extern void
repeatProps(uint32_t first, uint32_t last, uint32_t x) {
if(!utrie_setRange32(pTrie, (UChar32)first, (UChar32)(last+1), x, FALSE)) {
fprintf(stderr, "error: too many entries for the properties trie\n");
exit(U_BUFFER_OVERFLOW_ERROR);
}
}
U_CAPI void U_CALLCONV
upvec_compactToUTrieHandler(void *context,
UChar32 start, UChar32 end,
int32_t rowIndex, uint32_t *row, int32_t columns,
UErrorCode *pErrorCode) {
UPVecToUTrieContext *toUTrie=(UPVecToUTrieContext *)context;
if(start<UPVEC_FIRST_SPECIAL_CP) {
if(!utrie_setRange32(toUTrie->newTrie, start, end+1, (uint32_t)rowIndex, TRUE)) {
*pErrorCode=U_BUFFER_OVERFLOW_ERROR;
}
} else {
switch(start) {
case UPVEC_INITIAL_VALUE_CP:
toUTrie->initialValue=rowIndex;
break;
case UPVEC_START_REAL_VALUES_CP:
if(rowIndex>0xffff) {
/* too many rows for a 16-bit trie */
*pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
} else {
toUTrie->newTrie=utrie_open(NULL, NULL, toUTrie->capacity,
toUTrie->initialValue, toUTrie->initialValue,
toUTrie->latin1Linear);
if(toUTrie->newTrie==NULL) {
*pErrorCode=U_MEMORY_ALLOCATION_ERROR;
}
}
break;
default:
break;
}
}
}
extern void
storeRange(uint32_t start, uint32_t end, UStringPrepType type,UErrorCode* status){
uint16_t trieWord = 0;
if((int)(_SPREP_TYPE_THRESHOLD + type) > 0xFFFF){
fprintf(stderr,"trieWord cannot contain value greater than 0xFFFF.\n");
exit(U_ILLEGAL_CHAR_FOUND);
}
trieWord = (_SPREP_TYPE_THRESHOLD + type); /* the top 4 bits contain the value */
if(start == end){
uint32_t savedTrieWord = utrie_get32(sprepTrie, start, NULL);
if(savedTrieWord>0){
if(savedTrieWord < _SPREP_TYPE_THRESHOLD && type == USPREP_PROHIBITED){
/*
* A mapping is stored in the trie word
* and the only other possible type that a
* code point can have is USPREP_PROHIBITED
*
*/
/* turn on the 0th bit in the savedTrieWord */
savedTrieWord += 0x01;
/* the downcast is safe since we only save 16 bit values */
trieWord = (uint16_t)savedTrieWord;
/* make sure that the value of trieWord is less than the threshold */
if(trieWord < _SPREP_TYPE_THRESHOLD){
/* now set the value in the trie */
if(!utrie_set32(sprepTrie,start,trieWord)){
fprintf(stderr,"Could not set the value for code point.\n");
exit(U_ILLEGAL_ARGUMENT_ERROR);
}
/* value is set so just return */
return;
}else{
fprintf(stderr,"trieWord cannot contain value greater than threshold 0x%04X.\n",_SPREP_TYPE_THRESHOLD);
exit(U_ILLEGAL_CHAR_FOUND);
}
}else if(savedTrieWord != trieWord){
fprintf(stderr,"Value for codepoint \\U%08X already set!.\n", (int)start);
exit(U_ILLEGAL_ARGUMENT_ERROR);
}
/* if savedTrieWord == trieWord .. fall through and set the value */
}
if(!utrie_set32(sprepTrie,start,trieWord)){
fprintf(stderr,"Could not set the value for code point \\U%08X.\n", (int)start);
exit(U_ILLEGAL_ARGUMENT_ERROR);
}
}else{
if(!utrie_setRange32(sprepTrie, start, end+1, trieWord, FALSE)){
fprintf(stderr,"Value for certain codepoint already set.\n");
exit(U_ILLEGAL_CHAR_FOUND);
}
}
}
extern void
