+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
/*
*******************************************************************************
*
-* Copyright (C) 2005-2008, International Business Machines
+* Copyright (C) 2005-2014, International Business Machines
* Corporation and others. All Rights Reserved.
*
*******************************************************************************
* file name: swapimpl.cpp
-* encoding: US-ASCII
+* encoding: UTF-8
* tab size: 8 (not used)
* indentation:4
*
#include "unicode/utypes.h"
#include "unicode/putil.h"
#include "unicode/udata.h"
+
+/* Explicit include statement for std_string.h is needed
+ * for compilation on certain platforms. (e.g. AIX/VACPP)
+ */
+#include "unicode/std_string.h"
+
#include "cmemory.h"
#include "cstring.h"
#include "uinvchar.h"
#include "uarrsort.h"
#include "ucmndata.h"
#include "udataswp.h"
+#include "ulayout_props.h"
/* swapping implementations in common */
#include "ucol_swp.h"
#include "ucnv_bld.h"
#include "unormimp.h"
+#include "normalizer2impl.h"
#include "sprpimpl.h"
#include "propname.h"
#include "rbbidata.h"
-#include "triedict.h"
+#include "utrie.h"
+#include "utrie2.h"
+#include "dictionarydata.h"
/* swapping implementations in i18n */
+#if !UCONFIG_NO_NORMALIZATION
+#include "uspoof_impl.h"
+#endif
+
+U_NAMESPACE_USE
+
/* definitions */
-#define LENGTHOF(array) (int32_t)(sizeof(array)/sizeof((array)[0]))
+/* Unicode property (value) aliases data swapping --------------------------- */
+
+static int32_t U_CALLCONV
+upname_swap(const UDataSwapper *ds,
+ const void *inData, int32_t length, void *outData,
+ UErrorCode *pErrorCode) {
+ /* udata_swapDataHeader checks the arguments */
+ int32_t headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode);
+ if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
+ return 0;
+ }
+
+ /* check data format and format version */
+ const UDataInfo *pInfo=
+ reinterpret_cast<const UDataInfo *>(
+ static_cast<const char *>(inData)+4);
+ if(!(
+ pInfo->dataFormat[0]==0x70 && /* dataFormat="pnam" */
+ pInfo->dataFormat[1]==0x6e &&
+ pInfo->dataFormat[2]==0x61 &&
+ pInfo->dataFormat[3]==0x6d &&
+ pInfo->formatVersion[0]==2
+ )) {
+ udata_printError(ds, "upname_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as pnames.icu\n",
+ pInfo->dataFormat[0], pInfo->dataFormat[1],
+ pInfo->dataFormat[2], pInfo->dataFormat[3],
+ pInfo->formatVersion[0]);
+ *pErrorCode=U_UNSUPPORTED_ERROR;
+ return 0;
+ }
+
+ const uint8_t *inBytes=static_cast<const uint8_t *>(inData)+headerSize;
+ uint8_t *outBytes=static_cast<uint8_t *>(outData)+headerSize;
+
+ if(length>=0) {
+ length-=headerSize;
+ // formatVersion 2 initially has indexes[8], 32 bytes.
+ if(length<32) {
+ udata_printError(ds, "upname_swap(): too few bytes (%d after header) for pnames.icu\n",
+ (int)length);
+ *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
+ return 0;
+ }
+ }
+
+ const int32_t *inIndexes=reinterpret_cast<const int32_t *>(inBytes);
+ int32_t totalSize=udata_readInt32(ds, inIndexes[PropNameData::IX_TOTAL_SIZE]);
+ if(length>=0) {
+ if(length<totalSize) {
+ udata_printError(ds, "upname_swap(): too few bytes (%d after header, should be %d) "
+ "for pnames.icu\n",
+ (int)length, (int)totalSize);
+ *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
+ return 0;
+ }
+
+ int32_t numBytesIndexesAndValueMaps=
+ udata_readInt32(ds, inIndexes[PropNameData::IX_BYTE_TRIES_OFFSET]);
+
+ // Swap the indexes[] and the valueMaps[].
