organicmaps/icu
13
Fork 0
mirror of https://github.com/unicode-org/icu.git synced 2025-04-06 22:15:31 +00:00
Code Issues 1 Projects Releases Wiki Activity Actions 1

ICU-13030 merge from trunk

Browse source 
X-SVN-Rev: 40387
This commit is contained in:
Steven R. Loomis 2017-09-12 19:52:59 +00:00
commit eaa26ce352
1337 changed files with 34612 additions and 15866 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

776
icu4c/APIChangeReport.html
View file

File diff suppressed because it is too large Load diff

10
icu4c/readme.html
View file

@ -23,7 +23,7 @@
milestone - if the release note is a milestone release
-->
<body>
<body class="milestone">
<!-- <body> -->
<p class="only-draft"><b>Note:</b> This is a draft readme.</p>
@ -32,7 +32,7 @@
International Components for Unicode<br/>
<span class="only-rc">Release Candidate</span>
<span class="only-milestone">(Milestone Release)</span>
<abbr title="International Components for Unicode">ICU</abbr> 59.1 ReadMe
<abbr title="International Components for Unicode">ICU</abbr> 60 Milestone 1 ReadMe
</h1>
<!-- Shouldn't need to comment/uncomment this paragraph, just change the body class -->
@ -44,7 +44,7 @@
<p class="note only-rc">This is a release candidate version of ICU4C.
It is not recommended for production use.</p>
<p>Last updated: 2017-Apr-12<br/>
<p>Last updated: 2017-Jun-27<br/>
Copyright &copy; 2016 and later: Unicode, Inc. and others. License &amp; terms of use:
<a href="http://www.unicode.org/copyright.html">http://www.unicode.org/copyright.html</a><br/>
Copyright &copy; 1997-2016 International Business Machines Corporation and others.
@ -233,8 +233,8 @@
<h2><a name="News" href="#News" id="News">What Is New In This Release?</a></h2>
<p>See the <a href="http://site.icu-project.org/download/59">ICU 59 download page</a>
for an overview of this release, important changes, new features, bug fixes, known issues,
<p>See the <a href="http://site.icu-project.org/download/60">ICU 60 download page</a>
(coming soon) for an overview of this release, important changes, new features, bug fixes, known issues,
changes to supported platforms and build environments,
and migration issues for existing applications migrating from previous ICU releases.</p>

23
icu4c/source/common/brkiter.cpp
View file

@ -195,7 +195,7 @@ BreakIterator::getAvailableLocales(int32_t& count)
// ------------------------------------------
//
// Default constructor and destructor
// Constructors, destructor and assignment operator
//
//-------------------------------------------
@ -204,6 +204,19 @@ BreakIterator::BreakIterator()
*validLocale = *actualLocale = 0;
}
BreakIterator::BreakIterator(const BreakIterator &other) : UObject(other) {
uprv_strncpy(actualLocale, other.actualLocale, sizeof(actualLocale));
uprv_strncpy(validLocale, other.validLocale, sizeof(validLocale));
}
BreakIterator &BreakIterator::operator =(const BreakIterator &other) {
if (this != &other) {
uprv_strncpy(actualLocale, other.actualLocale, sizeof(actualLocale));
uprv_strncpy(validLocale, other.validLocale, sizeof(validLocale));
}
return *this;
}
BreakIterator::~BreakIterator()
{
}
@ -265,7 +278,7 @@ ICUBreakIteratorService::~ICUBreakIteratorService() {}
// defined in ucln_cmn.h
U_NAMESPACE_END
static icu::UInitOnce gInitOnce;
static icu::UInitOnce gInitOnceBrkiter;
static icu::ICULocaleService* gService = NULL;
@ -280,7 +293,7 @@ static UBool U_CALLCONV breakiterator_cleanup(void) {
delete gService;
gService = NULL;
}
gInitOnce.reset();
gInitOnceBrkiter.reset();
#endif
return TRUE;
}
@ -296,7 +309,7 @@ initService(void) {
static ICULocaleService*
getService(void)
{
umtx_initOnce(gInitOnce, &initService);
umtx_initOnce(gInitOnceBrkiter, &initService);
return gService;
}
@ -306,7 +319,7 @@ getService(void)
static inline UBool
hasService(void)
{
return !gInitOnce.isReset() && getService() != NULL;
return !gInitOnceBrkiter.isReset() && getService() != NULL;
}
// -------------------------------------

2
icu4c/source/common/caniter.cpp
View file

@ -405,7 +405,7 @@ UnicodeString* CanonicalIterator::getEquivalents(const UnicodeString &segment, i
//String[] finalResult = new String[result.size()];
UnicodeString *finalResult = NULL;
int32_t resultCount;
if((resultCount = result.count())) {
if((resultCount = result.count()) != 0) {
finalResult = new UnicodeString[resultCount];
if (finalResult == 0) {
status = U_MEMORY_ALLOCATION_ERROR;

8
icu4c/source/common/common.vcxproj
View file

@ -110,7 +110,7 @@
<Culture>0x0409</Culture>
</ResourceCompile>
<Link>
<OutputFile>..\..\bin\icuuc59.dll</OutputFile>
<OutputFile>..\..\bin\icuuc60.dll</OutputFile>
<SuppressStartupBanner>true</SuppressStartupBanner>
<ProgramDatabaseFile>.\..\..\lib\icuuc.pdb</ProgramDatabaseFile>
<EnableCOMDATFolding>true</EnableCOMDATFolding>
@ -152,7 +152,7 @@
<Culture>0x0409</Culture>
</ResourceCompile>
<Link>
<OutputFile>..\..\bin\icuuc59d.dll</OutputFile>
<OutputFile>..\..\bin\icuuc60d.dll</OutputFile>
<SuppressStartupBanner>true</SuppressStartupBanner>
<GenerateDebugInformation>true</GenerateDebugInformation>
<ProgramDatabaseFile>.\..\..\lib\icuucd.pdb</ProgramDatabaseFile>
@ -191,7 +191,7 @@
<Culture>0x0409</Culture>
</ResourceCompile>
<Link>
<OutputFile>..\..\bin64\icuuc59.dll</OutputFile>
<OutputFile>..\..\bin64\icuuc60.dll</OutputFile>
<SuppressStartupBanner>true</SuppressStartupBanner>
<ProgramDatabaseFile>.\..\..\lib64\icuuc.pdb</ProgramDatabaseFile>
<EnableCOMDATFolding>true</EnableCOMDATFolding>
@ -231,7 +231,7 @@
<Culture>0x0409</Culture>
</ResourceCompile>
<Link>
<OutputFile>..\..\bin64\icuuc59d.dll</OutputFile>
<OutputFile>..\..\bin64\icuuc60d.dll</OutputFile>
<SuppressStartupBanner>true</SuppressStartupBanner>
<GenerateDebugInformation>true</GenerateDebugInformation>
<ProgramDatabaseFile>.\..\..\lib64\icuucd.pdb</ProgramDatabaseFile>

616
icu4c/source/common/edits.cpp
View file

@ -17,10 +17,10 @@ namespace {
const int32_t MAX_UNCHANGED_LENGTH = 0x1000;
const int32_t MAX_UNCHANGED = MAX_UNCHANGED_LENGTH - 1;
// 0wwwcccccccccccc with w=1..6 records ccc+1 replacements of w:w text units.
// No length change.
const int32_t MAX_SHORT_WIDTH = 6;
const int32_t MAX_SHORT_CHANGE_LENGTH = 0xfff;
// 0mmmnnnccccccccc with m=1..6 records ccc+1 replacements of m:n text units.
const int32_t MAX_SHORT_CHANGE_OLD_LENGTH = 6;
const int32_t MAX_SHORT_CHANGE_NEW_LENGTH = 7;
const int32_t SHORT_CHANGE_NUM_MASK = 0x1ff;
const int32_t MAX_SHORT_CHANGE = 0x6fff;
// 0111mmmmmmnnnnnn records a replacement of m text units with n.
@ -33,20 +33,85 @@ const int32_t LENGTH_IN_2TRAIL = 62;
} // namespace
Edits::~Edits() {
if(array != stackArray) {
void Edits::releaseArray() U_NOEXCEPT {
if (array != stackArray) {
uprv_free(array);
}
}
void Edits::reset() {
length = delta = 0;
Edits &Edits::copyArray(const Edits &other) {
if (U_FAILURE(errorCode_)) {
length = delta = numChanges = 0;
return *this;
}
if (length > capacity) {
uint16_t *newArray = (uint16_t *)uprv_malloc((size_t)length * 2);
if (newArray == nullptr) {
length = delta = numChanges = 0;
errorCode_ = U_MEMORY_ALLOCATION_ERROR;
return *this;
}
releaseArray();
array = newArray;
capacity = length;
}
if (length > 0) {
uprv_memcpy(array, other.array, (size_t)length * 2);
}
return *this;
}
Edits &Edits::moveArray(Edits &src) U_NOEXCEPT {
if (U_FAILURE(errorCode_)) {
length = delta = numChanges = 0;
return *this;
}
releaseArray();
if (length > STACK_CAPACITY) {
array = src.array;
capacity = src.capacity;
src.array = src.stackArray;
src.capacity = STACK_CAPACITY;
src.reset();
return *this;
}
array = stackArray;
capacity = STACK_CAPACITY;
if (length > 0) {
uprv_memcpy(array, src.array, (size_t)length * 2);
}
return *this;
}
Edits &Edits::operator=(const Edits &other) {
length = other.length;
delta = other.delta;
numChanges = other.numChanges;
errorCode_ = other.errorCode_;
return copyArray(other);
}
Edits &Edits::operator=(Edits &&src) U_NOEXCEPT {
length = src.length;
delta = src.delta;
numChanges = src.numChanges;
errorCode_ = src.errorCode_;
return moveArray(src);
}
Edits::~Edits() {
releaseArray();
}
void Edits::reset() U_NOEXCEPT {
length = delta = numChanges = 0;
errorCode_ = U_ZERO_ERROR;
}
void Edits::addUnchanged(int32_t unchangedLength) {
if(U_FAILURE(errorCode) || unchangedLength == 0) { return; }
if(U_FAILURE(errorCode_) || unchangedLength == 0) { return; }
if(unchangedLength < 0) {
errorCode = U_ILLEGAL_ARGUMENT_ERROR;
errorCode_ = U_ILLEGAL_ARGUMENT_ERROR;
return;
}
// Merge into previous unchanged-text record, if any.
@ -72,38 +137,41 @@ void Edits::addUnchanged(int32_t unchangedLength) {
}
void Edits::addReplace(int32_t oldLength, int32_t newLength) {
if(U_FAILURE(errorCode)) { return; }
if(oldLength == newLength && 0 < oldLength && oldLength <= MAX_SHORT_WIDTH) {
// Replacement of short oldLength text units by same-length new text.
// Merge into previous short-replacement record, if any.
int32_t last = lastUnit();
if(MAX_UNCHANGED < last && last < MAX_SHORT_CHANGE &&
(last >> 12) == oldLength && (last & 0xfff) < MAX_SHORT_CHANGE_LENGTH) {
setLastUnit(last + 1);
return;
}
append(oldLength << 12);
return;
}
if(U_FAILURE(errorCode_)) { return; }
if(oldLength < 0 || newLength < 0) {
errorCode = U_ILLEGAL_ARGUMENT_ERROR;
errorCode_ = U_ILLEGAL_ARGUMENT_ERROR;
return;
}
if (oldLength == 0 && newLength == 0) {
return;
}
++numChanges;
int32_t newDelta = newLength - oldLength;
if (newDelta != 0) {
if ((newDelta > 0 && delta >= 0 && newDelta > (INT32_MAX - delta)) ||
(newDelta < 0 && delta < 0 && newDelta < (INT32_MIN - delta))) {
// Integer overflow or underflow.
errorCode = U_INDEX_OUTOFBOUNDS_ERROR;
errorCode_ = U_INDEX_OUTOFBOUNDS_ERROR;
return;
}
delta += newDelta;
}
if(0 < oldLength && oldLength <= MAX_SHORT_CHANGE_OLD_LENGTH &&
newLength <= MAX_SHORT_CHANGE_NEW_LENGTH) {
// Merge into previous same-lengths short-replacement record, if any.
int32_t u = (oldLength << 12) | (newLength << 9);
int32_t last = lastUnit();
if(MAX_UNCHANGED < last && last < MAX_SHORT_CHANGE &&
(last & ~SHORT_CHANGE_NUM_MASK) == u &&
(last & SHORT_CHANGE_NUM_MASK) < SHORT_CHANGE_NUM_MASK) {
setLastUnit(last + 1);
return;
}
append(u);
return;
}
int32_t head = 0x7000;
if (oldLength < LENGTH_IN_1TRAIL && newLength < LENGTH_IN_1TRAIL) {
head |= oldLength << 6;
@ -149,7 +217,7 @@ UBool Edits::growArray() {
} else if (capacity == INT32_MAX) {
// Not U_BUFFER_OVERFLOW_ERROR because that could be confused on a string transform API
// with a result-string-buffer overflow.
errorCode = U_INDEX_OUTOFBOUNDS_ERROR;
errorCode_ = U_INDEX_OUTOFBOUNDS_ERROR;
return FALSE;
} else if (capacity >= (INT32_MAX / 2)) {
newCapacity = INT32_MAX;
@ -158,18 +226,16 @@ UBool Edits::growArray() {
}
// Grow by at least 5 units so that a maximal change record will fit.
if ((newCapacity - capacity) < 5) {
errorCode = U_INDEX_OUTOFBOUNDS_ERROR;
errorCode_ = U_INDEX_OUTOFBOUNDS_ERROR;
return FALSE;
}
uint16_t *newArray = (uint16_t *)uprv_malloc((size_t)newCapacity * 2);
if (newArray == NULL) {
errorCode = U_MEMORY_ALLOCATION_ERROR;
errorCode_ = U_MEMORY_ALLOCATION_ERROR;
return FALSE;
}
uprv_memcpy(newArray, array, (size_t)length * 2);
if (array != stackArray) {
uprv_free(array);
}
releaseArray();
array = newArray;
capacity = newCapacity;
return TRUE;
@ -177,27 +243,161 @@ UBool Edits::growArray() {
UBool Edits::copyErrorTo(UErrorCode &outErrorCode) {
if (U_FAILURE(outErrorCode)) { return TRUE; }
if (U_SUCCESS(errorCode)) { return FALSE; }
outErrorCode = errorCode;
if (U_SUCCESS(errorCode_)) { return FALSE; }
outErrorCode = errorCode_;
return TRUE;
}
UBool Edits::hasChanges() const {
if (delta != 0) {
return TRUE;
}
for (int32_t i = 0; i < length; ++i) {
if (array[i] > MAX_UNCHANGED) {
return TRUE;
Edits &Edits::mergeAndAppend(const Edits &ab, const Edits &bc, UErrorCode &errorCode) {
if (copyErrorTo(errorCode)) { return *this; }
// Picture string a --(Edits ab)--> string b --(Edits bc)--> string c.
// Parallel iteration over both Edits.
Iterator abIter = ab.getFineIterator();
Iterator bcIter = bc.getFineIterator();
UBool abHasNext = TRUE, bcHasNext = TRUE;
// Copy iterator state into local variables, so that we can modify and subdivide spans.
// ab old & new length, bc old & new length
int32_t aLength = 0, ab_bLength = 0, bc_bLength = 0, cLength = 0;
// When we have different-intermediate-length changes, we accumulate a larger change.
int32_t pending_aLength = 0, pending_cLength = 0;
for (;;) {
// At this point, for each of the two iterators:
// Either we are done with the locally cached current edit,
// and its intermediate-string length has been reset,
// or we will continue to work with a truncated remainder of this edit.
//
// If the current edit is done, and the iterator has not yet reached the end,
// then we fetch the next edit. This is true for at least one of the iterators.
//
// Normally it does not matter whether we fetch from ab and then bc or vice versa.
// However, the result is observably different when
// ab deletions meet bc insertions at the same intermediate-string index.
// Some users expect the bc insertions to come first, so we fetch from bc first.
if (bc_bLength == 0) {
if (bcHasNext && (bcHasNext = bcIter.next(errorCode))) {
bc_bLength = bcIter.oldLength();
cLength = bcIter.newLength();
if (bc_bLength == 0) {
// insertion
if (ab_bLength == 0 || !abIter.hasChange()) {
addReplace(pending_aLength, pending_cLength + cLength);
pending_aLength = pending_cLength = 0;
} else {
pending_cLength += cLength;
}
continue;
}
}
// else see if the other iterator is done, too.
}
if (ab_bLength == 0) {
if (abHasNext && (abHasNext = abIter.next(errorCode))) {
aLength = abIter.oldLength();
ab_bLength = abIter.newLength();
if (ab_bLength == 0) {
// deletion
if (bc_bLength == bcIter.oldLength() || !bcIter.hasChange()) {
addReplace(pending_aLength + aLength, pending_cLength);
pending_aLength = pending_cLength = 0;
} else {
pending_aLength += aLength;
}
continue;
}
} else if (bc_bLength == 0) {
// Both iterators are done at the same time:
// The intermediate-string lengths match.
break;
} else {
// The ab output string is shorter than the bc input string.
if (!copyErrorTo(errorCode)) {
errorCode = U_ILLEGAL_ARGUMENT_ERROR;
}
return *this;
}
}
if (bc_bLength == 0) {
// The bc input string is shorter than the ab output string.
if (!copyErrorTo(errorCode)) {
errorCode = U_ILLEGAL_ARGUMENT_ERROR;
}
return *this;
}
// Done fetching: ab_bLength > 0 && bc_bLength > 0
// The current state has two parts:
// - Past: We accumulate a longer ac edit in the "pending" variables.
// - Current: We have copies of the current ab/bc edits in local variables.
// At least one side is newly fetched.
// One side might be a truncated remainder of an edit we fetched earlier.
if (!abIter.hasChange() && !bcIter.hasChange()) {
// An unchanged span all the way from string a to string c.
if (pending_aLength != 0 || pending_cLength != 0) {
addReplace(pending_aLength, pending_cLength);
pending_aLength = pending_cLength = 0;
}
int32_t unchangedLength = aLength <= cLength ? aLength : cLength;
addUnchanged(unchangedLength);
ab_bLength = aLength -= unchangedLength;
bc_bLength = cLength -= unchangedLength;
// At least one of the unchanged spans is now empty.
continue;
}
if (!abIter.hasChange() && bcIter.hasChange()) {
// Unchanged a->b but changed b->c.
if (ab_bLength >= bc_bLength) {
// Split the longer unchanged span into change + remainder.
addReplace(pending_aLength + bc_bLength, pending_cLength + cLength);
pending_aLength = pending_cLength = 0;
aLength = ab_bLength -= bc_bLength;
bc_bLength = 0;
continue;
}
// Handle the shorter unchanged span below like a change.
} else if (abIter.hasChange() && !bcIter.hasChange()) {
// Changed a->b and then unchanged b->c.
if (ab_bLength <= bc_bLength) {
// Split the longer unchanged span into change + remainder.
addReplace(pending_aLength + aLength, pending_cLength + ab_bLength);
pending_aLength = pending_cLength = 0;
cLength = bc_bLength -= ab_bLength;
ab_bLength = 0;
continue;
}
// Handle the shorter unchanged span below like a change.
} else { // both abIter.hasChange() && bcIter.hasChange()
if (ab_bLength == bc_bLength) {
// Changes on both sides up to the same position. Emit & reset.
addReplace(pending_aLength + aLength, pending_cLength + cLength);
pending_aLength = pending_cLength = 0;
ab_bLength = bc_bLength = 0;
continue;
}
}
// Accumulate the a->c change, reset the shorter side,
// keep a remainder of the longer one.
pending_aLength += aLength;
pending_cLength += cLength;
if (ab_bLength < bc_bLength) {
bc_bLength -= ab_bLength;
cLength = ab_bLength = 0;
} else { // ab_bLength > bc_bLength
ab_bLength -= bc_bLength;
aLength = bc_bLength = 0;
}
}
return FALSE;
if (pending_aLength != 0 || pending_cLength != 0) {
addReplace(pending_aLength, pending_cLength);
}
copyErrorTo(errorCode);
return *this;
}
Edits::Iterator::Iterator(const uint16_t *a, int32_t len, UBool oc, UBool crs) :
array(a), index(0), length(len), remaining(0),
onlyChanges_(oc), coarse(crs),
changed(FALSE), oldLength_(0), newLength_(0),
dir(0), changed(FALSE), oldLength_(0), newLength_(0),
srcIndex(0), replIndex(0), destIndex(0) {}
int32_t Edits::Iterator::readLength(int32_t head) {
@ -219,7 +419,7 @@ int32_t Edits::Iterator::readLength(int32_t head) {
}
}
void Edits::Iterator::updateIndexes() {
void Edits::Iterator::updateNextIndexes() {
srcIndex += oldLength_;
if (changed) {
replIndex += newLength_;
@ -227,22 +427,52 @@ void Edits::Iterator::updateIndexes() {
destIndex += newLength_;
}
void Edits::Iterator::updatePreviousIndexes() {
srcIndex -= oldLength_;
if (changed) {
replIndex -= newLength_;
}
destIndex -= newLength_;
}
UBool Edits::Iterator::noNext() {
// No change beyond the string.
// No change before or beyond the string.
dir = 0;
changed = FALSE;
oldLength_ = newLength_ = 0;
return FALSE;
}
UBool Edits::Iterator::next(UBool onlyChanges, UErrorCode &errorCode) {
// Forward iteration: Update the string indexes to the limit of the current span,
// and post-increment-read array units to assemble a new span.
// Leaves the array index one after the last unit of that span.
if (U_FAILURE(errorCode)) { return FALSE; }
// We have an errorCode in case we need to start guarding against integer overflows.
// It is also convenient for caller loops if we bail out when an error was set elsewhere.
updateIndexes();
if (remaining > 0) {
// Fine-grained iterator: Continue a sequence of equal-length changes.
--remaining;
return TRUE;
if (dir > 0) {
updateNextIndexes();
} else {
if (dir < 0) {
// Turn around from previous() to next().
// Post-increment-read the same span again.
if (remaining > 0) {
// Fine-grained iterator:
// Stay on the current one of a sequence of compressed changes.
++index; // next() rests on the index after the sequence unit.
dir = 1;
return TRUE;
}
}
dir = 1;
}
if (remaining >= 1) {
// Fine-grained iterator: Continue a sequence of compressed changes.
if (remaining > 1) {
--remaining;
return TRUE;
}
remaining = 0;
}
if (index >= length) {
return noNext();
@ -258,7 +488,7 @@ UBool Edits::Iterator::next(UBool onlyChanges, UErrorCode &errorCode) {
}
newLength_ = oldLength_;
if (onlyChanges) {
updateIndexes();
updateNextIndexes();
if (index >= length) {
return noNext();
}
@ -270,14 +500,19 @@ UBool Edits::Iterator::next(UBool onlyChanges, UErrorCode &errorCode) {
}
changed = TRUE;
if (u <= MAX_SHORT_CHANGE) {
int32_t oldLen = u >> 12;
int32_t newLen = (u >> 9) & MAX_SHORT_CHANGE_NEW_LENGTH;
int32_t num = (u & SHORT_CHANGE_NUM_MASK) + 1;
if (coarse) {
int32_t w = u >> 12;
int32_t len = (u & 0xfff) + 1;
oldLength_ = newLength_ = len * w;
oldLength_ = num * oldLen;
newLength_ = num * newLen;
} else {
// Split a sequence of equal-length changes that was compressed into one unit.
oldLength_ = newLength_ = u >> 12;
remaining = u & 0xfff;
// Split a sequence of changes that was compressed into one unit.
oldLength_ = oldLen;
newLength_ = newLen;
if (num > 1) {
remaining = num; // This is the first of two or more changes.
}
return TRUE;
}
} else {
@ -292,55 +527,250 @@ UBool Edits::Iterator::next(UBool onlyChanges, UErrorCode &errorCode) {
while (index < length && (u = array[index]) > MAX_UNCHANGED) {
++index;
if (u <= MAX_SHORT_CHANGE) {
int32_t w = u >> 12;
int32_t len = (u & 0xfff) + 1;
len = len * w;
oldLength_ += len;
newLength_ += len;
int32_t num = (u & SHORT_CHANGE_NUM_MASK) + 1;
oldLength_ += (u >> 12) * num;
newLength_ += ((u >> 9) & MAX_SHORT_CHANGE_NEW_LENGTH) * num;
} else {
U_ASSERT(u <= 0x7fff);
int32_t oldLen = readLength((u >> 6) & 0x3f);
int32_t newLen = readLength(u & 0x3f);
oldLength_ += oldLen;
newLength_ += newLen;
oldLength_ += readLength((u >> 6) & 0x3f);
newLength_ += readLength(u & 0x3f);
}
}
return TRUE;
}
UBool Edits::Iterator::findSourceIndex(int32_t i, UErrorCode &errorCode) {
if (U_FAILURE(errorCode) || i < 0) { return FALSE; }
if (i < srcIndex) {
// Reset the iterator to the start.
index = remaining = oldLength_ = newLength_ = srcIndex = replIndex = destIndex = 0;
} else if (i < (srcIndex + oldLength_)) {
// The index is in the current span.
return TRUE;
}
while (next(FALSE, errorCode)) {
if (i < (srcIndex + oldLength_)) {
// The index is in the current span.
return TRUE;
}
if (remaining > 0) {
// Is the index in one of the remaining compressed edits?
// srcIndex is the start of the current span, before the remaining ones.
int32_t len = (remaining + 1) * oldLength_;
if (i < (srcIndex + len)) {
int32_t n = (i - srcIndex) / oldLength_; // 1 <= n <= remaining
len = n * oldLength_;
srcIndex += len;
replIndex += len;
destIndex += len;
remaining -= n;
UBool Edits::Iterator::previous(UErrorCode &errorCode) {
// Backward iteration: Pre-decrement-read array units to assemble a new span,
// then update the string indexes to the start of that span.
// Leaves the array index on the head unit of that span.
if (U_FAILURE(errorCode)) { return FALSE; }
// We have an errorCode in case we need to start guarding against integer overflows.
// It is also convenient for caller loops if we bail out when an error was set elsewhere.
if (dir >= 0) {
if (dir > 0) {
// Turn around from next() to previous().
// Set the string indexes to the span limit and
// pre-decrement-read the same span again.
if (remaining > 0) {
// Fine-grained iterator:
// Stay on the current one of a sequence of compressed changes.
--index; // previous() rests on the sequence unit.
dir = -1;
return TRUE;
}
updateNextIndexes();
}
dir = -1;
}
if (remaining > 0) {
// Fine-grained iterator: Continue a sequence of compressed changes.
int32_t u = array[index];
U_ASSERT(MAX_UNCHANGED < u && u <= MAX_SHORT_CHANGE);
if (remaining <= (u & SHORT_CHANGE_NUM_MASK)) {
++remaining;
updatePreviousIndexes();
return TRUE;
}
remaining = 0;
}
if (index <= 0) {
return noNext();
}
int32_t u = array[--index];
if (u <= MAX_UNCHANGED) {
// Combine adjacent unchanged ranges.
changed = FALSE;
oldLength_ = u + 1;
while (index > 0 && (u = array[index - 1]) <= MAX_UNCHANGED) {
--index;
oldLength_ += u + 1;
}
newLength_ = oldLength_;
// No need to handle onlyChanges as long as previous() is called only from findIndex().
updatePreviousIndexes();
return TRUE;
}
changed = TRUE;
if (u <= MAX_SHORT_CHANGE) {
int32_t oldLen = u >> 12;
int32_t newLen = (u >> 9) & MAX_SHORT_CHANGE_NEW_LENGTH;
int32_t num = (u & SHORT_CHANGE_NUM_MASK) + 1;
if (coarse) {
oldLength_ = num * oldLen;
newLength_ = num * newLen;
} else {
// Split a sequence of changes that was compressed into one unit.
oldLength_ = oldLen;
newLength_ = newLen;
if (num > 1) {
remaining = 1; // This is the last of two or more changes.
}
updatePreviousIndexes();
return TRUE;
}
} else {
if (u <= 0x7fff) {
// The change is encoded in u alone.
oldLength_ = readLength((u >> 6) & 0x3f);
newLength_ = readLength(u & 0x3f);
} else {
// Back up to the head of the change, read the lengths,
// and reset the index to the head again.
U_ASSERT(index > 0);
while ((u = array[--index]) > 0x7fff) {}
U_ASSERT(u > MAX_SHORT_CHANGE);
int32_t headIndex = index++;
oldLength_ = readLength((u >> 6) & 0x3f);
newLength_ = readLength(u & 0x3f);
index = headIndex;
}
if (!coarse) {
updatePreviousIndexes();
return TRUE;
}
}
// Combine adjacent changes.
while (index > 0 && (u = array[index - 1]) > MAX_UNCHANGED) {
--index;
if (u <= MAX_SHORT_CHANGE) {
int32_t num = (u & SHORT_CHANGE_NUM_MASK) + 1;
oldLength_ += (u >> 12) * num;
newLength_ += ((u >> 9) & MAX_SHORT_CHANGE_NEW_LENGTH) * num;
} else if (u <= 0x7fff) {
// Read the lengths, and reset the index to the head again.
int32_t headIndex = index++;
oldLength_ += readLength((u >> 6) & 0x3f);
newLength_ += readLength(u & 0x3f);
index = headIndex;
}
}
updatePreviousIndexes();
return TRUE;
}
int32_t Edits::Iterator::findIndex(int32_t i, UBool findSource, UErrorCode &errorCode) {
if (U_FAILURE(errorCode) || i < 0) { return -1; }
int32_t spanStart, spanLength;
if (findSource) { // find source index
spanStart = srcIndex;
spanLength = oldLength_;
} else { // find destination index
spanStart = destIndex;
spanLength = newLength_;
}
if (i < spanStart) {
if (i >= (spanStart / 2)) {
// Search backwards.
for (;;) {
UBool hasPrevious = previous(errorCode);
U_ASSERT(hasPrevious); // because i>=0 and the first span starts at 0
(void)hasPrevious; // avoid unused-variable warning
spanStart = findSource ? srcIndex : destIndex;
if (i >= spanStart) {
// The index is in the current span.
return 0;
}
if (remaining > 0) {
// Is the index in one of the remaining compressed edits?
// spanStart is the start of the current span, first of the remaining ones.
spanLength = findSource ? oldLength_ : newLength_;
int32_t u = array[index];
U_ASSERT(MAX_UNCHANGED < u && u <= MAX_SHORT_CHANGE);
int32_t num = (u & SHORT_CHANGE_NUM_MASK) + 1 - remaining;
int32_t len = num * spanLength;
if (i >= (spanStart - len)) {
int32_t n = ((spanStart - i - 1) / spanLength) + 1;
// 1 <= n <= num
srcIndex -= n * oldLength_;
replIndex -= n * newLength_;
destIndex -= n * newLength_;
remaining += n;
return 0;
}
// Skip all of these edits at once.
srcIndex -= num * oldLength_;
replIndex -= num * newLength_;
destIndex -= num * newLength_;
remaining = 0;
}
}
}
// Reset the iterator to the start.
dir = 0;
index = remaining = oldLength_ = newLength_ = srcIndex = replIndex = destIndex = 0;
} else if (i < (spanStart + spanLength)) {
// The index is in the current span.
return 0;
}
while (next(FALSE, errorCode)) {
if (findSource) {
spanStart = srcIndex;
spanLength = oldLength_;
} else {
spanStart = destIndex;
spanLength = newLength_;
}
if (i < (spanStart + spanLength)) {
// The index is in the current span.
return 0;
}
if (remaining > 1) {
// Is the index in one of the remaining compressed edits?
// spanStart is the start of the current span, first of the remaining ones.
int32_t len = remaining * spanLength;
if (i < (spanStart + len)) {
int32_t n = (i - spanStart) / spanLength; // 1 <= n <= remaining - 1
srcIndex += n * oldLength_;
replIndex += n * newLength_;
destIndex += n * newLength_;
remaining -= n;
return 0;
}
// Make next() skip all of these edits at once.
oldLength_ = newLength_ = len;
oldLength_ *= remaining;
newLength_ *= remaining;
remaining = 0;
}
}
return FALSE;
return 1;
}
int32_t Edits::Iterator::destinationIndexFromSourceIndex(int32_t i, UErrorCode &errorCode) {
int32_t where = findIndex(i, TRUE, errorCode);
if (where < 0) {
// Error or before the string.
return 0;
}
if (where > 0 || i == srcIndex) {
// At or after string length, or at start of the found span.
return destIndex;
}
if (changed) {
// In a change span, map to its end.
return destIndex + newLength_;
} else {
// In an unchanged span, offset 1:1 within it.
return destIndex + (i - srcIndex);
}
}
int32_t Edits::Iterator::sourceIndexFromDestinationIndex(int32_t i, UErrorCode &errorCode) {
int32_t where = findIndex(i, FALSE, errorCode);
if (where < 0) {
// Error or before the string.
return 0;
}
if (where > 0 || i == destIndex) {
// At or after string length, or at start of the found span.
return srcIndex;
}
if (changed) {
// In a change span, map to its end.
return srcIndex + oldLength_;
} else {
// In an unchanged span, offset within it.
return srcIndex + (i - destIndex);
}
}
U_NAMESPACE_END

