1 /*
2 *******************************************************************************
3 *
4 * Copyright (C) 2005-2014, International Business Machines
5 * Corporation and others. All Rights Reserved.
6 *
7 *******************************************************************************
8 * file name: swapimpl.cpp
9 * encoding: US-ASCII
10 * tab size: 8 (not used)
11 * indentation:4
12 *
13 * created on: 2005may05
14 * created by: Markus W. Scherer
15 *
16 * Data file swapping functions moved here from the common library
17 * because some data is hardcoded in ICU4C and needs not be swapped any more.
18 * Moving the functions here simplifies testing (for code coverage) because
19 * we need not jump through hoops (like adding snapshots of these files
20 * to testdata).
21 *
22 * The declarations for these functions remain in the internal header files
23 * in icu/source/common/
24 */
25
26 #include "unicode/utypes.h"
27 #include "unicode/putil.h"
28 #include "unicode/udata.h"
29
30 /* Explicit include statement for std_string.h is needed
31 * for compilation on certain platforms. (e.g. AIX/VACPP)
32 */
33 #include "unicode/std_string.h"
34
35 #include "cmemory.h"
36 #include "cstring.h"
37 #include "uinvchar.h"
38 #include "uassert.h"
39 #include "uarrsort.h"
40 #include "ucmndata.h"
41 #include "udataswp.h"
42
43 /* swapping implementations in common */
44
45 #include "uresdata.h"
46 #include "ucnv_io.h"
47 #include "uprops.h"
48 #include "ucase.h"
49 #include "ubidi_props.h"
50 #include "ucol_swp.h"
51 #include "ucnv_bld.h"
52 #include "unormimp.h"
53 #include "normalizer2impl.h"
54 #include "sprpimpl.h"
55 #include "propname.h"
56 #include "rbbidata.h"
57 #include "utrie2.h"
58 #include "dictionarydata.h"
59
60 /* swapping implementations in i18n */
61
62 #if !UCONFIG_NO_NORMALIZATION
63 #include "uspoof_impl.h"
64 #endif
65
66 U_NAMESPACE_USE
67
68 /* definitions */
69
70 /* Unicode property (value) aliases data swapping --------------------------- */
71
72 static int32_t U_CALLCONV
upname_swap(const UDataSwapper * ds,const void * inData,int32_t length,void * outData,UErrorCode * pErrorCode)73 upname_swap(const UDataSwapper *ds,
74 const void *inData, int32_t length, void *outData,
75 UErrorCode *pErrorCode) {
76 /* udata_swapDataHeader checks the arguments */
77 int32_t headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode);
78 if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
79 return 0;
80 }
81
82 /* check data format and format version */
83 const UDataInfo *pInfo=
84 reinterpret_cast<const UDataInfo *>(
85 static_cast<const char *>(inData)+4);
86 if(!(
87 pInfo->dataFormat[0]==0x70 && /* dataFormat="pnam" */
88 pInfo->dataFormat[1]==0x6e &&
89 pInfo->dataFormat[2]==0x61 &&
90 pInfo->dataFormat[3]==0x6d &&
91 pInfo->formatVersion[0]==2
92 )) {
93 udata_printError(ds, "upname_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as pnames.icu\n",
94 pInfo->dataFormat[0], pInfo->dataFormat[1],
95 pInfo->dataFormat[2], pInfo->dataFormat[3],
96 pInfo->formatVersion[0]);
97 *pErrorCode=U_UNSUPPORTED_ERROR;
98 return 0;
99 }
100
101 const uint8_t *inBytes=static_cast<const uint8_t *>(inData)+headerSize;
102 uint8_t *outBytes=static_cast<uint8_t *>(outData)+headerSize;
103
104 if(length>=0) {
105 length-=headerSize;
106 // formatVersion 2 initially has indexes[8], 32 bytes.
107 if(length<32) {
108 udata_printError(ds, "upname_swap(): too few bytes (%d after header) for pnames.icu\n",
109 (int)length);
110 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
111 return 0;
112 }
113 }
114
115 const int32_t *inIndexes=reinterpret_cast<const int32_t *>(inBytes);
116 int32_t totalSize=udata_readInt32(ds, inIndexes[PropNameData::IX_TOTAL_SIZE]);
117 if(length>=0) {
118 if(length<totalSize) {
119 udata_printError(ds, "upname_swap(): too few bytes (%d after header, should be %d) "
120 "for pnames.icu\n",
121 (int)length, (int)totalSize);
122 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
123 return 0;
124 }
125
126 int32_t numBytesIndexesAndValueMaps=
127 udata_readInt32(ds, inIndexes[PropNameData::IX_BYTE_TRIES_OFFSET]);
128
129 // Swap the indexes[] and the valueMaps[].
