]>
Commit | Line | Data |
---|---|---|
0f5d89e8 A |
1 | // © 2017 and later: Unicode, Inc. and others. |
2 | // License & terms of use: http://www.unicode.org/copyright.html | |
3 | ||
4 | #include "unicode/utypes.h" | |
5 | ||
6 | #if !UCONFIG_NO_FORMATTING | |
7 | ||
8 | #include "umutex.h" | |
9 | #include "ucln_cmn.h" | |
10 | #include "ucln_in.h" | |
11 | #include "number_modifiers.h" | |
12 | ||
13 | using namespace icu; | |
14 | using namespace icu::number; | |
15 | using namespace icu::number::impl; | |
16 | ||
17 | namespace { | |
18 | ||
19 | // TODO: This is copied from simpleformatter.cpp | |
20 | const int32_t ARG_NUM_LIMIT = 0x100; | |
21 | ||
22 | // These are the default currency spacing UnicodeSets in CLDR. | |
23 | // Pre-compute them for performance. | |
24 | // The Java unit test testCurrencySpacingPatternStability() will start failing if these change in CLDR. | |
25 | icu::UInitOnce gDefaultCurrencySpacingInitOnce = U_INITONCE_INITIALIZER; | |
26 | ||
27 | UnicodeSet *UNISET_DIGIT = nullptr; | |
28 | UnicodeSet *UNISET_NOTS = nullptr; | |
29 | ||
30 | UBool U_CALLCONV cleanupDefaultCurrencySpacing() { | |
31 | delete UNISET_DIGIT; | |
32 | UNISET_DIGIT = nullptr; | |
33 | delete UNISET_NOTS; | |
34 | UNISET_NOTS = nullptr; | |
35 | gDefaultCurrencySpacingInitOnce.reset(); | |
36 | return TRUE; | |
37 | } | |
38 | ||
39 | void U_CALLCONV initDefaultCurrencySpacing(UErrorCode &status) { | |
40 | ucln_i18n_registerCleanup(UCLN_I18N_CURRENCY_SPACING, cleanupDefaultCurrencySpacing); | |
41 | UNISET_DIGIT = new UnicodeSet(UnicodeString(u"[:digit:]"), status); | |
42 | UNISET_NOTS = new UnicodeSet(UnicodeString(u"[:^S:]"), status); | |
43 | if (UNISET_DIGIT == nullptr || UNISET_NOTS == nullptr) { | |
44 | status = U_MEMORY_ALLOCATION_ERROR; | |
45 | return; | |
46 | } | |
47 | UNISET_DIGIT->freeze(); | |
48 | UNISET_NOTS->freeze(); | |
49 | } | |
50 | ||
51 | } // namespace | |
52 | ||
53 | ||
54 | Modifier::~Modifier() = default; | |
55 | ||
3d1f044b A |
56 | Modifier::Parameters::Parameters() |
57 | : obj(nullptr) {} | |
58 | ||
59 | Modifier::Parameters::Parameters( | |
60 | const ModifierStore* _obj, int8_t _signum, StandardPlural::Form _plural) | |
61 | : obj(_obj), signum(_signum), plural(_plural) {} | |
62 | ||
63 | ModifierStore::~ModifierStore() = default; | |
64 | ||
65 | AdoptingModifierStore::~AdoptingModifierStore() { | |
66 | for (const Modifier *mod : mods) { | |
67 | delete mod; | |
68 | } | |
69 | } | |
70 | ||
0f5d89e8 A |
71 | |
72 | int32_t ConstantAffixModifier::apply(NumberStringBuilder &output, int leftIndex, int rightIndex, | |
73 | UErrorCode &status) const { | |
74 | // Insert the suffix first since inserting the prefix will change the rightIndex | |
75 | int length = output.insert(rightIndex, fSuffix, fField, status); | |
76 | length += output.insert(leftIndex, fPrefix, fField, status); | |
77 | return length; | |
78 | } | |
79 | ||
3d1f044b | 80 | int32_t ConstantAffixModifier::getPrefixLength() const { |
0f5d89e8 A |
81 | return fPrefix.length(); |
82 | } | |
83 | ||
3d1f044b | 84 | int32_t ConstantAffixModifier::getCodePointCount() const { |
0f5d89e8 A |
85 | return fPrefix.countChar32() + fSuffix.countChar32(); |
86 | } | |
87 | ||
88 | bool ConstantAffixModifier::isStrong() const { | |
89 | return fStrong; | |
90 | } | |
