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 | UPRV_UNREACHABLE; |
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 UNUM_SIGN_FIELD; |
135 | case TYPE_PLUS_SIGN: |
136 | return UNUM_SIGN_FIELD; |
137 | case TYPE_PERCENT: |
138 | return UNUM_PERCENT_FIELD; |
139 | case TYPE_PERMILLE: |
140 | return UNUM_PERMILL_FIELD; |
141 | case TYPE_CURRENCY_SINGLE: |
142 | return UNUM_CURRENCY_FIELD; |
143 | case TYPE_CURRENCY_DOUBLE: |
144 | return UNUM_CURRENCY_FIELD; |
145 | case TYPE_CURRENCY_TRIPLE: |
146 | return UNUM_CURRENCY_FIELD; |
147 | case TYPE_CURRENCY_QUAD: |
148 | return UNUM_CURRENCY_FIELD; |
149 | case TYPE_CURRENCY_QUINT: |
150 | return UNUM_CURRENCY_FIELD; |
151 | case TYPE_CURRENCY_OVERFLOW: |
152 | return UNUM_CURRENCY_FIELD; |
153 | default: |
154 | UPRV_UNREACHABLE; |
155 | } |
156 | } |
157 | |
158 | int32_t |
159 | AffixUtils::unescape(const UnicodeString &affixPattern, FormattedStringBuilder &output, int32_t position, |
160 | const SymbolProvider &provider, Field field, UErrorCode &status) { |
161 | int32_t length = 0; |
162 | AffixTag tag; |
163 | while (hasNext(tag, affixPattern)) { |
164 | tag = nextToken(tag, affixPattern, status); |
165 | if (U_FAILURE(status)) { return length; } |
166 | if (tag.type == TYPE_CURRENCY_OVERFLOW) { |
167 | // Don't go to the provider for this special case |
168 | length += output.insertCodePoint(position + length, 0xFFFD, UNUM_CURRENCY_FIELD, status); |
169 | } else if (tag.type < 0) { |
170 | length += output.insert( |
171 | position + length, provider.getSymbol(tag.type), getFieldForType(tag.type), status); |
172 | } else { |
173 | length += output.insertCodePoint(position + length, tag.codePoint, field, status); |
174 | } |
175 | } |
176 | return length; |
177 | } |
178 | |
179 | int32_t AffixUtils::unescapedCodePointCount(const UnicodeString &affixPattern, |
180 | const SymbolProvider &provider, UErrorCode &status) { |
181 | int32_t length = 0; |
182 | AffixTag tag; |
183 | while (hasNext(tag, affixPattern)) { |
184 | tag = nextToken(tag, affixPattern, status); |
185 | if (U_FAILURE(status)) { return length; } |
186 | if (tag.type == TYPE_CURRENCY_OVERFLOW) { |
187 | length += 1; |
188 | } else if (tag.type < 0) { |
189 | length += provider.getSymbol(tag.type).length(); |
190 | } else { |
191 | length += U16_LENGTH(tag.codePoint); |
192 | } |
193 | } |
194 | return length; |
195 | } |
196 | |
197 | bool |
198 | AffixUtils::containsType(const UnicodeString &affixPattern, AffixPatternType type, UErrorCode &status) { |
199 | if (affixPattern.length() == 0) { |
200 | return false; |
201 | } |
202 | AffixTag tag; |
203 | while (hasNext(tag, affixPattern)) { |
204 | tag = nextToken(tag, affixPattern, status); |
205 | if (U_FAILURE(status)) { return false; } |
206 | if (tag.type == type) { |
207 | return true; |
208 | } |
209 | } |
210 | return false; |
211 | } |
212 | |
213 | bool AffixUtils::hasCurrencySymbols(const UnicodeString &affixPattern, UErrorCode &status) { |
214 | if (affixPattern.length() == 0) { |
215 | return false; |
216 | } |
217 | AffixTag tag; |
218 | while (hasNext(tag, affixPattern)) { |
219 | tag = nextToken(tag, affixPattern, status); |
220 | if (U_FAILURE(status)) { return false; } |
221 | if (tag.type < 0 && getFieldForType(tag.type) == UNUM_CURRENCY_FIELD) { |
222 | return true; |
223 | } |
224 | } |
225 | return false; |
226 | } |
227 | |
228 | UnicodeString AffixUtils::replaceType(const UnicodeString &affixPattern, AffixPatternType type, |