73
icu4c/source/common/filterednormalizer2.cpp
View file

@ -20,11 +20,13 @@
#if !UCONFIG_NO_NORMALIZATION
#include "unicode/edits.h"
#include "unicode/normalizer2.h"
#include "unicode/uniset.h"
#include "unicode/unistr.h"
#include "unicode/unorm.h"
#include "cpputils.h"
#include "ustr_imp.h" // U_EDITS_NO_RESET
U_NAMESPACE_BEGIN
@ -85,6 +87,52 @@ FilteredNormalizer2::normalize(const UnicodeString &src,
return dest;
}
void
FilteredNormalizer2::normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
Edits *edits, UErrorCode &errorCode) const {
if (U_FAILURE(errorCode)) {
return;
}
if (edits != nullptr && (options & U_EDITS_NO_RESET) == 0) {
edits->reset();
}
options |= U_EDITS_NO_RESET; // Do not reset for each span.
normalizeUTF8(options, src.data(), src.length(), sink, edits, USET_SPAN_SIMPLE, errorCode);
}
void
FilteredNormalizer2::normalizeUTF8(uint32_t options, const char *src, int32_t length,
ByteSink &sink, Edits *edits,
USetSpanCondition spanCondition,
UErrorCode &errorCode) const {
while (length > 0) {
int32_t spanLength = set.spanUTF8(src, length, spanCondition);
if (spanCondition == USET_SPAN_NOT_CONTAINED) {
if (spanLength != 0) {
if (edits != nullptr) {
edits->addUnchanged(spanLength);
}
if ((options & U_OMIT_UNCHANGED_TEXT) == 0) {
sink.Append(src, spanLength);
}
}
spanCondition = USET_SPAN_SIMPLE;
} else {
if (spanLength != 0) {
// Not norm2.normalizeSecondAndAppend() because we do not want
// to modify the non-filter part of dest.
norm2.normalizeUTF8(options, StringPiece(src, spanLength), sink, edits, errorCode);
if (U_FAILURE(errorCode)) {
break;
}
}
spanCondition = USET_SPAN_NOT_CONTAINED;
}
src += spanLength;
length -= spanLength;
}
}
UnicodeString &
FilteredNormalizer2::normalizeSecondAndAppend(UnicodeString &first,
const UnicodeString &second,
@ -196,6 +244,31 @@ FilteredNormalizer2::isNormalized(const UnicodeString &s, UErrorCode &errorCode)
return TRUE;
}
UBool
FilteredNormalizer2::isNormalizedUTF8(StringPiece sp, UErrorCode &errorCode) const {
if(U_FAILURE(errorCode)) {
return FALSE;
}
const char *s = sp.data();
int32_t length = sp.length();
USetSpanCondition spanCondition = USET_SPAN_SIMPLE;
while (length > 0) {
int32_t spanLength = set.spanUTF8(s, length, spanCondition);
if (spanCondition == USET_SPAN_NOT_CONTAINED) {
spanCondition = USET_SPAN_SIMPLE;
} else {
if (!norm2.isNormalizedUTF8(StringPiece(s, spanLength), errorCode) ||
U_FAILURE(errorCode)) {
return FALSE;
}
spanCondition = USET_SPAN_NOT_CONTAINED;
}
s += spanLength;
length -= spanLength;
}
return TRUE;
}
UNormalizationCheckResult
FilteredNormalizer2::quickCheck(const UnicodeString &s, UErrorCode &errorCode) const {
uprv_checkCanGetBuffer(s, errorCode);

4
icu4c/source/common/loadednormalizer2impl.cpp
View file

@ -62,7 +62,7 @@ LoadedNormalizer2Impl::isAcceptable(void * /*context*/,
pInfo->dataFormat[1]==0x72 &&
pInfo->dataFormat[2]==0x6d &&
pInfo->dataFormat[3]==0x32 &&
pInfo->formatVersion[0]==2
pInfo->formatVersion[0]==3
) {
// Normalizer2Impl *me=(Normalizer2Impl *)context;
// uprv_memcpy(me->dataVersion, pInfo->dataVersion, 4);
@ -84,7 +84,7 @@ LoadedNormalizer2Impl::load(const char *packageName, const char *name, UErrorCod
const uint8_t *inBytes=(const uint8_t *)udata_getMemory(memory);
const int32_t *inIndexes=(const int32_t *)inBytes;
int32_t indexesLength=inIndexes[IX_NORM_TRIE_OFFSET]/4;
if(indexesLength<=IX_MIN_MAYBE_YES) {
if(indexesLength<=IX_MIN_LCCC_CP) {
errorCode=U_INVALID_FORMAT_ERROR; // Not enough indexes.
return;
}

6
icu4c/source/common/locavailable.cpp
View file

@ -35,7 +35,7 @@ U_NAMESPACE_BEGIN
static icu::Locale* availableLocaleList = NULL;
static int32_t availableLocaleListCount;
static icu::UInitOnce gInitOnce = U_INITONCE_INITIALIZER;
static icu::UInitOnce gInitOnceLocale = U_INITONCE_INITIALIZER;
U_NAMESPACE_END
@ -50,7 +50,7 @@ static UBool U_CALLCONV locale_available_cleanup(void)
availableLocaleList = NULL;
}
availableLocaleListCount = 0;
gInitOnce.reset();
gInitOnceLocale.reset();
return TRUE;
}
@ -81,7 +81,7 @@ void U_CALLCONV locale_available_init() {
const Locale* U_EXPORT2
Locale::getAvailableLocales(int32_t& count)
{
umtx_initOnce(gInitOnce, &locale_available_init);
umtx_initOnce(gInitOnceLocale, &locale_available_init);
count = availableLocaleListCount;
return availableLocaleList;
}

68
icu4c/source/common/locmap.cpp
View file

@ -190,7 +190,10 @@ ILCID_POSIX_ELEMENT_ARRAY(0x0423, be, be_BY)
ILCID_POSIX_ELEMENT_ARRAY(0x0402, bg, bg_BG)
ILCID_POSIX_ELEMENT_ARRAY(0x0466, bin, bin_NG)
ILCID_POSIX_SUBTABLE(bin) {
{0x66, "bin"},
{0x0466, "bin_NG"}
};
ILCID_POSIX_SUBTABLE(bn) {
{0x45, "bn"},
@ -214,7 +217,13 @@ ILCID_POSIX_SUBTABLE(ca) {
};
ILCID_POSIX_ELEMENT_ARRAY(0x0483, co, co_FR)
ILCID_POSIX_ELEMENT_ARRAY(0x045c, chr,chr_US)
ILCID_POSIX_SUBTABLE(chr) {
{0x05c, "chr"},
{0x7c5c, "chr_Cher"},
{0x045c, "chr_Cher_US"},
{0x045c, "chr_US"}
};
// ICU has chosen different names for these.
ILCID_POSIX_SUBTABLE(ckb) {
@ -263,10 +272,10 @@ ILCID_POSIX_SUBTABLE(en) {
{0x2C09, "en_TT"},
{0x0409, "en_US"},
{0x007f, "en_US_POSIX"}, /* duplicate for round-tripping */
{0x2409, "en_VI"}, /* Virgin Islands AKA Caribbean Islands (en_CB). On Windows8+ This is 0x1000 or dynamically assigned */
{0x2409, "en_029"},
{0x1c09, "en_ZA"},
{0x3009, "en_ZW"},
{0x2409, "en_029"},
{0x2409, "en_VI"}, /* Virgin Islands AKA Caribbean Islands (en_CB). On Windows8+ This is 0x1000 or dynamically assigned */
{0x0409, "en_AS"}, /* Alias for en_US. Leave last. On Windows8+ This is 0x1000 or dynamically assigned */
{0x0409, "en_GU"}, /* Alias for en_US. Leave last. On Windows8+ This is 0x1000 or dynamically assigned */
{0x0409, "en_MH"}, /* Alias for en_US. Leave last. On Windows8+ This is 0x1000 or dynamically assigned */
@ -419,7 +428,12 @@ ILCID_POSIX_SUBTABLE(hsb) {
ILCID_POSIX_ELEMENT_ARRAY(0x040e, hu, hu_HU)
ILCID_POSIX_ELEMENT_ARRAY(0x042b, hy, hy_AM)
ILCID_POSIX_ELEMENT_ARRAY(0x0469, ibb, ibb_NG)
ILCID_POSIX_SUBTABLE(ibb) {
{0x69, "ibb"},
{0x0469, "ibb_NG"}
};
ILCID_POSIX_ELEMENT_ARRAY(0x0421, id, id_ID)
ILCID_POSIX_ELEMENT_ARRAY(0x0470, ig, ig_NG)
ILCID_POSIX_ELEMENT_ARRAY(0x0478, ii, ii_CN)
@ -458,13 +472,18 @@ ILCID_POSIX_ELEMENT_ARRAY(0x0471, kr, kr_NG)
ILCID_POSIX_SUBTABLE(ks) { /* We could add PK and CN too */
{0x60, "ks"},
{0x0860, "ks_IN"}, /* Documentation doesn't mention script */
{0x0460, "ks_Arab_IN"},
{0x0860, "ks_Deva_IN"}
};
ILCID_POSIX_ELEMENT_ARRAY(0x0440, ky, ky_KG) /* Kyrgyz is spoken in Kyrgyzstan */
ILCID_POSIX_ELEMENT_ARRAY(0x0476, la, la_IT) /* TODO: Verify the country */
ILCID_POSIX_SUBTABLE(la) {
{0x76, "la"},
{0x0476, "la_001"},
{0x0476, "la_IT"} /*Left in for compatibility*/
};
ILCID_POSIX_ELEMENT_ARRAY(0x046e, lb, lb_LU)
ILCID_POSIX_ELEMENT_ARRAY(0x0454, lo, lo_LA)
ILCID_POSIX_ELEMENT_ARRAY(0x0427, lt, lt_LT)
@ -535,15 +554,19 @@ ILCID_POSIX_SUBTABLE(or_IN) {
{0x0448, "or_IN"},
};
ILCID_POSIX_SUBTABLE(pa) {
{0x46, "pa"},
{0x0446, "pa_IN"},
{0x0846, "pa_PK"},
{0x0846, "pa_Arab_PK"}
{0x0846, "pa_Arab_PK"},
{0x0846, "pa_PK"}
};
ILCID_POSIX_SUBTABLE(pap) {
{0x79, "pap"},
{0x0479, "pap_029"},
{0x0479, "pap_AN"} /*Left in for compatibility*/
};
ILCID_POSIX_ELEMENT_ARRAY(0x0479, pap, pap_AN)
ILCID_POSIX_ELEMENT_ARRAY(0x0415, pl, pl_PL)
ILCID_POSIX_ELEMENT_ARRAY(0x0463, ps, ps_AF)
@ -619,9 +642,11 @@ ILCID_POSIX_ELEMENT_ARRAY(0x0485, sah,sah_RU)
ILCID_POSIX_SUBTABLE(sd) {
{0x59, "sd"},
{0x0459, "sd_IN"},
{0x0459, "sd_Deva_IN"},
{0x0859, "sd_PK"}
{0x0459, "sd_IN"},
{0x0859, "sd_Arab_PK"},
{0x0859, "sd_PK"},
{0x7c59, "sd_Arab"}
};
ILCID_POSIX_SUBTABLE(se) {
@ -645,9 +670,8 @@ ILCID_POSIX_ELEMENT_ARRAY(0x045b, si, si_LK)
ILCID_POSIX_ELEMENT_ARRAY(0x041b, sk, sk_SK)
ILCID_POSIX_ELEMENT_ARRAY(0x0424, sl, sl_SI)
ILCID_POSIX_SUBTABLE(so) { /* TODO: Verify the country */
ILCID_POSIX_SUBTABLE(so) {
{0x77, "so"},
{0x0477, "so_ET"},
{0x0477, "so_SO"}
};
@ -739,7 +763,12 @@ ILCID_POSIX_SUBTABLE(ve) { /* TODO: Verify the country */
ILCID_POSIX_ELEMENT_ARRAY(0x042a, vi, vi_VN)
ILCID_POSIX_ELEMENT_ARRAY(0x0488, wo, wo_SN)
ILCID_POSIX_ELEMENT_ARRAY(0x0434, xh, xh_ZA)
ILCID_POSIX_ELEMENT_ARRAY(0x043d, yi, yi)
ILCID_POSIX_SUBTABLE(yi) {
{0x003d, "yi"},
{0x043d, "yi_001"}
};
ILCID_POSIX_ELEMENT_ARRAY(0x046a, yo, yo_NG)
// Windows & ICU tend to different names for some of these
@ -1033,6 +1062,8 @@ uprv_convertToPosix(uint32_t hostid, char *posixID, int32_t posixIDCapacity, UEr
const char *pPosixID = NULL;
#ifdef USE_WINDOWS_LCID_MAPPING_API
char locName[LOCALE_NAME_MAX_LENGTH] = {}; // ICU name can't be longer than Windows name
// Note: Windows primary lang ID 0x92 in LCID is used for Central Kurdish and
// GetLocaleInfo() maps such LCID to "ku". However, CLDR uses "ku" for
// Northern Kurdish and "ckb" for Central Kurdish. For this reason, we cannot
@ -1040,7 +1071,6 @@ uprv_convertToPosix(uint32_t hostid, char *posixID, int32_t posixIDCapacity, UEr
if ((hostid & 0x3FF) != 0x92) {
int32_t tmpLen = 0;
UChar windowsLocaleName[LOCALE_NAME_MAX_LENGTH]; // ULOC_FULLNAME_CAPACITY > LOCALE_NAME_MAX_LENGTH
char locName[LOCALE_NAME_MAX_LENGTH]; // ICU name can't be longer than Windows name
// Note: LOCALE_ALLOW_NEUTRAL_NAMES was enabled in Windows7+, prior versions did not handle neutral (no-region) locale names.
tmpLen = LCIDToLocaleName(hostid, (PWSTR)windowsLocaleName, UPRV_LENGTHOF(windowsLocaleName), LOCALE_ALLOW_NEUTRAL_NAMES);
@ -1102,7 +1132,7 @@ uprv_convertToPosix(uint32_t hostid, char *posixID, int32_t posixIDCapacity, UEr
}
if (pPosixID) {
int32_t resLen = uprv_strlen(pPosixID);
int32_t resLen = static_cast<int32_t>(uprv_strlen(pPosixID));
int32_t copyLen = resLen <= posixIDCapacity ? resLen : posixIDCapacity;
uprv_memcpy(posixID, pPosixID, copyLen);
if (resLen < posixIDCapacity) {
@ -1176,7 +1206,7 @@ uprv_convertToLCIDPlatform(const char* localeID)
char asciiBCP47Tag[LOCALE_NAME_MAX_LENGTH] = {};
// this will change it from de_DE@collation=phonebook to de-DE-u-co-phonebk form
int32_t bcp47Len = uloc_toLanguageTag(mylocaleID, asciiBCP47Tag, UPRV_LENGTHOF(asciiBCP47Tag), FALSE, &myStatus);
(void)uloc_toLanguageTag(mylocaleID, asciiBCP47Tag, UPRV_LENGTHOF(asciiBCP47Tag), FALSE, &myStatus);
if (U_SUCCESS(myStatus))
{

1736
icu4c/source/common/norm2_nfc_data.h
View file

File diff suppressed because it is too large Load diff

54
icu4c/source/common/norm2allmodes.h
View file

@ -5,7 +5,7 @@
* Copyright (C) 2014, International Business Machines
* Corporation and others. All Rights Reserved.
*******************************************************************************
* loadednormalizer2impl.h
* norm2allmodes.h
*
* created on: 2014sep07
* created by: Markus W. Scherer
@ -18,10 +18,12 @@
#if !UCONFIG_NO_NORMALIZATION
#include "unicode/edits.h"
#include "unicode/normalizer2.h"
#include "unicode/unistr.h"
#include "cpputils.h"
#include "normalizer2impl.h"
#include "ustr_imp.h" // U_EDITS_NO_RESET
U_NAMESPACE_BEGIN
@ -210,8 +212,8 @@ private:
virtual UNormalizationCheckResult getQuickCheck(UChar32 c) const {
return impl.isDecompYes(impl.getNorm16(c)) ? UNORM_YES : UNORM_NO;
}
virtual UBool hasBoundaryBefore(UChar32 c) const { return impl.hasDecompBoundary(c, TRUE); }
virtual UBool hasBoundaryAfter(UChar32 c) const { return impl.hasDecompBoundary(c, FALSE); }
virtual UBool hasBoundaryBefore(UChar32 c) const { return impl.hasDecompBoundaryBefore(c); }
virtual UBool hasBoundaryAfter(UChar32 c) const { return impl.hasDecompBoundaryAfter(c); }
virtual UBool isInert(UChar32 c) const { return impl.isDecompInert(c); }
};
@ -224,19 +226,35 @@ public:
private:
virtual void
normalize(const UChar *src, const UChar *limit,
ReorderingBuffer &buffer, UErrorCode &errorCode) const {
ReorderingBuffer &buffer, UErrorCode &errorCode) const override {
impl.compose(src, limit, onlyContiguous, TRUE, buffer, errorCode);
}
using Normalizer2WithImpl::normalize; // Avoid warning about hiding base class function.
void
normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
Edits *edits, UErrorCode &errorCode) const override {
if (U_FAILURE(errorCode)) {
return;
}
if (edits != nullptr && (options & U_EDITS_NO_RESET) == 0) {
edits->reset();
}
const uint8_t *s = reinterpret_cast<const uint8_t *>(src.data());
impl.composeUTF8(options, onlyContiguous, s, s + src.length(),
&sink, edits, errorCode);
sink.Flush();
}
virtual void
normalizeAndAppend(const UChar *src, const UChar *limit, UBool doNormalize,
UnicodeString &safeMiddle,
ReorderingBuffer &buffer, UErrorCode &errorCode) const {
ReorderingBuffer &buffer, UErrorCode &errorCode) const override {
impl.composeAndAppend(src, limit, doNormalize, onlyContiguous, safeMiddle, buffer, errorCode);
}
virtual UBool
isNormalized(const UnicodeString &s, UErrorCode &errorCode) const {
isNormalized(const UnicodeString &s, UErrorCode &errorCode) const override {
if(U_FAILURE(errorCode)) {
return FALSE;
}
@ -252,8 +270,16 @@ private:
}
return impl.compose(sArray, sArray+s.length(), onlyContiguous, FALSE, buffer, errorCode);
}
virtual UBool
isNormalizedUTF8(StringPiece sp, UErrorCode &errorCode) const override {
if(U_FAILURE(errorCode)) {
return FALSE;
}
const uint8_t *s = reinterpret_cast<const uint8_t *>(sp.data());
return impl.composeUTF8(0, onlyContiguous, s, s + sp.length(), nullptr, nullptr, errorCode);
}
virtual UNormalizationCheckResult
quickCheck(const UnicodeString &s, UErrorCode &errorCode) const {
quickCheck(const UnicodeString &s, UErrorCode &errorCode) const override {
if(U_FAILURE(errorCode)) {
return UNORM_MAYBE;
}
@ -267,21 +293,21 @@ private:
return qcResult;
}
virtual const UChar *
spanQuickCheckYes(const UChar *src, const UChar *limit, UErrorCode &) const {
spanQuickCheckYes(const UChar *src, const UChar *limit, UErrorCode &) const override {
return impl.composeQuickCheck(src, limit, onlyContiguous, NULL);
}
using Normalizer2WithImpl::spanQuickCheckYes; // Avoid warning about hiding base class function.
virtual UNormalizationCheckResult getQuickCheck(UChar32 c) const {
virtual UNormalizationCheckResult getQuickCheck(UChar32 c) const override {
return impl.getCompQuickCheck(impl.getNorm16(c));
}
virtual UBool hasBoundaryBefore(UChar32 c) const {
virtual UBool hasBoundaryBefore(UChar32 c) const override {
return impl.hasCompBoundaryBefore(c);
}
virtual UBool hasBoundaryAfter(UChar32 c) const {
return impl.hasCompBoundaryAfter(c, onlyContiguous, FALSE);
virtual UBool hasBoundaryAfter(UChar32 c) const override {
return impl.hasCompBoundaryAfter(c, onlyContiguous);
}
virtual UBool isInert(UChar32 c) const {
return impl.hasCompBoundaryAfter(c, onlyContiguous, TRUE);
virtual UBool isInert(UChar32 c) const override {
return impl.isCompInert(c, onlyContiguous);
}
const UBool onlyContiguous;

