1 | // © 2017 and later: Unicode, Inc. and others. |
2 | // License & terms of use: http://www.unicode.org/copyright.html |
3 | |
4 | // ucptrie.cpp (modified from utrie2.cpp) |
5 | // created: 2017dec29 Markus W. Scherer |
6 | |
7 | // #define UCPTRIE_DEBUG |
8 | #ifdef UCPTRIE_DEBUG |
9 | # include <stdio.h> |
10 | #endif |
11 | |
12 | #include "unicode/utypes.h" |
13 | #include "unicode/ucptrie.h" |
14 | #include "unicode/utf.h" |
15 | #include "unicode/utf8.h" |
16 | #include "unicode/utf16.h" |
17 | #include "cmemory.h" |
18 | #include "uassert.h" |
19 | #include "ucptrie_impl.h" |
20 | |
21 | U_CAPI UCPTrie * U_EXPORT2 |
22 | ucptrie_openFromBinary(UCPTrieType type, UCPTrieValueWidth valueWidth, |
23 | const void *data, int32_t length, int32_t *pActualLength, |
24 | UErrorCode *pErrorCode) { |
25 | if (U_FAILURE(*pErrorCode)) { |
26 | return nullptr; |
27 | } |
28 | |
29 | if (length <= 0 || (U_POINTER_MASK_LSB(data, 3) != 0) || |
30 | type < UCPTRIE_TYPE_ANY || UCPTRIE_TYPE_SMALL < type || |
31 | valueWidth < UCPTRIE_VALUE_BITS_ANY || UCPTRIE_VALUE_BITS_8 < valueWidth) { |
32 | *pErrorCode = U_ILLEGAL_ARGUMENT_ERROR; |
33 | return nullptr; |
34 | } |
35 | |
36 | // Enough data for a trie header? |
37 | if (length < (int32_t)sizeof(UCPTrieHeader)) { |
38 | *pErrorCode = U_INVALID_FORMAT_ERROR; |
39 | return nullptr; |
40 | } |
41 | |
42 | // Check the signature. |
43 | const UCPTrieHeader * = (const UCPTrieHeader *)data; |
44 | if (header->signature != UCPTRIE_SIG) { |
45 | *pErrorCode = U_INVALID_FORMAT_ERROR; |
46 | return nullptr; |
47 | } |
48 | |
49 | int32_t options = header->options; |
50 | int32_t typeInt = (options >> 6) & 3; |
51 | int32_t valueWidthInt = options & UCPTRIE_OPTIONS_VALUE_BITS_MASK; |
52 | if (typeInt > UCPTRIE_TYPE_SMALL || valueWidthInt > UCPTRIE_VALUE_BITS_8 || |
53 | (options & UCPTRIE_OPTIONS_RESERVED_MASK) != 0) { |
54 | *pErrorCode = U_INVALID_FORMAT_ERROR; |
55 | return nullptr; |
56 | } |
57 | UCPTrieType actualType = (UCPTrieType)typeInt; |
58 | UCPTrieValueWidth actualValueWidth = (UCPTrieValueWidth)valueWidthInt; |
59 | if (type < 0) { |
60 | type = actualType; |
61 | } |
62 | if (valueWidth < 0) { |
63 | valueWidth = actualValueWidth; |
64 | } |
65 | if (type != actualType || valueWidth != actualValueWidth) { |
66 | *pErrorCode = U_INVALID_FORMAT_ERROR; |
67 | return nullptr; |
68 | } |
69 | |
70 | // Get the length values and offsets. |
71 | UCPTrie tempTrie; |
72 | uprv_memset(&tempTrie, 0, sizeof(tempTrie)); |
73 | tempTrie.indexLength = header->indexLength; |
74 | tempTrie.dataLength = |
75 | ((options & UCPTRIE_OPTIONS_DATA_LENGTH_MASK) << 4) | header->dataLength; |
76 | tempTrie.index3NullOffset = header->index3NullOffset; |
77 | tempTrie.dataNullOffset = |
78 | ((options & UCPTRIE_OPTIONS_DATA_NULL_OFFSET_MASK) << 8) | header->dataNullOffset; |
79 | |
80 | tempTrie.highStart = header->shiftedHighStart << UCPTRIE_SHIFT_2; |
81 | tempTrie.shifted12HighStart = (tempTrie.highStart + 0xfff) >> 12; |
82 | tempTrie.type = type; |
83 | tempTrie.valueWidth = valueWidth; |
84 | |
85 | // Calculate the actual length. |