229 | char16_t replacementChar, UErrorCode &status) { |
230 | UnicodeString output(affixPattern); // copy |
231 | if (affixPattern.length() == 0) { |
232 | return output; |
233 | } |
234 | AffixTag tag; |
235 | while (hasNext(tag, affixPattern)) { |
236 | tag = nextToken(tag, affixPattern, status); |
237 | if (U_FAILURE(status)) { return output; } |
238 | if (tag.type == type) { |
239 | output.replace(tag.offset - 1, 1, replacementChar); |
240 | } |
241 | } |
242 | return output; |
243 | } |
244 | |
245 | bool AffixUtils::containsOnlySymbolsAndIgnorables(const UnicodeString& affixPattern, |
246 | const UnicodeSet& ignorables, UErrorCode& status) { |
247 | if (affixPattern.length() == 0) { |
248 | return true; |
249 | } |
250 | AffixTag tag; |
251 | while (hasNext(tag, affixPattern)) { |
252 | tag = nextToken(tag, affixPattern, status); |
253 | if (U_FAILURE(status)) { return false; } |
254 | if (tag.type == TYPE_CODEPOINT && !ignorables.contains(tag.codePoint)) { |
255 | return false; |
256 | } |
257 | } |
258 | return true; |
259 | } |
260 | |
261 | void AffixUtils::iterateWithConsumer(const UnicodeString& affixPattern, TokenConsumer& consumer, |
262 | UErrorCode& status) { |
263 | if (affixPattern.length() == 0) { |
264 | return; |
265 | } |
266 | AffixTag tag; |
267 | while (hasNext(tag, affixPattern)) { |
268 | tag = nextToken(tag, affixPattern, status); |
269 | if (U_FAILURE(status)) { return; } |
270 | consumer.consumeToken(tag.type, tag.codePoint, status); |
271 | if (U_FAILURE(status)) { return; } |
272 | } |
273 | } |
274 | |
275 | AffixTag AffixUtils::nextToken(AffixTag tag, const UnicodeString &patternString, UErrorCode &status) { |
276 | int32_t offset = tag.offset; |
277 | int32_t state = tag.state; |
278 | for (; offset < patternString.length();) { |
279 | UChar32 cp = patternString.char32At(offset); |
280 | int32_t count = U16_LENGTH(cp); |
281 | |
282 | switch (state) { |
283 | case STATE_BASE: |
284 | switch (cp) { |
285 | case u'\'': |
286 | state = STATE_FIRST_QUOTE; |
287 | offset += count; |
288 | // continue to the next code point |
289 | break; |
290 | case u'-': |
291 | return makeTag(offset + count, TYPE_MINUS_SIGN, STATE_BASE, 0); |
292 | case u'+': |
293 | return makeTag(offset + count, TYPE_PLUS_SIGN, STATE_BASE, 0); |
294 | case u'%': |
295 | return makeTag(offset + count, TYPE_PERCENT, STATE_BASE, 0); |
296 | case u'‰': |
297 | return makeTag(offset + count, TYPE_PERMILLE, STATE_BASE, 0); |
298 | case u'¤': |
299 | state = STATE_FIRST_CURR; |
300 | offset += count; |
301 | // continue to the next code point |
302 | break; |
303 | default: |
304 | return makeTag(offset + count, TYPE_CODEPOINT, STATE_BASE, cp); |
305 | } |
306 | break; |
307 | case STATE_FIRST_QUOTE: |
308 | if (cp == u'\'') { |
309 | return makeTag(offset + count, TYPE_CODEPOINT, STATE_BASE, cp); |
310 | } else { |
311 | return makeTag(offset + count, TYPE_CODEPOINT, STATE_INSIDE_QUOTE, cp); |
312 | } |
313 | case STATE_INSIDE_QUOTE: |
314 | if (cp == u'\'') { |
315 | state = STATE_AFTER_QUOTE; |
316 | offset += count; |
317 | // continue to the next code point |
318 | break; |
319 | } else { |
320 | return makeTag(offset + count, TYPE_CODEPOINT, STATE_INSIDE_QUOTE, cp); |
321 | } |
322 | case STATE_AFTER_QUOTE: |
323 | if (cp == u'\'') { |
324 | return makeTag(offset + count, TYPE_CODEPOINT, STATE_INSIDE_QUOTE, cp); |
325 | } else { |
326 | state = STATE_BASE; |
327 | // re-evaluate this code point |
328 | break; |
329 | } |
330 | case STATE_FIRST_CURR: |
331 | if (cp == u'¤') { |