64
icu4c/source/common/normalizer2.cpp
View file

@ -20,6 +20,7 @@
#if !UCONFIG_NO_NORMALIZATION
#include "unicode/edits.h"
#include "unicode/normalizer2.h"
#include "unicode/unistr.h"
#include "unicode/unorm.h"
@ -29,6 +30,7 @@
#include "normalizer2impl.h"
#include "uassert.h"
#include "ucln_cmn.h"
#include "ustr_imp.h" // U_EDITS_NO_RESET
using icu::Normalizer2Impl;
@ -42,6 +44,20 @@ U_NAMESPACE_BEGIN
Normalizer2::~Normalizer2() {}
void
Normalizer2::normalizeUTF8(uint32_t /*options*/, StringPiece src, ByteSink &sink,
Edits *edits, UErrorCode &errorCode) const {
if (U_FAILURE(errorCode)) {
return;
}
if (edits != nullptr) {
errorCode = U_UNSUPPORTED_ERROR;
return;
}
UnicodeString src16 = UnicodeString::fromUTF8(src);
normalize(src16, errorCode).toUTF8(sink);
}
UBool
Normalizer2::getRawDecomposition(UChar32, UnicodeString &) const {
return FALSE;
@ -57,6 +73,11 @@ Normalizer2::getCombiningClass(UChar32 /*c*/) const {
return 0;
}
UBool
Normalizer2::isNormalizedUTF8(StringPiece s, UErrorCode &errorCode) const {
return U_SUCCESS(errorCode) && isNormalized(UnicodeString::fromUTF8(s), errorCode);
}
// Normalizer2 implementation for the old UNORM_NONE.
class NoopNormalizer2 : public Normalizer2 {
virtual ~NoopNormalizer2();
@ -64,7 +85,7 @@ class NoopNormalizer2 : public Normalizer2 {
virtual UnicodeString &
normalize(const UnicodeString &src,
UnicodeString &dest,
UErrorCode &errorCode) const {
UErrorCode &errorCode) const override {
if(U_SUCCESS(errorCode)) {
if(&dest!=&src) {
dest=src;
@ -74,10 +95,27 @@ class NoopNormalizer2 : public Normalizer2 {
}
return dest;
}
virtual void
normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
Edits *edits, UErrorCode &errorCode) const override {
if(U_SUCCESS(errorCode)) {
if (edits != nullptr) {
if ((options & U_EDITS_NO_RESET) == 0) {
edits->reset();
}
edits->addUnchanged(src.length());
}
if ((options & U_OMIT_UNCHANGED_TEXT) == 0) {
sink.Append(src.data(), src.length());
}
sink.Flush();
}
}
virtual UnicodeString &
normalizeSecondAndAppend(UnicodeString &first,
const UnicodeString &second,
UErrorCode &errorCode) const {
UErrorCode &errorCode) const override {
if(U_SUCCESS(errorCode)) {
if(&first!=&second) {
first.append(second);
@ -90,7 +128,7 @@ class NoopNormalizer2 : public Normalizer2 {
virtual UnicodeString &
append(UnicodeString &first,
const UnicodeString &second,
UErrorCode &errorCode) const {
UErrorCode &errorCode) const override {
if(U_SUCCESS(errorCode)) {
if(&first!=&second) {
first.append(second);
@ -101,25 +139,29 @@ class NoopNormalizer2 : public Normalizer2 {
return first;
}
virtual UBool
getDecomposition(UChar32, UnicodeString &) const {
getDecomposition(UChar32, UnicodeString &) const override {
return FALSE;
}
// No need to override the default getRawDecomposition().
virtual UBool
isNormalized(const UnicodeString &, UErrorCode &) const {
return TRUE;
isNormalized(const UnicodeString &, UErrorCode &errorCode) const override {
return U_SUCCESS(errorCode);
}
virtual UBool
isNormalizedUTF8(StringPiece, UErrorCode &errorCode) const override {
return U_SUCCESS(errorCode);
}
virtual UNormalizationCheckResult
quickCheck(const UnicodeString &, UErrorCode &) const {
quickCheck(const UnicodeString &, UErrorCode &) const override {
return UNORM_YES;
}
virtual int32_t
spanQuickCheckYes(const UnicodeString &s, UErrorCode &) const {
spanQuickCheckYes(const UnicodeString &s, UErrorCode &) const override {
return s.length();
}
virtual UBool hasBoundaryBefore(UChar32) const { return TRUE; }
virtual UBool hasBoundaryAfter(UChar32) const { return TRUE; }
virtual UBool isInert(UChar32) const { return TRUE; }
virtual UBool hasBoundaryBefore(UChar32) const override { return TRUE; }
virtual UBool hasBoundaryAfter(UChar32) const override { return TRUE; }
virtual UBool isInert(UChar32) const override { return TRUE; }
};
NoopNormalizer2::~NoopNormalizer2() {}

1770
icu4c/source/common/normalizer2impl.cpp
View file

File diff suppressed because it is too large Load diff

302
icu4c/source/common/normalizer2impl.h
View file

@ -35,6 +35,11 @@ U_NAMESPACE_BEGIN
struct CanonIterData;
class ByteSink;
class Edits;
class InitCanonIterData;
class LcccContext;
class U_COMMON_API Hangul {
public:
/* Korean Hangul and Jamo constants */
@ -63,9 +68,9 @@ public:
return HANGUL_BASE<=c && c<HANGUL_LIMIT;
}
static inline UBool
isHangulWithoutJamoT(UChar c) {
isHangulLV(UChar32 c) {
c-=HANGUL_BASE;
return c<HANGUL_COUNT && c%JAMO_T_COUNT==0;
return 0<=c && c<HANGUL_COUNT && c%JAMO_T_COUNT==0;
}
static inline UBool isJamoL(UChar32 c) {
return (uint32_t)(c-JAMO_L_BASE)<JAMO_L_COUNT;
@ -73,6 +78,14 @@ public:
static inline UBool isJamoV(UChar32 c) {
return (uint32_t)(c-JAMO_V_BASE)<JAMO_V_COUNT;
}
static inline UBool isJamoT(UChar32 c) {
int32_t t=c-JAMO_T_BASE;
return 0<t && t<JAMO_T_COUNT; // not JAMO_T_BASE itself
}
static UBool isJamo(UChar32 c) {
return JAMO_L_BASE<=c && c<=JAMO_T_END &&
(c<=JAMO_L_END || (JAMO_V_BASE<=c && c<=JAMO_V_END) || JAMO_T_BASE<c);
}
/**
* Decomposes c, which must be a Hangul syllable, into buffer
@ -117,10 +130,13 @@ class Normalizer2Impl;
class U_COMMON_API ReorderingBuffer : public UMemory {
public:
/** Constructs only; init() should be called. */
ReorderingBuffer(const Normalizer2Impl &ni, UnicodeString &dest) :
impl(ni), str(dest),
start(NULL), reorderStart(NULL), limit(NULL),
remainingCapacity(0), lastCC(0) {}
/** Constructs, removes the string contents, and initializes for a small initial capacity. */
ReorderingBuffer(const Normalizer2Impl &ni, UnicodeString &dest, UErrorCode &errorCode);
~ReorderingBuffer() {
if(start!=NULL) {
str.releaseBuffer((int32_t)(limit-start));
@ -135,11 +151,7 @@ public:
uint8_t getLastCC() const { return lastCC; }
UBool equals(const UChar *start, const UChar *limit) const;
// For Hangul composition, replacing the Leading consonant Jamo with the syllable.
void setLastChar(UChar c) {
*(limit-1)=c;
}
UBool equals(const uint8_t *otherStart, const uint8_t *otherLimit) const;
UBool append(UChar32 c, uint8_t cc, UErrorCode &errorCode) {
return (c<=0xffff) ?
@ -218,6 +230,12 @@ private:
UChar *codePointStart, *codePointLimit;
};
/**
* Low-level implementation of the Unicode Normalization Algorithm.
* For the data structure and details see the documentation at the end of
* this normalizer2impl.h and in the design doc at
* http://site.icu-project.org/design/normalization/custom
*/
class U_COMMON_API Normalizer2Impl : public UObject {
public:
Normalizer2Impl() : normTrie(NULL), fCanonIterData(NULL) {
@ -234,8 +252,6 @@ public:
// low-level properties ------------------------------------------------ ***
const UTrie2 *getNormTrie() const { return normTrie; }
UBool ensureCanonIterData(UErrorCode &errorCode) const;
uint16_t getNorm16(UChar32 c) const { return UTRIE2_GET16(normTrie, c); }
@ -255,15 +271,22 @@ public:
uint8_t getCC(uint16_t norm16) const {
if(norm16>=MIN_NORMAL_MAYBE_YES) {
return (uint8_t)norm16;
return getCCFromNormalYesOrMaybe(norm16);
}
if(norm16<minNoNo || limitNoNo<=norm16) {
return 0;
}
return getCCFromNoNo(norm16);
}
static uint8_t getCCFromNormalYesOrMaybe(uint16_t norm16) {
return (uint8_t)(norm16 >> OFFSET_SHIFT);
}
static uint8_t getCCFromYesOrMaybe(uint16_t norm16) {
return norm16>=MIN_NORMAL_MAYBE_YES ? (uint8_t)norm16 : 0;
return norm16>=MIN_NORMAL_MAYBE_YES ? getCCFromNormalYesOrMaybe(norm16) : 0;
}
uint8_t getCCFromYesOrMaybeCP(UChar32 c) const {
if (c < minCompNoMaybeCP) { return 0; }
return getCCFromYesOrMaybe(getNorm16(c));
}
/**
@ -272,10 +295,8 @@ public:
* @return The lccc(c) in bits 15..8 and tccc(c) in bits 7..0.
*/
uint16_t getFCD16(UChar32 c) const {
if(c<0) {
if(c<minDecompNoCP) {
return 0;
} else if(c<0x180) {
return tccc180[c];
} else if(c<=0xffff) {
if(!singleLeadMightHaveNonZeroFCD16(c)) { return 0; }
}
@ -291,9 +312,7 @@ public:
*/
uint16_t nextFCD16(const UChar *&s, const UChar *limit) const {
UChar32 c=*s++;
if(c<0x180) {
return tccc180[c];
} else if(!singleLeadMightHaveNonZeroFCD16(c)) {
if(c<minDecompNoCP || !singleLeadMightHaveNonZeroFCD16(c)) {
return 0;
}
UChar c2;
@ -311,8 +330,8 @@ public:
*/
uint16_t previousFCD16(const UChar *start, const UChar *&s) const {
UChar32 c=*--s;
if(c<0x180) {
return tccc180[c];
if(c<minDecompNoCP) {
return 0;
}
if(!U16_IS_TRAIL(c)) {
if(!singleLeadMightHaveNonZeroFCD16(c)) {
@ -328,8 +347,6 @@ public:
return getFCD16FromNormData(c);
}
/** Returns the FCD data for U+0000<=c<U+0180. */
uint16_t getFCD16FromBelow180(UChar32 c) const { return tccc180[c]; }
/** Returns TRUE if the single-or-lead code unit c might have non-zero FCD data. */
UBool singleLeadMightHaveNonZeroFCD16(UChar32 lead) const {
// 0<=lead<=0xffff
@ -340,9 +357,6 @@ public:
/** Returns the FCD value from the regular normalization data. */
uint16_t getFCD16FromNormData(UChar32 c) const;
void makeCanonIterDataFromNorm16(UChar32 start, UChar32 end, uint16_t norm16,
CanonIterData &newData, UErrorCode &errorCode) const;
/**
* Gets the decomposition for one code point.
* @param c code point
@ -367,14 +381,25 @@ public:
UBool getCanonStartSet(UChar32 c, UnicodeSet &set) const;
enum {
MIN_CCC_LCCC_CP=0x300
};
// Fixed norm16 values.
MIN_YES_YES_WITH_CC=0xfe02,
JAMO_VT=0xfe00,
MIN_NORMAL_MAYBE_YES=0xfc00,
JAMO_L=2, // offset=1 hasCompBoundaryAfter=FALSE
INERT=1, // offset=0 hasCompBoundaryAfter=TRUE
// norm16 bit 0 is comp-boundary-after.
HAS_COMP_BOUNDARY_AFTER=1,
OFFSET_SHIFT=1,
// For algorithmic one-way mappings, norm16 bits 2..1 indicate the
// tccc (0, 1, >1) for quick FCC boundary-after tests.
DELTA_TCCC_0=0,
DELTA_TCCC_1=2,
DELTA_TCCC_GT_1=4,
DELTA_TCCC_MASK=6,
DELTA_SHIFT=3,
enum {
MIN_YES_YES_WITH_CC=0xff01,
JAMO_VT=0xff00,
MIN_NORMAL_MAYBE_YES=0xfe00,
JAMO_L=1,
MAX_DELTA=0x40
};
@ -394,21 +419,32 @@ public:
IX_MIN_COMP_NO_MAYBE_CP,
// Norm16 value thresholds for quick check combinations and types of extra data.
IX_MIN_YES_NO, // Mappings & compositions in [minYesNo..minYesNoMappingsOnly[.
/** Mappings & compositions in [minYesNo..minYesNoMappingsOnly[. */
IX_MIN_YES_NO,
/** Mappings are comp-normalized. */
IX_MIN_NO_NO,
IX_LIMIT_NO_NO,
IX_MIN_MAYBE_YES,
IX_MIN_YES_NO_MAPPINGS_ONLY, // Mappings only in [minYesNoMappingsOnly..minNoNo[.
/** Mappings only in [minYesNoMappingsOnly..minNoNo[. */
IX_MIN_YES_NO_MAPPINGS_ONLY,
/** Mappings are not comp-normalized but have a comp boundary before. */
IX_MIN_NO_NO_COMP_BOUNDARY_BEFORE,
/** Mappings do not have a comp boundary before. */
IX_MIN_NO_NO_COMP_NO_MAYBE_CC,
/** Mappings to the empty string. */
IX_MIN_NO_NO_EMPTY,
IX_RESERVED15,
IX_MIN_LCCC_CP,
IX_RESERVED19,
IX_COUNT
};
enum {
MAPPING_HAS_CCC_LCCC_WORD=0x80,
MAPPING_HAS_RAW_MAPPING=0x40,
MAPPING_NO_COMP_BOUNDARY_AFTER=0x20,
// unused bit 0x20,
MAPPING_LENGTH_MASK=0x1f
};
@ -457,6 +493,12 @@ public:
UnicodeString &safeMiddle,
ReorderingBuffer &buffer,
UErrorCode &errorCode) const;
/** sink==nullptr: isNormalized() */
UBool composeUTF8(uint32_t options, UBool onlyContiguous,
const uint8_t *src, const uint8_t *limit,
ByteSink *sink, icu::Edits *edits, UErrorCode &errorCode) const;
const UChar *makeFCD(const UChar *src, const UChar *limit,
ReorderingBuffer *buffer, UErrorCode &errorCode) const;
void makeFCDAndAppend(const UChar *src, const UChar *limit,
@ -465,27 +507,42 @@ public:
ReorderingBuffer &buffer,
UErrorCode &errorCode) const;
UBool hasDecompBoundary(UChar32 c, UBool before) const;
UBool hasDecompBoundaryBefore(UChar32 c) const;
UBool norm16HasDecompBoundaryBefore(uint16_t norm16) const;
UBool hasDecompBoundaryAfter(UChar32 c) const;
UBool norm16HasDecompBoundaryAfter(uint16_t norm16) const;
UBool isDecompInert(UChar32 c) const { return isDecompYesAndZeroCC(getNorm16(c)); }
UBool hasCompBoundaryBefore(UChar32 c) const {
return c<minCompNoMaybeCP || hasCompBoundaryBefore(c, getNorm16(c));
return c<minCompNoMaybeCP || norm16HasCompBoundaryBefore(getNorm16(c));
}
UBool hasCompBoundaryAfter(UChar32 c, UBool onlyContiguous) const {
return norm16HasCompBoundaryAfter(getNorm16(c), onlyContiguous);
}
UBool isCompInert(UChar32 c, UBool onlyContiguous) const {
uint16_t norm16=getNorm16(c);
return isCompYesAndZeroCC(norm16) &&
(norm16 & HAS_COMP_BOUNDARY_AFTER) != 0 &&
(!onlyContiguous || isInert(norm16) || *getMapping(norm16) <= 0x1ff);
}
UBool hasCompBoundaryAfter(UChar32 c, UBool onlyContiguous, UBool testInert) const;
UBool hasFCDBoundaryBefore(UChar32 c) const { return c<MIN_CCC_LCCC_CP || getFCD16(c)<=0xff; }
UBool hasFCDBoundaryAfter(UChar32 c) const {
uint16_t fcd16=getFCD16(c);
return fcd16<=1 || (fcd16&0xff)==0;
}
UBool hasFCDBoundaryBefore(UChar32 c) const { return hasDecompBoundaryBefore(c); }
UBool hasFCDBoundaryAfter(UChar32 c) const { return hasDecompBoundaryAfter(c); }
UBool isFCDInert(UChar32 c) const { return getFCD16(c)<=1; }
private:
friend class InitCanonIterData;
friend class LcccContext;
UBool isMaybe(uint16_t norm16) const { return minMaybeYes<=norm16 && norm16<=JAMO_VT; }
UBool isMaybeOrNonZeroCC(uint16_t norm16) const { return norm16>=minMaybeYes; }
static UBool isInert(uint16_t norm16) { return norm16==0; }
static UBool isJamoL(uint16_t norm16) { return norm16==1; }
static UBool isInert(uint16_t norm16) { return norm16==INERT; }
static UBool isJamoL(uint16_t norm16) { return norm16==JAMO_L; }
static UBool isJamoVT(uint16_t norm16) { return norm16==JAMO_VT; }
UBool isHangul(uint16_t norm16) const { return norm16==minYesNo; }
uint16_t hangulLVT() const { return minYesNoMappingsOnly|HAS_COMP_BOUNDARY_AFTER; }
UBool isHangulLV(uint16_t norm16) const { return norm16==minYesNo; }
UBool isHangulLVT(uint16_t norm16) const {
return norm16==hangulLVT();
}
UBool isCompYesAndZeroCC(uint16_t norm16) const { return norm16<minNoNo; }
// UBool isCompYes(uint16_t norm16) const {
// return norm16>=MIN_YES_YES_WITH_CC || norm16<minNoNo;
@ -504,7 +561,7 @@ private:
/**
* A little faster and simpler than isDecompYesAndZeroCC() but does not include
* the MaybeYes which combine-forward and have ccc=0.
* (Standard Unicode 5.2 normalization does not have such characters.)
* (Standard Unicode 10 normalization does not have such characters.)
*/
UBool isMostDecompYesAndZeroCC(uint16_t norm16) const {
return norm16<minYesNo || norm16==MIN_NORMAL_MAYBE_YES || norm16==JAMO_VT;
@ -514,7 +571,7 @@ private:
// For use with isCompYes().
// Perhaps the compiler can combine the two tests for MIN_YES_YES_WITH_CC.
// static uint8_t getCCFromYes(uint16_t norm16) {
// return norm16>=MIN_YES_YES_WITH_CC ? (uint8_t)norm16 : 0;
// return norm16>=MIN_YES_YES_WITH_CC ? getCCFromNormalYesOrMaybe(norm16) : 0;
// }
uint8_t getCCFromNoNo(uint16_t norm16) const {
const uint16_t *mapping=getMapping(norm16);
@ -525,30 +582,47 @@ private:
}
}
// requires that the [cpStart..cpLimit[ character passes isCompYesAndZeroCC()
uint8_t getTrailCCFromCompYesAndZeroCC(const UChar *cpStart, const UChar *cpLimit) const;
uint8_t getTrailCCFromCompYesAndZeroCC(uint16_t norm16) const {
if(norm16<=minYesNo) {
return 0; // yesYes and Hangul LV have ccc=tccc=0
} else {
// For Hangul LVT we harmlessly fetch a firstUnit with tccc=0 here.
return (uint8_t)(*getMapping(norm16)>>8); // tccc from yesNo
}
}
uint8_t getPreviousTrailCC(const UChar *start, const UChar *p) const;
uint8_t getPreviousTrailCC(const uint8_t *start, const uint8_t *p) const;
// Requires algorithmic-NoNo.
UChar32 mapAlgorithmic(UChar32 c, uint16_t norm16) const {
return c+norm16-(minMaybeYes-MAX_DELTA-1);
return c+(norm16>>DELTA_SHIFT)-centerNoNoDelta;
}
UChar32 getAlgorithmicDelta(uint16_t norm16) const {
return (norm16>>DELTA_SHIFT)-centerNoNoDelta;
}
// Requires minYesNo<norm16<limitNoNo.
const uint16_t *getMapping(uint16_t norm16) const { return extraData+norm16; }
const uint16_t *getMapping(uint16_t norm16) const { return extraData+(norm16>>OFFSET_SHIFT); }
const uint16_t *getCompositionsListForDecompYes(uint16_t norm16) const {
if(norm16==0 || MIN_NORMAL_MAYBE_YES<=norm16) {
if(norm16<JAMO_L || MIN_NORMAL_MAYBE_YES<=norm16) {
return NULL;
} else if(norm16<minMaybeYes) {
return extraData+norm16; // for yesYes; if Jamo L: harmless empty list
return getMapping(norm16); // for yesYes; if Jamo L: harmless empty list
} else {
return maybeYesCompositions+norm16-minMaybeYes;
}
}
const uint16_t *getCompositionsListForComposite(uint16_t norm16) const {
const uint16_t *list=extraData+norm16; // composite has both mapping & compositions list
// A composite has both mapping & compositions list.
const uint16_t *list=getMapping(norm16);
return list+ // mapping pointer
1+ // +1 to skip the first unit with the mapping lenth
1+ // +1 to skip the first unit with the mapping length
(*list&MAPPING_LENGTH_MASK); // + mapping length
}
const uint16_t *getCompositionsListForMaybe(uint16_t norm16) const {
// minMaybeYes<=norm16<MIN_NORMAL_MAYBE_YES
return maybeYesCompositions+((norm16-minMaybeYes)>>OFFSET_SHIFT);
}
/**
* @param c code point must have compositions
* @return compositions list pointer
@ -563,46 +637,78 @@ private:
UChar32 minNeedDataCP,
ReorderingBuffer *buffer,
UErrorCode &errorCode) const;
UBool decomposeShort(const UChar *src, const UChar *limit,
ReorderingBuffer &buffer, UErrorCode &errorCode) const;
const UChar *decomposeShort(const UChar *src, const UChar *limit,
UBool stopAtCompBoundary, UBool onlyContiguous,
ReorderingBuffer &buffer, UErrorCode &errorCode) const;
UBool decompose(UChar32 c, uint16_t norm16,
ReorderingBuffer &buffer, UErrorCode &errorCode) const;
const uint8_t *decomposeShort(const uint8_t *src, const uint8_t *limit,
UBool stopAtCompBoundary, UBool onlyContiguous,
ReorderingBuffer &buffer, UErrorCode &errorCode) const;
static int32_t combine(const uint16_t *list, UChar32 trail);
void addComposites(const uint16_t *list, UnicodeSet &set) const;
void recompose(ReorderingBuffer &buffer, int32_t recomposeStartIndex,
UBool onlyContiguous) const;
UBool hasCompBoundaryBefore(UChar32 c, uint16_t norm16) const;
const UChar *findPreviousCompBoundary(const UChar *start, const UChar *p) const;
const UChar *findNextCompBoundary(const UChar *p, const UChar *limit) const;
UBool hasCompBoundaryBefore(UChar32 c, uint16_t norm16) const {
return c<minCompNoMaybeCP || norm16HasCompBoundaryBefore(norm16);
}
UBool norm16HasCompBoundaryBefore(uint16_t norm16) const {
return norm16 < minNoNoCompNoMaybeCC || isAlgorithmicNoNo(norm16);
}
UBool hasCompBoundaryBefore(const UChar *src, const UChar *limit) const;
UBool hasCompBoundaryBefore(const uint8_t *src, const uint8_t *limit) const;
UBool hasCompBoundaryAfter(const UChar *start, const UChar *p,
UBool onlyContiguous) const;
UBool hasCompBoundaryAfter(const uint8_t *start, const uint8_t *p,
UBool onlyContiguous) const;
UBool norm16HasCompBoundaryAfter(uint16_t norm16, UBool onlyContiguous) const {
return (norm16 & HAS_COMP_BOUNDARY_AFTER) != 0 &&
(!onlyContiguous || isTrailCC01ForCompBoundaryAfter(norm16));
}
/** For FCC: Given norm16 HAS_COMP_BOUNDARY_AFTER, does it have tccc<=1? */
UBool isTrailCC01ForCompBoundaryAfter(uint16_t norm16) const {
return isInert(norm16) || (isDecompNoAlgorithmic(norm16) ?
(norm16 & DELTA_TCCC_MASK) <= DELTA_TCCC_1 : *getMapping(norm16) <= 0x1ff);
}
const UChar *findPreviousCompBoundary(const UChar *start, const UChar *p, UBool onlyContiguous) const;
const UChar *findNextCompBoundary(const UChar *p, const UChar *limit, UBool onlyContiguous) const;
const UChar *findPreviousFCDBoundary(const UChar *start, const UChar *p) const;
const UChar *findNextFCDBoundary(const UChar *p, const UChar *limit) const;
void makeCanonIterDataFromNorm16(UChar32 start, UChar32 end, const uint16_t norm16,
CanonIterData &newData, UErrorCode &errorCode) const;
int32_t getCanonValue(UChar32 c) const;
const UnicodeSet &getCanonStartSet(int32_t n) const;
// UVersionInfo dataVersion;
// Code point thresholds for quick check codes.
UChar32 minDecompNoCP;
UChar32 minCompNoMaybeCP;
// BMP code point thresholds for quick check loops looking at single UTF-16 code units.
UChar minDecompNoCP;
UChar minCompNoMaybeCP;
UChar minLcccCP;
// Norm16 value thresholds for quick check combinations and types of extra data.
uint16_t minYesNo;
uint16_t minYesNoMappingsOnly;
uint16_t minNoNo;
uint16_t minNoNoCompBoundaryBefore;
uint16_t minNoNoCompNoMaybeCC;
uint16_t minNoNoEmpty;
uint16_t limitNoNo;
uint16_t centerNoNoDelta;
uint16_t minMaybeYes;
const UTrie2 *normTrie;
const uint16_t *maybeYesCompositions;
const uint16_t *extraData; // mappings and/or compositions for yesYes, yesNo & noNo characters
const uint8_t *smallFCD; // [0x100] one bit per 32 BMP code points, set if any FCD!=0
uint8_t tccc180[0x180]; // tccc values for U+0000..U+017F
public: // CanonIterData is public to allow access from C callback functions.
UInitOnce fCanonIterDataInitOnce;
CanonIterData *fCanonIterData;
};
@ -658,13 +764,14 @@ unorm_getFCD16(UChar32 c);
/**
* Format of Normalizer2 .nrm data files.
* Format version 2.0.
* Format version 3.0.
*
* Normalizer2 .nrm data files provide data for the Unicode Normalization algorithms.
* ICU ships with data files for standard Unicode Normalization Forms
* NFC and NFD (nfc.nrm), NFKC and NFKD (nfkc.nrm) and NFKC_Casefold (nfkc_cf.nrm).
* Custom (application-specific) data can be built into additional .nrm files
* with the gennorm2 build tool.
* ICU ships with one such file, uts46.nrm, for the implementation of UTS #46.
*
* Normalizer2.getInstance() causes a .nrm file to be loaded, unless it has been
* cached already. Internally, Normalizer2Impl.load() reads the .nrm file.
@ -695,14 +802,20 @@ unorm_getFCD16(UChar32 c);
* with a decomposition mapping, that is, with NF*D_QC=No.
* minCompNoMaybeCP=indexes[IX_MIN_COMP_NO_MAYBE_CP] is the lowest code point
* with NF*C_QC=No (has a one-way mapping) or Maybe (combines backward).
* minLcccCP=indexes[IX_MIN_LCCC_CP] (index 18, new in formatVersion 3)
* is the lowest code point with lccc!=0.
*
* The next five indexes are thresholds of 16-bit trie values for ranges of
* The next eight indexes are thresholds of 16-bit trie values for ranges of
* values indicating multiple normalization properties.
* They are listed here in threshold order, not in the order they are stored in the indexes.
* minYesNo=indexes[IX_MIN_YES_NO];
* minYesNoMappingsOnly=indexes[IX_MIN_YES_NO_MAPPINGS_ONLY];
* minNoNo=indexes[IX_MIN_NO_NO];
* minNoNoCompBoundaryBefore=indexes[IX_MIN_NO_NO_COMP_BOUNDARY_BEFORE];
* minNoNoCompNoMaybeCC=indexes[IX_MIN_NO_NO_COMP_NO_MAYBE_CC];
* minNoNoEmpty=indexes[IX_MIN_NO_NO_EMPTY];
* limitNoNo=indexes[IX_LIMIT_NO_NO];
* minMaybeYes=indexes[IX_MIN_MAYBE_YES];
* minYesNoMappingsOnly=indexes[IX_MIN_YES_NO_MAPPINGS_ONLY];
* See the normTrie description below and the design doc for details.
*
* UTrie2 normTrie; -- see utrie2_impl.h and utrie2.h
@ -710,12 +823,14 @@ unorm_getFCD16(UChar32 c);
* The trie holds the main normalization data. Each code point is mapped to a 16-bit value.
* Rather than using independent bits in the value (which would require more than 16 bits),
* information is extracted primarily via range checks.
* Except, format version 3 uses bit 0 for hasCompBoundaryAfter().
* For example, a 16-bit value norm16 in the range minYesNo<=norm16<minNoNo
* means that the character has NF*C_QC=Yes and NF*D_QC=No properties,
* which means it has a two-way (round-trip) decomposition mapping.
* Values in the range 2<=norm16<limitNoNo are also directly indexes into the extraData
* pointing to mappings, compositions lists, or both.
* Value norm16==0 means that the character is normalization-inert, that is,
* Value norm16==INERT (0 in versions 1 & 2, 1 in version 3)
* means that the character is normalization-inert, that is,
* it does not have a mapping, does not participate in composition, has a zero
* canonical combining class, and forms a boundary where text before it and after it
* can be normalized independently.
@ -729,7 +844,7 @@ unorm_getFCD16(UChar32 c);
* The trie has a value for each lead surrogate code unit representing the "worst case"
* properties of the 1024 supplementary characters whose UTF-16 form starts with
* the lead surrogate. If all of the 1024 supplementary characters are normalization-inert,
* then their lead surrogate code unit has the trie value 0.
* then their lead surrogate code unit has the trie value INERT.
* When the lead surrogate unit's value exceeds the quick check minimum during processing,
* the properties for the full supplementary code point need to be looked up.
*
@ -738,6 +853,7 @@ unorm_getFCD16(UChar32 c);
*
* There is only one byte offset for the end of these two arrays.
* The split between them is given by the constant and variable mentioned above.
* In version 3, the difference must be shifted right by OFFSET_SHIFT.
*
* The maybeYesCompositions array contains compositions lists for characters that
* combine both forward (as starters in composition pairs)
@ -754,6 +870,8 @@ unorm_getFCD16(UChar32 c);
* followed by only mappings for "NoNo" characters.
* (Referring to pairs of NFC/NFD quick check values.)
* The norm16 values of those characters are directly indexes into the extraData array.
* In version 3, the norm16 values must be shifted right by OFFSET_SHIFT
* for accessing extraData.
*
* The data structures for compositions lists and mappings are described in the design doc.
*
@ -784,6 +902,50 @@ unorm_getFCD16(UChar32 c);
* This is fully equivalent with formatVersion 1's MAPPING_PLUS_COMPOSITION_LIST flag.
* It is needed for the new (in ICU 49) composePair(), not for other normalization.
* - Addition of the smallFCD[] bit set.
*
* Changes from format version 2 to format version 3 (ICU 60) ------------------
*
* - norm16 bit 0 indicates hasCompBoundaryAfter(),
* except that for contiguous composition (FCC) the tccc must be checked as well.
* Data indexes and ccc values are shifted left by one (OFFSET_SHIFT).
* Thresholds like minNoNo are tested before shifting.
*
* - Algorithmic mapping deltas are shifted left by two more bits (total DELTA_SHIFT),
* to make room for two bits (three values) indicating whether the tccc is 0, 1, or greater.
* See DELTA_TCCC_MASK etc.
* This helps with fetching tccc/FCD values and FCC hasCompBoundaryAfter().
* minMaybeYes is 8-aligned so that the DELTA_TCCC_MASK bits can be tested directly.
*
* - Algorithmic mappings are only used for mapping to "comp yes and ccc=0" characters,
* and ASCII characters are mapped algorithmically only to other ASCII characters.
* This helps with hasCompBoundaryBefore() and compose() fast paths.
* It is never necessary any more to loop for algorithmic mappings.
*
* - Addition of indexes[IX_MIN_NO_NO_COMP_BOUNDARY_BEFORE],
* indexes[IX_MIN_NO_NO_COMP_NO_MAYBE_CC], and indexes[IX_MIN_NO_NO_EMPTY],
* and separation of the noNo extraData into distinct ranges.
* With this, the noNo norm16 value indicates whether the mapping is
* compose-normalized, not normalized but hasCompBoundaryBefore(),
* not even that, or maps to an empty string.
* hasCompBoundaryBefore() can be determined solely from the norm16 value.
*
* - The norm16 value for Hangul LVT is now different from that for Hangul LV,
* so that hasCompBoundaryAfter() need not check for the syllable type.
* For Hangul LV, minYesNo continues to be used (no comp-boundary-after).
* For Hangul LVT, minYesNoMappingsOnly|HAS_COMP_BOUNDARY_AFTER is used.
* The extraData units at these indexes are set to firstUnit=2 and firstUnit=3, respectively,
* to simplify some code.
*
* - The extraData firstUnit bit 5 is no longer necessary
* (norm16 bit 0 used instead of firstUnit MAPPING_NO_COMP_BOUNDARY_AFTER),
* is reserved again, and always set to 0.
*
* - Addition of indexes[IX_MIN_LCCC_CP], the first code point where lccc!=0.
* This used to be hardcoded to U+0300, but in data like NFKC_Casefold it is lower:
* U+00AD Soft Hyphen maps to an empty string,
* which is artificially assigned "worst case" values lccc=1 and tccc=255.
*
* - A mapping to an empty string has explicit lccc=1 and tccc=255 values.
*/
#endif /* !UCONFIG_NO_NORMALIZATION */