86 | int32_t actualLength = (int32_t)sizeof(UCPTrieHeader) + tempTrie.indexLength * 2; |
87 | if (valueWidth == UCPTRIE_VALUE_BITS_16) { |
88 | actualLength += tempTrie.dataLength * 2; |
89 | } else if (valueWidth == UCPTRIE_VALUE_BITS_32) { |
90 | actualLength += tempTrie.dataLength * 4; |
91 | } else { |
92 | actualLength += tempTrie.dataLength; |
93 | } |
94 | if (length < actualLength) { |
95 | *pErrorCode = U_INVALID_FORMAT_ERROR; // Not enough bytes. |
96 | return nullptr; |
97 | } |
98 | |
99 | // Allocate the trie. |
100 | UCPTrie *trie = (UCPTrie *)uprv_malloc(sizeof(UCPTrie)); |
101 | if (trie == nullptr) { |
102 | *pErrorCode = U_MEMORY_ALLOCATION_ERROR; |
103 | return nullptr; |
104 | } |
105 | uprv_memcpy(trie, &tempTrie, sizeof(tempTrie)); |
106 | #ifdef UCPTRIE_DEBUG |
107 | trie->name = "fromSerialized" ; |
108 | #endif |
109 | |
110 | // Set the pointers to its index and data arrays. |
111 | const uint16_t *p16 = (const uint16_t *)(header + 1); |
112 | trie->index = p16; |
113 | p16 += trie->indexLength; |
114 | |
115 | // Get the data. |
116 | int32_t nullValueOffset = trie->dataNullOffset; |
117 | if (nullValueOffset >= trie->dataLength) { |
118 | nullValueOffset = trie->dataLength - UCPTRIE_HIGH_VALUE_NEG_DATA_OFFSET; |
119 | } |
120 | switch (valueWidth) { |
121 | case UCPTRIE_VALUE_BITS_16: |
122 | trie->data.ptr16 = p16; |
123 | trie->nullValue = trie->data.ptr16[nullValueOffset]; |
124 | break; |
125 | case UCPTRIE_VALUE_BITS_32: |
126 | trie->data.ptr32 = (const uint32_t *)p16; |
127 | trie->nullValue = trie->data.ptr32[nullValueOffset]; |
128 | break; |
129 | case UCPTRIE_VALUE_BITS_8: |
130 | trie->data.ptr8 = (const uint8_t *)p16; |
131 | trie->nullValue = trie->data.ptr8[nullValueOffset]; |
132 | break; |
133 | default: |
134 | // Unreachable because valueWidth was checked above. |
135 | *pErrorCode = U_INVALID_FORMAT_ERROR; |
136 | return nullptr; |
137 | } |
138 | |
139 | if (pActualLength != nullptr) { |
140 | *pActualLength = actualLength; |
141 | } |
142 | return trie; |
143 | } |
144 | |
145 | U_CAPI void U_EXPORT2 |
146 | ucptrie_close(UCPTrie *trie) { |
147 | uprv_free(trie); |
148 | } |
149 | |
150 | U_CAPI UCPTrieType U_EXPORT2 |
151 | ucptrie_getType(const UCPTrie *trie) { |
152 | return (UCPTrieType)trie->type; |
153 | } |
154 | |
155 | U_CAPI UCPTrieValueWidth U_EXPORT2 |
156 | ucptrie_getValueWidth(const UCPTrie *trie) { |
157 | return (UCPTrieValueWidth)trie->valueWidth; |
158 | } |
159 | |
160 | U_CAPI int32_t U_EXPORT2 |
161 | ucptrie_internalSmallIndex(const UCPTrie *trie, UChar32 c) { |
162 | int32_t i1 = c >> UCPTRIE_SHIFT_1; |
163 | if (trie->type == UCPTRIE_TYPE_FAST) { |
164 | U_ASSERT(0xffff < c && c < trie->highStart); |
165 | i1 += UCPTRIE_BMP_INDEX_LENGTH - UCPTRIE_OMITTED_BMP_INDEX_1_LENGTH; |
166 | } else { |
167 | U_ASSERT((uint32_t)c < (uint32_t)trie->highStart && trie->highStart > UCPTRIE_SMALL_LIMIT); |
168 | i1 += UCPTRIE_SMALL_INDEX_LENGTH; |
169 | } |
170 | int32_t i3Block = trie->index[ |
171 | (int32_t)trie->index[i1] + ((c >> UCPTRIE_SHIFT_2) & UCPTRIE_INDEX_2_MASK)]; |
172 | int32_t i3 = (c >> UCPTRIE_SHIFT_3) & UCPTRIE_INDEX_3_MASK; |
173 | int32_t dataBlock; |
