]>
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 "number_affixutils.h" | |
9 | #include "unicode/utf16.h" | |
10 | #include "unicode/uniset.h" | |
11 | ||
12 | using namespace icu; | |
13 | using namespace icu::number; | |
14 | using namespace icu::number::impl; | |
15 | ||
16 | TokenConsumer::~TokenConsumer() = default; | |
17 | SymbolProvider::~SymbolProvider() = default; | |
18 | ||
19 | int32_t AffixUtils::estimateLength(const UnicodeString &patternString, UErrorCode &status) { | |
20 | AffixPatternState state = STATE_BASE; | |
21 | int32_t offset = 0; | |
22 | int32_t length = 0; | |
23 | for (; offset < patternString.length();) { | |
24 | UChar32 cp = patternString.char32At(offset); | |
25 | ||
26 | switch (state) { | |
27 | case STATE_BASE: | |
28 | if (cp == u'\'') { | |
29 | // First quote | |
30 | state = STATE_FIRST_QUOTE; | |
31 | } else { | |
32 | // Unquoted symbol | |
33 | length++; | |
34 | } | |
35 | break; | |
36 | case STATE_FIRST_QUOTE: | |
37 | if (cp == u'\'') { | |
38 | // Repeated quote | |
39 | length++; | |
40 | state = STATE_BASE; | |
41 | } else { | |
42 | // Quoted code point | |
43 | length++; | |
44 | state = STATE_INSIDE_QUOTE; | |
45 | } | |
46 | break; | |
47 | case STATE_INSIDE_QUOTE: | |
48 | if (cp == u'\'') { | |
49 | // End of quoted sequence | |
50 | state = STATE_AFTER_QUOTE; | |
51 | } else { | |
52 | // Quoted code point | |
53 | length++; | |
54 | } | |
55 | break; | |
56 | case STATE_AFTER_QUOTE: | |
57 | if (cp == u'\'') { | |
58 | // Double quote inside of quoted sequence | |
59 | length++; | |
60 | state = STATE_INSIDE_QUOTE; | |
61 | } else { | |
62 | // Unquoted symbol | |
63 | length++; | |
64 | } | |
65 | break; | |
66 | default: | |
67 | U_ASSERT(false); | |
68 | } | |
69 | ||
70 | offset += U16_LENGTH(cp); | |
71 | } | |
72 | ||
73 | switch (state) { | |
74 | case STATE_FIRST_QUOTE: | |
75 | case STATE_INSIDE_QUOTE: | |
76 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
77 | break; | |
78 | default: | |
79 | break; | |
80 | } | |
81 | ||
82 | return length; | |
83 | } | |
84 | ||
85 | UnicodeString AffixUtils::escape(const UnicodeString &input) { | |
86 | AffixPatternState state = STATE_BASE; | |
87 | int32_t offset = 0; | |
88 | UnicodeString output; | |
89 | for (; offset < input.length();) { | |
90 | UChar32 cp = input.char32At(offset); | |
91 | ||
92 | switch (cp) { | |
93 | case u'\'': | |
94 | output.append(u"''", -1); | |
95 | break; | |
96 | ||
97 | case u'-': | |
98 | case u'+': | |
99 | case u'%': | |
100 | case u'‰': | |
101 | case u'¤': | |
102 | if (state == STATE_BASE) { | |
103 | output.append(u'\''); | |
104 | output.append(cp); | |
105 | state = STATE_INSIDE_QUOTE; | |
106 | } else { | |
107 | output.append(cp); | |
108 | } | |
109 | break; | |
110 | ||
111 | default: | |
112 | if (state == STATE_INSIDE_QUOTE) { | |
113 | output.append(u'\''); | |
114 | output.append(cp); | |
115 | state = STATE_BASE; | |
116 | } else { | |
117 | output.append(cp); | |
118 | } | |
119 | break; | |
120 | } | |
121 | offset += U16_LENGTH(cp); | |
122 | } | |
123 | ||
124 | if (state == STATE_INSIDE_QUOTE) { | |
125 | output.append(u'\''); | |
126 | } | |
127 | ||
128 | return output; | |
129 | } | |
130 | ||
131 | Field AffixUtils::getFieldForType(AffixPatternType type) { | |
132 | switch (type) { | |
133 | case TYPE_MINUS_SIGN: | |
134 | return Field::UNUM_SIGN_FIELD; | |
135 | case TYPE_PLUS_SIGN: | |
136 | return Field::UNUM_SIGN_FIELD; | |
137 | case TYPE_PERCENT: | |
138 | return Field::UNUM_PERCENT_FIELD; | |
139 | case TYPE_PERMILLE: | |
140 | return Field::UNUM_PERMILL_FIELD; | |
141 | case TYPE_CURRENCY_SINGLE: | |