91 | ||
3d1f044b A |
92 | bool ConstantAffixModifier::containsField(UNumberFormatFields field) const { |
93 | (void)field; | |
94 | // This method is not currently used. | |
95 | UPRV_UNREACHABLE; | |
96 | } | |
97 | ||
98 | void ConstantAffixModifier::getParameters(Parameters& output) const { | |
99 | (void)output; | |
100 | // This method is not currently used. | |
101 | UPRV_UNREACHABLE; | |
102 | } | |
103 | ||
104 | bool ConstantAffixModifier::semanticallyEquivalent(const Modifier& other) const { | |
105 | auto* _other = dynamic_cast<const ConstantAffixModifier*>(&other); | |
106 | if (_other == nullptr) { | |
107 | return false; | |
108 | } | |
109 | return fPrefix == _other->fPrefix | |
110 | && fSuffix == _other->fSuffix | |
111 | && fField == _other->fField | |
112 | && fStrong == _other->fStrong; | |
113 | } | |
114 | ||
115 | ||
0f5d89e8 | 116 | SimpleModifier::SimpleModifier(const SimpleFormatter &simpleFormatter, Field field, bool strong) |
3d1f044b A |
117 | : SimpleModifier(simpleFormatter, field, strong, {}) {} |
118 | ||
119 | SimpleModifier::SimpleModifier(const SimpleFormatter &simpleFormatter, Field field, bool strong, | |
120 | const Modifier::Parameters parameters) | |
121 | : fCompiledPattern(simpleFormatter.compiledPattern), fField(field), fStrong(strong), | |
122 | fParameters(parameters) { | |
0f5d89e8 A |
123 | int32_t argLimit = SimpleFormatter::getArgumentLimit( |
124 | fCompiledPattern.getBuffer(), fCompiledPattern.length()); | |
125 | if (argLimit == 0) { | |
126 | // No arguments in compiled pattern | |
127 | fPrefixLength = fCompiledPattern.charAt(1) - ARG_NUM_LIMIT; | |
128 | U_ASSERT(2 + fPrefixLength == fCompiledPattern.length()); | |
129 | // Set suffixOffset = -1 to indicate no arguments in compiled pattern. | |
130 | fSuffixOffset = -1; | |
131 | fSuffixLength = 0; | |
132 | } else { | |
133 | U_ASSERT(argLimit == 1); | |
134 | if (fCompiledPattern.charAt(1) != 0) { | |
3d1f044b | 135 | // Found prefix |
0f5d89e8 A |
136 | fPrefixLength = fCompiledPattern.charAt(1) - ARG_NUM_LIMIT; |
137 | fSuffixOffset = 3 + fPrefixLength; | |
138 | } else { | |
3d1f044b | 139 | // No prefix |
0f5d89e8 A |
140 | fPrefixLength = 0; |
141 | fSuffixOffset = 2; | |
142 | } | |
143 | if (3 + fPrefixLength < fCompiledPattern.length()) { | |
3d1f044b | 144 | // Found suffix |
0f5d89e8 A |
145 | fSuffixLength = fCompiledPattern.charAt(fSuffixOffset) - ARG_NUM_LIMIT; |
146 | } else { | |
3d1f044b | 147 | // No suffix |
0f5d89e8 A |
148 | fSuffixLength = 0; |
149 | } | |
150 | } | |
151 | } | |
152 | ||
153 | SimpleModifier::SimpleModifier() | |
154 | : fField(UNUM_FIELD_COUNT), fStrong(false), fPrefixLength(0), fSuffixLength(0) { | |
155 | } | |
156 | ||
157 | int32_t SimpleModifier::apply(NumberStringBuilder &output, int leftIndex, int rightIndex, | |
158 | UErrorCode &status) const { | |
3d1f044b | 159 | return formatAsPrefixSuffix(output, leftIndex, rightIndex, status); |
0f5d89e8 A |
160 | } |
161 | ||
3d1f044b | 162 | int32_t SimpleModifier::getPrefixLength() const { |
0f5d89e8 A |
163 | return fPrefixLength; |
164 | } | |
165 | ||
3d1f044b | 166 | int32_t SimpleModifier::getCodePointCount() const { |
0f5d89e8 A |
167 | int32_t count = 0; |
168 | if (fPrefixLength > 0) { | |
169 | count += fCompiledPattern.countChar32(2, fPrefixLength); | |
170 | } | |