174 | if ((i3Block & 0x8000) == 0) { |
175 | // 16-bit indexes |
176 | dataBlock = trie->index[i3Block + i3]; |
177 | } else { |
178 | // 18-bit indexes stored in groups of 9 entries per 8 indexes. |
179 | i3Block = (i3Block & 0x7fff) + (i3 & ~7) + (i3 >> 3); |
180 | i3 &= 7; |
181 | dataBlock = ((int32_t)trie->index[i3Block++] << (2 + (2 * i3))) & 0x30000; |
182 | dataBlock |= trie->index[i3Block + i3]; |
183 | } |
184 | return dataBlock + (c & UCPTRIE_SMALL_DATA_MASK); |
185 | } |
186 | |
187 | U_CAPI int32_t U_EXPORT2 |
188 | ucptrie_internalSmallU8Index(const UCPTrie *trie, int32_t lt1, uint8_t t2, uint8_t t3) { |
189 | UChar32 c = (lt1 << 12) | (t2 << 6) | t3; |
190 | if (c >= trie->highStart) { |
191 | // Possible because the UTF-8 macro compares with shifted12HighStart which may be higher. |
192 | return trie->dataLength - UCPTRIE_HIGH_VALUE_NEG_DATA_OFFSET; |
193 | } |
194 | return ucptrie_internalSmallIndex(trie, c); |
195 | } |
196 | |
197 | U_CAPI int32_t U_EXPORT2 |
198 | ucptrie_internalU8PrevIndex(const UCPTrie *trie, UChar32 c, |
199 | const uint8_t *start, const uint8_t *src) { |
200 | int32_t i, length; |
201 | // Support 64-bit pointers by avoiding cast of arbitrary difference. |
202 | if ((src - start) <= 7) { |
203 | i = length = (int32_t)(src - start); |
204 | } else { |
205 | i = length = 7; |
206 | start = src - 7; |
207 | } |
208 | c = utf8_prevCharSafeBody(start, 0, &i, c, -1); |
209 | i = length - i; // Number of bytes read backward from src. |
210 | int32_t idx = _UCPTRIE_CP_INDEX(trie, 0xffff, c); |
211 | return (idx << 3) | i; |
212 | } |
213 | |
214 | namespace { |
215 | |
216 | inline uint32_t getValue(UCPTrieData data, UCPTrieValueWidth valueWidth, int32_t dataIndex) { |
217 | switch (valueWidth) { |
218 | case UCPTRIE_VALUE_BITS_16: |
219 | return data.ptr16[dataIndex]; |
220 | case UCPTRIE_VALUE_BITS_32: |
221 | return data.ptr32[dataIndex]; |
222 | case UCPTRIE_VALUE_BITS_8: |
223 | return data.ptr8[dataIndex]; |
224 | default: |
225 | // Unreachable if the trie is properly initialized. |
226 | return 0xffffffff; |
227 | } |
228 | } |
229 | |
230 | } // namespace |
231 | |
232 | U_CAPI uint32_t U_EXPORT2 |
233 | ucptrie_get(const UCPTrie *trie, UChar32 c) { |
234 | int32_t dataIndex; |
235 | if ((uint32_t)c <= 0x7f) { |
236 | // linear ASCII |
237 | dataIndex = c; |
238 | } else { |
239 | UChar32 fastMax = trie->type == UCPTRIE_TYPE_FAST ? 0xffff : UCPTRIE_SMALL_MAX; |
240 | dataIndex = _UCPTRIE_CP_INDEX(trie, fastMax, c); |
241 | } |
242 | return getValue(trie->data, (UCPTrieValueWidth)trie->valueWidth, dataIndex); |
243 | } |
244 | |
245 | namespace { |
246 | |
247 | constexpr int32_t MAX_UNICODE = 0x10ffff; |
248 | |
249 | inline uint32_t maybeFilterValue(uint32_t value, uint32_t trieNullValue, uint32_t nullValue, |
250 | UCPMapValueFilter *filter, const void *context) { |
251 | if (value == trieNullValue) { |
252 | value = nullValue; |
253 | } else if (filter != nullptr) { |
254 | value = filter(context, value); |
255 | } |
256 | return value; |
257 | } |
258 | |
259 | UChar32 getRange(const void *t, UChar32 start, |
260 | UCPMapValueFilter *filter, const void *context, uint32_t *pValue) { |