130 ds->swapArray32(ds, inBytes, numBytesIndexesAndValueMaps, outBytes, pErrorCode);
131
132 // Copy the rest of the data.
133 if(inBytes!=outBytes) {
134 uprv_memcpy(outBytes+numBytesIndexesAndValueMaps,
135 inBytes+numBytesIndexesAndValueMaps,
136 totalSize-numBytesIndexesAndValueMaps);
137 }
138
139 // We need not swap anything else:
140 //
141 // The ByteTries are already byte-serialized, and are fixed on ASCII.
142 // (On an EBCDIC machine, the input string is converted to lowercase ASCII
143 // while matching.)
144 //
145 // The name groups are mostly invariant characters, but since we only
146 // generate, and keep in subversion, ASCII versions of pnames.icu,
147 // and since only ICU4J uses the pnames.icu data file
148 // (the data is hardcoded in ICU4C) and ICU4J uses ASCII data files,
149 // we just copy those bytes too.
150 }
151
152 return headerSize+totalSize;
153 }
154
155 /* Unicode properties data swapping ----------------------------------------- */
156
157 static int32_t U_CALLCONV
uprops_swap(const UDataSwapper * ds,const void * inData,int32_t length,void * outData,UErrorCode * pErrorCode)158 uprops_swap(const UDataSwapper *ds,
159 const void *inData, int32_t length, void *outData,
160 UErrorCode *pErrorCode) {
161 const UDataInfo *pInfo;
162 int32_t headerSize, i;
163
164 int32_t dataIndexes[UPROPS_INDEX_COUNT];
165 const int32_t *inData32;
166
167 /* udata_swapDataHeader checks the arguments */
168 headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode);
169 if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
170 return 0;
171 }
172
173 /* check data format and format version */
174 pInfo=(const UDataInfo *)((const char *)inData+4);
175 if(!(
176 pInfo->dataFormat[0]==0x55 && /* dataFormat="UPro" */
177 pInfo->dataFormat[1]==0x50 &&
178 pInfo->dataFormat[2]==0x72 &&
179 pInfo->dataFormat[3]==0x6f &&
180 (3<=pInfo->formatVersion[0] && pInfo->formatVersion[0]<=7) &&
181 (pInfo->formatVersion[0]>=7 ||
182 (pInfo->formatVersion[2]==UTRIE_SHIFT &&
183 pInfo->formatVersion[3]==UTRIE_INDEX_SHIFT))
184 )) {
185 udata_printError(ds, "uprops_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not a Unicode properties file\n",
186 pInfo->dataFormat[0], pInfo->dataFormat[1],
187 pInfo->dataFormat[2], pInfo->dataFormat[3],
188 pInfo->formatVersion[0]);
189 *pErrorCode=U_UNSUPPORTED_ERROR;
190 return 0;
191 }
192
193 /* the properties file must contain at least the indexes array */
194 if(length>=0 && (length-headerSize)<(int32_t)sizeof(dataIndexes)) {
195 udata_printError(ds, "uprops_swap(): too few bytes (%d after header) for a Unicode properties file\n",
196 length-headerSize);
197 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
198 return 0;
199 }
200
201 /* read the indexes */
202 inData32=(const int32_t *)((const char *)inData+headerSize);
203 for(i=0; i<UPROPS_INDEX_COUNT; ++i) {
204 dataIndexes[i]=udata_readInt32(ds, inData32[i]);
205 }
206
207 /*
208 * comments are copied from the data format description in genprops/store.c
209 * indexes[] constants are in uprops.h
210 */
211 int32_t dataTop;
212 if(length>=0) {
213 int32_t *outData32;
214
215 /*
216 * In formatVersion 7, UPROPS_DATA_TOP_INDEX has the post-header data size.
217 * In earlier formatVersions, it is 0 and a lower dataIndexes entry
218 * has the top of the last item.