332 | state = STATE_SECOND_CURR; |
333 | offset += count; |
334 | // continue to the next code point |
335 | break; |
336 | } else { |
337 | return makeTag(offset, TYPE_CURRENCY_SINGLE, STATE_BASE, 0); |
338 | } |
339 | case STATE_SECOND_CURR: |
340 | if (cp == u'¤') { |
341 | state = STATE_THIRD_CURR; |
342 | offset += count; |
343 | // continue to the next code point |
344 | break; |
345 | } else { |
346 | return makeTag(offset, TYPE_CURRENCY_DOUBLE, STATE_BASE, 0); |
347 | } |
348 | case STATE_THIRD_CURR: |
349 | if (cp == u'¤') { |
350 | state = STATE_FOURTH_CURR; |
351 | offset += count; |
352 | // continue to the next code point |
353 | break; |
354 | } else { |
355 | return makeTag(offset, TYPE_CURRENCY_TRIPLE, STATE_BASE, 0); |
356 | } |
357 | case STATE_FOURTH_CURR: |
358 | if (cp == u'¤') { |
359 | state = STATE_FIFTH_CURR; |
360 | offset += count; |
361 | // continue to the next code point |
362 | break; |
363 | } else { |
364 | return makeTag(offset, TYPE_CURRENCY_QUAD, STATE_BASE, 0); |
365 | } |
366 | case STATE_FIFTH_CURR: |
367 | if (cp == u'¤') { |
368 | state = STATE_OVERFLOW_CURR; |
369 | offset += count; |
370 | // continue to the next code point |
371 | break; |
372 | } else { |
373 | return makeTag(offset, TYPE_CURRENCY_QUINT, STATE_BASE, 0); |
374 | } |
375 | case STATE_OVERFLOW_CURR: |
376 | if (cp == u'¤') { |
377 | offset += count; |
378 | // continue to the next code point and loop back to this state |
379 | break; |
380 | } else { |
381 | return makeTag(offset, TYPE_CURRENCY_OVERFLOW, STATE_BASE, 0); |
382 | } |
383 | default: |
384 | UPRV_UNREACHABLE; |
385 | } |
386 | } |
387 | // End of string |
388 | switch (state) { |
389 | case STATE_BASE: |
390 | // No more tokens in string. |
391 | return {-1}; |
392 | case STATE_FIRST_QUOTE: |
393 | case STATE_INSIDE_QUOTE: |
394 | // For consistent behavior with the JDK and ICU 58, set an error here. |
395 | status = U_ILLEGAL_ARGUMENT_ERROR; |
396 | return {-1}; |
397 | case STATE_AFTER_QUOTE: |
398 | // No more tokens in string. |
399 | return {-1}; |
400 | case STATE_FIRST_CURR: |
401 | return makeTag(offset, TYPE_CURRENCY_SINGLE, STATE_BASE, 0); |
402 | case STATE_SECOND_CURR: |
403 | return makeTag(offset, TYPE_CURRENCY_DOUBLE, STATE_BASE, 0); |
404 | case STATE_THIRD_CURR: |
405 | return makeTag(offset, TYPE_CURRENCY_TRIPLE, STATE_BASE, 0); |
406 | case STATE_FOURTH_CURR: |
407 | return makeTag(offset, TYPE_CURRENCY_QUAD, STATE_BASE, 0); |
408 | case STATE_FIFTH_CURR: |
409 | return makeTag(offset, TYPE_CURRENCY_QUINT, STATE_BASE, 0); |
410 | case STATE_OVERFLOW_CURR: |
411 | return makeTag(offset, TYPE_CURRENCY_OVERFLOW, STATE_BASE, 0); |
412 | default: |
413 | UPRV_UNREACHABLE; |
414 | } |
415 | } |
416 | |
417 | bool AffixUtils::hasNext(const AffixTag &tag, const UnicodeString &string) { |
418 | // First check for the {-1} and default initializer syntax. |
419 | if (tag.offset < 0) { |
420 | return false; |
421 | } else if (tag.offset == 0) { |
422 | return string.length() > 0; |
423 | } |
424 | // The rest of the fields are safe to use now. |
425 | // Special case: the last character in string is an end quote. |
426 | if (tag.state == STATE_INSIDE_QUOTE && tag.offset == string.length() - 1 && |
427 | string.charAt(tag.offset) == u'\'') { |
428 | return false; |
429 | } else if (tag.state != STATE_BASE) { |
430 | return true; |
431 | } else { |
432 | return tag.offset < string.length(); |
433 | } |
434 | } |
435 | |
436 | #endif /* #if !UCONFIG_NO_FORMATTING */ |
437 | |