94
icu4c/source/common/putil.cpp
View file

@ -939,30 +939,30 @@ static CharString *gSearchTZFileResult = NULL;
* This function is not thread safe - it uses a global, gSearchTZFileResult, to hold its results.
*/
static char* searchForTZFile(const char* path, DefaultTZInfo* tzInfo) {
DIR* dirp = opendir(path);
DIR* subDirp = NULL;
DIR* dirp = NULL;
struct dirent* dirEntry = NULL;
char* result = NULL;
UErrorCode status = U_ZERO_ERROR;
/* Save the current path */
CharString curpath(path, -1, status);
if (U_FAILURE(status)) {
goto cleanupAndReturn;
}
dirp = opendir(path);
if (dirp == NULL) {
return result;
goto cleanupAndReturn;
}
if (gSearchTZFileResult == NULL) {
gSearchTZFileResult = new CharString;
if (gSearchTZFileResult == NULL) {
return NULL;
goto cleanupAndReturn;
}
ucln_common_registerCleanup(UCLN_COMMON_PUTIL, putil_cleanup);
}
/* Save the current path */
UErrorCode status = U_ZERO_ERROR;
CharString curpath(path, -1, status);
if (U_FAILURE(status)) {
return NULL;
}
/* Check each entry in the directory. */
while((dirEntry = readdir(dirp)) != NULL) {
const char* dirName = dirEntry->d_name;
@ -971,15 +971,16 @@ static char* searchForTZFile(const char* path, DefaultTZInfo* tzInfo) {
CharString newpath(curpath, status);
newpath.append(dirName, -1, status);
if (U_FAILURE(status)) {
return NULL;
break;
}
DIR* subDirp = NULL;
if ((subDirp = opendir(newpath.data())) != NULL) {
/* If this new path is a directory, make a recursive call with the newpath. */
closedir(subDirp);
newpath.append('/', status);
if (U_FAILURE(status)) {
return NULL;
break;
}
result = searchForTZFile(newpath.data(), tzInfo);
/*
@ -1003,7 +1004,7 @@ static char* searchForTZFile(const char* path, DefaultTZInfo* tzInfo) {
gSearchTZFileResult->clear();
gSearchTZFileResult->append(zoneid, -1, status);
if (U_FAILURE(status)) {
return NULL;
break;
}
result = gSearchTZFileResult->data();
/* Get out after the first one found. */
@ -1012,7 +1013,11 @@ static char* searchForTZFile(const char* path, DefaultTZInfo* tzInfo) {
}
}
}
closedir(dirp);
cleanupAndReturn:
if (dirp) {
closedir(dirp);
}
return result;
}
#endif
@ -1045,7 +1050,7 @@ uprv_getWindowsTimeZone()
hr = timezone->GetTimeZone(timezoneString.GetAddressOf());
if (SUCCEEDED(hr))
{
int32_t length = wcslen(timezoneString.GetRawBuffer(NULL));
int32_t length = static_cast<int32_t>(wcslen(timezoneString.GetRawBuffer(NULL)));
char* asciiId = (char*)uprv_calloc(length + 1, sizeof(char));
if (asciiId != nullptr)
{
@ -1064,6 +1069,7 @@ uprv_getWindowsTimeZone()
U_CAPI const char* U_EXPORT2
uprv_tzname(int n)
{
(void)n; // Avoid unreferenced parameter warning.
const char *tzid = NULL;
#if U_PLATFORM_USES_ONLY_WIN32_API
#if U_PLATFORM_HAS_WINUWP_API > 0
@ -1287,7 +1293,7 @@ u_setDataDirectory(const char *directory) {
#if (U_FILE_SEP_CHAR != U_FILE_ALT_SEP_CHAR)
{
char *p;
while(p = uprv_strchr(newDataDir, U_FILE_ALT_SEP_CHAR)) {
while((p = uprv_strchr(newDataDir, U_FILE_ALT_SEP_CHAR)) != NULL) {
*p = U_FILE_SEP_CHAR;
}
}
@ -1445,7 +1451,7 @@ static void setTimeZoneFilesDir(const char *path, UErrorCode &status) {
gTimeZoneFilesDirectory->append(path, status);
#if (U_FILE_SEP_CHAR != U_FILE_ALT_SEP_CHAR)
char *p = gTimeZoneFilesDirectory->data();
while (p = uprv_strchr(p, U_FILE_ALT_SEP_CHAR)) {
while ((p = uprv_strchr(p, U_FILE_ALT_SEP_CHAR)) != NULL) {
*p = U_FILE_SEP_CHAR;
}
#endif
@ -1809,6 +1815,8 @@ The leftmost codepage (.xxx) wins.
}
// Now normalize the resulting name
correctedPOSIXLocale = static_cast<char *>(uprv_malloc(POSIX_LOCALE_CAPACITY + 1));
/* TODO: Should we just exit on memory allocation failure? */
if (correctedPOSIXLocale)
{
int32_t posixLen = uloc_canonicalize(modifiedWindowsLocale, correctedPOSIXLocale, POSIX_LOCALE_CAPACITY, &status);
@ -2326,19 +2334,16 @@ u_getVersion(UVersionInfo versionArray) {
* icucfg.h dependent code
*/
#if U_ENABLE_DYLOAD
#if HAVE_DLOPEN && !U_PLATFORM_USES_ONLY_WIN32_API
#if U_ENABLE_DYLOAD && HAVE_DLOPEN && !U_PLATFORM_USES_ONLY_WIN32_API
#if HAVE_DLFCN_H
#ifdef __MVS__
#ifndef __SUSV3
#define __SUSV3 1
#endif
#endif
#include <dlfcn.h>
#endif
#endif /* HAVE_DLFCN_H */
U_INTERNAL void * U_EXPORT2
uprv_dl_open(const char *libName, UErrorCode *status) {
@ -2378,38 +2383,10 @@ uprv_dlsym_func(void *lib, const char* sym, UErrorCode *status) {
return uret.fp;
}
#else
#elif U_ENABLE_DYLOAD && U_PLATFORM_USES_ONLY_WIN32_API && !U_PLATFORM_HAS_WINUWP_API
/* null (nonexistent) implementation. */
U_INTERNAL void * U_EXPORT2
uprv_dl_open(const char *libName, UErrorCode *status) {
if(U_FAILURE(*status)) return NULL;
*status = U_UNSUPPORTED_ERROR;
return NULL;
}
U_INTERNAL void U_EXPORT2
uprv_dl_close(void *lib, UErrorCode *status) {
if(U_FAILURE(*status)) return;
*status = U_UNSUPPORTED_ERROR;
return;
}
U_INTERNAL UVoidFunction* U_EXPORT2
uprv_dlsym_func(void *lib, const char* sym, UErrorCode *status) {
if(U_SUCCESS(*status)) {
*status = U_UNSUPPORTED_ERROR;
}
return (UVoidFunction*)NULL;
}
#endif
#elif U_PLATFORM_USES_ONLY_WIN32_API
/* Windows API implementation. */
// Note: UWP does not expose/allow these APIs, so the UWP version gets the null implementation. */
U_INTERNAL void * U_EXPORT2
uprv_dl_open(const char *libName, UErrorCode *status) {
@ -2436,7 +2413,6 @@ uprv_dl_close(void *lib, UErrorCode *status) {
return;
}
U_INTERNAL UVoidFunction* U_EXPORT2
uprv_dlsym_func(void *lib, const char* sym, UErrorCode *status) {
HMODULE handle = (HMODULE)lib;
@ -2458,10 +2434,9 @@ uprv_dlsym_func(void *lib, const char* sym, UErrorCode *status) {
return addr;
}
#else
/* No dynamic loading set. */
/* No dynamic loading, null (nonexistent) implementation. */
U_INTERNAL void * U_EXPORT2
uprv_dl_open(const char *libName, UErrorCode *status) {
@ -2479,7 +2454,6 @@ uprv_dl_close(void *lib, UErrorCode *status) {
return;
}
U_INTERNAL UVoidFunction* U_EXPORT2
uprv_dlsym_func(void *lib, const char* sym, UErrorCode *status) {
(void)lib;
@ -2490,7 +2464,7 @@ uprv_dlsym_func(void *lib, const char* sym, UErrorCode *status) {
return (UVoidFunction*)NULL;
}
#endif /* U_ENABLE_DYLOAD */
#endif
/*
* Hey, Emacs, please set the following:

33
icu4c/source/common/putilimp.h
View file

@ -72,15 +72,6 @@
typedef size_t uintptr_t;
#endif
/**
* \def U_HAVE_MSVC_2003_OR_EARLIER
* Flag for workaround of MSVC 2003 optimization bugs
* @internal
*/
#if !defined(U_HAVE_MSVC_2003_OR_EARLIER) && defined(_MSC_VER) && (_MSC_VER < 1400)
#define U_HAVE_MSVC_2003_OR_EARLIER
#endif
/*===========================================================================*/
/** @{ Information about POSIX support */
/*===========================================================================*/
@ -120,15 +111,15 @@ typedef size_t uintptr_t;
/* Use the predefined value. */
#elif U_PLATFORM == U_PF_ANDROID
# define U_TIMEZONE timezone
#elif defined(__UCLIBC__)
// uClibc does not have __timezone or _timezone.
#elif defined(_NEWLIB_VERSION)
# define U_TIMEZONE _timezone
#elif defined(__GLIBC__)
// glibc
# define U_TIMEZONE __timezone
#elif U_PLATFORM_IS_LINUX_BASED
# if defined(__UCLIBC__)
/* uClibc does not have __timezone or _timezone. */
# elif defined(_NEWLIB_VERSION)
# define U_TIMEZONE _timezone
# elif defined(__GLIBC__)
/* glibc */
# define U_TIMEZONE __timezone
# endif
// not defined
#elif U_PLATFORM_USES_ONLY_WIN32_API
# define U_TIMEZONE _timezone
#elif U_PLATFORM == U_PF_BSD && !defined(__NetBSD__)
@ -214,7 +205,7 @@ typedef size_t uintptr_t;
/**
* \def U_HAVE_STD_ATOMICS
* Defines whether the standard C++11 <atomic> is available.
* ICU will use this when avialable,
* ICU will use this when available,
* otherwise will fall back to compiler or platform specific alternatives.
* @internal
*/
@ -239,7 +230,7 @@ typedef size_t uintptr_t;
/**
* \def U_HAVE_CLANG_ATOMICS
* Defines whether Clang c11 style built-in atomics are avaialable.
* Defines whether Clang c11 style built-in atomics are available.
* These are used in preference to gcc atomics when both are available.
*/
#ifdef U_HAVE_CLANG_ATOMICS
@ -277,7 +268,7 @@ typedef size_t uintptr_t;
/**
* Platform utilities isolates the platform dependencies of the
* libarary. For each platform which this code is ported to, these
* library. For each platform which this code is ported to, these
* functions may have to be re-implemented.
*/
@ -425,7 +416,7 @@ U_INTERNAL const char* U_EXPORT2 uprv_getDefaultCodepage(void);
/**
* Please use uloc_getDefault() instead.
* Return the default locale ID string by querying ths system, or
* Return the default locale ID string by querying the system, or
* zero if one cannot be found.
* This function can call setlocale() on Unix platforms. Please read the
* platform documentation on setlocale() before calling this function.