+ ds->swapArray32(ds, inBytes, numBytesIndexesAndValueMaps, outBytes, pErrorCode);
+
+ // Copy the rest of the data.
+ if(inBytes!=outBytes) {
+ uprv_memcpy(outBytes+numBytesIndexesAndValueMaps,
+ inBytes+numBytesIndexesAndValueMaps,
+ totalSize-numBytesIndexesAndValueMaps);
+ }
+
+ // We need not swap anything else:
+ //
+ // The ByteTries are already byte-serialized, and are fixed on ASCII.
+ // (On an EBCDIC machine, the input string is converted to lowercase ASCII
+ // while matching.)
+ //
+ // The name groups are mostly invariant characters, but since we only
+ // generate, and keep in subversion, ASCII versions of pnames.icu,
+ // and since only ICU4J uses the pnames.icu data file
+ // (the data is hardcoded in ICU4C) and ICU4J uses ASCII data files,
+ // we just copy those bytes too.
+ }
+
+ return headerSize+totalSize;
+}
/* Unicode properties data swapping ----------------------------------------- */
-U_CAPI int32_t U_EXPORT2
+static int32_t U_CALLCONV
uprops_swap(const UDataSwapper *ds,
const void *inData, int32_t length, void *outData,
UErrorCode *pErrorCode) {
pInfo->dataFormat[1]==0x50 &&
pInfo->dataFormat[2]==0x72 &&
pInfo->dataFormat[3]==0x6f &&
- (pInfo->formatVersion[0]==3 || pInfo->formatVersion[0]==4 || pInfo->formatVersion[0]==5) &&
- pInfo->formatVersion[2]==UTRIE_SHIFT &&
- pInfo->formatVersion[3]==UTRIE_INDEX_SHIFT
+ (3<=pInfo->formatVersion[0] && pInfo->formatVersion[0]<=7) &&
+ (pInfo->formatVersion[0]>=7 ||
+ (pInfo->formatVersion[2]==UTRIE_SHIFT &&
+ pInfo->formatVersion[3]==UTRIE_INDEX_SHIFT))
)) {
udata_printError(ds, "uprops_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not a Unicode properties file\n",
pInfo->dataFormat[0], pInfo->dataFormat[1],
* comments are copied from the data format description in genprops/store.c
* indexes[] constants are in uprops.h
*/
+ int32_t dataTop;
if(length>=0) {
int32_t *outData32;
- if((length-headerSize)<(4*dataIndexes[UPROPS_RESERVED_INDEX])) {
+ /*
+ * In formatVersion 7, UPROPS_DATA_TOP_INDEX has the post-header data size.
+ * In earlier formatVersions, it is 0 and a lower dataIndexes entry
+ * has the top of the last item.