261 | if ((uint32_t)start > MAX_UNICODE) { |
262 | return U_SENTINEL; |
263 | } |
264 | const UCPTrie *trie = reinterpret_cast<const UCPTrie *>(t); |
265 | UCPTrieValueWidth valueWidth = (UCPTrieValueWidth)trie->valueWidth; |
266 | if (start >= trie->highStart) { |
267 | if (pValue != nullptr) { |
268 | int32_t di = trie->dataLength - UCPTRIE_HIGH_VALUE_NEG_DATA_OFFSET; |
269 | uint32_t value = getValue(trie->data, valueWidth, di); |
270 | if (filter != nullptr) { value = filter(context, value); } |
271 | *pValue = value; |
272 | } |
273 | return MAX_UNICODE; |
274 | } |
275 | |
276 | uint32_t nullValue = trie->nullValue; |
277 | if (filter != nullptr) { nullValue = filter(context, nullValue); } |
278 | const uint16_t *index = trie->index; |
279 | |
280 | int32_t prevI3Block = -1; |
281 | int32_t prevBlock = -1; |
282 | UChar32 c = start; |
283 | uint32_t trieValue, value = nullValue; |
284 | bool haveValue = false; |
285 | do { |
286 | int32_t i3Block; |
287 | int32_t i3; |
288 | int32_t i3BlockLength; |
289 | int32_t dataBlockLength; |
290 | if (c <= 0xffff && (trie->type == UCPTRIE_TYPE_FAST || c <= UCPTRIE_SMALL_MAX)) { |
291 | i3Block = 0; |
292 | i3 = c >> UCPTRIE_FAST_SHIFT; |
293 | i3BlockLength = trie->type == UCPTRIE_TYPE_FAST ? |
294 | UCPTRIE_BMP_INDEX_LENGTH : UCPTRIE_SMALL_INDEX_LENGTH; |
295 | dataBlockLength = UCPTRIE_FAST_DATA_BLOCK_LENGTH; |
296 | } else { |
297 | // Use the multi-stage index. |
298 | int32_t i1 = c >> UCPTRIE_SHIFT_1; |
299 | if (trie->type == UCPTRIE_TYPE_FAST) { |
300 | U_ASSERT(0xffff < c && c < trie->highStart); |
301 | i1 += UCPTRIE_BMP_INDEX_LENGTH - UCPTRIE_OMITTED_BMP_INDEX_1_LENGTH; |
302 | } else { |
303 | U_ASSERT(c < trie->highStart && trie->highStart > UCPTRIE_SMALL_LIMIT); |
304 | i1 += UCPTRIE_SMALL_INDEX_LENGTH; |
305 | } |
306 | i3Block = trie->index[ |
307 | (int32_t)trie->index[i1] + ((c >> UCPTRIE_SHIFT_2) & UCPTRIE_INDEX_2_MASK)]; |
308 | if (i3Block == prevI3Block && (c - start) >= UCPTRIE_CP_PER_INDEX_2_ENTRY) { |
309 | // The index-3 block is the same as the previous one, and filled with value. |
310 | U_ASSERT((c & (UCPTRIE_CP_PER_INDEX_2_ENTRY - 1)) == 0); |
311 | c += UCPTRIE_CP_PER_INDEX_2_ENTRY; |
312 | continue; |
313 | } |
314 | prevI3Block = i3Block; |
315 | if (i3Block == trie->index3NullOffset) { |
316 | // This is the index-3 null block. |
317 | if (haveValue) { |
318 | if (nullValue != value) { |
319 | return c - 1; |
320 | } |
321 | } else { |
322 | trieValue = trie->nullValue; |
323 | value = nullValue; |
324 | if (pValue != nullptr) { *pValue = nullValue; } |
325 | haveValue = true; |
326 | } |
327 | prevBlock = trie->dataNullOffset; |
328 | c = (c + UCPTRIE_CP_PER_INDEX_2_ENTRY) & ~(UCPTRIE_CP_PER_INDEX_2_ENTRY - 1); |
329 | continue; |
330 | } |
331 | i3 = (c >> UCPTRIE_SHIFT_3) & UCPTRIE_INDEX_3_MASK; |
332 | i3BlockLength = UCPTRIE_INDEX_3_BLOCK_LENGTH; |
333 | dataBlockLength = UCPTRIE_SMALL_DATA_BLOCK_LENGTH; |
334 | } |
335 | // Enumerate data blocks for one index-3 block. |
336 | do { |
337 | int32_t block; |
338 | if ((i3Block & 0x8000) == 0) { |
339 | block = index[i3Block + i3]; |
340 | } else { |