142 | return Field::UNUM_CURRENCY_FIELD; | |
143 | case TYPE_CURRENCY_DOUBLE: | |
144 | return Field::UNUM_CURRENCY_FIELD; | |
145 | case TYPE_CURRENCY_TRIPLE: | |
146 | return Field::UNUM_CURRENCY_FIELD; | |
147 | case TYPE_CURRENCY_QUAD: | |
148 | return Field::UNUM_CURRENCY_FIELD; | |
149 | case TYPE_CURRENCY_QUINT: | |
150 | return Field::UNUM_CURRENCY_FIELD; | |
151 | case TYPE_CURRENCY_OVERFLOW: | |
152 | return Field::UNUM_CURRENCY_FIELD; | |
153 | default: | |
154 | U_ASSERT(false); | |
155 | return Field::UNUM_FIELD_COUNT; // suppress "control reaches end of non-void function" | |
156 | } | |
157 | } | |
158 | ||
159 | int32_t | |
160 | AffixUtils::unescape(const UnicodeString &affixPattern, NumberStringBuilder &output, int32_t position, | |
161 | const SymbolProvider &provider, UErrorCode &status) { | |
162 | int32_t length = 0; | |
163 | AffixTag tag; | |
164 | while (hasNext(tag, affixPattern)) { | |
165 | tag = nextToken(tag, affixPattern, status); | |
166 | if (U_FAILURE(status)) { return length; } | |
167 | if (tag.type == TYPE_CURRENCY_OVERFLOW) { | |
168 | // Don't go to the provider for this special case | |
169 | length += output.insertCodePoint(position + length, 0xFFFD, UNUM_CURRENCY_FIELD, status); | |
170 | } else if (tag.type < 0) { | |
171 | length += output.insert( | |
172 | position + length, provider.getSymbol(tag.type), getFieldForType(tag.type), status); | |
173 | } else { | |
174 | length += output.insertCodePoint(position + length, tag.codePoint, UNUM_FIELD_COUNT, status); | |
175 | } | |
176 | } | |
177 | return length; | |
178 | } | |
179 | ||
180 | int32_t AffixUtils::unescapedCodePointCount(const UnicodeString &affixPattern, | |
181 | const SymbolProvider &provider, UErrorCode &status) { | |
182 | int32_t length = 0; | |
183 | AffixTag tag; | |
184 | while (hasNext(tag, affixPattern)) { | |
185 | tag = nextToken(tag, affixPattern, status); | |
186 | if (U_FAILURE(status)) { return length; } | |
187 | if (tag.type == TYPE_CURRENCY_OVERFLOW) { | |
188 | length += 1; | |
189 | } else if (tag.type < 0) { | |
190 | length += provider.getSymbol(tag.type).length(); | |
191 | } else { | |
192 | length += U16_LENGTH(tag.codePoint); | |
193 | } | |
194 | } | |
195 | return length; | |
196 | } | |
197 | ||
198 | bool | |
199 | AffixUtils::containsType(const UnicodeString &affixPattern, AffixPatternType type, UErrorCode &status) { | |
200 | if (affixPattern.length() == 0) { | |
201 | return false; | |
202 | } | |
203 | AffixTag tag; | |
204 | while (hasNext(tag, affixPattern)) { | |
205 | tag = nextToken(tag, affixPattern, status); | |
206 | if (U_FAILURE(status)) { return false; } | |
207 | if (tag.type == type) { | |
208 | return true; | |
209 | } | |
210 | } | |
211 | return false; | |
212 | } | |
213 | ||
214 | bool AffixUtils::hasCurrencySymbols(const UnicodeString &affixPattern, UErrorCode &status) { | |
215 | if (affixPattern.length() == 0) { | |
216 | return false; | |
217 | } | |
218 | AffixTag tag; | |
219 | while (hasNext(tag, affixPattern)) { | |
220 | tag = nextToken(tag, affixPattern, status); | |
221 | if (U_FAILURE(status)) { return false; } | |
222 | if (tag.type < 0 && getFieldForType(tag.type) == UNUM_CURRENCY_FIELD) { | |
223 | return true; | |
224 | } | |
225 | } | |
226 | return false; | |
227 | } | |
228 | ||
229 | UnicodeString AffixUtils::replaceType(const UnicodeString &affixPattern, AffixPatternType type, | |
230 | char16_t replacementChar, UErrorCode &status) { | |
231 | UnicodeString output(affixPattern); // copy | |
232 | if (affixPattern.length() == 0) { | |
233 | return output; | |
234 | }; | |
235 | AffixTag tag; | |