+ */
+ for(i=UPROPS_DATA_TOP_INDEX; i>0 && (dataTop=dataIndexes[i])==0; --i) {}
+
+ if((length-headerSize)<(4*dataTop)) {
udata_printError(ds, "uprops_swap(): too few bytes (%d after header) for a Unicode properties file\n",
length-headerSize);
*pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
/* copy everything for inaccessible data (padding) */
if(inData32!=outData32) {
- uprv_memcpy(outData32, inData32, 4*dataIndexes[UPROPS_RESERVED_INDEX]);
+ uprv_memcpy(outData32, inData32, 4*(size_t)dataTop);
}
/* swap the indexes[16] */
* swap the main properties UTrie
* PT serialized properties trie, see utrie.h (byte size: 4*(i0-16))
*/
- utrie_swap(ds,
+ utrie_swapAnyVersion(ds,
inData32+UPROPS_INDEX_COUNT,
4*(dataIndexes[UPROPS_PROPS32_INDEX]-UPROPS_INDEX_COUNT),
outData32+UPROPS_INDEX_COUNT,
* swap the additional UTrie
* i3 additionalTrieIndex; -- 32-bit unit index to the additional trie for more properties
*/
- utrie_swap(ds,
+ utrie_swapAnyVersion(ds,
inData32+dataIndexes[UPROPS_ADDITIONAL_TRIE_INDEX],
4*(dataIndexes[UPROPS_ADDITIONAL_VECTORS_INDEX]-dataIndexes[UPROPS_ADDITIONAL_TRIE_INDEX]),
outData32+dataIndexes[UPROPS_ADDITIONAL_TRIE_INDEX],
*/
ds->swapArray32(ds,
inData32+dataIndexes[UPROPS_ADDITIONAL_VECTORS_INDEX],
- 4*(dataIndexes[UPROPS_RESERVED_INDEX]-dataIndexes[UPROPS_ADDITIONAL_VECTORS_INDEX]),
+ 4*(dataIndexes[UPROPS_SCRIPT_EXTENSIONS_INDEX]-dataIndexes[UPROPS_ADDITIONAL_VECTORS_INDEX]),
outData32+dataIndexes[UPROPS_ADDITIONAL_VECTORS_INDEX],
pErrorCode);
+
+ // swap the Script_Extensions data
+ // SCX const uint16_t scriptExtensions[2*(i7-i6)];
+ ds->swapArray16(ds,
+ inData32+dataIndexes[UPROPS_SCRIPT_EXTENSIONS_INDEX],
+ 4*(dataIndexes[UPROPS_RESERVED_INDEX_7]-dataIndexes[UPROPS_SCRIPT_EXTENSIONS_INDEX]),
+ outData32+dataIndexes[UPROPS_SCRIPT_EXTENSIONS_INDEX],
+ pErrorCode);
}
- /* i6 reservedItemIndex; -- 32-bit unit index to the top of the properties vectors table */
- return headerSize+4*dataIndexes[UPROPS_RESERVED_INDEX];
+ /* i7 reservedIndex7; -- 32-bit unit index to the top of the Script_Extensions data */
+ return headerSize+4*dataIndexes[UPROPS_RESERVED_INDEX_7];
}
/* Unicode case mapping data swapping --------------------------------------- */
-U_CAPI int32_t U_EXPORT2
+static int32_t U_CALLCONV
ucase_swap(const UDataSwapper *ds,
const void *inData, int32_t length, void *outData,
UErrorCode *pErrorCode) {
pInfo->dataFormat[1]==UCASE_FMT_1 &&
pInfo->dataFormat[2]==UCASE_FMT_2 &&
pInfo->dataFormat[3]==UCASE_FMT_3 &&
- pInfo->formatVersion[0]==1 &&
- pInfo->formatVersion[2]==UTRIE_SHIFT &&
- pInfo->formatVersion[3]==UTRIE_INDEX_SHIFT
+ ((pInfo->formatVersion[0]==1 &&
+ pInfo->formatVersion[2]==UTRIE_SHIFT &&
+ pInfo->formatVersion[3]==UTRIE_INDEX_SHIFT) ||
+ (2<=pInfo->formatVersion[0] && pInfo->formatVersion[0]<=4))
)) {
udata_printError(ds, "ucase_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as case mapping data\n",
pInfo->dataFormat[0], pInfo->dataFormat[1],
/* swap the UTrie */
count=indexes[UCASE_IX_TRIE_SIZE];
- utrie_swap(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
+ utrie_swapAnyVersion(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
offset+=count;
/* swap the uint16_t exceptions[] and unfold[] */
/* Unicode bidi/shaping data swapping --------------------------------------- */
-U_CAPI int32_t U_EXPORT2
+static int32_t U_CALLCONV