341 | // 18-bit indexes stored in groups of 9 entries per 8 indexes. |
342 | int32_t group = (i3Block & 0x7fff) + (i3 & ~7) + (i3 >> 3); |
343 | int32_t gi = i3 & 7; |
344 | block = ((int32_t)index[group++] << (2 + (2 * gi))) & 0x30000; |
345 | block |= index[group + gi]; |
346 | } |
347 | if (block == prevBlock && (c - start) >= dataBlockLength) { |
348 | // The block is the same as the previous one, and filled with value. |
349 | U_ASSERT((c & (dataBlockLength - 1)) == 0); |
350 | c += dataBlockLength; |
351 | } else { |
352 | int32_t dataMask = dataBlockLength - 1; |
353 | prevBlock = block; |
354 | if (block == trie->dataNullOffset) { |
355 | // This is the data null block. |
356 | if (haveValue) { |
357 | if (nullValue != value) { |
358 | return c - 1; |
359 | } |
360 | } else { |
361 | trieValue = trie->nullValue; |
362 | value = nullValue; |
363 | if (pValue != nullptr) { *pValue = nullValue; } |
364 | haveValue = true; |
365 | } |
366 | c = (c + dataBlockLength) & ~dataMask; |
367 | } else { |
368 | int32_t di = block + (c & dataMask); |
369 | uint32_t trieValue2 = getValue(trie->data, valueWidth, di); |
370 | if (haveValue) { |
371 | if (trieValue2 != trieValue) { |
372 | if (filter == nullptr || |
373 | maybeFilterValue(trieValue2, trie->nullValue, nullValue, |
374 | filter, context) != value) { |
375 | return c - 1; |
376 | } |
377 | trieValue = trieValue2; // may or may not help |
378 | } |
379 | } else { |
380 | trieValue = trieValue2; |
381 | value = maybeFilterValue(trieValue2, trie->nullValue, nullValue, |
382 | filter, context); |
383 | if (pValue != nullptr) { *pValue = value; } |
384 | haveValue = true; |
385 | } |
386 | while ((++c & dataMask) != 0) { |
387 | trieValue2 = getValue(trie->data, valueWidth, ++di); |
388 | if (trieValue2 != trieValue) { |
389 | if (filter == nullptr || |
390 | maybeFilterValue(trieValue2, trie->nullValue, nullValue, |
391 | filter, context) != value) { |
392 | return c - 1; |
393 | } |
394 | trieValue = trieValue2; // may or may not help |
395 | } |
396 | } |
397 | } |
398 | } |
399 | } while (++i3 < i3BlockLength); |
400 | } while (c < trie->highStart); |
401 | U_ASSERT(haveValue); |
402 | int32_t di = trie->dataLength - UCPTRIE_HIGH_VALUE_NEG_DATA_OFFSET; |
403 | uint32_t highValue = getValue(trie->data, valueWidth, di); |
404 | if (maybeFilterValue(highValue, trie->nullValue, nullValue, |
405 | filter, context) != value) { |
406 | return c - 1; |
407 | } else { |
408 | return MAX_UNICODE; |
409 | } |
410 | } |
411 | |
412 | } // namespace |
413 | |
414 | U_CFUNC UChar32 |
415 | ucptrie_internalGetRange(UCPTrieGetRange *getRange, |
416 | const void *trie, UChar32 start, |
417 | UCPMapRangeOption option, uint32_t surrogateValue, |
418 | UCPMapValueFilter *filter, const void *context, uint32_t *pValue) { |
419 | if (option == UCPMAP_RANGE_NORMAL) { |
420 | return getRange(trie, start, filter, context, pValue); |
421 | } |
422 | uint32_t value; |
423 | if (pValue == nullptr) { |
424 | // We need to examine the range value even if the caller does not want it. |
425 | pValue = &value; |
426 | } |
427 | UChar32 surrEnd = option == UCPMAP_RANGE_FIXED_ALL_SURROGATES ? 0xdfff : 0xdbff; |