171 | if (fSuffixLength > 0) { | |
172 | count += fCompiledPattern.countChar32(1 + fSuffixOffset, fSuffixLength); | |
173 | } | |
174 | return count; | |
175 | } | |
176 | ||
177 | bool SimpleModifier::isStrong() const { | |
178 | return fStrong; | |
179 | } | |
180 | ||
3d1f044b A |
181 | bool SimpleModifier::containsField(UNumberFormatFields field) const { |
182 | (void)field; | |
183 | // This method is not currently used. | |
184 | UPRV_UNREACHABLE; | |
185 | } | |
186 | ||
187 | void SimpleModifier::getParameters(Parameters& output) const { | |
188 | output = fParameters; | |
189 | } | |
190 | ||
191 | bool SimpleModifier::semanticallyEquivalent(const Modifier& other) const { | |
192 | auto* _other = dynamic_cast<const SimpleModifier*>(&other); | |
193 | if (_other == nullptr) { | |
194 | return false; | |
195 | } | |
196 | if (fParameters.obj != nullptr) { | |
197 | return fParameters.obj == _other->fParameters.obj; | |
198 | } | |
199 | return fCompiledPattern == _other->fCompiledPattern | |
200 | && fField == _other->fField | |
201 | && fStrong == _other->fStrong; | |
202 | } | |
203 | ||
204 | ||
0f5d89e8 A |
205 | int32_t |
206 | SimpleModifier::formatAsPrefixSuffix(NumberStringBuilder &result, int32_t startIndex, int32_t endIndex, | |
3d1f044b A |
207 | UErrorCode &status) const { |
208 | if (fSuffixOffset == -1 && fPrefixLength + fSuffixLength > 0) { | |
0f5d89e8 | 209 | // There is no argument for the inner number; overwrite the entire segment with our string. |
3d1f044b | 210 | return result.splice(startIndex, endIndex, fCompiledPattern, 2, 2 + fPrefixLength, fField, status); |
0f5d89e8 A |
211 | } else { |
212 | if (fPrefixLength > 0) { | |
3d1f044b | 213 | result.insert(startIndex, fCompiledPattern, 2, 2 + fPrefixLength, fField, status); |
0f5d89e8 A |
214 | } |
215 | if (fSuffixLength > 0) { | |
216 | result.insert( | |
217 | endIndex + fPrefixLength, | |
218 | fCompiledPattern, | |
219 | 1 + fSuffixOffset, | |
220 | 1 + fSuffixOffset + fSuffixLength, | |
3d1f044b | 221 | fField, |
0f5d89e8 A |
222 | status); |
223 | } | |
224 | return fPrefixLength + fSuffixLength; | |
225 | } | |
226 | } | |
227 | ||
3d1f044b A |
228 | |
229 | int32_t | |
230 | SimpleModifier::formatTwoArgPattern(const SimpleFormatter& compiled, NumberStringBuilder& result, | |
231 | int32_t index, int32_t* outPrefixLength, int32_t* outSuffixLength, | |
232 | Field field, UErrorCode& status) { | |
233 | const UnicodeString& compiledPattern = compiled.compiledPattern; | |
234 | int32_t argLimit = SimpleFormatter::getArgumentLimit( | |
235 | compiledPattern.getBuffer(), compiledPattern.length()); | |
236 | if (argLimit != 2) { | |
237 | status = U_INTERNAL_PROGRAM_ERROR; | |
238 | return 0; | |
239 | } | |
240 | int32_t offset = 1; // offset into compiledPattern | |
241 | int32_t length = 0; // chars added to result | |
242 | ||
243 | int32_t prefixLength = compiledPattern.charAt(offset); | |
244 | offset++; | |
245 | if (prefixLength < ARG_NUM_LIMIT) { | |
246 | // No prefix | |
247 | prefixLength = 0; | |
248 | } else { | |
249 | prefixLength -= ARG_NUM_LIMIT; | |
250 | result.insert(index + length, compiledPattern, offset, offset + prefixLength, field, status); | |
251 | offset += prefixLength; | |
252 | length += prefixLength; | |
253 | offset++; | |
254 | } | |
255 | ||