generateData(const char *dataDir, UBool csource) {
static int32_t indexes[UBIDI_IX_TOP]={
UBIDI_IX_TOP
};
static uint8_t trieBlock[40000];
static uint8_t jgArray[0x300]; /* at most for U+0600..U+08FF */
const uint32_t *row;
UChar32 start, end, prev, jgStart;
int32_t i;
UNewDataMemory *pData;
UNewTrie *pTrie;
UErrorCode errorCode=U_ZERO_ERROR;
int32_t trieSize;
long dataLength;
makeMirror();
pTrie=utrie_open(NULL, NULL, 20000, 0, 0, TRUE);
if(pTrie==NULL) {
fprintf(stderr, "genbidi error: unable to create a UNewTrie\n");
exit(U_MEMORY_ALLOCATION_ERROR);
}
prev=jgStart=0;
for(i=0; (row=upvec_getRow(pv, i, &start, &end))!=NULL && start<UPVEC_FIRST_SPECIAL_CP; ++i) {
/* store most values from vector column 0 in the trie */
if(!utrie_setRange32(pTrie, start, end+1, *row, TRUE)) {
fprintf(stderr, "genbidi error: unable to set trie value (overflow)\n");
exit(U_BUFFER_OVERFLOW_ERROR);
}
/* store Joining_Group values from vector column 1 in a simple byte array */
if(row[1]!=0) {
if(start<0x600 || 0x8ff<end) {
fprintf(stderr, "genbidi error: Joining_Group for out-of-range code points U+%04lx..U+%04lx\n",
(long)start, (long)end);
exit(U_ILLEGAL_ARGUMENT_ERROR);
}
if(prev==0) {
/* first code point with any value */
prev=jgStart=start;
} else {
/* add No_Joining_Group for code points between prev and start */
while(prev<start) {
jgArray[prev++ -jgStart]=0;
}
}
/* set Joining_Group value for start..end */
while(prev<=end) {
jgArray[prev++ -jgStart]=(uint8_t)row[1];
}
}
}
/* finish jgArray, pad to multiple of 4 */
while((prev-jgStart)&3) {
jgArray[prev++ -jgStart]=0;
}
indexes[UBIDI_IX_JG_START]=jgStart;
indexes[UBIDI_IX_JG_LIMIT]=prev;
trieSize=utrie_serialize(pTrie, trieBlock, sizeof(trieBlock), NULL, TRUE, &errorCode);
if(U_FAILURE(errorCode)) {
fprintf(stderr, "genbidi error: utrie_serialize failed: %s (length %ld)\n", u_errorName(errorCode), (long)trieSize);
exit(errorCode);
}
indexes[UBIDI_IX_TRIE_SIZE]=trieSize;
indexes[UBIDI_IX_MIRROR_LENGTH]=mirrorTop;
indexes[UBIDI_IX_LENGTH]=
(int32_t)sizeof(indexes)+
trieSize+
4*mirrorTop+
(prev-jgStart);
if(beVerbose) {
printf("trie size in bytes: %5d\n", (int)trieSize);
printf("size in bytes of mirroring table: %5d\n", (int)(4*mirrorTop));
printf("length of Joining_Group array: %5d (U+%04x..U+%04x)\n", (int)(prev-jgStart), (int)jgStart, (int)(prev-1));
printf("data size: %5d\n", (int)indexes[UBIDI_IX_LENGTH]);
}
indexes[UBIDI_MAX_VALUES_INDEX]=
((int32_t)U_CHAR_DIRECTION_COUNT-1)|
(((int32_t)U_JT_COUNT-1)<<UBIDI_JT_SHIFT)|
(((int32_t)U_JG_COUNT-1)<<UBIDI_MAX_JG_SHIFT);
if(csource) {
/* write .c file for hardcoded data */
UTrie trie={ NULL };
UTrie2 *trie2;
FILE *f;
utrie_unserialize(&trie, trieBlock, trieSize, &errorCode);
if(U_FAILURE(errorCode)) {
fprintf(
stderr,
"genbidi error: failed to utrie_unserialize(ubidi.icu trie) - %s\n",
u_errorName(errorCode));
exit(errorCode);
}
/* use UTrie2 */
dataInfo.formatVersion[0]=2;
dataInfo.formatVersion[2]=0;
dataInfo.formatVersion[3]=0;
trie2=utrie2_fromUTrie(&trie, 0, &errorCode);