236 | while (hasNext(tag, affixPattern)) { | |
237 | tag = nextToken(tag, affixPattern, status); | |
238 | if (U_FAILURE(status)) { return output; } | |
239 | if (tag.type == type) { | |
240 | output.replace(tag.offset - 1, 1, replacementChar); | |
241 | } | |
242 | } | |
243 | return output; | |
244 | } | |
245 | ||
246 | bool AffixUtils::containsOnlySymbolsAndIgnorables(const UnicodeString& affixPattern, | |
247 | const UnicodeSet& ignorables, UErrorCode& status) { | |
248 | if (affixPattern.length() == 0) { | |
249 | return true; | |
250 | }; | |
251 | AffixTag tag; | |
252 | while (hasNext(tag, affixPattern)) { | |
253 | tag = nextToken(tag, affixPattern, status); | |
254 | if (U_FAILURE(status)) { return false; } | |
255 | if (tag.type == TYPE_CODEPOINT && !ignorables.contains(tag.codePoint)) { | |
256 | return false; | |
257 | } | |
258 | } | |
259 | return true; | |
260 | } | |
261 | ||
262 | void AffixUtils::iterateWithConsumer(const UnicodeString& affixPattern, TokenConsumer& consumer, | |
263 | UErrorCode& status) { | |
264 | if (affixPattern.length() == 0) { | |
265 | return; | |
266 | }; | |
267 | AffixTag tag; | |
268 | while (hasNext(tag, affixPattern)) { | |
269 | tag = nextToken(tag, affixPattern, status); | |
270 | if (U_FAILURE(status)) { return; } | |
271 | consumer.consumeToken(tag.type, tag.codePoint, status); | |
272 | if (U_FAILURE(status)) { return; } | |
273 | } | |
274 | } | |
275 | ||
276 | AffixTag AffixUtils::nextToken(AffixTag tag, const UnicodeString &patternString, UErrorCode &status) { | |
277 | int32_t offset = tag.offset; | |
278 | int32_t state = tag.state; | |
279 | for (; offset < patternString.length();) { | |
280 | UChar32 cp = patternString.char32At(offset); | |
281 | int32_t count = U16_LENGTH(cp); | |
282 | ||
283 | switch (state) { | |
284 | case STATE_BASE: | |
285 | switch (cp) { | |
286 | case u'\'': | |
287 | state = STATE_FIRST_QUOTE; | |
288 | offset += count; | |
289 | // continue to the next code point | |
290 | break; | |
291 | case u'-': | |
292 | return makeTag(offset + count, TYPE_MINUS_SIGN, STATE_BASE, 0); | |
293 | case u'+': | |
294 | return makeTag(offset + count, TYPE_PLUS_SIGN, STATE_BASE, 0); | |
295 | case u'%': | |
296 | return makeTag(offset + count, TYPE_PERCENT, STATE_BASE, 0); | |
297 | case u'‰': | |
298 | return makeTag(offset + count, TYPE_PERMILLE, STATE_BASE, 0); | |
299 | case u'¤': | |
300 | state = STATE_FIRST_CURR; | |
301 | offset += count; | |
302 | // continue to the next code point | |
303 | break; | |
304 | default: | |
305 | return makeTag(offset + count, TYPE_CODEPOINT, STATE_BASE, cp); | |
306 | } | |
307 | break; | |
308 | case STATE_FIRST_QUOTE: | |
309 | if (cp == u'\'') { | |
310 | return makeTag(offset + count, TYPE_CODEPOINT, STATE_BASE, cp); | |
311 | } else { | |
312 | return makeTag(offset + count, TYPE_CODEPOINT, STATE_INSIDE_QUOTE, cp); | |
313 | } | |
314 | case STATE_INSIDE_QUOTE: | |
315 | if (cp == u'\'') { | |
316 | state = STATE_AFTER_QUOTE; | |
317 | offset += count; | |
318 | // continue to the next code point | |
319 | break; | |
320 | } else { | |
321 | return makeTag(offset + count, TYPE_CODEPOINT, STATE_INSIDE_QUOTE, cp); | |
322 | } | |
323 | case STATE_AFTER_QUOTE: | |
324 | if (cp == u'\'') { | |
325 | return makeTag(offset + count, TYPE_CODEPOINT, STATE_INSIDE_QUOTE, cp); | |
326 | } else { | |
327 | state = STATE_BASE; | |
328 | // re-evaluate this code point | |
329 | break; | |
330 | } | |
331 | case STATE_FIRST_CURR: | |
332 | if (cp == u'¤') { | |
333 | state = STATE_SECOND_CURR; | |
334 | offset += count; | |
335 | // continue to the next code point | |
336 | break; | |
337 | } else { | |