219 */
220 for(i=UPROPS_DATA_TOP_INDEX; i>0 && (dataTop=dataIndexes[i])==0; --i) {}
221
222 if((length-headerSize)<(4*dataTop)) {
223 udata_printError(ds, "uprops_swap(): too few bytes (%d after header) for a Unicode properties file\n",
224 length-headerSize);
225 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
226 return 0;
227 }
228
229 outData32=(int32_t *)((char *)outData+headerSize);
230
231 /* copy everything for inaccessible data (padding) */
232 if(inData32!=outData32) {
233 uprv_memcpy(outData32, inData32, 4*dataTop);
234 }
235
236 /* swap the indexes[16] */
237 ds->swapArray32(ds, inData32, 4*UPROPS_INDEX_COUNT, outData32, pErrorCode);
238
239 /*
240 * swap the main properties UTrie
241 * PT serialized properties trie, see utrie.h (byte size: 4*(i0-16))
242 */
243 utrie2_swapAnyVersion(ds,
244 inData32+UPROPS_INDEX_COUNT,
245 4*(dataIndexes[UPROPS_PROPS32_INDEX]-UPROPS_INDEX_COUNT),
246 outData32+UPROPS_INDEX_COUNT,
247 pErrorCode);
248
249 /*
250 * swap the properties and exceptions words
251 * P const uint32_t props32[i1-i0];
252 * E const uint32_t exceptions[i2-i1];
253 */
254 ds->swapArray32(ds,
255 inData32+dataIndexes[UPROPS_PROPS32_INDEX],
256 4*(dataIndexes[UPROPS_EXCEPTIONS_TOP_INDEX]-dataIndexes[UPROPS_PROPS32_INDEX]),
257 outData32+dataIndexes[UPROPS_PROPS32_INDEX],
258 pErrorCode);
259
260 /*
261 * swap the UChars
262 * U const UChar uchars[2*(i3-i2)];
263 */
264 ds->swapArray16(ds,
265 inData32+dataIndexes[UPROPS_EXCEPTIONS_TOP_INDEX],
266 4*(dataIndexes[UPROPS_ADDITIONAL_TRIE_INDEX]-dataIndexes[UPROPS_EXCEPTIONS_TOP_INDEX]),
267 outData32+dataIndexes[UPROPS_EXCEPTIONS_TOP_INDEX],
268 pErrorCode);
269
270 /*
271 * swap the additional UTrie
272 * i3 additionalTrieIndex; -- 32-bit unit index to the additional trie for more properties
273 */
274 utrie2_swapAnyVersion(ds,
275 inData32+dataIndexes[UPROPS_ADDITIONAL_TRIE_INDEX],
276 4*(dataIndexes[UPROPS_ADDITIONAL_VECTORS_INDEX]-dataIndexes[UPROPS_ADDITIONAL_TRIE_INDEX]),
277 outData32+dataIndexes[UPROPS_ADDITIONAL_TRIE_INDEX],
278 pErrorCode);
279
280 /*
281 * swap the properties vectors
282 * PV const uint32_t propsVectors[(i6-i4)/i5][i5]==uint32_t propsVectors[i6-i4];
283 */
284 ds->swapArray32(ds,
285 inData32+dataIndexes[UPROPS_ADDITIONAL_VECTORS_INDEX],
286 4*(dataIndexes[UPROPS_SCRIPT_EXTENSIONS_INDEX]-dataIndexes[UPROPS_ADDITIONAL_VECTORS_INDEX]),
287 outData32+dataIndexes[UPROPS_ADDITIONAL_VECTORS_INDEX],
288 pErrorCode);
289
290 // swap the Script_Extensions data
291 // SCX const uint16_t scriptExtensions[2*(i7-i6)];
292 ds->swapArray16(ds,
293 inData32+dataIndexes[UPROPS_SCRIPT_EXTENSIONS_INDEX],
294 4*(dataIndexes[UPROPS_RESERVED_INDEX_7]-dataIndexes[UPROPS_SCRIPT_EXTENSIONS_INDEX]),
295 outData32+dataIndexes[UPROPS_SCRIPT_EXTENSIONS_INDEX],
296 pErrorCode);
297 }
298
299 /* i7 reservedIndex7; -- 32-bit unit index to the top of the Script_Extensions data */
300 return headerSize+4*dataIndexes[UPROPS_RESERVED_INDEX_7];
301 }
302
303 /* Unicode case mapping data swapping --------------------------------------- */
304
305 static int32_t U_CALLCONV
ucase_swap(const UDataSwapper * ds,const void * inData,int32_t length,void * outData,UErrorCode * pErrorCode)306 ucase_swap(const UDataSwapper *ds,
307 const void *inData, int32_t length, void *outData,
308 UErrorCode *pErrorCode) {
309 const UDataInfo *pInfo;
310 int32_t headerSize;
311
312 const uint8_t *inBytes;
313 uint8_t *outBytes;
314
315 const int32_t *inIndexes;
316 int32_t indexes[16];
317
318 int32_t i, offset, count, size;
319
320 /* udata_swapDataHeader checks the arguments */