ubidi_swap(const UDataSwapper *ds,
const void *inData, int32_t length, void *outData,
UErrorCode *pErrorCode) {
pInfo->dataFormat[1]==UBIDI_FMT_1 &&
pInfo->dataFormat[2]==UBIDI_FMT_2 &&
pInfo->dataFormat[3]==UBIDI_FMT_3 &&
- pInfo->formatVersion[0]==1 &&
- pInfo->formatVersion[2]==UTRIE_SHIFT &&
- pInfo->formatVersion[3]==UTRIE_INDEX_SHIFT
+ ((pInfo->formatVersion[0]==1 &&
+ pInfo->formatVersion[2]==UTRIE_SHIFT &&
+ pInfo->formatVersion[3]==UTRIE_INDEX_SHIFT) ||
+ pInfo->formatVersion[0]==2)
)) {
udata_printError(ds, "ubidi_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as bidi/shaping data\n",
pInfo->dataFormat[0], pInfo->dataFormat[1],
/* swap the UTrie */
count=indexes[UBIDI_IX_TRIE_SIZE];
- utrie_swap(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
+ utrie_swapAnyVersion(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
offset+=count;
/* swap the uint32_t mirrors[] */
ds->swapArray32(ds, inBytes+offset, count, outBytes+offset, pErrorCode);
offset+=count;
- /* just skip the uint8_t jgArray[] */
+ /* just skip the uint8_t jgArray[] and jgArray2[] */
count=indexes[UBIDI_IX_JG_LIMIT]-indexes[UBIDI_IX_JG_START];
offset+=count;
+ count=indexes[UBIDI_IX_JG_LIMIT2]-indexes[UBIDI_IX_JG_START2];
+ offset+=count;
U_ASSERT(offset==size);
}
#if !UCONFIG_NO_NORMALIZATION
-U_CAPI int32_t U_EXPORT2
+static int32_t U_CALLCONV
unorm_swap(const UDataSwapper *ds,
const void *inData, int32_t length, void *outData,
UErrorCode *pErrorCode) {
#endif
+// Unicode text layout properties data swapping --------------------------------
+
+static int32_t U_CALLCONV
+ulayout_swap(const UDataSwapper *ds,
+ const void *inData, int32_t length, void *outData,
+ UErrorCode *pErrorCode) {
+ // udata_swapDataHeader checks the arguments.
+ int32_t headerSize = udata_swapDataHeader(ds, inData, length, outData, pErrorCode);
+ if (pErrorCode == nullptr || U_FAILURE(*pErrorCode)) {
+ return 0;
+ }
+
+ // Check data format and format version.
+ const UDataInfo *pInfo = (const UDataInfo *)((const char *)inData + 4);
+ if (!(
+ pInfo->dataFormat[0] == ULAYOUT_FMT_0 && // dataFormat="Layo"
+ pInfo->dataFormat[1] == ULAYOUT_FMT_1 &&
+ pInfo->dataFormat[2] == ULAYOUT_FMT_2 &&
+ pInfo->dataFormat[3] == ULAYOUT_FMT_3 &&
+ pInfo->formatVersion[0] == 1)) {
+ udata_printError(ds,
+ "ulayout_swap(): data format %02x.%02x.%02x.%02x (format version %02x) "
+ "is not recognized as text layout properties data\n",
+ pInfo->dataFormat[0], pInfo->dataFormat[1],
+ pInfo->dataFormat[2], pInfo->dataFormat[3],
+ pInfo->formatVersion[0]);
+ *pErrorCode = U_UNSUPPORTED_ERROR;
+ return 0;
+ }
+
+ const uint8_t *inBytes = (const uint8_t *)inData + headerSize;
+ uint8_t *outBytes = (uint8_t *)outData + headerSize;
+
+ const int32_t *inIndexes = (const int32_t *)inBytes;
+
+ if (length >= 0) {
+ length -= headerSize;
+ if (length < 12 * 4) {
+ udata_printError(ds,
+ "ulayout_swap(): too few bytes (%d after header) for text layout properties data\n",
+ length);
+ *pErrorCode = U_INDEX_OUTOFBOUNDS_ERROR;
+ return 0;
+ }
+ }
+
+ int32_t indexesLength = udata_readInt32(ds, inIndexes[ULAYOUT_IX_INDEXES_LENGTH]);
+ if (indexesLength < 12) {
+ udata_printError(ds,
+ "ulayout_swap(): too few indexes (%d) for text layout properties data\n",
+ indexesLength);
+ *pErrorCode = U_INDEX_OUTOFBOUNDS_ERROR;
+ return 0;
+ }
+
+ // Read the data offsets before swapping anything.