49
icu4c/source/common/rbbi.cpp
View file

@ -213,6 +213,8 @@ RuleBasedBreakIterator::operator=(const RuleBasedBreakIterator& that) {
if (this == &that) {
return *this;
}
BreakIterator::operator=(that);
reset(); // Delete break cache information
fBreakType = that.fBreakType;
if (fLanguageBreakEngines != NULL) {
@ -311,16 +313,19 @@ RuleBasedBreakIterator::operator==(const BreakIterator& that) const {
return FALSE;
}
// The base class BreakIterator carries no state that participates in equality,
// and does not implement an equality function that would otherwise be
// checked at this point.
const RuleBasedBreakIterator& that2 = (const RuleBasedBreakIterator&) that;
if (!utext_equals(fText, that2.fText)) {
// The two break iterators are operating on different text,
// or have a different interation position.
// or have a different iteration position.
// Note that fText's position is always the same as the break iterator's position.
return FALSE;
};
// TODO: need a check for when in a dictionary region at different offsets.
if (that2.fData == fData ||
(fData != NULL && that2.fData != NULL && *that2.fData == *fData)) {
// The two break iterators are using the same rules.
@ -1078,7 +1083,7 @@ int32_t RuleBasedBreakIterator::handleNext(const RBBIStateTable *statetable) {
// Note: the 16 in UTRIE_GET16 refers to the size of the data being returned,
// not the size of the character going in, which is a UChar32.
//
UTRIE_GET16(&fData->fTrie, c, category);
category = UTRIE2_GET16(fData->fTrie, c);
// Check the dictionary bit in the character's category.
// Counter is only used by dictionary based iterators (subclasses).
@ -1275,7 +1280,7 @@ int32_t RuleBasedBreakIterator::handlePrevious(const RBBIStateTable *statetable)
// Note: the 16 in UTRIE_GET16 refers to the size of the data being returned,
// not the size of the character going in, which is a UChar32.
//
UTRIE_GET16(&fData->fTrie, c, category);
category = UTRIE2_GET16(fData->fTrie, c);
// Check the dictionary bit in the character's category.
// Counter is only used by dictionary based iterators (subclasses).
@ -1510,26 +1515,6 @@ BreakIterator * RuleBasedBreakIterator::createBufferClone(void * /*stackBuffer*
}
//-------------------------------------------------------------------------------
//
// isDictionaryChar Return true if the category lookup for this char
// indicates that it is in the set of dictionary lookup
// chars.
//
// This function is intended for use by dictionary based
// break iterators.
//
//-------------------------------------------------------------------------------
/*UBool RuleBasedBreakIterator::isDictionaryChar(UChar32 c) {
if (fData == NULL) {
return FALSE;
}
uint16_t category;
UTRIE_GET16(&fData->fTrie, c, category);
return (category & 0x4000) != 0;
}*/
//-------------------------------------------------------------------------------
//
// checkDictionary This function handles all processing of characters in
@ -1569,7 +1554,7 @@ int32_t RuleBasedBreakIterator::checkDictionary(int32_t startPos,
int32_t foundBreakCount = 0;
UChar32 c = utext_current32(fText);
UTRIE_GET16(&fData->fTrie, c, category);
category = UTRIE2_GET16(fData->fTrie, c);
// Is the character we're starting on a dictionary character? If so, we
// need to back up to include the entire run; otherwise the results of
@ -1581,7 +1566,7 @@ int32_t RuleBasedBreakIterator::checkDictionary(int32_t startPos,
do {
utext_next32(fText); // TODO: recast to work directly with postincrement.
c = utext_current32(fText);
UTRIE_GET16(&fData->fTrie, c, category);
category = UTRIE2_GET16(fData->fTrie, c);
} while (c != U_SENTINEL && (category & 0x4000));
// Back up to the last dictionary character
rangeEnd = (int32_t)UTEXT_GETNATIVEINDEX(fText);
@ -1597,7 +1582,7 @@ int32_t RuleBasedBreakIterator::checkDictionary(int32_t startPos,
else {
do {
c = UTEXT_PREVIOUS32(fText);
UTRIE_GET16(&fData->fTrie, c, category);
category = UTRIE2_GET16(fData->fTrie, c);
}
while (c != U_SENTINEL && (category & 0x4000));
// Back up to the last dictionary character
@ -1611,7 +1596,7 @@ int32_t RuleBasedBreakIterator::checkDictionary(int32_t startPos,
}
rangeStart = (int32_t)UTEXT_GETNATIVEINDEX(fText);;
}
UTRIE_GET16(&fData->fTrie, c, category);
category = UTRIE2_GET16(fData->fTrie, c);
}
// Loop through the text, looking for ranges of dictionary characters.
@ -1622,13 +1607,13 @@ int32_t RuleBasedBreakIterator::checkDictionary(int32_t startPos,
if (reverse) {
utext_setNativeIndex(fText, rangeStart);
c = utext_current32(fText);
UTRIE_GET16(&fData->fTrie, c, category);
category = UTRIE2_GET16(fData->fTrie, c);
}
while(U_SUCCESS(status)) {
while((current = (int32_t)UTEXT_GETNATIVEINDEX(fText)) < rangeEnd && (category & 0x4000) == 0) {
utext_next32(fText); // TODO: tweak for post-increment operation
c = utext_current32(fText);
UTRIE_GET16(&fData->fTrie, c, category);
category = UTRIE2_GET16(fData->fTrie, c);
}
if (current >= rangeEnd) {
break;
@ -1646,7 +1631,7 @@ int32_t RuleBasedBreakIterator::checkDictionary(int32_t startPos,
// Reload the loop variables for the next go-round
c = utext_current32(fText);
UTRIE_GET16(&fData->fTrie, c, category);
category = UTRIE2_GET16(fData->fTrie, c);
}
// If we found breaks, build a new break cache. The first and last entries must

60
icu4c/source/common/rbbidata.cpp
View file

@ -23,23 +23,6 @@
#include "uassert.h"
//-----------------------------------------------------------------------------------
//
// Trie access folding function. Copied as-is from properties code in uchar.c
//
//-----------------------------------------------------------------------------------
U_CDECL_BEGIN
static int32_t U_CALLCONV
getFoldingOffset(uint32_t data) {
/* if bit 15 is set, then the folding offset is in bits 14..0 of the 16-bit trie result */
if(data&0x8000) {
return (int32_t)(data&0x7fff);
} else {
return 0;
}
}
U_CDECL_END
U_NAMESPACE_BEGIN
//-----------------------------------------------------------------------------
@ -71,9 +54,8 @@ RBBIDataWrapper::RBBIDataWrapper(UDataMemory* udm, UErrorCode &status) {
dh->info.dataFormat[0] == 0x42 && // dataFormat="Brk "
dh->info.dataFormat[1] == 0x72 &&
dh->info.dataFormat[2] == 0x6b &&
dh->info.dataFormat[3] == 0x20)
// Note: info.fFormatVersion is duplicated in the RBBIDataHeader, and is
// validated when checking that.
dh->info.dataFormat[3] == 0x20 &&
isDataVersionAcceptable(dh->info.formatVersion))
) {
status = U_INVALID_FORMAT_ERROR;
return;
@ -84,6 +66,11 @@ RBBIDataWrapper::RBBIDataWrapper(UDataMemory* udm, UErrorCode &status) {
fUDataMem = udm;
}
UBool RBBIDataWrapper::isDataVersionAcceptable(const UVersionInfo version) {
return RBBI_DATA_FORMAT_VERSION[0] == version[0];
}
//-----------------------------------------------------------------------------
//
// init(). Does most of the work of construction, shared between the
@ -98,6 +85,7 @@ void RBBIDataWrapper::init0() {
fSafeRevTable = NULL;
fRuleSource = NULL;
fRuleStatusTable = NULL;
fTrie = NULL;
fUDataMem = NULL;
fRefCount = 0;
fDontFreeData = TRUE;
@ -108,8 +96,7 @@ void RBBIDataWrapper::init(const RBBIDataHeader *data, UErrorCode &status) {
return;
}
fHeader = data;
if (fHeader->fMagic != 0xb1a0 || fHeader->fFormatVersion[0] != 3)
{
if (fHeader->fMagic != 0xb1a0 || !isDataVersionAcceptable(fHeader->fFormatVersion)) {
status = U_INVALID_FORMAT_ERROR;
return;
}
@ -132,15 +119,14 @@ void RBBIDataWrapper::init(const RBBIDataHeader *data, UErrorCode &status) {
}
utrie_unserialize(&fTrie,
(uint8_t *)data + fHeader->fTrie,
fHeader->fTrieLen,
&status);
fTrie = utrie2_openFromSerialized(UTRIE2_16_VALUE_BITS,
(uint8_t *)data + fHeader->fTrie,
fHeader->fTrieLen,
NULL, // *actual length
&status);
if (U_FAILURE(status)) {
return;
}
fTrie.getFoldingOffset=getFoldingOffset;
fRuleSource = (UChar *)((char *)data + fHeader->fRuleSource);
fRuleString.setTo(TRUE, fRuleSource, -1);
@ -165,6 +151,8 @@ void RBBIDataWrapper::init(const RBBIDataHeader *data, UErrorCode &status) {
//-----------------------------------------------------------------------------
RBBIDataWrapper::~RBBIDataWrapper() {
U_ASSERT(fRefCount == 0);
utrie2_close(fTrie);
fTrie = NULL;
if (fUDataMem) {
udata_close(fUDataMem);
} else if (!fDontFreeData) {
@ -323,7 +311,7 @@ ubrk_swap(const UDataSwapper *ds, const void *inData, int32_t length, void *outD
pInfo->dataFormat[1]==0x72 &&
pInfo->dataFormat[2]==0x6b &&
pInfo->dataFormat[3]==0x20 &&
pInfo->formatVersion[0]==3 )) {
RBBIDataWrapper::isDataVersionAcceptable(pInfo->formatVersion) )) {
udata_printError(ds, "ubrk_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized\n",
pInfo->dataFormat[0], pInfo->dataFormat[1],
pInfo->dataFormat[2], pInfo->dataFormat[3],
@ -344,17 +332,11 @@ ubrk_swap(const UDataSwapper *ds, const void *inData, int32_t length, void *outD
//
// Get the RRBI Data Header, and check that it appears to be OK.
//
// Note: ICU 3.2 and earlier, RBBIDataHeader::fDataFormat was actually
// an int32_t with a value of 1. Starting with ICU 3.4,
// RBBI's fDataFormat matches the dataFormat field from the
// UDataInfo header, four int8_t bytes. The value is {3,1,0,0}
//
const uint8_t *inBytes =(const uint8_t *)inData+headerSize;
RBBIDataHeader *rbbiDH = (RBBIDataHeader *)inBytes;
if (ds->readUInt32(rbbiDH->fMagic) != 0xb1a0 ||
rbbiDH->fFormatVersion[0] != 3 ||
ds->readUInt32(rbbiDH->fLength) < sizeof(RBBIDataHeader))
{
!RBBIDataWrapper::isDataVersionAcceptable(rbbiDH->fFormatVersion) ||
ds->readUInt32(rbbiDH->fLength) < sizeof(RBBIDataHeader)) {
udata_printError(ds, "ubrk_swap(): RBBI Data header is invalid.\n");
*status=U_UNSUPPORTED_ERROR;
return 0;
@ -451,8 +433,8 @@ ubrk_swap(const UDataSwapper *ds, const void *inData, int32_t length, void *outD
}
// Trie table for character categories
utrie_swap(ds, inBytes+ds->readUInt32(rbbiDH->fTrie), ds->readUInt32(rbbiDH->fTrieLen),
outBytes+ds->readUInt32(rbbiDH->fTrie), status);
utrie2_swap(ds, inBytes+ds->readUInt32(rbbiDH->fTrie), ds->readUInt32(rbbiDH->fTrieLen),
outBytes+ds->readUInt32(rbbiDH->fTrie), status);
// Source Rules Text. It's UChar data
ds->swapArray16(ds, inBytes+ds->readUInt32(rbbiDH->fRuleSource), ds->readUInt32(rbbiDH->fRuleSourceLen),

15
icu4c/source/common/rbbidata.h
View file

@ -51,22 +51,23 @@ ubrk_swap(const UDataSwapper *ds,
#include "unicode/uobject.h"
#include "unicode/unistr.h"
#include "unicode/uversion.h"
#include "umutex.h"
#include "utrie.h"
#include "utrie2.h"
U_NAMESPACE_BEGIN
// The current RBBI data format version.
static const uint8_t RBBI_DATA_FORMAT_VERSION[] = {4, 0, 0, 0};
/*
* The following structs map exactly onto the raw data from ICU common data file.
*/
struct RBBIDataHeader {
uint32_t fMagic; /* == 0xbla0 */
uint8_t fFormatVersion[4]; /* Data Format. Same as the value in struct UDataInfo */
UVersionInfo fFormatVersion; /* Data Format. Same as the value in struct UDataInfo */
/* if there is one associated with this data. */
/* (version originates in rbbi, is copied to UDataInfo) */
/* For ICU 3.2 and earlier, this field was */
/* uint32_t fVersion */
/* with a value of 1. */
uint32_t fLength; /* Total length in bytes of this RBBI Data, */
/* including all sections, not just the header. */
uint32_t fCatCount; /* Number of character categories. */
@ -152,6 +153,8 @@ public:
RBBIDataWrapper(UDataMemory* udm, UErrorCode &status);
~RBBIDataWrapper();
static UBool isDataVersionAcceptable(const UVersionInfo version);
void init0();
void init(const RBBIDataHeader *data, UErrorCode &status);
RBBIDataWrapper *addReference();
@ -181,7 +184,7 @@ public:
/* number of int32_t values in the rule status table. Used to sanity check indexing */
int32_t fStatusMaxIdx;
UTrie fTrie;
UTrie2 *fTrie;
private:
u_atomic_int32_t fRefCount;

8
icu4c/source/common/rbbirb.cpp
View file

@ -177,10 +177,10 @@ RBBIDataHeader *RBBIRuleBuilder::flattenData() {
data->fMagic = 0xb1a0;
data->fFormatVersion[0] = 3;
data->fFormatVersion[1] = 1;
data->fFormatVersion[2] = 0;
data->fFormatVersion[3] = 0;
data->fFormatVersion[0] = RBBI_DATA_FORMAT_VERSION[0];
data->fFormatVersion[1] = RBBI_DATA_FORMAT_VERSION[1];
data->fFormatVersion[2] = RBBI_DATA_FORMAT_VERSION[2];
data->fFormatVersion[3] = RBBI_DATA_FORMAT_VERSION[3];
data->fLength = totalSize;
data->fCatCount = fSetBuilder->getNumCharCategories();

78
icu4c/source/common/rbbisetb.cpp
View file

@ -35,7 +35,7 @@
#if !UCONFIG_NO_BREAK_ITERATION
#include "unicode/uniset.h"
#include "utrie.h"
#include "utrie2.h"
#include "uvector.h"
#include "uassert.h"
#include "cmemory.h"
@ -44,43 +44,6 @@
#include "rbbisetb.h"
#include "rbbinode.h"
//------------------------------------------------------------------------
//
// getFoldedRBBIValue Call-back function used during building of Trie table.
// Folding value: just store the offset (16 bits)
// if there is any non-0 entry.
// (It'd really be nice if the Trie builder would provide a
// simple default, so this function could go away from here.)
//
//------------------------------------------------------------------------
/* folding value: just store the offset (16 bits) if there is any non-0 entry */
U_CDECL_BEGIN
static uint32_t U_CALLCONV
getFoldedRBBIValue(UNewTrie *trie, UChar32 start, int32_t offset) {
uint32_t value;
UChar32 limit;
UBool inBlockZero;
limit=start+0x400;
while(start<limit) {
value=utrie_get32(trie, start, &inBlockZero);
if(inBlockZero) {
start+=UTRIE_DATA_BLOCK_LENGTH;
} else if(value!=0) {
return (uint32_t)(offset|0x8000);
} else {
++start;
}
}
return 0;
}
U_CDECL_END
U_NAMESPACE_BEGIN
//------------------------------------------------------------------------
@ -116,7 +79,7 @@ RBBISetBuilder::~RBBISetBuilder()
delete r;
}
utrie_close(fTrie);
utrie2_close(fTrie);
}
@ -287,33 +250,30 @@ void RBBISetBuilder::build() {
// 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
fTrie = utrie2_open(0, // Initial value for all code points
0, // errorValue
fStatus);
for (rlRange = fRangeList; rlRange!=0; rlRange=rlRange->fNext) {
utrie_setRange32(fTrie, rlRange->fStartChar, rlRange->fEndChar+1, rlRange->fNum, TRUE);
utrie2_setRange32(fTrie, rlRange->fStartChar, rlRange->fEndChar, rlRange->fNum, TRUE, fStatus);
}
}
//-----------------------------------------------------------------------------------
//
// getTrieSize() Return the size that will be required to serialize the Trie.
//
//-----------------------------------------------------------------------------------
int32_t RBBISetBuilder::getTrieSize() /*const*/ {
fTrieSize = utrie_serialize(fTrie,
NULL, // Buffer
0, // Capacity
getFoldedRBBIValue,
TRUE, // Reduce to 16 bits
fStatus);
utrie2_freeze(fTrie, UTRIE2_16_VALUE_BITS, fStatus);
fTrieSize = utrie2_serialize(fTrie,
NULL, // Buffer
0, // Capacity
fStatus);
if (*fStatus == U_BUFFER_OVERFLOW_ERROR) {
*fStatus = U_ZERO_ERROR;
}
// RBBIDebugPrintf("Trie table size is %d\n", trieSize);
return fTrieSize;
}
@ -327,12 +287,10 @@ int32_t RBBISetBuilder::getTrieSize() /*const*/ {
//
//-----------------------------------------------------------------------------------
void RBBISetBuilder::serializeTrie(uint8_t *where) {
utrie_serialize(fTrie,
where, // Buffer
fTrieSize, // Capacity
getFoldedRBBIValue,
TRUE, // Reduce to 16 bits
fStatus);
utrie2_serialize(fTrie,
where, // Buffer
fTrieSize, // Capacity
fStatus);
}
//------------------------------------------------------------------------

5
icu4c/source/common/rbbisetb.h
View file

@ -15,10 +15,9 @@
#include "unicode/utypes.h"
#include "unicode/uobject.h"
#include "rbbirb.h"
#include "utrie2.h"
#include "uvector.h"
struct UNewTrie;
U_NAMESPACE_BEGIN
//
@ -109,7 +108,7 @@ private:
RangeDescriptor *fRangeList; // Head of the linked list of RangeDescriptors
UNewTrie *fTrie; // The mapping TRIE that is the end result of processing
UTrie2 *fTrie; // The mapping TRIE that is the end result of processing
uint32_t fTrieSize; // the Unicode Sets.
// Groups correspond to character categories -

12
icu4c/source/common/ucase.h
View file

@ -61,7 +61,7 @@ enum {
/**
* Bit mask for getting just the options from a string compare options word
* that are relevant for case-insensitive string comparison.
* See uchar.h. Also include _STRNCMP_STYLE and U_COMPARE_CODE_POINT_ORDER.
* See stringoptions.h. Also include _STRNCMP_STYLE and U_COMPARE_CODE_POINT_ORDER.
* @internal
*/
#define _STRCASECMP_OPTIONS_MASK 0xffff
@ -69,10 +69,16 @@ enum {
/**
* Bit mask for getting just the options from a string compare options word
* that are relevant for case folding (of a single string or code point).
* See uchar.h.
*
* Currently only bit 0 for U_FOLD_CASE_EXCLUDE_SPECIAL_I.
* It is conceivable that at some point we might use one more bit for using uppercase sharp s.
* It is conceivable that at some point we might want the option to use only simple case foldings
* when operating on strings.
*
* See stringoptions.h.
* @internal
*/
#define _FOLD_CASE_OPTIONS_MASK 0xff
#define _FOLD_CASE_OPTIONS_MASK 7
/* single-code point functions */

55
icu4c/source/common/ucasemap.cpp
View file

@ -164,7 +164,7 @@ appendResult(uint8_t *dest, int32_t destIndex, int32_t destCapacity,
/* (not) original code point */
if(edits!=NULL) {
edits->addUnchanged(cpLength);
if(options & UCASEMAP_OMIT_UNCHANGED_TEXT) {
if(options & U_OMIT_UNCHANGED_TEXT) {
return destIndex;
}
}
@ -282,7 +282,7 @@ appendUnchanged(uint8_t *dest, int32_t destIndex, int32_t destCapacity,
if(length>0) {
if(edits!=NULL) {
edits->addUnchanged(length);
if(options & UCASEMAP_OMIT_UNCHANGED_TEXT) {
if(options & U_OMIT_UNCHANGED_TEXT) {
return destIndex;
}
}
@ -381,7 +381,7 @@ ucasemap_internalUTF8ToTitle(
const uint8_t *src, int32_t srcLength,
icu::Edits *edits,
UErrorCode &errorCode) {
if(U_FAILURE(errorCode)) {
if (!ustrcase_checkTitleAdjustmentOptions(options, errorCode)) {
return 0;
}
@ -408,45 +408,38 @@ ucasemap_internalUTF8ToTitle(
}
/*
* Unicode 4 & 5 section 3.13 Default Case Operations:
*
* R3 toTitlecase(X): Find the word boundaries based on Unicode Standard Annex
* #29, "Text Boundaries." Between each pair of word boundaries, find the first
* cased character F. If F exists, map F to default_title(F); then map each
* subsequent character C to default_lower(C).
*
* In this implementation, segment [prev..index[ into 3 parts:
* a) uncased characters (copy as-is) [prev..titleStart[
* b) first case letter (titlecase) [titleStart..titleLimit[
* Segment [prev..index[ into 3 parts:
* a) skipped characters (copy as-is) [prev..titleStart[
* b) first letter (titlecase) [titleStart..titleLimit[
* c) subsequent characters (lowercase) [titleLimit..index[
*/
if(prev<index) {
/* find and copy uncased characters [prev..titleStart[ */
/* find and copy skipped characters [prev..titleStart[ */
int32_t titleStart=prev;
int32_t titleLimit=prev;
UChar32 c;
U8_NEXT(src, titleLimit, index, c);
if((options&U_TITLECASE_NO_BREAK_ADJUSTMENT)==0 && UCASE_NONE==ucase_getType(c)) {
/* Adjust the titlecasing index (titleStart) to the next cased character. */
for(;;) {
if ((options&U_TITLECASE_NO_BREAK_ADJUSTMENT)==0) {
// Adjust the titlecasing index to the next cased character,
// or to the next letter/number/symbol/private use.
// Stop with titleStart<titleLimit<=index
// if there is a character to be titlecased,
// or else stop with titleStart==titleLimit==index.
UBool toCased = (options&U_TITLECASE_ADJUST_TO_CASED) != 0;
while (toCased ? UCASE_NONE==ucase_getType(c) : !ustrcase_isLNS(c)) {
titleStart=titleLimit;
if(titleLimit==index) {
/*
* only uncased characters in [prev..index[
* stop with titleStart==titleLimit==index
*/
break;
}
U8_NEXT(src, titleLimit, index, c);
if(UCASE_NONE!=ucase_getType(c)) {
break; /* cased letter at [titleStart..titleLimit[ */
}
}
destIndex=appendUnchanged(dest, destIndex, destCapacity,
src+prev, titleStart-prev, options, edits);
if(destIndex<0) {
errorCode=U_INDEX_OUTOFBOUNDS_ERROR;
return 0;
if (prev < titleStart) {
destIndex=appendUnchanged(dest, destIndex, destCapacity,
src+prev, titleStart-prev, options, edits);
if(destIndex<0) {
errorCode=U_INDEX_OUTOFBOUNDS_ERROR;
return 0;
}
}
}
@ -666,7 +659,7 @@ int32_t toUpper(uint32_t options,
edits->addUnchanged(oldLength);
}
// Write unchanged text?
change = (options & UCASEMAP_OMIT_UNCHANGED_TEXT) == 0;
change = (options & U_OMIT_UNCHANGED_TEXT) == 0;
}
}
@ -827,7 +820,7 @@ ucasemap_mapUTF8(int32_t caseLocale, uint32_t options, UCASEMAP_BREAK_ITERATOR_P
return 0;
}
if(edits!=NULL) {
if (edits != nullptr && (options & U_EDITS_NO_RESET) == 0) {
edits->reset();
}
destLength=stringCaseMapper(caseLocale, options, UCASEMAP_BREAK_ITERATOR

58
icu4c/source/common/ucasemap_imp.h
View file

@ -9,16 +9,26 @@
#include "unicode/utypes.h"
#include "unicode/ucasemap.h"
#include "unicode/uchar.h"
#include "ucase.h"
#ifndef U_COMPARE_IGNORE_CASE
/* see also unorm.h */
/**
* Option bit for unorm_compare:
* Perform case-insensitive comparison.
* Bit mask for the titlecasing iterator options bit field.
* Currently only 3 out of 8 values are used:
* 0 (words), U_TITLECASE_WHOLE_STRING, U_TITLECASE_SENTENCES.
* See stringoptions.h.
* @internal
*/
#define U_COMPARE_IGNORE_CASE 0x10000
#endif
#define U_TITLECASE_ITERATOR_MASK 0xe0
/**
* Bit mask for the titlecasing index adjustment options bit set.
* Currently two bits are defined:
* U_TITLECASE_NO_BREAK_ADJUSTMENT, U_TITLECASE_ADJUST_TO_CASED.
* See stringoptions.h.
* @internal
*/
#define U_TITLECASE_ADJUSTMENT_MASK 0x600
/**
* Internal API, used by u_strcasecmp() etc.
@ -32,7 +42,7 @@ u_strcmpFold(const UChar *s1, int32_t length1,
UErrorCode *pErrorCode);
/**
* Interanl API, used for detecting length of
* Internal API, used for detecting length of
* shared prefix case-insensitively.
* @param s1 input string 1
* @param length1 length of string 1, or -1 (NULL terminated)
@ -61,6 +71,40 @@ uprv_haveProperties(UErrorCode *pErrorCode);
#ifdef __cplusplus
U_NAMESPACE_BEGIN
/** Returns TRUE if the options are valid. Otherwise FALSE, and sets an error. */
inline UBool ustrcase_checkTitleAdjustmentOptions(uint32_t options, UErrorCode &errorCode) {
if (U_FAILURE(errorCode)) { return FALSE; }
if ((options & U_TITLECASE_ADJUSTMENT_MASK) == U_TITLECASE_ADJUSTMENT_MASK) {
// Both options together.
errorCode = U_ILLEGAL_ARGUMENT_ERROR;
return FALSE;
}
return TRUE;
}
inline UBool ustrcase_isLNS(UChar32 c) {
// Letter, number, symbol,
// or a private use code point because those are typically used as letters or numbers.
// Consider modifier letters only if they are cased.
const uint32_t LNS = (U_GC_L_MASK|U_GC_N_MASK|U_GC_S_MASK|U_GC_CO_MASK) & ~U_GC_LM_MASK;
int gc = u_charType(c);
return (U_MASK(gc) & LNS) != 0 || (gc == U_MODIFIER_LETTER && ucase_getType(c) != UCASE_NONE);
}
#if !UCONFIG_NO_BREAK_ITERATION
/** Returns nullptr if error. Pass in either locale or locID, not both. */
U_CFUNC
BreakIterator *ustrcase_getTitleBreakIterator(
const Locale *locale, const char *locID, uint32_t options, BreakIterator *iter,
LocalPointer<BreakIterator> &ownedIter, UErrorCode &errorCode);
#endif
U_NAMESPACE_END
#include "unicode/unistr.h" // for UStringCaseMapper
/*

18
icu4c/source/common/ucasemap_titlecase_brkiter.cpp
View file

@ -42,11 +42,8 @@ int32_t CaseMap::utf8ToTitle(
UText utext=UTEXT_INITIALIZER;
utext_openUTF8(&utext, src, srcLength, &errorCode);
LocalPointer<BreakIterator> ownedIter;
iter = ustrcase_getTitleBreakIterator(nullptr, locale, options, iter, ownedIter, errorCode);
if(iter==NULL) {
iter=BreakIterator::createWordInstance(Locale(locale), errorCode);
ownedIter.adoptInstead(iter);
}
if(U_FAILURE(errorCode)) {
utext_close(&utext);
return 0;
}
@ -88,12 +85,19 @@ ucasemap_utf8ToTitle(UCaseMap *csm,
}
UText utext=UTEXT_INITIALIZER;
utext_openUTF8(&utext, (const char *)src, srcLength, pErrorCode);
if(csm->iter==NULL) {
csm->iter=BreakIterator::createWordInstance(Locale(csm->locale), *pErrorCode);
}
if (U_FAILURE(*pErrorCode)) {
return 0;
}
if(csm->iter==NULL) {
LocalPointer<BreakIterator> ownedIter;
BreakIterator *iter = ustrcase_getTitleBreakIterator(
nullptr, csm->locale, csm->options, nullptr, ownedIter, *pErrorCode);
if (iter == nullptr) {
utext_close(&utext);
return 0;
}
csm->iter = ownedIter.orphan();
}
csm->iter->setText(&utext, *pErrorCode);
int32_t length=ucasemap_mapUTF8(
csm->caseLocale, csm->options, csm->iter,