321 headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode);
322 if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
323 return 0;
324 }
325
326 /* check data format and format version */
327 pInfo=(const UDataInfo *)((const char *)inData+4);
328 if(!(
329 pInfo->dataFormat[0]==UCASE_FMT_0 && /* dataFormat="cAsE" */
330 pInfo->dataFormat[1]==UCASE_FMT_1 &&
331 pInfo->dataFormat[2]==UCASE_FMT_2 &&
332 pInfo->dataFormat[3]==UCASE_FMT_3 &&
333 ((pInfo->formatVersion[0]==1 &&
334 pInfo->formatVersion[2]==UTRIE_SHIFT &&
335 pInfo->formatVersion[3]==UTRIE_INDEX_SHIFT) ||
336 pInfo->formatVersion[0]==2 || pInfo->formatVersion[0]==3)
337 )) {
338 udata_printError(ds, "ucase_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as case mapping data\n",
339 pInfo->dataFormat[0], pInfo->dataFormat[1],
340 pInfo->dataFormat[2], pInfo->dataFormat[3],
341 pInfo->formatVersion[0]);
342 *pErrorCode=U_UNSUPPORTED_ERROR;
343 return 0;
344 }
345
346 inBytes=(const uint8_t *)inData+headerSize;
347 outBytes=(uint8_t *)outData+headerSize;
348
349 inIndexes=(const int32_t *)inBytes;
350
351 if(length>=0) {
352 length-=headerSize;
353 if(length<16*4) {
354 udata_printError(ds, "ucase_swap(): too few bytes (%d after header) for case mapping data\n",
355 length);
356 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
357 return 0;
358 }
359 }
360
361 /* read the first 16 indexes (ICU 3.2/format version 1: UCASE_IX_TOP==16, might grow) */
362 for(i=0; i<16; ++i) {
363 indexes[i]=udata_readInt32(ds, inIndexes[i]);
364 }
365
366 /* get the total length of the data */
367 size=indexes[UCASE_IX_LENGTH];
368
369 if(length>=0) {
370 if(length<size) {
371 udata_printError(ds, "ucase_swap(): too few bytes (%d after header) for all of case mapping data\n",
372 length);
373 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
374 return 0;
375 }
376
377 /* copy the data for inaccessible bytes */
378 if(inBytes!=outBytes) {
379 uprv_memcpy(outBytes, inBytes, size);
380 }
381
382 offset=0;
383
384 /* swap the int32_t indexes[] */
385 count=indexes[UCASE_IX_INDEX_TOP]*4;
386 ds->swapArray32(ds, inBytes, count, outBytes, pErrorCode);
387 offset+=count;
388
389 /* swap the UTrie */
390 count=indexes[UCASE_IX_TRIE_SIZE];
391 utrie2_swapAnyVersion(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
392 offset+=count;
393
394 /* swap the uint16_t exceptions[] and unfold[] */
395 count=(indexes[UCASE_IX_EXC_LENGTH]+indexes[UCASE_IX_UNFOLD_LENGTH])*2;
396 ds->swapArray16(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
397 offset+=count;
398
399 U_ASSERT(offset==size);
400 }
401
402 return headerSize+size;
403 }
404
405 /* Unicode bidi/shaping data swapping --------------------------------------- */
406
407 static int32_t U_CALLCONV
ubidi_swap(const UDataSwapper * ds,const void * inData,int32_t length,void * outData,UErrorCode * pErrorCode)408 ubidi_swap(const UDataSwapper *ds,
409 const void *inData, int32_t length, void *outData,
410 UErrorCode *pErrorCode) {
411 const UDataInfo *pInfo;
412 int32_t headerSize;
413
414 const uint8_t *inBytes;
415 uint8_t *outBytes;
416
417 const int32_t *inIndexes;
418 int32_t indexes[16];
419
420 int32_t i, offset, count, size;
421
422 /* udata_swapDataHeader checks the arguments */
423 headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode);
424 if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
425 return 0;
426 }
427
428 /* check data format and format version */
429 pInfo=(const UDataInfo *)((const char *)inData+4);
430 if(!(
431 pInfo->dataFormat[0]==UBIDI_FMT_0 && /* dataFormat="BiDi" */
432 pInfo->dataFormat[1]==UBIDI_FMT_1 &&
433 pInfo->dataFormat[2]==UBIDI_FMT_2 &&