256 | int32_t infixLength = compiledPattern.charAt(offset); | |
257 | offset++; | |
258 | if (infixLength < ARG_NUM_LIMIT) { | |
259 | // No infix | |
260 | infixLength = 0; | |
261 | } else { | |
262 | infixLength -= ARG_NUM_LIMIT; | |
263 | result.insert(index + length, compiledPattern, offset, offset + infixLength, field, status); | |
264 | offset += infixLength; | |
265 | length += infixLength; | |
266 | offset++; | |
267 | } | |
268 | ||
269 | int32_t suffixLength; | |
270 | if (offset == compiledPattern.length()) { | |
271 | // No suffix | |
272 | suffixLength = 0; | |
273 | } else { | |
274 | suffixLength = compiledPattern.charAt(offset) - ARG_NUM_LIMIT; | |
275 | offset++; | |
276 | result.insert(index + length, compiledPattern, offset, offset + suffixLength, field, status); | |
277 | length += suffixLength; | |
278 | } | |
279 | ||
280 | *outPrefixLength = prefixLength; | |
281 | *outSuffixLength = suffixLength; | |
282 | ||
283 | return length; | |
284 | } | |
285 | ||
286 | ||
0f5d89e8 A |
287 | int32_t ConstantMultiFieldModifier::apply(NumberStringBuilder &output, int leftIndex, int rightIndex, |
288 | UErrorCode &status) const { | |
289 | int32_t length = output.insert(leftIndex, fPrefix, status); | |
290 | if (fOverwrite) { | |
291 | length += output.splice( | |
292 | leftIndex + length, | |
293 | rightIndex + length, | |
294 | UnicodeString(), 0, 0, | |
295 | UNUM_FIELD_COUNT, status); | |
296 | } | |
297 | length += output.insert(rightIndex + length, fSuffix, status); | |
298 | return length; | |
299 | } | |
300 | ||
3d1f044b | 301 | int32_t ConstantMultiFieldModifier::getPrefixLength() const { |
0f5d89e8 A |
302 | return fPrefix.length(); |
303 | } | |
304 | ||
3d1f044b | 305 | int32_t ConstantMultiFieldModifier::getCodePointCount() const { |
0f5d89e8 A |
306 | return fPrefix.codePointCount() + fSuffix.codePointCount(); |
307 | } | |
308 | ||
309 | bool ConstantMultiFieldModifier::isStrong() const { | |
310 | return fStrong; | |
311 | } | |
312 | ||
3d1f044b A |
313 | bool ConstantMultiFieldModifier::containsField(UNumberFormatFields field) const { |
314 | return fPrefix.containsField(field) || fSuffix.containsField(field); | |
315 | } | |
316 | ||
317 | void ConstantMultiFieldModifier::getParameters(Parameters& output) const { | |
318 | output = fParameters; | |
319 | } | |
320 | ||
321 | bool ConstantMultiFieldModifier::semanticallyEquivalent(const Modifier& other) const { | |
322 | auto* _other = dynamic_cast<const ConstantMultiFieldModifier*>(&other); | |
323 | if (_other == nullptr) { | |
324 | return false; | |
325 | } | |
326 | if (fParameters.obj != nullptr) { | |
327 | return fParameters.obj == _other->fParameters.obj; | |
328 | } | |
329 | return fPrefix.contentEquals(_other->fPrefix) | |
330 | && fSuffix.contentEquals(_other->fSuffix) | |
331 | && fOverwrite == _other->fOverwrite | |
332 | && fStrong == _other->fStrong; | |
333 | } | |
334 | ||
335 | ||
0f5d89e8 A |
336 | CurrencySpacingEnabledModifier::CurrencySpacingEnabledModifier(const NumberStringBuilder &prefix, |
337 | const NumberStringBuilder &suffix, | |
338 | bool overwrite, | |
339 | bool strong, | |
340 | const DecimalFormatSymbols &symbols, | |
341 | UErrorCode &status) | |
342 | : ConstantMultiFieldModifier(prefix, suffix, overwrite, strong) { | |