5
icu4c/source/common/uchar.cpp
View file

@ -729,8 +729,5 @@ upropsvec_addPropertyStarts(const USetAdder *sa, UErrorCode *pErrorCode) {
}
/* add the start code point of each same-value range of the properties vectors trie */
if(propsVectorsColumns>0) {
/* if propsVectorsColumns==0 then the properties vectors trie may not be there at all */
utrie2_enum(&propsVectorsTrie, NULL, _enumPropertyStartsRange, sa);
}
utrie2_enum(&propsVectorsTrie, NULL, _enumPropertyStartsRange, sa);
}

460
icu4c/source/common/uchar_props_data.h
View file

@ -1788,9 +1788,9 @@ static const uint16_t propsVectorsTrie_index[29236]={
0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,
0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xef7,0xeeb,0xeeb,0xeeb,0xeeb,0xeeb,
0xeeb,0xeeb,0xeeb,0xeeb,0xefa,0xefa,0xeee,0xeee,0xef1,0xf00,0xefd,0x10e,0x10e,0x10e,0x10e,0x10e,
0x189f,0x189f,0x189f,0x189f,0x189f,0x189f,0x189f,0x189f,0x189f,0x189f,0x189f,0x25e,0x25e,0x25e,0x25e,0x25e,
0x18a2,0x18a2,0x18a2,0x18a2,0x18a2,0x18a2,0x18a2,0x18a2,0x18a2,0x18a2,0x18a2,0x25e,0x25e,0x25e,0x25e,0x25e,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0x17e5,0x17e5,0x17e5,0x17e5,0x17e5,0x17e5,0x17e5,0x17e5,0x17e5,0x237,0x237,0x237,0x237,0x237,0x237,0x237,
0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x237,0x237,0x237,0x237,0x237,0x237,0x237,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0xaec,0xaec,0xaef,0xaef,0xaec,0xaec,0xaec,0xaec,0xaec,0xaec,0xaec,0xaec,0x72,0x72,0x72,0x72,
0x1563,0x1563,0x1563,0x1563,0x1c5,0x1c5,0x1c5,0x1c5,0x1c5,0x1c5,0x1c5,0x1c5,0x1c5,0x1c5,0x1c5,0x1560,
@ -1805,9 +1805,9 @@ static const uint16_t propsVectorsTrie_index[29236]={
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0x147c,0x147c,0x147c,0x147c,0x147c,0x147c,0x147c,0x147c,0x147c,0x147c,0x1bf,0x1bf,0x1bf,0x1bf,0x1bf,0x1bf,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0x188d,0x1890,0x1890,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,
0x1890,0x1893,0x1893,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,0x25b,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0x174f,0x174f,0x174f,0x174f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,
0x1752,0x1752,0x1752,0x1752,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,0x21f,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0x1248,0x1248,0x1248,0x1248,0x1248,0x1248,0x1248,0x1248,0x1248,0x1248,0x1248,0x1248,0x1248,0x1248,0x1248,0x183,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
@ -1818,14 +1818,14 @@ static const uint16_t propsVectorsTrie_index[29236]={
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0x252,0x252,0x252,0x252,0x252,0x252,0x252,0x252,0x252,0x252,0x252,0x252,0x252,0x252,0x252,0x252,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,
0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,
0x16bc,0x16bc,0x16bc,0x16bc,0x207,0x207,0x207,0x207,0x207,0x207,0x207,0x207,0x207,0x207,0x207,0x207,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0xdd7,0xdd7,0xdd4,0xdd4,0xdd4,0xdd7,0xdb,0xdb,0xdb,0xdb,0xdb,0xdb,0xdb,0xdb,0xdb,0xdb,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0x22e,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,
0x22e,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0x17e8,0x17e8,0x23a,0x17e8,0x17e8,0x23a,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x23a,0x23a,0x23a,0x23a,0x23a,
0x17eb,0x17eb,0x23a,0x17eb,0x17eb,0x23a,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x23a,0x23a,0x23a,0x23a,0x23a,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0x933,0x933,
3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,
@ -1841,7 +1841,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x38d,0x1b,0x1b,0x39c,0x39c,0x1b,0x1b,0x39f,0x39f,0x393,0xd83,0x1b,0x1b,0x1b,0x1b,0x1b,
0x1b,0x1b,0x1b,0x38d,0x1b,0x1b,0x1b,0x1b,0x3a5,0x3a5,0x1b,0x3a5,0x3a2,0x3a2,0x38d,0x38d,
0x1b,0x1b,0x91e,0x91e,0x91e,0x91e,0x91e,0x91e,0x91e,0x91e,0x91e,0x91e,0x3a2,0x3a2,0x399,0x399,
0x396,0x396,0x396,0x396,0x396,0x399,0x396,0x110a,0x1848,0x1845,0x1b,0x1b,0x1e,0xc93,0x3a8,0xc96,
0x396,0x396,0x396,0x396,0x396,0x399,0x396,0x110a,0x184b,0x1848,0x1b,0x1b,0x1e,0xc93,0x3a8,0xc96,
0x1e,0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x1e,0x1e,0x1e,0x1e,0x3b4,0x3b4,0x1e,0x1e,0x3b4,
0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x1e,0x3b4,0x3b4,
0x3b4,0x3b4,0x3b4,0x3b4,0x3b4,0x1e,0x3b4,0x3b7,0x1e,0x3b4,0x3b7,0x1e,0x3b4,0x3b4,0x1e,0x1e,
@ -1856,7 +1856,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x3ba,0x3ba,0x21,0x3ba,0x3ba,0x3c3,0x21,0x3c3,0x3c3,0x3c0,0x21,0x21,0x3c6,0x21,0x21,0x21,
0x21,0x21,0x21,0x21,0x21,0x21,0x21,0x21,0x21,0x21,0x21,0x21,0x3c6,0xc9f,0xc99,0xc99,
0x21,0x21,0x924,0x924,0x924,0x924,0x924,0x924,0x924,0x924,0x924,0x924,0x13c8,0xc9c,0x21,0x21,
0x21,0x21,0x21,0x21,0x21,0x16cb,0x184b,0x184b,0x184b,0x184e,0x184e,0x184e,0x24,0x3c9,0x3d8,0x3d8,
0x21,0x21,0x21,0x21,0x21,0x16ce,0x184e,0x184e,0x184e,0x1851,0x1851,0x1851,0x24,0x3c9,0x3d8,0x3d8,
0x24,0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x24,0x24,0x3de,0x3de,0x24,0x24,0x3de,
0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x3de,0x24,0x3de,0x3de,
0x3de,0x3de,0x3de,0x3de,0x3de,0x24,0x3de,0x3de,0x24,0xca2,0x3de,0x3de,0x3de,0x3de,0x24,0x24,
@ -1877,9 +1877,9 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x405,0x2a,0x405,0x405,0x405,0x405,0x405,0x405,0x405,0x405,0x405,0x405,0x14b8,0x405,0x405,0x405,
0x405,0x405,0x2a,0x2a,0x2a,0xf75,0x3f6,0x3f6,0x3f6,0x402,0x402,0x402,0x402,0x2a,0x3f6,0x3f6,
0x3f9,0x2a,0x3f6,0x3f6,0x3f6,0x3fc,0x2a,0x2a,0x2a,0x2a,0x2a,0x2a,0x2a,0x3f6,0x3f6,0x2a,
0xf75,0xf75,0x16ce,0x2a,0x2a,0x2a,0x2a,0x2a,0x405,0x405,0xf6f,0xf6f,0x2a,0x2a,0x3ff,0x3ff,
0xf75,0xf75,0x16d1,0x2a,0x2a,0x2a,0x2a,0x2a,0x405,0x405,0xf6f,0xf6f,0x2a,0x2a,0x3ff,0x3ff,
0x3ff,0x3ff,0x3ff,0x3ff,0x3ff,0x3ff,0x3ff,0x3ff,0x2a,0x2a,0x2a,0x2a,0x2a,0x2a,0x2a,0x2a,
0xf72,0xf72,0xf72,0xf72,0xf72,0xf72,0xf72,0xf72,0x178b,0x14bb,0x411,0x411,0x2d,0x417,0x417,0x417,
0xf72,0xf72,0xf72,0xf72,0xf72,0xf72,0xf72,0xf72,0x178e,0x14bb,0x411,0x411,0x2d,0x417,0x417,0x417,
0x417,0x417,0x417,0x417,0x417,0x2d,0x417,0x417,0x417,0x2d,0x417,0x417,0x417,0x417,0x417,0x417,
0x417,0x417,0x417,0x417,0x417,0x417,0x417,0x417,0x417,0x2d,0x417,0x417,0x417,0x417,0x417,0x417,
0x417,0x417,0x417,0x417,0x2d,0x417,0x417,0x417,0x417,0x417,0x2d,0x2d,0xcab,0xcae,0x411,0x408,
@ -1887,12 +1887,12 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x2d,0x2d,0x2d,0x2d,0x2d,0x408,0x408,0x2d,0x2d,0x2d,0x2d,0x2d,0x2d,0x2d,0x417,0x2d,
0x417,0x417,0xeb8,0xeb8,0x2d,0x2d,0x40e,0x40e,0x40e,0x40e,0x40e,0x40e,0x40e,0x40e,0x40e,0x40e,
0x2d,0xebb,0xebb,0x2d,0x2d,0x2d,0x2d,0x2d,0x2d,0x2d,0x2d,0x2d,0x2d,0x2d,0x2d,0x2d,
0x1851,0x14be,0x423,0x423,0x30,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x30,0x429,0x429,
0x1854,0x14be,0x423,0x423,0x30,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x30,0x429,0x429,
0x429,0x30,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,
0x423,0x41a,0x41a,0x41a,0xf78,0x30,0x423,0x423,0x423,0x30,0x426,0x426,0x426,0x41d,0x12e7,0x178e,
0x30,0x30,0x30,0x30,0x1791,0x1791,0x1791,0x41a,0x178e,0x178e,0x178e,0x178e,0x178e,0x178e,0x178e,0x16d1,
0x423,0x41a,0x41a,0x41a,0xf78,0x30,0x423,0x423,0x423,0x30,0x426,0x426,0x426,0x41d,0x12e7,0x1791,
0x30,0x30,0x30,0x30,0x1794,0x1794,0x1794,0x41a,0x1791,0x1791,0x1791,0x1791,0x1791,0x1791,0x1791,0x16d4,
0x429,0x429,0xf78,0xf78,0x30,0x30,0x420,0x420,0x420,0x420,0x420,0x420,0x420,0x420,0x420,0x420,
0xf7b,0xf7b,0xf7b,0xf7b,0xf7b,0xf7b,0x178e,0x178e,0x178e,0xf7e,0xf81,0xf81,0xf81,0xf81,0xf81,0xf81,
0xf7b,0xf7b,0xf7b,0xf7b,0xf7b,0xf7b,0x1791,0x1791,0x1791,0xf7e,0xf81,0xf81,0xf81,0xf81,0xf81,0xf81,
0x33,0x33,0x9f6,0x9f6,0x33,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,
0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x33,0x33,0x33,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,
0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x9fc,0x33,0x9fc,
@ -1942,8 +1942,8 @@ static const uint16_t propsVectorsTrie_index[29236]={
0xa35,0xa35,0xa35,0x42,0x42,0x12f3,0x12f3,0xd92,0xd95,0xa2f,0xa38,0xa2c,0xa2c,0xa2c,0xa2c,0xa38,
0xa38,0xa32,0xa32,0xa32,0xa32,0xa32,0xa32,0xa32,0xa32,0xa32,0xa29,0xa29,0xa29,0xa29,0xa29,0xa29,
0xa29,0xa29,0xa29,0xa29,0xa29,0x42,0x42,0x42,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,
0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0x16d7,0x45,0x45,
0x16d4,0x16d4,0x16d4,0x16d4,0x16d4,0x16d4,0x45,0x45,0xa4d,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,
0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0xa3b,0x16da,0x45,0x45,
0x16d7,0x16d7,0x16d7,0x16d7,0x16d7,0x16d7,0x45,0x45,0xa4d,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,
0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,0xa50,
0xa50,0xa50,0xa50,0xa4a,0xa47,0x48,0x48,0x48,0xa56,0xa56,0xa56,0xa56,0xa56,0xa56,0xa56,0xa56,
0xa56,0xa56,0xa56,0xa53,0xa53,0xa53,0xa56,0xa56,0xa56,0x14c4,0x14c4,0x14c4,0x14c4,0x14c4,0x14c4,0x14c4,
@ -1977,7 +1977,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x504,0x504,0x504,0x504,0x504,0x501,0x501,0x501,0x501,0x504,0xab6,0xab6,0xb97,0xb9d,0xb9d,0xb9a,
0xb9a,0xb9a,0xb9a,0xda7,0xebe,0xebe,0xebe,0xebe,0x10f8,0x60,0x60,0x60,0x60,0x60,0x60,0x60,
0x60,0x60,0x60,0x60,0x60,0x60,0x60,0x60,0x534,0x534,0x534,0xabf,0xec7,0xfbd,0xfbd,0xfbd,
0xfbd,0x1254,0x16dd,0x16dd,0x63,0x63,0x63,0x63,0x6c3,0x6c3,0x6c3,0x6c3,0x6c6,0x6c6,0x6c6,0x6c6,
0xfbd,0x1254,0x16e0,0x16e0,0x63,0x63,0x63,0x63,0x6c3,0x6c3,0x6c3,0x6c3,0x6c6,0x6c6,0x6c6,0x6c6,
0x6c6,0x6c6,0x540,0x540,0x53d,0x53d,0x53d,0x53d,0x567,0x567,0x567,0x567,0x567,0xac8,0xac8,0x66,
0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,
0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x66,0x56a,0x56a,0x56a,0x56a,0x56a,0x56a,0x56a,0x56a,
@ -1994,7 +1994,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x7e9,0x7e9,0x7e9,0x7e9,0x7e9,0x7e9,0x7ef,0x7e9,0x7e9,0x7e9,0x7e9,0x7e9,0x7ec,0xc3f,0xc3f,0x75,
0x75,0x90f,0x90f,0x8d9,0x8d9,0x7f2,0x7f5,0xc3c,0x78,0x78,0x78,0x78,0x78,0x807,0x807,0x807,
0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,
0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x10e6,0x18c0,0x78,0x7b,0x80a,0x80a,0x80a,
0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x807,0x10e6,0x18c3,0x78,0x7b,0x80a,0x80a,0x80a,
0x80a,0x80a,0x80a,0x80a,0x80a,0x80a,0x80a,0x80a,0x80a,0x80a,0x80a,0x80a,0x80a,0x80a,0x80a,0x7b,
0x8e2,0x8e2,0x8e5,0x8e5,0x8e5,0x8e5,0x8e5,0x8e5,0x8e5,0x8e5,0x8e5,0x8e5,0x8e5,0x8e5,0x8e5,0x8e5,
0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,0xaf5,
@ -2038,15 +2038,15 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x9c,
0x9c,0x9c,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x9c,0x9c,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,
0x9c,0x9c,0x71d,0x71d,0x71d,0x71d,0x71d,0x71d,0x9c,0x9c,0x71d,0x71d,0x71d,0x9c,0x9c,0x9c,
0xb1f,0xb1f,0xb1f,0xb1f,0x9f,0x9f,0x9f,0x9f,0x9f,0x9f,0x9f,0x9f,0x9f,0x185d,0x185d,0x185d,
0xb1f,0xb1f,0xb1f,0xb1f,0x9f,0x9f,0x9f,0x9f,0x9f,0x9f,0x9f,0x9f,0x9f,0x1860,0x1860,0x1860,
0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,0xb25,
0xb25,0xb25,0xb25,0xa2,0xa2,0xa2,0xa2,0xa2,0x162c,0x162c,0x162c,0x162c,0x162c,0x162c,0x162c,0x162c,
0x162c,0x162c,0x162c,0x162c,0x162c,0x162c,0x162c,0x162c,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,
0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xb2e,0xa5,0xa5,
0xa5,0xa5,0xa5,0xa5,0xa5,0xa5,0xa5,0xa5,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xa8,
0xa8,0xfc9,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,
0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0x16e3,0x16e3,0x16e3,0x16e3,0x16e3,0x16e3,0x16e3,0x16e3,
0x16e3,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,
0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0x16e6,0x16e6,0x16e6,0x16e6,0x16e6,0x16e6,0x16e6,0x16e6,
0x16e6,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,
0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xa8,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,
0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xab,0xb4f,0xb4f,
0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb52,0xb52,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,
@ -2088,8 +2088,8 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0xc3,0xc3,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,
0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,
0x14e5,0x14e5,0xc3,0xc3,0xc3,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,
0x14e5,0xc3,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x1860,0xc3,0xc3,0xc3,0xc3,0xc3,
0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0x16e6,0x16e6,0x16e6,0x16e6,0xc3,0xc3,0xc3,0xc3,
0x14e5,0xc3,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x14e5,0x1863,0xc3,0xc3,0xc3,0xc3,0xc3,
0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0x16e9,0x16e9,0x16e9,0x16e9,0xc3,0xc3,0xc3,0xc3,
0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xc3,0xd11,0xd11,0xd11,0xd11,
0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xc6,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,
0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xd11,0xc6,
@ -2108,20 +2108,20 @@ static const uint16_t propsVectorsTrie_index[29236]={
0xd1a,0xd1a,0xcf,0xd17,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,
0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,0xd23,
0xd23,0xd23,0xd2,0xd2,0xd20,0xd20,0xd20,0xd20,0xd20,0xd20,0xd20,0xd20,0xd20,0xd20,0xd2,0xd2,
0xd2,0xd2,0xd2,0xd2,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,
0x181e,0x181e,0x181e,0x181e,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd5,0xd5,0xd26,0xd5,0xd26,0xd26,
0xd2,0xd2,0xd2,0xd2,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,
0x1821,0x1821,0x1821,0x1821,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd5,0xd5,0xd26,0xd5,0xd26,0xd26,
0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,
0xd26,0xd26,0xd26,0xd26,0xd26,0xd26,0xd5,0xd26,0xd26,0xd5,0xd5,0xd5,0xd26,0xd5,0xd5,0xd26,
0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,
0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd29,0xd8,0xd8,0xd8,0xd8,0xd8,0xd8,0xd8,0xd8,0xd8,
0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0x14e8,0x14e8,0x1797,0x1797,0xde,
0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0xdda,0x14e8,0x14e8,0x179a,0x179a,0xde,
0x10c5,0x10c5,0x10c5,0x10c5,0x10c5,0x10c5,0x10c5,0x10c5,0x10c5,0x10c5,0x10c5,0x10c5,0x135,0x135,0x135,0x135,
0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,
0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xdec,0xde3,0xde3,0xde9,0xde9,0xde3,0xe1,0xe1,0xde6,0xde6,
0x10f5,0x10f5,0x10f5,0x10f5,0xe4,0xe4,0xe4,0xe4,0xe4,0xe4,0xe4,0xe4,0xe4,0xe4,0xe4,0xe4,
0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,0xc4e,
0xfe4,0xfe4,0xfe4,0xfe4,0xfe4,0xfe4,0xfe4,0x14eb,0x14eb,0x14eb,0x14eb,0x14eb,0x14eb,0x14eb,0x14eb,0x14eb,
0x14eb,0x14eb,0x14eb,0x14eb,0x14eb,0x14ee,0x1863,0x1863,0x1863,0x1863,0xe7,0x179a,0x1314,0x1137,0xee5,0xee5,
0x14eb,0x14eb,0x14eb,0x14eb,0x14eb,0x14ee,0x1866,0x1866,0x1866,0x1866,0xe7,0x179d,0x1314,0x1137,0xee5,0xee5,
0xdfe,0xdfb,0xdfe,0xdfb,0xdfb,0xdf2,0xdf2,0xdf2,0xdf2,0xdf2,0xdf2,0x1140,0x113d,0x1140,0x113d,0x113a,
0x113a,0x113a,0x13da,0x13d7,0xea,0xea,0xea,0xea,0xea,0xdf8,0xdf5,0xdf5,0xdf5,0xdf2,0xdf8,0xdf5,
0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,0xe01,
@ -2146,8 +2146,8 @@ static const uint16_t propsVectorsTrie_index[29236]={
0xe31,0xe31,0xfc,0xfc,0xfc,0xfc,0xfc,0xfc,0xe2e,0xe2e,0xe2e,0xe2e,0xe2e,0xe2e,0xe2e,0xe2e,
0xe2e,0xe2e,0x1143,0xfc,0xfc,0xfc,0xe2b,0xe2b,0xe3a,0xe3a,0xe3a,0xe3a,0xff,0xff,0xff,0xff,
0xe3a,0xe3a,0xe3a,0xe3a,0xe3a,0xe3a,0xe3a,0xe3a,0xe37,0xe3a,0xe3a,0xe3a,0xe3a,0xe3a,0xff,0xff,
0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x14f7,0x14fd,0x14fa,0x1842,0x179d,0x1866,0x1866,0x1866,
0x1866,0x1866,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,
0xff,0xff,0xff,0xff,0xff,0xff,0xff,0xff,0x14f7,0x14fd,0x14fa,0x1845,0x17a0,0x1869,0x1869,0x1869,
0x1869,0x1869,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,
0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,0x102,
0x102,0x102,0x102,0x102,0xe61,0xe61,0xe61,0xe5e,0xe5e,0xe55,0xe55,0xe5e,0xe5b,0xe5b,0xe5b,0xe5b,
0x105,0x105,0x105,0x105,0x12b1,0x12b1,0x12b1,0x12b1,0x12b1,0x12b1,0x12b4,0x12b4,0x12b7,0x12b4,0x159,0x159,
@ -2160,7 +2160,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0xf09,0xf0f,0xf24,0xf24,0xf0c,0xf09,0xf09,0xf09,0xf09,0xf09,0xf09,0xf09,0xf09,0xf09,0xf09,0xf06,
0xf06,0xf06,0xf06,0xf06,0xf06,0xf06,0xf06,0xf06,0xf09,0xf09,0xf09,0xf09,0xf09,0xf09,0xf09,0xf09,
0xf09,0x111,0x111,0x111,0x1320,0x131d,0x1320,0x131d,0x1320,0x131d,0x1320,0x131d,0x1320,0x131d,0x13ef,0x1509,
0x1509,0x1509,0x17a0,0x114,0x1509,0x1509,0x16ef,0x16ef,0x16ef,0x16e9,0x16ef,0x16e9,0x114,0x114,0x114,0x114,
0x1509,0x1509,0x17a3,0x114,0x1509,0x1509,0x16f2,0x16f2,0x16f2,0x16ec,0x16f2,0x16ec,0x114,0x114,0x114,0x114,
0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,
0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x114,0x1506,
0x13f2,0x13f2,0x131d,0x1020,0x1020,0x1020,0x1020,0x1020,0xf33,0xf33,0xf33,0xf33,0xf33,0xf33,0xf33,0xf33,
@ -2169,7 +2169,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0xf3f,0xf3f,0xf3f,0xf3f,0xf3f,0xf3f,0xf3f,0xf3f,0xf3f,0xf3f,0xf3f,0xf3f,0xf3f,0xf3f,0xf39,0xf39,
0xf39,0xf39,0x114f,0x114f,0x11a,0x11a,0x11a,0xf3c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,
0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,0x150c,
0x150c,0x16f2,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,
0x150c,0x16f5,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,
0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,0x11d,
0x11d,0x11d,0x11d,0x11d,0xf48,0xf48,0xf48,0x1512,0x1512,0x1512,0x1512,0x1512,0x1512,0x1512,0x1512,0x1512,
0x1512,0x1512,0x1512,0x120,0xf45,0xf45,0xf45,0xf45,0x150f,0x120,0x120,0x120,0x120,0x120,0x120,0x120,
@ -2181,7 +2181,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x126,0x1047,0x1047,0x1047,0x104d,0x104d,0x104d,0x104d,0x104d,0x104d,0x104d,0x104d,0x104d,0x104d,0x1050,0x1050,
0x1050,0x1050,0x1050,0x1050,0x1062,0x1062,0x1062,0x1062,0x1062,0x1062,0x1062,0x1062,0x1062,0x1062,0x1065,0x1065,
0x129,0x129,0x129,0x129,0x129,0x129,0x129,0x129,0x129,0x129,0x129,0x129,0x129,0x129,0x129,0x129,
0x129,0x129,0x129,0x129,0x108c,0x108c,0x108c,0x108c,0x1086,0x17a3,0x12c,0x12c,0x12c,0x12c,0x12c,0x12c,
0x129,0x129,0x129,0x129,0x108c,0x108c,0x108c,0x108c,0x1086,0x17a6,0x12c,0x12c,0x12c,0x12c,0x12c,0x12c,
0x12c,0x12c,0x1092,0x1092,0x1089,0x1089,0x1089,0x1089,0x1089,0x1089,0x1089,0x1089,0x1089,0x1089,0x12c,0x12c,
0x12c,0x12c,0x12c,0x12c,0x10b0,0x10b0,0x10b0,0x10b0,0x10b0,0x10b0,0x10b0,0x10a4,0x10a4,0x10a4,0x10a4,0x10a4,
0x10a4,0x10a4,0x10a4,0x10a4,0x10a4,0x10a4,0x10aa,0x10ad,0x12f,0x12f,0x12f,0x12f,0x12f,0x12f,0x12f,0x12f,
@ -2215,12 +2215,12 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x150,0x150,0x150,0x150,0x1170,0x1170,0x1170,0x1170,0x1170,0x1170,0x1170,0x117f,0x1182,0x1182,0x1182,0x1182,
0x1170,0x1170,0x150,0x150,0x1569,0x1569,0x1569,0x1569,0x1569,0x1569,0x1569,0x1569,0x1569,0x1569,0x1569,0x1569,
0x1569,0x1569,0x1566,0x1cb,0x12c6,0x12a5,0x12c0,0x12c0,0x12c0,0x12c0,0x12c0,0x12c0,0x12c0,0x12a8,0x12a8,0x12a8,
0x12a8,0x12c0,0x12a8,0x12a8,0x12a8,0x12a8,0x12ae,0x1494,0x149a,0x1497,0x1491,0x18e1,0x16bf,0x16bf,0x153,0x153,
0x12a8,0x12c0,0x12a8,0x12a8,0x12a8,0x12a8,0x12ae,0x1494,0x149a,0x1497,0x1491,0x18e4,0x16bf,0x16bf,0x153,0x153,
0x153,0x153,0x153,0x153,0x1197,0x1197,0x1197,0x1197,0x1197,0x1197,0x1197,0x1197,0x1197,0x1197,0x1197,0x1197,
0x1197,0x1197,0x1197,0x1197,0x118e,0x118e,0x1191,0x119a,0x1194,0x1194,0x1194,0x119a,0x156,0x156,0x156,0x156,
0x156,0x156,0x156,0x156,0x119d,0x119d,0x119d,0x119d,0x119d,0x119d,0x119d,0x119d,0x119d,0x119d,0x119d,0x119d,
0x119d,0x119d,0x119d,0x119d,0x119d,0x12cc,0x11a3,0x12cf,0x11a3,0x11a3,0x11a3,0x11a3,0x11a0,0x11a0,0x11a0,0x11a3,
0x16f8,0x16fb,0x15c,0x15c,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,
0x16fb,0x16fe,0x15c,0x15c,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,
0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,0x1293,
0x1293,0x15f,0x15f,0x15f,0x11b8,0x11ac,0x11ac,0x11ac,0x11ac,0x11ac,0x11ac,0x11af,0x11be,0x11be,0x11ac,0x11ac,
0x11ac,0x11ac,0x162,0x12ba,0x11b2,0x11b2,0x11b2,0x11b2,0x11b2,0x11b2,0x11b2,0x11b2,0x11b2,0x11b2,0x162,0x162,
@ -2248,18 +2248,18 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x126f,0x126f,0x126f,0x1272,0x1272,0x1272,0x1251,0x186,0x1374,0x127b,0x1374,0x1374,0x1374,0x1374,0x1374,0x1374,
0x1374,0x1374,0x1374,0x1374,0x1374,0x127b,0x1374,0x127b,0x1371,0x1371,0x1371,0x1371,0x1371,0x1371,0x1371,0x1371,