434 pInfo->dataFormat[3]==UBIDI_FMT_3 &&
435 ((pInfo->formatVersion[0]==1 &&
436 pInfo->formatVersion[2]==UTRIE_SHIFT &&
437 pInfo->formatVersion[3]==UTRIE_INDEX_SHIFT) ||
438 pInfo->formatVersion[0]==2)
439 )) {
440 udata_printError(ds, "ubidi_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as bidi/shaping data\n",
441 pInfo->dataFormat[0], pInfo->dataFormat[1],
442 pInfo->dataFormat[2], pInfo->dataFormat[3],
443 pInfo->formatVersion[0]);
444 *pErrorCode=U_UNSUPPORTED_ERROR;
445 return 0;
446 }
447
448 inBytes=(const uint8_t *)inData+headerSize;
449 outBytes=(uint8_t *)outData+headerSize;
450
451 inIndexes=(const int32_t *)inBytes;
452
453 if(length>=0) {
454 length-=headerSize;
455 if(length<16*4) {
456 udata_printError(ds, "ubidi_swap(): too few bytes (%d after header) for bidi/shaping data\n",
457 length);
458 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
459 return 0;
460 }
461 }
462
463 /* read the first 16 indexes (ICU 3.4/format version 1: UBIDI_IX_TOP==16, might grow) */
464 for(i=0; i<16; ++i) {
465 indexes[i]=udata_readInt32(ds, inIndexes[i]);
466 }
467
468 /* get the total length of the data */
469 size=indexes[UBIDI_IX_LENGTH];
470
471 if(length>=0) {
472 if(length<size) {
473 udata_printError(ds, "ubidi_swap(): too few bytes (%d after header) for all of bidi/shaping data\n",
474 length);
475 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
476 return 0;
477 }
478
479 /* copy the data for inaccessible bytes */
480 if(inBytes!=outBytes) {
481 uprv_memcpy(outBytes, inBytes, size);
482 }
483
484 offset=0;
485
486 /* swap the int32_t indexes[] */
487 count=indexes[UBIDI_IX_INDEX_TOP]*4;
488 ds->swapArray32(ds, inBytes, count, outBytes, pErrorCode);
489 offset+=count;
490
491 /* swap the UTrie */
492 count=indexes[UBIDI_IX_TRIE_SIZE];
493 utrie2_swapAnyVersion(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
494 offset+=count;
495
496 /* swap the uint32_t mirrors[] */
497 count=indexes[UBIDI_IX_MIRROR_LENGTH]*4;
498 ds->swapArray32(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
499 offset+=count;
500
501 /* just skip the uint8_t jgArray[] and jgArray2[] */
502 count=indexes[UBIDI_IX_JG_LIMIT]-indexes[UBIDI_IX_JG_START];
503 offset+=count;
504 count=indexes[UBIDI_IX_JG_LIMIT2]-indexes[UBIDI_IX_JG_START2];
505 offset+=count;
506
507 U_ASSERT(offset==size);
508 }
509
510 return headerSize+size;
511 }
512
513 /* Unicode normalization data swapping -------------------------------------- */
514
515 #if !UCONFIG_NO_NORMALIZATION
516
517 static int32_t U_CALLCONV
unorm_swap(const UDataSwapper * ds,const void * inData,int32_t length,void * outData,UErrorCode * pErrorCode)518 unorm_swap(const UDataSwapper *ds,
519 const void *inData, int32_t length, void *outData,
520 UErrorCode *pErrorCode) {
521 const UDataInfo *pInfo;
522 int32_t headerSize;
523
524 const uint8_t *inBytes;
525 uint8_t *outBytes;
526
527 const int32_t *inIndexes;
528 int32_t indexes[32];
529
530 int32_t i, offset, count, size;
531
532 /* udata_swapDataHeader checks the arguments */
533 headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode);
534 if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
535 return 0;
536 }
537
538 /* check data format and format version */
539 pInfo=(const UDataInfo *)((const char *)inData+4);
540 if(!(
541 pInfo->dataFormat[0]==0x4e && /* dataFormat="Norm" */
542 pInfo->dataFormat[1]==0x6f &&
543 pInfo->dataFormat[2]==0x72 &&
544 pInfo->dataFormat[3]==0x6d &&
545 pInfo->formatVersion[0]==2
546 )) {
547 udata_printError(ds, "unorm_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as unorm.icu\n",