if(U_FAILURE(errorCode)) {
fprintf(
stderr,
"genbidi error: utrie2_fromUTrie() failed - %s\n",
u_errorName(errorCode));
exit(errorCode);
}
{
/* delete lead surrogate code unit values */
UChar lead;
trie2=utrie2_cloneAsThawed(trie2, &errorCode);
for(lead=0xd800; lead<0xdc00; ++lead) {
utrie2_set32ForLeadSurrogateCodeUnit(trie2, lead, trie2->initialValue, &errorCode);
}
utrie2_freeze(trie2, UTRIE2_16_VALUE_BITS, &errorCode);
if(U_FAILURE(errorCode)) {
fprintf(
stderr,
"genbidi error: deleting lead surrogate code unit values failed - %s\n",
u_errorName(errorCode));
exit(errorCode);
}
}
f=usrc_create(dataDir, "ubidi_props_data.c");
if(f!=NULL) {
usrc_writeArray(f,
"static const UVersionInfo ubidi_props_dataVersion={",
dataInfo.dataVersion, 8, 4,
"};\n\n");
usrc_writeArray(f,
"static const int32_t ubidi_props_indexes[UBIDI_IX_TOP]={",
indexes, 32, UBIDI_IX_TOP,
"};\n\n");
usrc_writeUTrie2Arrays(f,
"static const uint16_t ubidi_props_trieIndex[%ld]={\n", NULL,
trie2,
"\n};\n\n");
usrc_writeArray(f,
"static const uint32_t ubidi_props_mirrors[%ld]={\n",
mirrors, 32, mirrorTop,
"\n};\n\n");
usrc_writeArray(f,
"static const uint8_t ubidi_props_jgArray[%ld]={\n",
jgArray, 8, prev-jgStart,
"\n};\n\n");
fputs(
"static const UBiDiProps ubidi_props_singleton={\n"
" NULL,\n"
" ubidi_props_indexes,\n"
" ubidi_props_mirrors,\n"
" ubidi_props_jgArray,\n",
f);
usrc_writeUTrie2Struct(f,
" {\n",
trie2, "ubidi_props_trieIndex", NULL,
" },\n");
usrc_writeArray(f, " { ", dataInfo.formatVersion, 8, 4, " }\n");
fputs("};\n", f);
fclose(f);
}
utrie2_close(trie2);
} else {
/* write the data */
pData=udata_create(dataDir, UBIDI_DATA_TYPE, UBIDI_DATA_NAME, &dataInfo,
haveCopyright ? U_COPYRIGHT_STRING : NULL, &errorCode);
if(U_FAILURE(errorCode)) {
fprintf(stderr, "genbidi: unable to create data memory, %s\n", u_errorName(errorCode));
exit(errorCode);
}
udata_writeBlock(pData, indexes, sizeof(indexes));
udata_writeBlock(pData, trieBlock, trieSize);
udata_writeBlock(pData, mirrors, 4*mirrorTop);
udata_writeBlock(pData, jgArray, prev-jgStart);
/* finish up */
dataLength=udata_finish(pData, &errorCode);
if(U_FAILURE(errorCode)) {
fprintf(stderr, "genbidi: error %d writing the output file\n", errorCode);
exit(errorCode);
}
if(dataLength!=indexes[UBIDI_IX_LENGTH]) {
fprintf(stderr, "genbidi: data length %ld != calculated size %d\n",
dataLength, (int)indexes[UBIDI_IX_LENGTH]);
exit(U_INTERNAL_PROGRAM_ERROR);
}
}
utrie_close(pTrie);
upvec_close(pv);
}
//------------------------------------------------------------------------
//
// build Build the list of non-overlapping character ranges
// from the Unicode Sets.
//
//------------------------------------------------------------------------
void RBBISetBuilder::build() {
RBBINode *usetNode;
RangeDescriptor *rlRange;
if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "usets")) {printSets();}
//
// Initialize the process by creating a single range encompassing all characters
// that is in no sets.