0x1371,0x1371,0x1401,0x1401,0x186,0x186,0x186,0x186,0x1377,0x1377,0x1371,0x1371,0x1371,0x1371,0x1371,0x1371,
0x1371,0x1278,0x1371,0x1278,0x1278,0x1371,0x1377,0x127e,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,
0x1821,0x1821,0x1821,0x1821,0x1821,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,
0x1371,0x1278,0x1371,0x1278,0x1278,0x1371,0x1377,0x127e,0x1824,0x1824,0x1824,0x1824,0x1824,0x1824,0x1824,0x1824,
0x1824,0x1824,0x1824,0x1824,0x1824,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,
0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,
0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x186,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,
0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,0x1329,
0x1329,0x1329,0x1329,0x1329,0x129f,0x1392,0x138f,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,
0x189,0x189,0x189,0x189,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x129c,0x1299,
0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x1299,0x129c,
0x1299,0x1299,0x1392,0x1392,0x1392,0x1392,0x1392,0x138f,0x1392,0x1392,0x1392,0x1824,0x189,0x189,0x189,0x189,
0x1299,0x1299,0x1392,0x1392,0x1392,0x1392,0x1392,0x138f,0x1392,0x1392,0x1392,0x1827,0x189,0x189,0x189,0x189,
0x1296,0x1296,0x1296,0x1296,0x1296,0x1296,0x1296,0x1296,0x1296,0x189,0x189,0x189,0x189,0x189,0x189,0x189,
0x13bf,0x13bf,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,
0x18c6,0x18c6,0x18c6,0x18c6,0x18c6,0x18c6,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,
0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,
0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,
0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x189,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,
0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,0x1332,
@ -2282,16 +2282,16 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x1365,0x1365,0x1365,0x1365,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,
0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x1530,0x19b,0x19b,0x19b,0x19b,0x19b,0x19b,
0x19b,0x19b,0x19b,0x19b,0x13b9,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x1545,0x1545,0x1545,0x1545,0x1545,0x1548,
0x16b6,0x1548,0x1548,0x1548,0x177f,0x182d,0x182d,0x1869,0x1869,0x19e,0x19e,0x19e,0x19e,0x19e,0x19e,0x19e,
0x16b6,0x1548,0x1548,0x1548,0x1782,0x1830,0x1830,0x186c,0x186c,0x19e,0x19e,0x19e,0x19e,0x19e,0x19e,0x19e,
0x19e,0x19e,0x19e,0x19e,0x1548,0x1548,0x1548,0x1548,0x1548,0x1548,0x1545,0x1545,0x1545,0x1548,0x1545,0x16b3,
0x16b3,0x19e,0x19e,0x19e,0x1548,0x1545,0x1545,0x1548,0x182d,0x182d,0x182d,0x18cc,0x18cc,0x19e,0x19e,0x19e,
0x16b3,0x19e,0x19e,0x19e,0x1548,0x1545,0x1545,0x1548,0x1830,0x1830,0x1830,0x18cf,0x18cf,0x19e,0x19e,0x19e,
0x19e,0x19e,0x19e,0x19e,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,
0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1a1,0x1a1,0x1a1,0x1a1,0x1a1,0x1a1,0x1a1,0x1a1,
0x1a1,0x1a1,0x1a1,0x1a1,0x140d,0x154e,0x140d,0x140d,0x140d,0x140d,0x140d,0x140d,0x140d,0x140d,0x140d,0x140d,
0x140d,0x154e,0x154e,0x154e,0x154e,0x154e,0x154e,0x1704,0x1704,0x1a4,0x17af,0x17af,0x17af,0x17af,0x17af,0x17af,
0x17af,0x17af,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,
0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x17ac,0x17ac,0x17ac,0x17ac,0x17ac,0x17ac,0x17ac,0x17ac,
0x17ac,0x17ac,0x17ac,0x17ac,0x1413,0x1413,0x1413,0x1413,0x1a7,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,
0x140d,0x154e,0x154e,0x154e,0x154e,0x154e,0x154e,0x1707,0x1707,0x1a4,0x17b2,0x17b2,0x17b2,0x17b2,0x17b2,0x17b2,
0x17b2,0x17b2,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,
0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x1a4,0x17af,0x17af,0x17af,0x17af,0x17af,0x17af,0x17af,0x17af,
0x17af,0x17af,0x17af,0x17af,0x1413,0x1413,0x1413,0x1413,0x1a7,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,
0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,
0x1413,0x1413,0x1413,0x1413,0x1a7,0x1413,0x1413,0x1a7,0x1413,0x1a7,0x1a7,0x1413,0x1a7,0x1413,0x1413,0x1413,
0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1413,0x1a7,0x1413,0x1413,0x1413,0x1413,0x1a7,0x1413,0x1a7,0x1413,
@ -2313,17 +2313,17 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x1437,0x1437,0x1437,0x1437,0x1437,0x1437,0x1437,0x1437,0x1437,0x1437,0x1437,0x1434,0x142e,0x142e,0x1434,0x1434,
0x143d,0x143d,0x1437,0x143a,0x143a,0x1434,0x1431,0x1ad,0x1ad,0x1ad,0x1ad,0x1ad,0x1ad,0x1ad,0x1ad,0x1ad,
0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,
0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1b0,0x1b0,0x1b0,0x1b0,0x1707,0x1707,0x1440,0x1440,
0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,
0x1b0,0x1b0,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,
0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1440,0x1b0,0x1b0,0x1b0,0x1b0,0x170a,0x170a,0x1440,0x1440,
0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,
0x1b0,0x1b0,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,
0x144c,0x144c,0x144c,0x144c,0x144c,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,
0x144c,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,
0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,0x1449,
0x1449,0x1449,0x1449,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,0x1b3,
0x1b3,0x1b3,0x1b3,0x1446,0x1446,0x1446,0x1446,0x144f,0x144f,0x144f,0x144f,0x144f,0x144f,0x144f,0x144f,0x144f,
0x144f,0x144f,0x144f,0x144f,0x1461,0x1464,0x1467,0x1467,0x1464,0x146a,0x146a,0x1455,0x1458,0x170d,0x170a,0x170a,
0x170a,0x1554,0x1b6,0x1b6,0x145b,0x145b,0x145b,0x145b,0x145b,0x145b,0x145b,0x145b,0x145b,0x145b,0x1551,0x1713,
0x1716,0x1710,0x1719,0x1719,0x1470,0x1470,0x1470,0x1470,0x1470,0x1470,0x1470,0x1470,0x1470,0x1b9,0x1b9,0x1b9,
0x144f,0x144f,0x144f,0x144f,0x1461,0x1464,0x1467,0x1467,0x1464,0x146a,0x146a,0x1455,0x1458,0x1710,0x170d,0x170d,
0x170d,0x1554,0x1b6,0x1b6,0x145b,0x145b,0x145b,0x145b,0x145b,0x145b,0x145b,0x145b,0x145b,0x145b,0x1551,0x1716,
0x1719,0x1713,0x171c,0x171c,0x1470,0x1470,0x1470,0x1470,0x1470,0x1470,0x1470,0x1470,0x1470,0x1b9,0x1b9,0x1b9,
0x1b9,0x1b9,0x1b9,0x1b9,0x146d,0x146d,0x146d,0x146d,0x146d,0x146d,0x146d,0x146d,0x146d,0x146d,0x1b9,0x1b9,
0x1b9,0x1b9,0x1b9,0x1b9,0x1473,0x1473,0x1473,0x1473,0x1473,0x1473,0x1473,0x1473,0x1bc,0x1bc,0x1bc,0x1bc,
0x1bc,0x1bc,0x1bc,0x1bc,0x12c3,0x12c0,0x12c3,0x12ab,0x12c0,0x12c0,0x12c0,0x12c6,0x12c0,0x12c6,0x12c9,0x12c0,
@ -2344,21 +2344,21 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x157e,0x157e,0x157e,0x157e,0x157e,0x157e,0x157e,0x157e,0x157e,0x157e,0x157e,0x157e,0x157e,0x1d4,0x1d4,0x1d4,
0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,
0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,0x1d4,
0x171c,0x1581,0x1587,0x16c8,0x1d7,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1d7,0x1d7,0x1590,
0x171f,0x16c5,0x1587,0x16cb,0x1d7,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1d7,0x1d7,0x1590,
0x1590,0x1d7,0x1d7,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,
0x1590,0x1d7,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1590,0x1d7,0x1590,0x1590,0x1d7,0x1590,0x1590,0x1590,
0x1590,0x1590,0x1d7,0x1d7,0x16c5,0x1590,0x1581,0x1587,0x1581,0x1587,0x1587,0x1587,0x1587,0x1d7,0x1d7,0x1587,
0x1587,0x1d7,0x1d7,0x158a,0x158a,0x158d,0x1d7,0x1d7,0x171f,0x1d7,0x1d7,0x1d7,0x1d7,0x1d7,0x1d7,0x1581,
0x1590,0x1590,0x1d7,0x1d7,0x16c8,0x1590,0x1581,0x1587,0x1581,0x1587,0x1587,0x1587,0x1587,0x1d7,0x1d7,0x1587,
0x1587,0x1d7,0x1d7,0x158a,0x158a,0x158d,0x1d7,0x1d7,0x1722,0x1d7,0x1d7,0x1d7,0x1d7,0x1d7,0x1d7,0x1581,
0x1d7,0x1d7,0x1d7,0x1d7,0x1d7,0x1593,0x1590,0x1590,0x1590,0x1590,0x1587,0x1587,0x1d7,0x1d7,0x1584,0x1584,
0x1584,0x1584,0x1584,0x1584,0x1584,0x1d7,0x1d7,0x1d7,0x1584,0x1584,0x1584,0x1584,0x1584,0x1d7,0x1d7,0x1d7,
0x1d7,0x1d7,0x1d7,0x1d7,0x1d7,0x1d7,0x1d7,0x1d7,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,
0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x1da,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,
0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a8,0x15a2,0x15a2,0x15a2,0x1596,0x1596,0x1596,0x15a2,0x15a2,
0x1596,0x15a5,0x1599,0x1596,0x15ab,0x15ab,0x159f,0x15ab,0x15ab,0x159c,0x17b2,0x1da,0x15ba,0x15ba,0x15ba,0x15ae,
0x1596,0x15a5,0x1599,0x1596,0x15ab,0x15ab,0x159f,0x15ab,0x15ab,0x159c,0x17b5,0x1da,0x15ba,0x15ba,0x15ba,0x15ae,
0x15ae,0x15ae,0x15ae,0x15ae,0x15ae,0x15b1,0x15b4,0x1dd,0x1dd,0x1dd,0x1dd,0x1dd,0x15b7,0x15b7,0x15b7,0x15b7,
0x15b7,0x15b7,0x15b7,0x15b7,0x15b7,0x15b7,0x1dd,0x1dd,0x1dd,0x1dd,0x1dd,0x1dd,0x1722,0x1722,0x1722,0x1722,
0x15c6,0x15c3,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x174c,0x174c,0x174c,0x174c,
0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x15cc,0x15cc,0x15cc,0x15cc,
0x15b7,0x15b7,0x15b7,0x15b7,0x15b7,0x15b7,0x1dd,0x1dd,0x1dd,0x1dd,0x1dd,0x1dd,0x1725,0x1725,0x1725,0x1725,
0x15c6,0x15c3,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x1e0,0x174f,0x174f,0x174f,0x174f,
0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x15cc,0x15cc,0x15cc,0x15cc,
0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,
0x15cc,0x15cc,0x15cc,0x1e3,0x1e3,0x1e3,0x1e3,0x1e3,0x1e3,0x1e3,0x1e3,0x1e3,0x15cc,0x15cc,0x15cc,0x15cc,
0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,0x15cc,
@ -2391,8 +2391,8 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x204,0x204,0x204,0x204,0x204,0x1659,0x1659,0x1659,0x1659,0x204,0x204,0x204,0x1677,0x1677,0x1677,0x1677,
0x1677,0x1677,0x1677,0x1677,0x1677,0x1677,0x1677,0x1677,0x1677,0x1677,0x1677,0x165f,0x1671,0x1671,0x165f,0x165f,
0x165f,0x165f,0x20a,0x20a,0x1671,0x1671,0x1674,0x1674,0x165f,0x165f,0x1671,0x1665,0x1662,0x1668,0x167a,0x167a,
0x166b,0x166b,0x166e,0x166e,0x166e,0x167a,0x172b,0x172b,0x172b,0x172b,0x172b,0x172b,0x172b,0x172b,0x172b,0x172b,
0x172b,0x172b,0x172b,0x172b,0x1728,0x1728,0x1728,0x1728,0x1725,0x1725,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,
0x166b,0x166b,0x166e,0x166e,0x166e,0x167a,0x172e,0x172e,0x172e,0x172e,0x172e,0x172e,0x172e,0x172e,0x172e,0x172e,
0x172e,0x172e,0x172e,0x172e,0x172b,0x172b,0x172b,0x172b,0x1728,0x1728,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,
0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,
0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20a,0x20d,0x167d,0x167d,0x167d,
0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,0x167d,
@ -2408,92 +2408,92 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x213,0x213,0x213,0x213,0x213,0x213,0x213,0x213,0x168f,0x168f,0x168f,0x168f,0x168f,0x168f,0x168f,0x168f,
0x168f,0x168f,0x213,0x213,0x213,0x213,0x213,0x213,0x169e,0x169e,0x169e,0x169e,0x169e,0x169e,0x169e,0x169e,
0x169e,0x169e,0x169b,0x169b,0x169b,0x169b,0x169b,0x169b,0x169b,0x169b,0x169b,0x216,0x216,0x216,0x216,0x216,
0x216,0x216,0x216,0x216,0x216,0x216,0x216,0x16a4,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,
0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,0x173d,
0x173d,0x173d,0x219,0x219,0x219,0x172e,0x172e,0x172e,0x173a,0x173a,0x172e,0x172e,0x172e,0x172e,0x173a,0x172e,
0x172e,0x172e,0x172e,0x1731,0x219,0x219,0x219,0x219,0x1737,0x1737,0x1737,0x1737,0x1737,0x1737,0x1737,0x1737,
0x1737,0x1737,0x1734,0x1734,0x1740,0x1740,0x1740,0x1734,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x21c,
0x216,0x216,0x216,0x216,0x216,0x216,0x216,0x16a4,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,
0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,0x1740,
0x1740,0x1740,0x219,0x219,0x219,0x1731,0x1731,0x1731,0x173d,0x173d,0x1731,0x1731,0x1731,0x1731,0x173d,0x1731,
0x1731,0x1731,0x1731,0x1734,0x219,0x219,0x219,0x219,0x173a,0x173a,0x173a,0x173a,0x173a,0x173a,0x173a,0x173a,
0x173a,0x173a,0x1737,0x1737,0x1743,0x1743,0x1743,0x1737,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x21c,
0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,
0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,0x21c,
0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,0x1755,
0x1755,0x1755,0x1755,0x222,0x1755,0x1755,0x222,0x222,0x222,0x222,0x222,0x1752,0x1752,0x1752,0x1752,0x1752,
0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x225,0x1758,0x225,0x1758,0x1758,0x1758,0x1758,0x225,0x1758,
0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x225,0x1758,
0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x175b,0x225,0x225,0x225,0x225,0x225,0x225,
0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,0x1758,
0x1758,0x1758,0x1758,0x222,0x1758,0x1758,0x222,0x222,0x222,0x222,0x222,0x1755,0x1755,0x1755,0x1755,0x1755,
0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x225,0x175b,0x225,0x175b,0x175b,0x175b,0x175b,0x225,0x175b,
0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x225,0x175b,
0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175b,0x175e,0x225,0x225,0x225,0x225,0x225,0x225,
0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,0x15bd,
0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,
0x1767,0x1767,0x1767,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,
0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,
0x1764,0x1764,0x1764,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x228,
0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,
0x1761,0x1761,0x1761,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x175e,0x175e,0x175e,0x175e,0x175e,0x175e,
0x186c,0x186c,0x186c,0x186c,0x186c,0x186c,0x186c,0x186c,0x186c,0x186c,0x186c,0x186c,0x22b,0x22b,0x22b,0x22b,
0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1788,0x1836,0x1836,0x1836,0x1836,0x1833,0x1836,0x18d2,
0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1836,0x1833,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,
0x1836,0x18d2,0x18d2,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1833,0x1830,0x1833,0x1836,0x1836,0x22b,
0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1830,0x1833,0x1833,0x1833,0x1833,0x1833,0x18cf,0x22b,0x22b,0x22b,
0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x18cf,
0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x22b,0x22b,0x22b,0x22b,
0x1764,0x1764,0x1764,0x228,0x228,0x228,0x228,0x228,0x228,0x228,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,
0x186f,0x186f,0x186f,0x186f,0x186f,0x186f,0x186f,0x186f,0x186f,0x186f,0x186f,0x186f,0x22b,0x22b,0x22b,0x22b,
0x1788,0x1788,0x1788,0x1788,0x1788,0x1788,0x1788,0x1788,0x178b,0x1839,0x1839,0x1839,0x1839,0x1836,0x1839,0x18d5,
0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1839,0x1836,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,
0x1839,0x18d5,0x18d5,0x1839,0x1839,0x1839,0x1839,0x1839,0x1839,0x1839,0x1836,0x1833,0x1836,0x1839,0x1839,0x22b,
0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1833,0x1836,0x1836,0x1836,0x1836,0x1836,0x18d2,0x22b,0x22b,0x22b,
0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x18d2,
0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x22b,0x22b,0x22b,0x22b,
0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,
0x1785,0x1785,0x1785,0x1785,0x1785,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,0x1833,
0x1833,0x1833,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,
0x1788,0x1788,0x1788,0x1788,0x1788,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,0x1836,
0x1836,0x1836,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,
0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,
0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x1785,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,
0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x18cf,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,
0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x18cf,0x22b,
0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x1788,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,
0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x18d2,0x18d5,0x18d5,0x18d5,0x18d5,0x18d5,0x18d5,0x18d5,
0x18d5,0x18d5,0x18d5,0x18d5,0x18d5,0x18d5,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x18d2,0x22b,
0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,
0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x176a,0x176a,0x176a,0x176a,0x1767,0x176a,0x176a,0x176d,
0x1770,0x176d,0x176d,0x176a,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,
0x22e,0x22e,0x22e,0x1767,0x1767,0x1767,0x1767,0x1767,0x17c4,0x17c4,0x17c4,0x17c4,0x17bb,0x17bb,0x17bb,0x17b5,
0x17b8,0x17b8,0x17b8,0x231,0x231,0x231,0x231,0x231,0x17c1,0x17c1,0x17c1,0x17c1,0x17c1,0x17c1,0x17c1,0x17c1,
0x17c1,0x17c1,0x231,0x231,0x231,0x231,0x17be,0x17be,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,
0x17df,0x234,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,
0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17df,0x17dc,0x17ca,0x17ca,0x17ca,0x17ca,
0x17ca,0x17ca,0x17ca,0x234,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17dc,0x17cd,0x17df,0x17e2,0x17e2,0x17d6,
0x17d3,0x17d3,0x234,0x234,0x234,0x234,0x234,0x234,0x234,0x234,0x234,0x234,0x17d9,0x17d9,0x17d9,0x17d9,
0x17d9,0x17d9,0x17d9,0x17d9,0x17d9,0x17d9,0x17d0,0x17d0,0x17d0,0x17d0,0x17d0,0x17d0,0x17d0,0x17d0,0x17d0,0x17d0,
0x17d0,0x17d0,0x17d0,0x17d0,0x17d0,0x234,0x234,0x234,0x17ee,0x17f1,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,
0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x23a,
0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,
0x17e8,0x23a,0x23a,0x17e8,0x17e8,0x17e8,0x17e8,0x17e8,0x1839,0x18d5,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,
0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x22b,0x176d,0x176d,0x176d,0x176d,0x176a,0x176d,0x176d,0x1770,
0x1773,0x1770,0x1770,0x176d,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,0x22e,
0x22e,0x22e,0x22e,0x176a,0x176a,0x176a,0x176a,0x176a,0x17c7,0x17c7,0x17c7,0x17c7,0x17be,0x17be,0x17be,0x17b8,
0x17bb,0x17bb,0x17bb,0x231,0x231,0x231,0x231,0x231,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,
0x17c4,0x17c4,0x231,0x231,0x231,0x231,0x17c1,0x17c1,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,
0x17e2,0x234,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,
0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17e2,0x17df,0x17cd,0x17cd,0x17cd,0x17cd,
0x17cd,0x17cd,0x17cd,0x234,0x17cd,0x17cd,0x17cd,0x17cd,0x17cd,0x17cd,0x17df,0x17d0,0x17e2,0x17e5,0x17e5,0x17d9,
0x17d6,0x17d6,0x234,0x234,0x234,0x234,0x234,0x234,0x234,0x234,0x234,0x234,0x17dc,0x17dc,0x17dc,0x17dc,
0x17dc,0x17dc,0x17dc,0x17dc,0x17dc,0x17dc,0x17d3,0x17d3,0x17d3,0x17d3,0x17d3,0x17d3,0x17d3,0x17d3,0x17d3,0x17d3,
0x17d3,0x17d3,0x17d3,0x17d3,0x17d3,0x234,0x234,0x234,0x17f1,0x17f4,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,
0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x23a,
0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,
0x17eb,0x23a,0x23a,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x183c,0x18d8,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,
0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,
0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,
0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x17f7,0x240,0x240,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,
0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x240,0x17f4,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,0x17eb,
0x17eb,0x17f4,0x17eb,0x17eb,0x17f4,0x17eb,0x17eb,0x240,0x240,0x240,0x240,0x240,0x240,0x240,0x240,0x240,
0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x243,0x243,0x243,
0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x23d,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,
0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x17fa,0x240,0x240,0x17ee,0x17ee,0x17ee,0x17ee,0x17ee,0x17ee,
0x17ee,0x17ee,0x17ee,0x17ee,0x17ee,0x17ee,0x17ee,0x17ee,0x240,0x17f7,0x17ee,0x17ee,0x17ee,0x17ee,0x17ee,0x17ee,
0x17ee,0x17f7,0x17ee,0x17ee,0x17f7,0x17ee,0x17ee,0x240,0x240,0x240,0x240,0x240,0x240,0x240,0x240,0x240,
0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x243,0x243,0x243,
0x243,0x243,0x243,0x243,0x243,0x243,0x243,0x243,0x243,0x243,0x243,0x243,0x243,0x243,0x243,0x243,
0x1812,0x1812,0x1803,0x17fd,0x17fd,0x1812,0x1800,0x1815,0x1815,0x1815,0x1815,0x1818,0x1818,0x180c,0x1809,0x1806,
0x180f,0x180f,0x180f,0x180f,0x180f,0x180f,0x180f,0x180f,0x180f,0x180f,0x246,0x180c,0x246,0x1806,0x246,0x246,
0x1815,0x1815,0x1806,0x1800,0x1800,0x1815,0x1803,0x1818,0x1818,0x1818,0x1818,0x181b,0x181b,0x180f,0x180c,0x1809,
0x1812,0x1812,0x1812,0x1812,0x1812,0x1812,0x1812,0x1812,0x1812,0x1812,0x246,0x180f,0x246,0x1809,0x246,0x246,
0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,
0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,0x246,
0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,0x1821,
0x1821,0x1821,0x1821,0x1821,0x249,0x249,0x249,0x249,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,