338 | return makeTag(offset, TYPE_CURRENCY_SINGLE, STATE_BASE, 0); | |
339 | } | |
340 | case STATE_SECOND_CURR: | |
341 | if (cp == u'¤') { | |
342 | state = STATE_THIRD_CURR; | |
343 | offset += count; | |
344 | // continue to the next code point | |
345 | break; | |
346 | } else { | |
347 | return makeTag(offset, TYPE_CURRENCY_DOUBLE, STATE_BASE, 0); | |
348 | } | |
349 | case STATE_THIRD_CURR: | |
350 | if (cp == u'¤') { | |
351 | state = STATE_FOURTH_CURR; | |
352 | offset += count; | |
353 | // continue to the next code point | |
354 | break; | |
355 | } else { | |
356 | return makeTag(offset, TYPE_CURRENCY_TRIPLE, STATE_BASE, 0); | |
357 | } | |
358 | case STATE_FOURTH_CURR: | |
359 | if (cp == u'¤') { | |
360 | state = STATE_FIFTH_CURR; | |
361 | offset += count; | |
362 | // continue to the next code point | |
363 | break; | |
364 | } else { | |
365 | return makeTag(offset, TYPE_CURRENCY_QUAD, STATE_BASE, 0); | |
366 | } | |
367 | case STATE_FIFTH_CURR: | |
368 | if (cp == u'¤') { | |
369 | state = STATE_OVERFLOW_CURR; | |
370 | offset += count; | |
371 | // continue to the next code point | |
372 | break; | |
373 | } else { | |
374 | return makeTag(offset, TYPE_CURRENCY_QUINT, STATE_BASE, 0); | |
375 | } | |
376 | case STATE_OVERFLOW_CURR: | |
377 | if (cp == u'¤') { | |
378 | offset += count; | |
379 | // continue to the next code point and loop back to this state | |
380 | break; | |
381 | } else { | |
382 | return makeTag(offset, TYPE_CURRENCY_OVERFLOW, STATE_BASE, 0); | |
383 | } | |
384 | default: | |
385 | U_ASSERT(false); | |
386 | } | |
387 | } | |
388 | // End of string | |
389 | switch (state) { | |
390 | case STATE_BASE: | |
391 | // No more tokens in string. | |
392 | return {-1}; | |
393 | case STATE_FIRST_QUOTE: | |
394 | case STATE_INSIDE_QUOTE: | |
395 | // For consistent behavior with the JDK and ICU 58, set an error here. | |
396 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
397 | return {-1}; | |
398 | case STATE_AFTER_QUOTE: | |
399 | // No more tokens in string. | |
400 | return {-1}; | |
401 | case STATE_FIRST_CURR: | |
402 | return makeTag(offset, TYPE_CURRENCY_SINGLE, STATE_BASE, 0); | |
403 | case STATE_SECOND_CURR: | |
404 | return makeTag(offset, TYPE_CURRENCY_DOUBLE, STATE_BASE, 0); | |
405 | case STATE_THIRD_CURR: | |
406 | return makeTag(offset, TYPE_CURRENCY_TRIPLE, STATE_BASE, 0); | |
407 | case STATE_FOURTH_CURR: | |
408 | return makeTag(offset, TYPE_CURRENCY_QUAD, STATE_BASE, 0); | |
409 | case STATE_FIFTH_CURR: | |
410 | return makeTag(offset, TYPE_CURRENCY_QUINT, STATE_BASE, 0); | |
411 | case STATE_OVERFLOW_CURR: | |
412 | return makeTag(offset, TYPE_CURRENCY_OVERFLOW, STATE_BASE, 0); | |
413 | default: | |
414 | U_ASSERT(false); | |
415 | return {-1}; // suppress "control reaches end of non-void function" | |
416 | } | |
417 | } | |
418 | ||
419 | bool AffixUtils::hasNext(const AffixTag &tag, const UnicodeString &string) { | |
420 | // First check for the {-1} and default initializer syntax. | |
421 | if (tag.offset < 0) { | |
422 | return false; | |
423 | } else if (tag.offset == 0) { | |
424 | return string.length() > 0; | |
425 | } | |
426 | // The rest of the fields are safe to use now. | |
427 | // Special case: the last character in string is an end quote. | |
428 | if (tag.state == STATE_INSIDE_QUOTE && tag.offset == string.length() - 1 && | |
429 | string.charAt(tag.offset) == u'\'') { | |
430 | return false; | |
431 | } else if (tag.state != STATE_BASE) { | |
432 | return true; | |
433 | } else { | |
434 | return tag.offset < string.length(); | |
435 | } | |
436 | } | |
437 | ||
438 | #endif /* #if !UCONFIG_NO_FORMATTING */ |