//
fRangeList = new RangeDescriptor(*fStatus); // will check for status here
fRangeList->fStartChar = 0;
fRangeList->fEndChar = 0x10ffff;
if (U_FAILURE(*fStatus)) {
return;
}
//
// Find the set of non-overlapping ranges of characters
//
int ni;
for (ni=0; ; ni++) { // Loop over each of the UnicodeSets encountered in the input rules
usetNode = (RBBINode *)this->fRB->fUSetNodes->elementAt(ni);
if (usetNode==NULL) {
break;
}
UnicodeSet *inputSet = usetNode->fInputSet;
int32_t inputSetRangeCount = inputSet->getRangeCount();
int inputSetRangeIndex = 0;
rlRange = fRangeList;
for (;;) {
if (inputSetRangeIndex >= inputSetRangeCount) {
break;
}
UChar32 inputSetRangeBegin = inputSet->getRangeStart(inputSetRangeIndex);
UChar32 inputSetRangeEnd = inputSet->getRangeEnd(inputSetRangeIndex);
// skip over ranges from the range list that are completely
// below the current range from the input unicode set.
while (rlRange->fEndChar < inputSetRangeBegin) {
rlRange = rlRange->fNext;
}
// If the start of the range from the range list is before with
// the start of the range from the unicode set, split the range list range
// in two, with one part being before (wholly outside of) the unicode set
// and the other containing the rest.
// Then continue the loop; the post-split current range will then be skipped
// over
if (rlRange->fStartChar < inputSetRangeBegin) {
rlRange->split(inputSetRangeBegin, *fStatus);
if (U_FAILURE(*fStatus)) {
return;
}
continue;
}
// Same thing at the end of the ranges...
// If the end of the range from the range list doesn't coincide with
// the end of the range from the unicode set, split the range list
// range in two. The first part of the split range will be
// wholly inside the Unicode set.
if (rlRange->fEndChar > inputSetRangeEnd) {
rlRange->split(inputSetRangeEnd+1, *fStatus);
if (U_FAILURE(*fStatus)) {
return;
}
}
// The current rlRange is now entirely within the UnicodeSet range.
// Add this unicode set to the list of sets for this rlRange
if (rlRange->fIncludesSets->indexOf(usetNode) == -1) {
rlRange->fIncludesSets->addElement(usetNode, *fStatus);
if (U_FAILURE(*fStatus)) {
return;
}
}
// Advance over ranges that we are finished with.
if (inputSetRangeEnd == rlRange->fEndChar) {
inputSetRangeIndex++;
}
rlRange = rlRange->fNext;
}
}
if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "range")) { printRanges();}
//
// Group the above ranges, with each group consisting of one or more
// ranges that are in exactly the same set of original UnicodeSets.
// The groups are numbered, and these group numbers are the set of
// input symbols recognized by the run-time state machine.
//
// Numbering: # 0 (state table column 0) is unused.
// # 1 is reserved - table column 1 is for end-of-input
// # 2 is reserved - table column 2 is for beginning-in-input
// # 3 is the first range list.
//
RangeDescriptor *rlSearchRange;
for (rlRange = fRangeList; rlRange!=0; rlRange=rlRange->fNext) {
for (rlSearchRange=fRangeList; rlSearchRange != rlRange; rlSearchRange=rlSearchRange->fNext) {
if (rlRange->fIncludesSets->equals(*rlSearchRange->fIncludesSets)) {
rlRange->fNum = rlSearchRange->fNum;
break;
}
}
if (rlRange->fNum == 0) {
fGroupCount ++;
rlRange->fNum = fGroupCount+2;
rlRange->setDictionaryFlag();
addValToSets(rlRange->fIncludesSets, fGroupCount+2);
}
}
// Handle input sets that contain the special string {eof}.
// Column 1 of the state table is reserved for EOF on input.
// Column 2 is reserved for before-the-start-input.
// (This column can be optimized away later if there are no rule
// references to {bof}.)