548 pInfo->dataFormat[0], pInfo->dataFormat[1],
549 pInfo->dataFormat[2], pInfo->dataFormat[3],
550 pInfo->formatVersion[0]);
551 *pErrorCode=U_UNSUPPORTED_ERROR;
552 return 0;
553 }
554
555 inBytes=(const uint8_t *)inData+headerSize;
556 outBytes=(uint8_t *)outData+headerSize;
557
558 inIndexes=(const int32_t *)inBytes;
559
560 if(length>=0) {
561 length-=headerSize;
562 if(length<32*4) {
563 udata_printError(ds, "unorm_swap(): too few bytes (%d after header) for unorm.icu\n",
564 length);
565 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
566 return 0;
567 }
568 }
569
570 /* read the first 32 indexes (ICU 2.8/format version 2.2: _NORM_INDEX_TOP==32, might grow) */
571 for(i=0; i<32; ++i) {
572 indexes[i]=udata_readInt32(ds, inIndexes[i]);
573 }
574
575 /* calculate the total length of the data */
576 size=
577 32*4+ /* size of indexes[] */
578 indexes[_NORM_INDEX_TRIE_SIZE]+
579 indexes[_NORM_INDEX_UCHAR_COUNT]*2+
580 indexes[_NORM_INDEX_COMBINE_DATA_COUNT]*2+
581 indexes[_NORM_INDEX_FCD_TRIE_SIZE]+
582 indexes[_NORM_INDEX_AUX_TRIE_SIZE]+
583 indexes[_NORM_INDEX_CANON_SET_COUNT]*2;
584
585 if(length>=0) {
586 if(length<size) {
587 udata_printError(ds, "unorm_swap(): too few bytes (%d after header) for all of unorm.icu\n",
588 length);
589 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
590 return 0;
591 }
592
593 /* copy the data for inaccessible bytes */
594 if(inBytes!=outBytes) {
595 uprv_memcpy(outBytes, inBytes, size);
596 }
597
598 offset=0;
599
600 /* swap the indexes[] */
601 count=32*4;
602 ds->swapArray32(ds, inBytes, count, outBytes, pErrorCode);
603 offset+=count;
604
605 /* swap the main UTrie */
606 count=indexes[_NORM_INDEX_TRIE_SIZE];
607 utrie_swap(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
608 offset+=count;
609
610 /* swap the uint16_t extraData[] and the uint16_t combiningTable[] */
611 count=(indexes[_NORM_INDEX_UCHAR_COUNT]+indexes[_NORM_INDEX_COMBINE_DATA_COUNT])*2;
612 ds->swapArray16(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
613 offset+=count;
614
615 /* swap the FCD UTrie */
616 count=indexes[_NORM_INDEX_FCD_TRIE_SIZE];
617 if(count!=0) {
618 utrie_swap(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
619 offset+=count;
620 }
621
622 /* swap the aux UTrie */
623 count=indexes[_NORM_INDEX_AUX_TRIE_SIZE];
624 if(count!=0) {
625 utrie_swap(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
626 offset+=count;
627 }
628
629 /* swap the uint16_t combiningTable[] */
630 count=indexes[_NORM_INDEX_CANON_SET_COUNT]*2;
631 ds->swapArray16(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
632 offset+=count;
633 }
634
635 return headerSize+size;
636 }
637
638 #endif
639
640 /* Swap 'Test' data from gentest */
641 static int32_t U_CALLCONV
test_swap(const UDataSwapper * ds,const void * inData,int32_t length,void * outData,UErrorCode * pErrorCode)642 test_swap(const UDataSwapper *ds,
643 const void *inData, int32_t length, void *outData,
644 UErrorCode *pErrorCode) {
645 const UDataInfo *pInfo;
646 int32_t headerSize;
647
648 const uint8_t *inBytes;
649 uint8_t *outBytes;
650
651 int32_t offset;
652
653 /* udata_swapDataHeader checks the arguments */
654 headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode);
655 if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
656 udata_printError(ds, "test_swap(): data header swap failed %s\n", pErrorCode != NULL ? u_errorName(*pErrorCode) : "pErrorCode is NULL");
657 return 0;
658 }
659
660 /* check data format and format version */
661 pInfo=(const UDataInfo *)((const char *)inData+4);
662 if(!(
663 pInfo->dataFormat[0]==0x54 && /* dataFormat="Norm" */