+ int32_t indexes[ULAYOUT_IX_TRIES_TOP + 1];
+ for (int32_t i = ULAYOUT_IX_INPC_TRIE_TOP; i <= ULAYOUT_IX_TRIES_TOP; ++i) {
+ indexes[i] = udata_readInt32(ds, inIndexes[i]);
+ }
+ int32_t size = indexes[ULAYOUT_IX_TRIES_TOP];
+
+ if (length >= 0) {
+ if (length < size) {
+ udata_printError(ds,
+ "ulayout_swap(): too few bytes (%d after header) "
+ "for all of text layout properties data\n",
+ length);
+ *pErrorCode = U_INDEX_OUTOFBOUNDS_ERROR;
+ return 0;
+ }
+
+ // Copy the data for inaccessible bytes.
+ if (inBytes != outBytes) {
+ uprv_memcpy(outBytes, inBytes, size);
+ }
+
+ // Swap the int32_t indexes[].
+ int32_t offset = 0;
+ int32_t count = indexesLength * 4;
+ ds->swapArray32(ds, inBytes, count, outBytes, pErrorCode);
+ offset += count;
+
+ // Swap each trie.
+ for (int32_t i = ULAYOUT_IX_INPC_TRIE_TOP; i <= ULAYOUT_IX_TRIES_TOP; ++i) {
+ int32_t top = indexes[i];
+ count = top - offset;
+ U_ASSERT(count >= 0);
+ if (count >= 16) {
+ utrie_swapAnyVersion(ds, inBytes + offset, count, outBytes + offset, pErrorCode);
+ }
+ offset = top;
+ }
+
+ U_ASSERT(offset == size);
+ }
+
+ return headerSize + size;
+}
+
+/* Swap 'Test' data from gentest */
+static int32_t U_CALLCONV
+test_swap(const UDataSwapper *ds,
+ const void *inData, int32_t length, void *outData,
+ UErrorCode *pErrorCode) {
+ const UDataInfo *pInfo;
+ int32_t headerSize;
+
+ const uint8_t *inBytes;
+ uint8_t *outBytes;
+
+ int32_t offset;
+
+ /* udata_swapDataHeader checks the arguments */
+ headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode);
+ if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
+ udata_printError(ds, "test_swap(): data header swap failed %s\n", pErrorCode != NULL ? u_errorName(*pErrorCode) : "pErrorCode is NULL");
+ return 0;
+ }
+
+ /* check data format and format version */
+ pInfo=(const UDataInfo *)((const char *)inData+4);
+ if(!(
+ pInfo->dataFormat[0]==0x54 && /* dataFormat="Norm" */
+ pInfo->dataFormat[1]==0x65 &&
+ pInfo->dataFormat[2]==0x73 &&
+ pInfo->dataFormat[3]==0x74 &&
+ pInfo->formatVersion[0]==1
+ )) {
+ udata_printError(ds, "test_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as testdata\n",
+ pInfo->dataFormat[0], pInfo->dataFormat[1],
+ pInfo->dataFormat[2], pInfo->dataFormat[3],
+ pInfo->formatVersion[0]);
+ *pErrorCode=U_UNSUPPORTED_ERROR;
+ return 0;
+ }
+
+ inBytes=(const uint8_t *)inData+headerSize;
+ outBytes=(uint8_t *)outData+headerSize;
+
+ int32_t size16 = 2; // 16bit plus padding
+ int32_t sizeStr = 5; // 4 char inv-str plus null
+ int32_t size = size16 + sizeStr;
+
+ if(length>=0) {