// Add this column value (1 or 2) to the equivalent expression
// subtree for each UnicodeSet that contains the string {eof}
// Because {bof} and {eof} are not a characters in the normal sense,
// they doesn't affect the computation of ranges or TRIE.
static const UChar eofUString[] = {0x65, 0x6f, 0x66, 0};
static const UChar bofUString[] = {0x62, 0x6f, 0x66, 0};
UnicodeString eofString(eofUString);
UnicodeString bofString(bofUString);
for (ni=0; ; ni++) { // Loop over each of the UnicodeSets encountered in the input rules
usetNode = (RBBINode *)this->fRB->fUSetNodes->elementAt(ni);
if (usetNode==NULL) {
break;
}
UnicodeSet *inputSet = usetNode->fInputSet;
if (inputSet->contains(eofString)) {
addValToSet(usetNode, 1);
}
if (inputSet->contains(bofString)) {
addValToSet(usetNode, 2);
fSawBOF = TRUE;
}
}
if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "rgroup")) {printRangeGroups();}
if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "esets")) {printSets();}
//
// Build the Trie table for mapping UChar32 values to the corresponding
// range group number
//
fTrie = utrie_open(NULL, // Pre-existing trie to be filled in
NULL, // Data array (utrie will allocate one)
100000, // Max Data Length
0, // Initial value for all code points
0, // Lead surrogate unit value
TRUE); // Keep Latin 1 in separately
for (rlRange = fRangeList; rlRange!=0; rlRange=rlRange->fNext) {
utrie_setRange32(fTrie, rlRange->fStartChar, rlRange->fEndChar+1, rlRange->fNum, TRUE);
}
}
//------------------------------------------------------------------------
//
// build Build the list of non-overlapping character ranges
// from the Unicode Sets.
//
//------------------------------------------------------------------------
void RBBISetBuilder::build() {
RBBINode *usetNode;
RangeDescriptor *rlRange;
if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "usets")) {
printSets();
}
//
// Initialize the process by creating a single range encompassing all characters
// that is in no sets.
//
fRangeList = new RangeDescriptor(*fStatus); // will check for status here
fRangeList->fStartChar = 0;
fRangeList->fEndChar = 0x10ffff;
if (U_FAILURE(*fStatus)) {
return;
}
//
// Find the set of non-overlapping ranges of characters
//
int ni;
for (ni=0; ; ni++) {
usetNode = (RBBINode *)this->fRB->fUSetNodes->elementAt(ni);
if (usetNode==NULL) {
break;
}
UnicodeSet *inputSet = usetNode->fInputSet;
int32_t inputSetRangeCount = inputSet->getRangeCount();
int inputSetRangeIndex = 0;
rlRange = fRangeList;
for (;;) {
if (inputSetRangeIndex >= inputSetRangeCount) {
break;
}
UChar32 inputSetRangeBegin = inputSet->getRangeStart(inputSetRangeIndex);
UChar32 inputSetRangeEnd = inputSet->getRangeEnd(inputSetRangeIndex);
// skip over ranges from the range list that are completely
// below the current range from the input unicode set.
while (rlRange->fEndChar < inputSetRangeBegin) {
rlRange = rlRange->fNext;
}
// If the start of the range from the range list is before with
// the start of the range from the unicode set, split the range list range
// in two, with one part being before (wholly outside of) the unicode set
// and the other containing the rest.
// Then continue the loop; the post-split current range will then be skipped
// over
if (rlRange->fStartChar < inputSetRangeBegin) {
rlRange->split(inputSetRangeBegin, *fStatus);
if (U_FAILURE(*fStatus)) {
return;
}
continue;
}
// Same thing at the end of the ranges...
// If the end of the range from the range list doesn't coincide with
// the end of the range from the unicode set, split the range list
// range in two. The first part of the split range will be
// wholly inside the Unicode set.
if (rlRange->fEndChar > inputSetRangeEnd) {
rlRange->split(inputSetRangeEnd+1, *fStatus);
if (U_FAILURE(*fStatus)) {
return;
}
}
// The current rlRange is now entirely within the UnicodeSet range.
// Add this unicode set to the list of sets for this rlRange
if (rlRange->fIncludesSets->indexOf(usetNode) == -1) {
rlRange->fIncludesSets->addElement(usetNode, *fStatus);
if (U_FAILURE(*fStatus)) {
return;
}
}
// Advance over ranges that we are finished with.
if (inputSetRangeEnd == rlRange->fEndChar) {
inputSetRangeIndex++;
}
rlRange = rlRange->fNext;
}
}
if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "range")) {
printRanges();
}
//
// Group the above ranges, with each group consisting of one or more
// ranges that are in exactly the same set of original UnicodeSets.
// The groups are numbered, and these group numbers are the set of
// input symbols recognized by the run-time state machine.