664 pInfo->dataFormat[1]==0x65 &&
665 pInfo->dataFormat[2]==0x73 &&
666 pInfo->dataFormat[3]==0x74 &&
667 pInfo->formatVersion[0]==1
668 )) {
669 udata_printError(ds, "test_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as testdata\n",
670 pInfo->dataFormat[0], pInfo->dataFormat[1],
671 pInfo->dataFormat[2], pInfo->dataFormat[3],
672 pInfo->formatVersion[0]);
673 *pErrorCode=U_UNSUPPORTED_ERROR;
674 return 0;
675 }
676
677 inBytes=(const uint8_t *)inData+headerSize;
678 outBytes=(uint8_t *)outData+headerSize;
679
680 int32_t size16 = 2; // 16bit plus padding
681 int32_t sizeStr = 5; // 4 char inv-str plus null
682 int32_t size = size16 + sizeStr;
683
684 if(length>=0) {
685 if(length<size) {
686 udata_printError(ds, "test_swap(): too few bytes (%d after header, wanted %d) for all of testdata\n",
687 length, size);
688 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
689 return 0;
690 }
691
692 offset =0;
693 /* swap a 1 entry array */
694 ds->swapArray16(ds, inBytes+offset, size16, outBytes+offset, pErrorCode);
695 offset+=size16;
696 ds->swapInvChars(ds, inBytes+offset, sizeStr, outBytes+offset, pErrorCode);
697 }
698
699 return headerSize+size;
700 }
701
702 /* swap any data (except a .dat package) ------------------------------------ */
703
704 static const struct {
705 uint8_t dataFormat[4];
706 UDataSwapFn *swapFn;
707 } swapFns[]={
708 { { 0x52, 0x65, 0x73, 0x42 }, ures_swap }, /* dataFormat="ResB" */
709 #if !UCONFIG_NO_LEGACY_CONVERSION
710 { { 0x63, 0x6e, 0x76, 0x74 }, ucnv_swap }, /* dataFormat="cnvt" */
711 #endif
712 #if !UCONFIG_NO_CONVERSION
713 { { 0x43, 0x76, 0x41, 0x6c }, ucnv_swapAliases }, /* dataFormat="CvAl" */
714 #endif
715 #if !UCONFIG_NO_IDNA
716 { { 0x53, 0x50, 0x52, 0x50 }, usprep_swap }, /* dataFormat="SPRP" */
717 #endif
718 /* insert data formats here, descending by expected frequency of occurrence */
719 { { 0x55, 0x50, 0x72, 0x6f }, uprops_swap }, /* dataFormat="UPro" */
720
721 { { UCASE_FMT_0, UCASE_FMT_1, UCASE_FMT_2, UCASE_FMT_3 },
722 ucase_swap }, /* dataFormat="cAsE" */
723
724 { { UBIDI_FMT_0, UBIDI_FMT_1, UBIDI_FMT_2, UBIDI_FMT_3 },
725 ubidi_swap }, /* dataFormat="BiDi" */
726
727 #if !UCONFIG_NO_NORMALIZATION
728 { { 0x4e, 0x6f, 0x72, 0x6d }, unorm_swap }, /* dataFormat="Norm" */
729 { { 0x4e, 0x72, 0x6d, 0x32 }, unorm2_swap }, /* dataFormat="Nrm2" */
730 #endif
731 #if !UCONFIG_NO_COLLATION
732 { { 0x55, 0x43, 0x6f, 0x6c }, ucol_swap }, /* dataFormat="UCol" */
733 { { 0x49, 0x6e, 0x76, 0x43 }, ucol_swapInverseUCA },/* dataFormat="InvC" */
734 #endif
735 #if !UCONFIG_NO_BREAK_ITERATION
736 { { 0x42, 0x72, 0x6b, 0x20 }, ubrk_swap }, /* dataFormat="Brk " */
737 { { 0x44, 0x69, 0x63, 0x74 }, udict_swap }, /* dataFormat="Dict" */
738 #endif
739 { { 0x70, 0x6e, 0x61, 0x6d }, upname_swap }, /* dataFormat="pnam" */
740 { { 0x75, 0x6e, 0x61, 0x6d }, uchar_swapNames }, /* dataFormat="unam" */
741 #if !UCONFIG_NO_NORMALIZATION
742 { { 0x43, 0x66, 0x75, 0x20 }, uspoof_swap }, /* dataFormat="Cfu " */
743 #endif
744 { { 0x54, 0x65, 0x73, 0x74 }, test_swap } /* dataFormat="Test" */
745 };
746
747 U_CAPI int32_t U_EXPORT2
udata_swap(const UDataSwapper * ds,const void * inData,int32_t length,void * outData,UErrorCode * pErrorCode)748 udata_swap(const UDataSwapper *ds,
749 const void *inData, int32_t length, void *outData,
750 UErrorCode *pErrorCode) {
751 char dataFormatChars[4];
752 const UDataInfo *pInfo;
753 int32_t i, swappedLength;
754
755 if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
756 return 0;
757 }
758
759 /*
760 * Preflight the header first; checks for illegal arguments, too.