428 | UChar32 end = getRange(trie, start, filter, context, pValue); |
429 | if (end < 0xd7ff || start > surrEnd) { |
430 | return end; |
431 | } |
432 | // The range overlaps with surrogates, or ends just before the first one. |
433 | if (*pValue == surrogateValue) { |
434 | if (end >= surrEnd) { |
435 | // Surrogates followed by a non-surrogateValue range, |
436 | // or surrogates are part of a larger surrogateValue range. |
437 | return end; |
438 | } |
439 | } else { |
440 | if (start <= 0xd7ff) { |
441 | return 0xd7ff; // Non-surrogateValue range ends before surrogateValue surrogates. |
442 | } |
443 | // Start is a surrogate with a non-surrogateValue code *unit* value. |
444 | // Return a surrogateValue code *point* range. |
445 | *pValue = surrogateValue; |
446 | if (end > surrEnd) { |
447 | return surrEnd; // Surrogate range ends before non-surrogateValue rest of range. |
448 | } |
449 | } |
450 | // See if the surrogateValue surrogate range can be merged with |
451 | // an immediately following range. |
452 | uint32_t value2; |
453 | UChar32 end2 = getRange(trie, surrEnd + 1, filter, context, &value2); |
454 | if (value2 == surrogateValue) { |
455 | return end2; |
456 | } |
457 | return surrEnd; |
458 | } |
459 | |
460 | U_CAPI UChar32 U_EXPORT2 |
461 | ucptrie_getRange(const UCPTrie *trie, UChar32 start, |
462 | UCPMapRangeOption option, uint32_t surrogateValue, |
463 | UCPMapValueFilter *filter, const void *context, uint32_t *pValue) { |
464 | return ucptrie_internalGetRange(getRange, trie, start, |
465 | option, surrogateValue, |
466 | filter, context, pValue); |
467 | } |
468 | |
469 | U_CAPI int32_t U_EXPORT2 |
470 | ucptrie_toBinary(const UCPTrie *trie, |
471 | void *data, int32_t capacity, |
472 | UErrorCode *pErrorCode) { |
473 | if (U_FAILURE(*pErrorCode)) { |
474 | return 0; |
475 | } |
476 | |
477 | UCPTrieType type = (UCPTrieType)trie->type; |
478 | UCPTrieValueWidth valueWidth = (UCPTrieValueWidth)trie->valueWidth; |
479 | if (type < UCPTRIE_TYPE_FAST || UCPTRIE_TYPE_SMALL < type || |
480 | valueWidth < UCPTRIE_VALUE_BITS_16 || UCPTRIE_VALUE_BITS_8 < valueWidth || |
481 | capacity < 0 || |
482 | (capacity > 0 && (data == nullptr || (U_POINTER_MASK_LSB(data, 3) != 0)))) { |
483 | *pErrorCode = U_ILLEGAL_ARGUMENT_ERROR; |
484 | return 0; |
485 | } |
486 | |
487 | int32_t length = (int32_t)sizeof(UCPTrieHeader) + trie->indexLength * 2; |
488 | switch (valueWidth) { |
489 | case UCPTRIE_VALUE_BITS_16: |
490 | length += trie->dataLength * 2; |
491 | break; |
492 | case UCPTRIE_VALUE_BITS_32: |
493 | length += trie->dataLength * 4; |
494 | break; |
495 | case UCPTRIE_VALUE_BITS_8: |
496 | length += trie->dataLength; |
497 | break; |
498 | default: |
499 | // unreachable |
500 | break; |
501 | } |
502 | if (capacity < length) { |
503 | *pErrorCode = U_BUFFER_OVERFLOW_ERROR; |
504 | return length; |
505 | } |
506 | |
507 | char *bytes = (char *)data; |
508 | UCPTrieHeader * = (UCPTrieHeader *)bytes; |
509 | header->signature = UCPTRIE_SIG; // "Tri3" |
510 | header->options = (uint16_t)( |
511 | ((trie->dataLength & 0xf0000) >> 4) | |