//
RangeDescriptor *rlSearchRange;
for (rlRange = fRangeList; rlRange!=0; rlRange=rlRange->fNext) {
for (rlSearchRange=fRangeList; rlSearchRange != rlRange; rlSearchRange=rlSearchRange->fNext) {
if (rlRange->fIncludesSets->equals(*rlSearchRange->fIncludesSets)) {
rlRange->fNum = rlSearchRange->fNum;
break;
}
}
if (rlRange->fNum == 0) {
fGroupCount ++;
rlRange->fNum = fGroupCount;
rlRange->setDictionaryFlag();
addValToSets(rlRange->fIncludesSets, fGroupCount);
}
}
if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "rgroup")) {
printRangeGroups();
}
if (fRB->fDebugEnv && uprv_strstr(fRB->fDebugEnv, "esets")) {
printSets();
}
//
// Build the Trie table for mapping UChar32 values to the corresponding
// range group number
//
fTrie = utrie_open(NULL, // Pre-existing trie to be filled in
NULL, // Data array (utrie will allocate one)
100000, // Max Data Length
0, // Initial value for all code points
0, // Lead surrogate unit value
TRUE); // Keep Latin 1 in separately
for (rlRange = fRangeList; rlRange!=0; rlRange=rlRange->fNext) {
utrie_setRange32(fTrie, rlRange->fStartChar, rlRange->fEndChar+1, rlRange->fNum, TRUE);
}
}
static void
testTrieRanges(const char *testName,
const SetRange setRanges[], int32_t countSetRanges,
const CheckRange checkRanges[], int32_t countCheckRanges,
UBool dataIs32, UBool latin1Linear) {
union{
double bogus; /* needed for aligining the storage */
uint8_t storage[32768];
} storageHolder;
UTrieGetFoldingOffset *getFoldingOffset;
UNewTrieGetFoldedValue *getFoldedValue;
const CheckRange *enumRanges;
UNewTrie *newTrie;
UTrie trie={ 0 };
uint32_t value, value2;
UChar32 start, limit;
int32_t i, length;
UErrorCode errorCode;
UBool overwrite, ok;
log_verbose("\ntesting Trie '%s'\n", testName);
newTrie=utrie_open(NULL, NULL, 2000,
checkRanges[0].value, checkRanges[0].value,
latin1Linear);
/* set values from setRanges[] */
ok=TRUE;
for(i=0; i<countSetRanges; ++i) {
start=setRanges[i].start;
limit=setRanges[i].limit;
value=setRanges[i].value;
overwrite=setRanges[i].overwrite;
if((limit-start)==1 && overwrite) {
ok&=utrie_set32(newTrie, start, value);
} else {
ok&=utrie_setRange32(newTrie, start, limit, value, overwrite);
}
}
if(!ok) {
log_err("error: setting values into a trie failed (%s)\n", testName);
return;
}
/* verify that all these values are in the new Trie */
start=0;
for(i=0; i<countCheckRanges; ++i) {
limit=checkRanges[i].limit;
value=checkRanges[i].value;
while(start<limit) {
if(value!=utrie_get32(newTrie, start, NULL)) {
log_err("error: newTrie(%s)[U+%04lx]==0x%lx instead of 0x%lx\n",
testName, start, utrie_get32(newTrie, start, NULL), value);
}
++start;
}
}
if(dataIs32) {
getFoldingOffset=_testFoldingOffset32;
getFoldedValue=_testFoldedValue32;
} else {
getFoldingOffset=_testFoldingOffset16;
getFoldedValue=_testFoldedValue16;
}
/*
* code coverage for utrie.c/defaultGetFoldedValue(),
* pick some combination of parameters for selecting the UTrie defaults
*/
if(!dataIs32 && latin1Linear) {
getFoldingOffset=NULL;
getFoldedValue=NULL;
}
errorCode=U_ZERO_ERROR;
length=utrie_serialize(newTrie, storageHolder.storage, sizeof(storageHolder.storage),
getFoldedValue,
(UBool)!dataIs32,
&errorCode);
if(U_FAILURE(errorCode)) {