761 * Do not swap the header right away because the format-specific swapper
762 * will swap it, get the headerSize again, and also use the header
763 * information. Otherwise we would have to pass some of the information
764 * and not be able to use the UDataSwapFn signature.
765 */
766 udata_swapDataHeader(ds, inData, -1, NULL, pErrorCode);
767
768 /*
769 * If we wanted udata_swap() to also handle non-loadable data like a UTrie,
770 * then we could check here for further known magic values and structures.
771 */
772 if(U_FAILURE(*pErrorCode)) {
773 return 0; /* the data format was not recognized */
774 }
775
776 pInfo=(const UDataInfo *)((const char *)inData+4);
777
778 {
779 /* convert the data format from ASCII to Unicode to the system charset */
780 UChar u[4]={
781 pInfo->dataFormat[0], pInfo->dataFormat[1],
782 pInfo->dataFormat[2], pInfo->dataFormat[3]
783 };
784
785 if(uprv_isInvariantUString(u, 4)) {
786 u_UCharsToChars(u, dataFormatChars, 4);
787 } else {
788 dataFormatChars[0]=dataFormatChars[1]=dataFormatChars[2]=dataFormatChars[3]='?';
789 }
790 }
791
792 /* dispatch to the swap function for the dataFormat */
793 for(i=0; i<UPRV_LENGTHOF(swapFns); ++i) {
794 if(0==memcmp(swapFns[i].dataFormat, pInfo->dataFormat, 4)) {
795 swappedLength=swapFns[i].swapFn(ds, inData, length, outData, pErrorCode);
796
797 if(U_FAILURE(*pErrorCode)) {
798 udata_printError(ds, "udata_swap(): failure swapping data format %02x.%02x.%02x.%02x (\"%c%c%c%c\") - %s\n",
799 pInfo->dataFormat[0], pInfo->dataFormat[1],
800 pInfo->dataFormat[2], pInfo->dataFormat[3],
801 dataFormatChars[0], dataFormatChars[1],
802 dataFormatChars[2], dataFormatChars[3],
803 u_errorName(*pErrorCode));
804 } else if(swappedLength<(length-15)) {
805 /* swapped less than expected */
806 udata_printError(ds, "udata_swap() warning: swapped only %d out of %d bytes - data format %02x.%02x.%02x.%02x (\"%c%c%c%c\")\n",
807 swappedLength, length,
808 pInfo->dataFormat[0], pInfo->dataFormat[1],
809 pInfo->dataFormat[2], pInfo->dataFormat[3],
810 dataFormatChars[0], dataFormatChars[1],
811 dataFormatChars[2], dataFormatChars[3],
812 u_errorName(*pErrorCode));
813 }
814
815 return swappedLength;
816 }
817 }
818
819 /* the dataFormat was not recognized */
820 udata_printError(ds, "udata_swap(): unknown data format %02x.%02x.%02x.%02x (\"%c%c%c%c\")\n",
821 pInfo->dataFormat[0], pInfo->dataFormat[1],
822 pInfo->dataFormat[2], pInfo->dataFormat[3],
823 dataFormatChars[0], dataFormatChars[1],
824 dataFormatChars[2], dataFormatChars[3]);
825
826 *pErrorCode=U_UNSUPPORTED_ERROR;
827 return 0;
828 }
829