |
1334 | buffer.trim().toLower(); | |
374ca955 A |
1335 | for (int32_t i = 0; list[i]; ++i) { |
1336 | if (!buffer.compare(list[i], u_strlen(list[i]))) { | |
b75a7d8f | 1337 | return i; |
374ca955 | 1338 | } |
b75a7d8f A |
1339 | } |
1340 | return -1; | |
1341 | } | |
1342 | ||
1343 | // ------------------------------------- | |
1344 | // Checks the range of the source text to quote the special | |
1345 | // characters, { and ' and copy to target buffer. | |
1346 | ||
1347 | void | |
1348 | MessageFormat::copyAndFixQuotes(const UnicodeString& source, | |
1349 | int32_t start, | |
1350 | int32_t end, | |
1351 | UnicodeString& appendTo) | |
1352 | { | |
1353 | UBool gotLB = FALSE; | |
1354 | ||
1355 | for (int32_t i = start; i < end; ++i) { | |
1356 | UChar ch = source[i]; | |
1357 | if (ch == LEFT_CURLY_BRACE) { | |
1358 | appendTo += SINGLE_QUOTE; | |
1359 | appendTo += LEFT_CURLY_BRACE; | |
1360 | appendTo += SINGLE_QUOTE; | |
1361 | gotLB = TRUE; | |
1362 | } | |
1363 | else if (ch == RIGHT_CURLY_BRACE) { | |
1364 | if(gotLB) { | |
1365 | appendTo += RIGHT_CURLY_BRACE; | |
1366 | gotLB = FALSE; | |
1367 | } | |
1368 | else { | |
1369 | // orig code. | |
1370 | appendTo += SINGLE_QUOTE; | |
1371 | appendTo += RIGHT_CURLY_BRACE; | |
1372 | appendTo += SINGLE_QUOTE; | |
1373 | } | |
1374 | } | |
1375 | else if (ch == SINGLE_QUOTE) { | |
1376 | appendTo += SINGLE_QUOTE; | |
1377 | appendTo += SINGLE_QUOTE; | |
1378 | } | |
1379 | else { | |
1380 | appendTo += ch; | |
1381 | } | |
1382 | } | |
1383 | } | |
1384 | ||
1385 | /** | |
1386 | * Convenience method that ought to be in NumberFormat | |
1387 | */ | |
1388 | NumberFormat* | |
1389 | MessageFormat::createIntegerFormat(const Locale& locale, UErrorCode& status) const { | |
1390 | NumberFormat *temp = NumberFormat::createInstance(locale, status); | |
1391 | if (temp->getDynamicClassID() == DecimalFormat::getStaticClassID()) { | |
1392 | DecimalFormat *temp2 = (DecimalFormat*) temp; | |
1393 | temp2->setMaximumFractionDigits(0); | |
1394 | temp2->setDecimalSeparatorAlwaysShown(FALSE); | |
1395 | temp2->setParseIntegerOnly(TRUE); | |
1396 | } | |
1397 | ||
1398 | return temp; | |
1399 | } | |
1400 | ||
1401 | /** | |
1402 | * Return the default number format. Used to format a numeric | |
1403 | * argument when subformats[i].format is NULL. Returns NULL | |
1404 | * on failure. | |
1405 | * | |
1406 | * Semantically const but may modify *this. | |
1407 | */ | |
1408 | const NumberFormat* MessageFormat::getDefaultNumberFormat(UErrorCode& ec) const { | |
1409 | if (defaultNumberFormat == NULL) { | |
1410 | MessageFormat* t = (MessageFormat*) this; | |
1411 | t->defaultNumberFormat = NumberFormat::createInstance(fLocale, ec); | |
1412 | if (U_FAILURE(ec)) { | |
1413 | delete t->defaultNumberFormat; | |
1414 | t->defaultNumberFormat = NULL; | |
1415 | } else if (t->defaultNumberFormat == NULL) { | |
1416 | ec = U_MEMORY_ALLOCATION_ERROR; | |
1417 | } | |
1418 | } | |
1419 | return defaultNumberFormat; | |
1420 | } | |
1421 | ||
1422 | /** | |
1423 | * Return the default date format. Used to format a date | |
1424 | * argument when subformats[i].format is NULL. Returns NULL | |
1425 | * on failure. | |
1426 | * | |
1427 | * Semantically const but may modify *this. | |
1428 | */ | |
1429 | const DateFormat* MessageFormat::getDefaultDateFormat(UErrorCode& ec) const { | |
1430 | if (defaultDateFormat == NULL) { | |
1431 | MessageFormat* t = (MessageFormat*) this; | |
1432 | t->defaultDateFormat = DateFormat::createDateTimeInstance(DateFormat::kShort, DateFormat::kShort, fLocale); | |
1433 | if (t->defaultDateFormat == NULL) { | |
1434 | ec = U_MEMORY_ALLOCATION_ERROR; | |
1435 | } | |
1436 | } | |
1437 | return defaultDateFormat; | |
1438 | } | |
1439 | ||
1440 | U_NAMESPACE_END | |
1441 | ||
1442 | #endif /* #if !UCONFIG_NO_FORMATTING */ | |
1443 | ||
1444 | //eof |