log_err("error: utrie_serialize(%s) failed: %s\n", testName, u_errorName(errorCode));
utrie_close(newTrie);
return;
}
if (length >= (int32_t)sizeof(storageHolder.storage)) {
log_err("error: utrie_serialize(%s) needs more memory\n", testName);
utrie_close(newTrie);
return;
}
/* test linear Latin-1 range from utrie_getData() */
if(latin1Linear) {
uint32_t *data;
int32_t dataLength;
data=utrie_getData(newTrie, &dataLength);
start=0;
for(i=0; i<countCheckRanges && start<=0xff; ++i) {
limit=checkRanges[i].limit;
value=checkRanges[i].value;
while(start<limit && start<=0xff) {
if(value!=data[UTRIE_DATA_BLOCK_LENGTH+start]) {
log_err("error: newTrie(%s).latin1Data[U+%04lx]==0x%lx instead of 0x%lx\n",
testName, start, data[UTRIE_DATA_BLOCK_LENGTH+start], value);
}
++start;
}
}
}
utrie_close(newTrie);
errorCode=U_ZERO_ERROR;
if(!utrie_unserialize(&trie, storageHolder.storage, length, &errorCode)) {
log_err("error: utrie_unserialize() failed, %s\n", u_errorName(errorCode));
return;
}
if(getFoldingOffset!=NULL) {
trie.getFoldingOffset=getFoldingOffset;
}
if(dataIs32!=(trie.data32!=NULL)) {
log_err("error: trie serialization (%s) did not preserve 32-bitness\n", testName);
}
if(latin1Linear!=trie.isLatin1Linear) {
log_err("error: trie serialization (%s) did not preserve Latin-1-linearity\n", testName);
}
/* verify that all these values are in the unserialized Trie */
start=0;
for(i=0; i<countCheckRanges; ++i) {
limit=checkRanges[i].limit;
value=checkRanges[i].value;
if(start==0xd800) {
/* skip surrogates */
start=limit;
continue;
}
while(start<limit) {
if(start<=0xffff) {
if(dataIs32) {
value2=UTRIE_GET32_FROM_BMP(&trie, start);
} else {
value2=UTRIE_GET16_FROM_BMP(&trie, start);
}
if(value!=value2) {
log_err("error: unserialized trie(%s).fromBMP(U+%04lx)==0x%lx instead of 0x%lx\n",
testName, start, value2, value);
}
if(!U16_IS_LEAD(start)) {
if(dataIs32) {
value2=UTRIE_GET32_FROM_LEAD(&trie, start);
} else {
value2=UTRIE_GET16_FROM_LEAD(&trie, start);
}
if(value!=value2) {
log_err("error: unserialized trie(%s).fromLead(U+%04lx)==0x%lx instead of 0x%lx\n",
testName, start, value2, value);
}
}
}
if(dataIs32) {
UTRIE_GET32(&trie, start, value2);
} else {
UTRIE_GET16(&trie, start, value2);
}
if(value!=value2) {
log_err("error: unserialized trie(%s).get(U+%04lx)==0x%lx instead of 0x%lx\n",
testName, start, value2, value);
}
++start;
}
}
/* enumerate and verify all ranges */
enumRanges=checkRanges+1;
utrie_enum(&trie, _testEnumValue, _testEnumRange, &enumRanges);
/* test linear Latin-1 range */
if(trie.isLatin1Linear) {
if(trie.data32!=NULL) {
const uint32_t *latin1=UTRIE_GET32_LATIN1(&trie);
for(start=0; start<0x100; ++start) {
if(latin1[start]!=UTRIE_GET32_FROM_LEAD(&trie, start)) {
log_err("error: (%s) trie.latin1[U+%04lx]=0x%lx!=0x%lx=trie.get32(U+%04lx)\n",
testName, start, latin1[start], UTRIE_GET32_FROM_LEAD(&trie, start), start);
}
}
} else {
const uint16_t *latin1=UTRIE_GET16_LATIN1(&trie);
for(start=0; start<0x100; ++start) {
if(latin1[start]!=UTRIE_GET16_FROM_LEAD(&trie, start)) {
log_err("error: (%s) trie.latin1[U+%04lx]=0x%lx!=0x%lx=trie.get16(U+%04lx)\n",
testName, start, latin1[start], UTRIE_GET16_FROM_LEAD(&trie, start), start);
}
}
}
}
testTrieIteration(testName, &trie, checkRanges, countCheckRanges);
}