512 | ((trie->dataNullOffset & 0xf0000) >> 8) | |
513 | (trie->type << 6) | |
514 | valueWidth); |
515 | header->indexLength = (uint16_t)trie->indexLength; |
516 | header->dataLength = (uint16_t)trie->dataLength; |
517 | header->index3NullOffset = trie->index3NullOffset; |
518 | header->dataNullOffset = (uint16_t)trie->dataNullOffset; |
519 | header->shiftedHighStart = trie->highStart >> UCPTRIE_SHIFT_2; |
520 | bytes += sizeof(UCPTrieHeader); |
521 | |
522 | uprv_memcpy(bytes, trie->index, trie->indexLength * 2); |
523 | bytes += trie->indexLength * 2; |
524 | |
525 | switch (valueWidth) { |
526 | case UCPTRIE_VALUE_BITS_16: |
527 | uprv_memcpy(bytes, trie->data.ptr16, trie->dataLength * 2); |
528 | break; |
529 | case UCPTRIE_VALUE_BITS_32: |
530 | uprv_memcpy(bytes, trie->data.ptr32, trie->dataLength * 4); |
531 | break; |
532 | case UCPTRIE_VALUE_BITS_8: |
533 | uprv_memcpy(bytes, trie->data.ptr8, trie->dataLength); |
534 | break; |
535 | default: |
536 | // unreachable |
537 | break; |
538 | } |
539 | return length; |
540 | } |
541 | |
542 | namespace { |
543 | |
544 | #ifdef UCPTRIE_DEBUG |
545 | long countNull(const UCPTrie *trie) { |
546 | uint32_t nullValue=trie->nullValue; |
547 | int32_t length=trie->dataLength; |
548 | long count=0; |
549 | switch (trie->valueWidth) { |
550 | case UCPTRIE_VALUE_BITS_16: |
551 | for(int32_t i=0; i<length; ++i) { |
552 | if(trie->data.ptr16[i]==nullValue) { ++count; } |
553 | } |
554 | break; |
555 | case UCPTRIE_VALUE_BITS_32: |
556 | for(int32_t i=0; i<length; ++i) { |
557 | if(trie->data.ptr32[i]==nullValue) { ++count; } |
558 | } |
559 | break; |
560 | case UCPTRIE_VALUE_BITS_8: |
561 | for(int32_t i=0; i<length; ++i) { |
562 | if(trie->data.ptr8[i]==nullValue) { ++count; } |
563 | } |
564 | break; |
565 | default: |
566 | // unreachable |
567 | break; |
568 | } |
569 | return count; |
570 | } |
571 | |
572 | U_CFUNC void |
573 | ucptrie_printLengths(const UCPTrie *trie, const char *which) { |
574 | long indexLength=trie->indexLength; |
575 | long dataLength=(long)trie->dataLength; |
576 | long totalLength=(long)sizeof(UCPTrieHeader)+indexLength*2+ |
577 | dataLength*(trie->valueWidth==UCPTRIE_VALUE_BITS_16 ? 2 : |
578 | trie->valueWidth==UCPTRIE_VALUE_BITS_32 ? 4 : 1); |
579 | printf("**UCPTrieLengths(%s %s)** index:%6ld data:%6ld countNull:%6ld serialized:%6ld\n" , |
580 | which, trie->name, indexLength, dataLength, countNull(trie), totalLength); |
581 | } |
582 | #endif |
583 | |
584 | } // namespace |
585 | |
586 | // UCPMap ---- |
587 | // Initially, this is the same as UCPTrie. This may well change. |
588 | |
589 | U_CAPI uint32_t U_EXPORT2 |
590 | ucpmap_get(const UCPMap *map, UChar32 c) { |
591 | return ucptrie_get(reinterpret_cast<const UCPTrie *>(map), c); |
592 | } |
593 | |
594 | U_CAPI UChar32 U_EXPORT2 |
595 | ucpmap_getRange(const UCPMap *map, UChar32 start, |
596 | UCPMapRangeOption option, uint32_t surrogateValue, |
597 | UCPMapValueFilter *filter, const void *context, uint32_t *pValue) { |
598 | return ucptrie_getRange(reinterpret_cast<const UCPTrie *>(map), start, |
599 | option, surrogateValue, |
600 | filter, context, pValue); |
601 | } |
602 | |