+ if(length<size) {
+ udata_printError(ds, "test_swap(): too few bytes (%d after header, wanted %d) for all of testdata\n",
+ length, size);
+ *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
+ return 0;
+ }
+
+ offset =0;
+ /* swap a 1 entry array */
+ ds->swapArray16(ds, inBytes+offset, size16, outBytes+offset, pErrorCode);
+ offset+=size16;
+ ds->swapInvChars(ds, inBytes+offset, sizeStr, outBytes+offset, pErrorCode);
+ }
+
+ return headerSize+size;
+}
+
/* swap any data (except a .dat package) ------------------------------------ */
static const struct {
#if !UCONFIG_NO_NORMALIZATION
{ { 0x4e, 0x6f, 0x72, 0x6d }, unorm_swap }, /* dataFormat="Norm" */
+ { { 0x4e, 0x72, 0x6d, 0x32 }, unorm2_swap }, /* dataFormat="Nrm2" */
#endif
+
+ { { ULAYOUT_FMT_0, ULAYOUT_FMT_1, ULAYOUT_FMT_2, ULAYOUT_FMT_3 },
+ ulayout_swap }, // dataFormat="Layo"
+
#if !UCONFIG_NO_COLLATION
{ { 0x55, 0x43, 0x6f, 0x6c }, ucol_swap }, /* dataFormat="UCol" */
{ { 0x49, 0x6e, 0x76, 0x43 }, ucol_swapInverseUCA },/* dataFormat="InvC" */
#endif
#if !UCONFIG_NO_BREAK_ITERATION
{ { 0x42, 0x72, 0x6b, 0x20 }, ubrk_swap }, /* dataFormat="Brk " */
- { { 0x54, 0x72, 0x44, 0x63 }, triedict_swap }, /* dataFormat="TrDc " */
+ { { 0x44, 0x69, 0x63, 0x74 }, udict_swap }, /* dataFormat="Dict" */
#endif
{ { 0x70, 0x6e, 0x61, 0x6d }, upname_swap }, /* dataFormat="pnam" */
- { { 0x75, 0x6e, 0x61, 0x6d }, uchar_swapNames } /* dataFormat="unam" */
+ { { 0x75, 0x6e, 0x61, 0x6d }, uchar_swapNames }, /* dataFormat="unam" */
+#if !UCONFIG_NO_NORMALIZATION
+ { { 0x43, 0x66, 0x75, 0x20 }, uspoof_swap }, /* dataFormat="Cfu " */
+#endif
+ { { 0x54, 0x65, 0x73, 0x74 }, test_swap } /* dataFormat="Test" */
};
U_CAPI int32_t U_EXPORT2
UErrorCode *pErrorCode) {
char dataFormatChars[4];
const UDataInfo *pInfo;
- int32_t headerSize, i, swappedLength;
+ int32_t i, swappedLength;
if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
return 0;
* information. Otherwise we would have to pass some of the information
* and not be able to use the UDataSwapFn signature.
*/
- headerSize=udata_swapDataHeader(ds, inData, -1, NULL, pErrorCode);
+ udata_swapDataHeader(ds, inData, -1, NULL, pErrorCode);
/*
* If we wanted udata_swap() to also handle non-loadable data like a UTrie,
}
/* dispatch to the swap function for the dataFormat */
- for(i=0; i<LENGTHOF(swapFns); ++i) {
+ for(i=0; i<UPRV_LENGTHOF(swapFns); ++i) {
if(0==memcmp(swapFns[i].dataFormat, pInfo->dataFormat, 4)) {
swappedLength=swapFns[i].swapFn(ds, inData, length, outData, pErrorCode);
projects / apple / icu.git / blobdiff