343 | // Check for currency spacing. Do not build the UnicodeSets unless there is | |
344 | // a currency code point at a boundary. | |
345 | if (prefix.length() > 0 && prefix.fieldAt(prefix.length() - 1) == UNUM_CURRENCY_FIELD) { | |
346 | int prefixCp = prefix.getLastCodePoint(); | |
347 | UnicodeSet prefixUnicodeSet = getUnicodeSet(symbols, IN_CURRENCY, PREFIX, status); | |
348 | if (prefixUnicodeSet.contains(prefixCp)) { | |
349 | fAfterPrefixUnicodeSet = getUnicodeSet(symbols, IN_NUMBER, PREFIX, status); | |
350 | fAfterPrefixUnicodeSet.freeze(); | |
351 | fAfterPrefixInsert = getInsertString(symbols, PREFIX, status); | |
352 | } else { | |
353 | fAfterPrefixUnicodeSet.setToBogus(); | |
354 | fAfterPrefixInsert.setToBogus(); | |
355 | } | |
356 | } else { | |
357 | fAfterPrefixUnicodeSet.setToBogus(); | |
358 | fAfterPrefixInsert.setToBogus(); | |
359 | } | |
360 | if (suffix.length() > 0 && suffix.fieldAt(0) == UNUM_CURRENCY_FIELD) { | |
361 | int suffixCp = suffix.getLastCodePoint(); | |
362 | UnicodeSet suffixUnicodeSet = getUnicodeSet(symbols, IN_CURRENCY, SUFFIX, status); | |
363 | if (suffixUnicodeSet.contains(suffixCp)) { | |
364 | fBeforeSuffixUnicodeSet = getUnicodeSet(symbols, IN_NUMBER, SUFFIX, status); | |
365 | fBeforeSuffixUnicodeSet.freeze(); | |
366 | fBeforeSuffixInsert = getInsertString(symbols, SUFFIX, status); | |
367 | } else { | |
368 | fBeforeSuffixUnicodeSet.setToBogus(); | |
369 | fBeforeSuffixInsert.setToBogus(); | |
370 | } | |
371 | } else { | |
372 | fBeforeSuffixUnicodeSet.setToBogus(); | |
373 | fBeforeSuffixInsert.setToBogus(); | |
374 | } | |
375 | } | |
376 | ||
377 | int32_t CurrencySpacingEnabledModifier::apply(NumberStringBuilder &output, int leftIndex, int rightIndex, | |
378 | UErrorCode &status) const { | |
379 | // Currency spacing logic | |
380 | int length = 0; | |
381 | if (rightIndex - leftIndex > 0 && !fAfterPrefixUnicodeSet.isBogus() && | |
382 | fAfterPrefixUnicodeSet.contains(output.codePointAt(leftIndex))) { | |
383 | // TODO: Should we use the CURRENCY field here? | |
384 | length += output.insert(leftIndex, fAfterPrefixInsert, UNUM_FIELD_COUNT, status); | |
385 | } | |
386 | if (rightIndex - leftIndex > 0 && !fBeforeSuffixUnicodeSet.isBogus() && | |
387 | fBeforeSuffixUnicodeSet.contains(output.codePointBefore(rightIndex))) { | |
388 | // TODO: Should we use the CURRENCY field here? | |
389 | length += output.insert(rightIndex + length, fBeforeSuffixInsert, UNUM_FIELD_COUNT, status); | |
390 | } | |
391 | ||
392 | // Call super for the remaining logic | |
393 | length += ConstantMultiFieldModifier::apply(output, leftIndex, rightIndex + length, status); | |
394 | return length; | |
395 | } | |
396 | ||
397 | int32_t | |
398 | CurrencySpacingEnabledModifier::applyCurrencySpacing(NumberStringBuilder &output, int32_t prefixStart, | |
399 | int32_t prefixLen, int32_t suffixStart, | |
400 | int32_t suffixLen, | |
401 | const DecimalFormatSymbols &symbols, | |
402 | UErrorCode &status) { | |
403 | int length = 0; | |
404 | bool hasPrefix = (prefixLen > 0); | |
405 | bool hasSuffix = (suffixLen > 0); | |
406 | bool hasNumber = (suffixStart - prefixStart - prefixLen > 0); // could be empty string | |
407 | if (hasPrefix && hasNumber) { | |
408 | length += applyCurrencySpacingAffix(output, prefixStart + prefixLen, PREFIX, symbols, status); | |
409 | } | |
410 | if (hasSuffix && hasNumber) { | |
411 | length += applyCurrencySpacingAffix(output, suffixStart + length, SUFFIX, symbols, status); | |
412 | } | |
413 | return length; | |
414 | } | |
415 | ||
416 | int32_t | |
417 | CurrencySpacingEnabledModifier::applyCurrencySpacingAffix(NumberStringBuilder &output, int32_t index, | |
418 | EAffix affix, | |
419 | const DecimalFormatSymbols &symbols, | |
420 | UErrorCode &status) { | |
421 | // NOTE: For prefix, output.fieldAt(index-1) gets the last field type in the prefix. | |
422 | // This works even if the last code point in the prefix is 2 code units because the | |
423 | // field value gets populated to both indices in the field array. | |
424 | Field affixField = (affix == PREFIX) ? output.fieldAt(index - 1) : output.fieldAt(index); | |
425 | if (affixField != UNUM_CURRENCY_FIELD) { | |
426 | return 0; | |
427 | } | |
428 | int affixCp = (affix == PREFIX) ? output.codePointBefore(index) : output.codePointAt(index); | |
429 | UnicodeSet affixUniset = getUnicodeSet(symbols, IN_CURRENCY, affix, status); | |
430 | if (!affixUniset.contains(affixCp)) { | |
431 | return 0; | |
432 | } | |
433 | int numberCp = (affix == PREFIX) ? output.codePointAt(index) : output.codePointBefore(index); | |
434 | UnicodeSet numberUniset = getUnicodeSet(symbols, IN_NUMBER, affix, status); | |
435 | if (!numberUniset.contains(numberCp)) { | |
436 | return 0; | |
437 | } | |
438 | UnicodeString spacingString = getInsertString(symbols, affix, status); | |
439 | ||
440 | // NOTE: This next line *inserts* the spacing string, triggering an arraycopy. | |
441 | // It would be more efficient if this could be done before affixes were attached, | |
442 | // so that it could be prepended/appended instead of inserted. | |
443 | // However, the build code path is more efficient, and this is the most natural | |
444 | // place to put currency spacing in the non-build code path. | |
445 | // TODO: Should we use the CURRENCY field here? | |
446 | return output.insert(index, spacingString, UNUM_FIELD_COUNT, status); | |
447 | } | |
448 | ||
449 | UnicodeSet | |
450 | CurrencySpacingEnabledModifier::getUnicodeSet(const DecimalFormatSymbols &symbols, EPosition position, | |
451 | EAffix affix, UErrorCode &status) { | |
452 | // Ensure the static defaults are initialized: | |
453 | umtx_initOnce(gDefaultCurrencySpacingInitOnce, &initDefaultCurrencySpacing, status); | |
454 | if (U_FAILURE(status)) { | |
455 | return UnicodeSet(); | |
456 | } | |
457 | ||
458 | const UnicodeString& pattern = symbols.getPatternForCurrencySpacing( | |
459 | position == IN_CURRENCY ? UNUM_CURRENCY_MATCH : UNUM_CURRENCY_SURROUNDING_MATCH, | |
460 | affix == SUFFIX, | |
461 | status); | |
462 | if (pattern.compare(u"[:digit:]", -1) == 0) { | |
463 | return *UNISET_DIGIT; | |
464 | } else if (pattern.compare(u"[:^S:]", -1) == 0) { | |
465 | return *UNISET_NOTS; | |
466 | } else { | |
467 | return UnicodeSet(pattern, status); | |
468 | } | |
469 | } | |
470 | ||
471 | UnicodeString | |
472 | CurrencySpacingEnabledModifier::getInsertString(const DecimalFormatSymbols &symbols, EAffix affix, | |
473 | UErrorCode &status) { | |
474 | return symbols.getPatternForCurrencySpacing(UNUM_CURRENCY_INSERT, affix == SUFFIX, status); | |
475 | } | |
476 | ||
477 | #endif /* #if !UCONFIG_NO_FORMATTING */ |