0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,0x181e,
0x181e,0x181e,0x181e,0x181e,0x249,0x249,0x249,0x249,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,
0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,0x181b,
0x181b,0x181b,0x181b,0x181b,0x249,0x249,0x249,0x249,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,
0x183c,0x183c,0x183c,0x183c,0x183c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,
0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,
0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x24f,0x24f,0x24f,0x24f,0x24f,
0x24f,0x24f,0x24f,0x24f,0x24f,0x24f,0x24f,0x24f,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,
0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,
0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x252,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x255,
0x187b,0x187b,0x255,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,
0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x186f,0x186f,0x186f,0x186f,0x186f,0x186f,0x255,
0x255,0x255,0x186f,0x255,0x186f,0x186f,0x255,0x186f,0x186f,0x186f,0x1872,0x186f,0x1875,0x1875,0x187e,0x186f,
0x255,0x255,0x255,0x255,0x255,0x255,0x255,0x255,0x1878,0x1878,0x1878,0x1878,0x1878,0x1878,0x1878,0x1878,
0x1878,0x1878,0x255,0x255,0x255,0x255,0x255,0x255,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,
0x181e,0x181e,0x181e,0x181e,0x249,0x249,0x249,0x249,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,
0x183f,0x183f,0x183f,0x183f,0x183f,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,
0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x24c,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,
0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x24f,0x24f,0x24f,0x24f,0x24f,
0x24f,0x24f,0x24f,0x24f,0x24f,0x24f,0x24f,0x24f,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,
0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,
0x18de,0x18de,0x18de,0x18de,0x258,0x258,0x258,0x258,0x1896,0x1896,0x1896,0x1896,0x25b,0x25b,0x1899,0x1899,
0x1899,0x1899,0x1881,0x1881,0x1881,0x1881,0x1881,0x1881,0x1881,0x1881,0x1881,0x1881,0x1881,0x1881,0x1881,0x1893,
0x1884,0x1887,0x188a,0x189c,0x189c,0x25b,0x188d,0x188d,0x18ab,0x18ae,0x18bd,0x18bd,0x18ae,0x18b1,0x18ab,0x18a8,
0x261,0x261,0x261,0x261,0x261,0x261,0x261,0x261,0x1896,0x1881,0x1881,0x1881,0x1881,0x1881,0x1881,0x1893,
0x1893,0x1881,0x1881,0x1881,0x1896,0x1896,0x1896,0x1896,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,
0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x252,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x255,
0x187e,0x187e,0x255,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,
0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x187e,0x1872,0x1872,0x1872,0x1872,0x1872,0x1872,0x255,
0x255,0x255,0x1872,0x255,0x1872,0x1872,0x255,0x1872,0x1872,0x1872,0x1875,0x1872,0x1878,0x1878,0x1881,0x1872,
0x255,0x255,0x255,0x255,0x255,0x255,0x255,0x255,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,0x187b,
0x187b,0x187b,0x255,0x255,0x255,0x255,0x255,0x255,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,
0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,
0x18e1,0x18e1,0x18e1,0x18e1,0x258,0x258,0x258,0x258,0x1899,0x1899,0x1899,0x1899,0x25b,0x25b,0x189c,0x189c,
0x189c,0x189c,0x1884,0x1884,0x1884,0x1884,0x1884,0x1884,0x1884,0x1884,0x1884,0x1884,0x1884,0x1884,0x1884,0x1896,
0x1887,0x188a,0x188d,0x189f,0x189f,0x25b,0x1890,0x1890,0x18ae,0x18b1,0x18c0,0x18c0,0x18b1,0x18b4,0x18ae,0x18ab,
0x261,0x261,0x261,0x261,0x261,0x261,0x261,0x261,0x1899,0x1884,0x1884,0x1884,0x1884,0x1884,0x1884,0x1896,
0x1896,0x1884,0x1884,0x1884,0x1899,0x1899,0x1899,0x1899,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,
0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,
0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x18d8,0x27f,0x27f,0x27f,0x27f,0x27f,0x27f,0x27f,
0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x18db,0x27f,0x27f,0x27f,0x27f,0x27f,0x27f,0x27f,
0x27f,0x27f,0x27f,0x27f,0x27f,0x27f,0x27f,0x27f,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,
0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,0x264,
0x264,0x264,0x264,0x264,0x264,0x264,0x933,0x933,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,
0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0xaf8,0x267,0x267,
0x267,0x267,0x267,0x267,0x267,0x267,0x267,0x267,0x18c3,0x18c3,0x18c3,0x18c3,0x18c3,0x18c3,0x18c3,0x18c3,
0x18c3,0x18c3,0x18c3,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,
0x267,0x267,0x267,0x267,0x267,0x267,0x267,0x267,0x18c6,0x18c6,0x18c6,0x18c6,0x18c6,0x18c6,0x18c6,0x18c6,
0x18c6,0x18c6,0x18c6,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,
0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0x26a,0xc4b,0xc4b,0xc4b,0xc4b,0xc4b,0xc4b,0xc4b,0xc4b,
0xc4b,0xc4b,0xc4b,0x128d,0x128d,0x128d,0x26d,0x26d,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,
0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,0xe7c,
@ -2507,9 +2507,9 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x276,0x276,0x276,0x276,0x276,0x276,0x276,
0x276,0x276,0x276,0x276,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,
0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,
0x13bc,0x13bc,0x279,0x279,0x1782,0x1782,0x27c,0x27c,0x27c,0x27c,0x27c,0x27c,0x27c,0x27c,0x27c,0x27c,
0x27c,0x27c,0x27c,0x27c,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,
0x18d8,0x18d8,0x18d8,0x18d8,0x1101,0x378,0x378,0x384,0xc8d,0x387,0x387,0x387,0x387,0x387,0x387,0x387,
0x13bc,0x13bc,0x279,0x279,0x1785,0x1785,0x27c,0x27c,0x27c,0x27c,0x27c,0x27c,0x27c,0x27c,0x27c,0x27c,
0x27c,0x27c,0x27c,0x27c,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,
0x18db,0x18db,0x18db,0x18db,0x1101,0x378,0x378,0x384,0xc8d,0x387,0x387,0x387,0x387,0x387,0x387,0x387,
0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,
0x387,0x387,0x387,0x387,0x384,0x378,0x378,0x378,0x378,0x378,0x378,0x378,0x378,0x384,0x384,0x384,
0x384,0x37e,0x1104,0x12db,0x387,0x900,0x903,0x37b,0x37b,0x1101,0x12d8,0x12d8,0x38a,0x38a,0x38a,0x38a,
@ -2518,8 +2518,8 @@ static const uint16_t propsVectorsTrie_index[29236]={
0xeb5,0xd80,0xeb5,0xeb5,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x387,0x38a,0x387,0x387,
0x387,0x387,0x387,0x387,0x387,0x38a,0x387,0x387,0x38a,0x387,0x387,0x387,0x387,0x387,0x12d8,0x12db,
0x37b,0x387,0x384,0x384,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x12e4,0x429,0x429,
0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x12e4,0x1854,
0x1854,0xf81,0x41a,0x423,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,
0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x429,0x12e4,0x1857,
0x1857,0xf81,0x41a,0x423,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,
0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0x465,0xb79,0xb79,0xd8c,0xd8c,0x885,
0xd8f,0x13ce,0x13ce,0x13ce,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,
0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,0x468,
@ -2543,7 +2543,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0xaaa,0xaaa,0xb85,0xb85,0xb85,0xb85,0xb8b,0xcba,0xcb7,0xd9b,0xd9e,0xb88,0xd9e,0xd9e,0xd9e,0xd9e,
0xd9b,0xd9e,0xd9e,0xb82,0x4fb,0x4fb,0x4fb,0x4fb,0x4fb,0x4fb,0x4fb,0x4f8,0x4fe,0x714,0x4fb,0x978,
0x999,0xab3,0xab3,0xab3,0xb94,0xb94,0xda4,0xda4,0xda4,0xda4,0x1125,0x1128,0x1128,0x12f9,0x149d,0x14c7,
0x14ca,0x14ca,0x16da,0x1857,0x50a,0x50a,0x522,0x6a2,0x507,0x69f,0x50a,0x51f,0x507,0x6a2,0x519,0x522,
0x14ca,0x14ca,0x16dd,0x185a,0x50a,0x50a,0x522,0x6a2,0x507,0x69f,0x50a,0x51f,0x507,0x6a2,0x519,0x522,
0x522,0x522,0x519,0x519,0x522,0x522,0x522,0x6ab,0x507,0x522,0x6a5,0x507,0x516,0x522,0x522,0x522,
0x522,0x522,0x507,0x507,0x50d,0x69f,0x6a8,0x507,0x522,0x507,0x6ae,0x507,0x522,0x510,0x528,0x6b1,
0x522,0x522,0x513,0x519,0x522,0x522,0x525,0x522,0x519,0x51c,0x51c,0x51c,0x51c,0xabc,0xab9,0xcbd,
@ -2642,7 +2642,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d9,0x5d9,0x5d9,0x5d9,
0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,0x5d9,
0x5d9,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5d6,0x5dc,0x5dc,0x5dc,0x5dc,
0xfc6,0xfc6,0xfc6,0x14d6,0x14d6,0x14d6,0x14d6,0x14d6,0x14d6,0x14d6,0x16e0,0x16e0,0x831,0x837,0x837,0x843,
0xfc6,0xfc6,0xfc6,0x14d6,0x14d6,0x14d6,0x14d6,0x14d6,0x14d6,0x14d6,0x16e3,0x16e3,0x831,0x837,0x837,0x843,
0x843,0x834,0x82b,0x834,0x82b,0x834,0x82b,0x834,0x82b,0x834,0x82b,0x834,0x5eb,0x5eb,0x5e5,0x5eb,
0x5e5,0x5eb,0x5e5,0x5eb,0x5e5,0x5eb,0x5e5,0x5e8,0x5ee,0x5eb,0x5e5,0x5eb,0x5e5,0x5e8,0x5ee,0x5eb,
0x5e5,0x5eb,0x5e5,0x5e8,0x5ee,0x5eb,0x5e5,0x5e8,0x5ee,0x5eb,0x5e5,0x5e8,0x5ee,0x5eb,0x5e5,0x5eb,
@ -2795,7 +2795,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb46,0xb46,0xb46,0xb46,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,
0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3d,0xb3d,0xb3d,0xb3d,0xb3d,0xb3d,0xb3a,0xb3a,0xb3a,
0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,
0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0x16e3,0x16e3,0xb52,0xb49,0xb4f,0xb4f,
0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0xb3a,0x16e6,0x16e6,0xb52,0xb49,0xb4f,0xb4f,
0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,
0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb49,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb4f,0xb52,0xb52,
0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,0xb52,
@ -2903,7 +2903,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0xe9a,0xe9a,0xe9a,0xe9a,0xe9a,0xe9a,0xe9a,0xe9a,0xe9a,0xe9a,0xe9a,0xe9a,0x10f5,0x10f5,0x10f5,0x10f5,
0x10f5,0x10f5,0x10f5,0x10f5,0x10f5,0x10f5,0x10f5,0x10f5,0x10f5,0x10f5,0x10f5,0x10f5,0xecd,0xecd,0xecd,0xeca,
0xeca,0xeca,0xeca,0xeca,0x112e,0x137a,0x137a,0x137a,0x137a,0x12fc,0x12fc,0x12fc,0x137d,0x12ff,0x12ff,0x137d,
0x14cd,0x14cd,0x14cd,0x14cd,0x14d0,0x14d0,0x14d0,0x1794,0x1794,0x1794,0x1794,0x185a,0xee2,0xee2,0xee2,0xee2,
0x14cd,0x14cd,0x14cd,0x14cd,0x14d0,0x14d0,0x14d0,0x1797,0x1797,0x1797,0x1797,0x185d,0xee2,0xee2,0xee2,0xee2,
0xfdb,0xfdb,0xfdb,0xfdb,0xfdb,0xfdb,0xfdb,0xfdb,0xfdb,0xfdb,0xfdb,0xfdb,0xfde,0xfde,0xfde,0xfde,
0xfde,0xfde,0xfde,0xfde,0xfde,0xfde,0xfde,0xfde,0xfde,0xfde,0xfde,0xfde,0xf03,0xf03,0xf03,0xf03,
0xf15,0xf1e,0xf21,0xf1e,0xf21,0xf1e,0xf21,0xf1e,0xf21,0xf1e,0xf21,0xf1e,0xf1e,0xf1e,0xf21,0xf1e,
@ -2931,7 +2931,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x14e5,0x14e5,0x14e5,0x14e5,0xffc,0xffc,0xff9,0xff3,0xff9,0xff3,0xff9,0xff3,0xff9,0xff3,0xff0,0xff0,
0xff0,0xff0,0x1005,0x1002,0xff0,0x1149,0x13e0,0x13e3,0x13e3,0x13e0,0x13e0,0x13e0,0x13e0,0x13e0,0x13e6,0x13e6,
0x1500,0x14f4,0x14f4,0x14f1,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,0x1017,0x1014,0x1014,0x1023,
0x101a,0x1320,0x131d,0x16ec,0x1320,0x131d,0x13ef,0x13ec,0x1503,0x1503,0x1509,0x1503,0x1509,0x1503,0x1509,0x1503,
0x101a,0x1320,0x131d,0x16ef,0x1320,0x131d,0x13ef,0x13ec,0x1503,0x1503,0x1509,0x1503,0x1509,0x1503,0x1509,0x1503,
0x1509,0x1503,0x1509,0x1503,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,
0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,0x1023,0x101a,
0x1023,0x101a,0x1023,0x101a,0x101d,0x101a,0x101a,0x101a,0x101a,0x101a,0x101a,0x101a,0x101a,0x1023,0x101a,0x1023,
@ -2952,7 +2952,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x1080,0x107a,0x107d,0x10fe,0x1071,0x1071,0x1071,0x1077,0x13fe,0x13fe,0x13fe,0x13fe,0x13fe,0x13fe,0x13fe,0x13fe,
0x1074,0x1074,0x1077,0x1083,0x1080,0x107a,0x1080,0x107a,0x1080,0x107a,0x1080,0x107a,0x1080,0x107a,0x1080,0x107a,
0x1080,0x107a,0x1080,0x107a,0x1080,0x107a,0x1080,0x107a,0x1080,0x107a,0x1080,0x107a,0x1080,0x107a,0x1080,0x107a,
0x1080,0x107a,0x1080,0x107a,0x1518,0x1515,0x1518,0x1515,0x151b,0x151b,0x16f5,0x13fe,0x108c,0x108c,0x108f,0x108f,
0x1080,0x107a,0x1080,0x107a,0x1518,0x1515,0x1518,0x1515,0x151b,0x151b,0x16f8,0x13fe,0x108c,0x108c,0x108f,0x108f,
0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,
0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108f,0x108c,0x108c,0x108c,0x108c,
0x108c,0x108c,0x108c,0x108c,0x108c,0x108c,0x108c,0x108c,0x109e,0x109e,0x109e,0x109e,0x109e,0x109e,0x1095,0x1095,
@ -2968,8 +2968,8 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x10da,0x10da,0x10da,0x10da,0x10da,0x10da,0x10da,0x10da,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,
0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,
0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10dd,0x10e9,0x10e9,0x10e9,0x10e9,0x128a,0x128a,0x128a,0x128a,
0x128a,0x128a,0x128a,0x128a,0x1488,0x1773,0x1773,0x1773,0x1773,0x1773,0x1773,0x1773,0x1773,0x1773,0x18c3,0x18c3,
0x18c3,0x18c3,0x18c3,0x18c3,0x18c3,0x18c3,0x18c3,0x18c3,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,
0x128a,0x128a,0x128a,0x128a,0x1488,0x1776,0x1776,0x1776,0x1776,0x1776,0x1776,0x1776,0x1776,0x1776,0x18c6,0x18c6,
0x18c6,0x18c6,0x18c6,0x18c6,0x18c6,0x18c6,0x18c6,0x18c6,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,
0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x115e,0x1155,0x1155,
0x1158,0x1158,0x115e,0x1155,0x1155,0x1155,0x1155,0x1155,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,
0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,0x1161,
@ -3009,7 +3009,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x1302,0x1302,0x1302,0x1302,0x1257,0x125d,0x1284,0x1257,0x1257,0x1257,0x1257,0x1257,0x125d,0x1260,0x1284,0x1284,
0x1260,0x1284,0x1257,0x1260,0x1260,0x1263,0x1284,0x1257,0x1257,0x1284,0x125a,0x125a,0x1371,0x1371,0x1371,0x1371,
0x1371,0x1371,0x1371,0x1371,0x1371,0x1371,0x126c,0x126c,0x126c,0x126c,0x138c,0x136b,0x1275,0x138c,0x138c,0x138c,
0x138c,0x138c,0x138c,0x138c,0x138c,0x138c,0x138c,0x1821,0x1821,0x1821,0x1821,0x1821,0x1374,0x1374,0x127b,0x1374,
0x138c,0x138c,0x138c,0x138c,0x138c,0x138c,0x138c,0x1824,0x1824,0x1824,0x1824,0x1824,0x1374,0x1374,0x127b,0x1374,
0x1374,0x1374,0x127b,0x1374,0x1374,0x1374,0x1275,0x1275,0x1275,0x1275,0x1275,0x136e,0x1371,0x1371,0x1371,0x1371,
0x1371,0x1371,0x1371,0x1278,0x1371,0x1371,0x1371,0x1371,0x1371,0x1371,0x1371,0x1278,0x12a2,0x12a2,0x12a2,0x12a2,
0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,0x12a2,
@ -3022,9 +3022,9 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,0x1362,
0x1362,0x1362,0x1362,0x1362,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,
0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,0x1368,
0x1368,0x1368,0x1368,0x1368,0x1398,0x1395,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,
0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,
0x18c9,0x18c9,0x18c9,0x18c9,0x13a1,0x13a1,0x13a1,0x13a1,0x13a4,0x13a1,0x13a1,0x13a1,0x13a4,0x13a1,0x13a1,0x13a1,
0x1368,0x1368,0x1368,0x1368,0x1398,0x1395,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,
0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,
0x18cc,0x18cc,0x18cc,0x18cc,0x13a1,0x13a1,0x13a1,0x13a1,0x13a4,0x13a1,0x13a1,0x13a1,0x13a4,0x13a1,0x13a1,0x13a1,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x139e,0x139e,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x139e,0x13a1,0x13a1,0x13a1,0x139e,0x13a1,0x139e,0x13a1,0x139e,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a7,0x13a1,
0x13a1,0x13a1,0x13a1,0x139e,0x13a1,0x139e,0x139e,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a4,
@ -3034,17 +3034,17 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x13a1,0x13a1,0x139e,0x139e,0x139e,0x139e,0x139e,0x139e,0x139e,0x139e,0x139e,0x139e,0x139e,0x139e,0x1533,0x1533,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a4,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x13a1,0x153c,0x1536,0x1536,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x1776,0x1776,0x1776,
0x13a1,0x153c,0x1536,0x1536,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x1779,0x1779,0x1779,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x153c,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a4,0x13a1,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x13a1,0x13a1,0x13a1,0x13a4,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x153c,0x1776,0x1776,
0x13a1,0x13a1,0x13a1,0x13a4,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x153c,0x1779,0x1779,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a7,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x13a1,0x13a1,0x13a1,0x13a4,0x1536,0x1536,0x153c,0x153c,0x1536,0x153c,0x153c,0x153c,0x1533,0x1533,0x153c,0x153c,
0x13a1,0x13a1,0x13a7,0x13a7,0x13a7,0x16aa,0x13a1,0x13a7,0x13a1,0x13a1,0x13a7,0x1542,0x1542,0x153c,0x153c,0x1776,
0x1776,0x1776,0x1776,0x1776,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,
0x13a1,0x13a1,0x13a7,0x13a7,0x13a7,0x16aa,0x13a1,0x13a7,0x13a1,0x13a1,0x13a7,0x1542,0x1542,0x153c,0x153c,0x1779,
0x1779,0x1779,0x1779,0x1779,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,0x153c,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a4,0x13a1,0x13a4,0x13a1,0x13a1,
0x13a1,0x1536,0x1536,0x153c,0x16aa,0x153c,0x1536,0x153c,0x1776,0x1776,0x1776,0x1779,0x1779,0x1779,0x1779,0x1779,
0x13a1,0x1536,0x1536,0x153c,0x16aa,0x153c,0x1536,0x153c,0x1779,0x1779,0x1779,0x177c,0x177c,0x177c,0x177c,0x177c,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x153c,
0x13a1,0x153c,0x13a7,0x13a7,0x13a1,0x13a1,0x13a7,0x13a7,0x13a7,0x13a7,0x13a7,0x13a7,0x13a7,0x13a7,0x13a7,0x13a7,
@ -3053,17 +3053,17 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x13a7,0x13a7,0x13a7,0x13a7,0x13a7,0x13a1,0x13a1,0x13a1,0x13a7,0x13a1,0x13a1,0x13a1,0x13a1,0x13a7,0x13a7,0x13a7,
0x13a1,0x13a7,0x13a7,0x13a7,0x13a1,0x13a1,0x13a1,0x13a4,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x16aa,0x13a1,0x13a1,0x13a1,0x13a1,0x153c,0x1536,0x1776,0x1404,0x1404,0x1404,0x1404,0x1533,0x1533,0x1533,0x1533,
0x1533,0x1539,0x153c,0x1776,0x1776,0x1776,0x1776,0x16fe,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x16aa,0x13a1,0x13a1,0x13a1,0x13a1,0x153c,0x1536,0x1779,0x1404,0x1404,0x1404,0x1404,0x1533,0x1533,0x1533,0x1533,
0x1533,0x1539,0x153c,0x1779,0x1779,0x1779,0x1779,0x1701,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x153c,
0x153c,0x1536,0x1536,0x153c,0x1542,0x1542,0x153c,0x153c,0x153c,0x153c,0x182a,0x1536,0x1536,0x1536,0x1536,0x1536,
0x153c,0x1536,0x1536,0x153c,0x1542,0x1542,0x153c,0x153c,0x153c,0x153c,0x182d,0x1536,0x1536,0x1536,0x1536,0x1536,
0x1536,0x153c,0x1536,0x153c,0x1536,0x1536,0x1536,0x1536,0x153f,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x153c,
0x1536,0x1536,0x1536,0x153c,0x1533,0x1533,0x1533,0x1533,0x1533,0x1533,0x153c,0x13a1,0x13a1,0x13a1,0x13a1,0x13a1,
0x148e,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,
0x13ad,0x148e,0x13ad,0x13ad,0x13ad,0x148e,0x13ad,0x148e,0x13ad,0x148e,0x13ad,0x148e,0x13ad,0x13ad,0x13ad,0x148e,
0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x148e,0x148e,0x13ad,0x13ad,0x13ad,0x13ad,0x148e,0x13ad,0x148e,0x148e,
0x13ad,0x13ad,0x13ad,0x13ad,0x148e,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,0x13ad,
0x13ad,0x16b0,0x16b0,0x177c,0x177c,0x13b0,0x13b0,0x13b0,0x13ad,0x13ad,0x13ad,0x13b0,0x13b0,0x13b0,0x13b0,0x13b0,
0x13ad,0x16b0,0x16b0,0x177f,0x177f,0x13b0,0x13b0,0x13b0,0x13ad,0x13ad,0x13ad,0x13b0,0x13b0,0x13b0,0x13b0,0x13b0,
0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,0x162f,
0x13b6,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,
0x13b3,0x13b3,0x13b6,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,
@ -3071,7 +3071,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x13b3,0x13b3,0x13b3,0x13b3,0x13b9,0x13b9,0x13b9,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,0x13b3,
0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,
0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,0x13bc,
0x17a9,0x17a9,0x17a6,0x1701,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x1407,0x1407,0x1407,0x1407,0x1407,0x1407,
0x17ac,0x17ac,0x17a9,0x1704,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x1407,0x1407,0x1407,0x1407,0x1407,0x1407,
0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x140a,0x154b,
0x1416,0x1416,0x1416,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,
0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,0x1428,
@ -3088,7 +3088,7 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x1536,0x153c,0x153c,0x153c,0x1533,0x1533,0x1533,0x1533,0x1533,0x1533,0x1533,0x1533,0x153c,0x153c,0x153c,0x1536,
0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x153c,0x1536,0x1536,0x153c,0x153c,0x153c,0x153c,0x1536,0x1536,
0x1542,0x1536,0x1536,0x1536,0x1536,0x16ad,0x16ad,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,
0x1827,0x153c,0x1536,0x1536,0x153c,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x153c,0x153c,0x1536,
0x182a,0x153c,0x1536,0x1536,0x153c,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x153c,0x153c,0x1536,
0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x1536,0x153c,0x1536,0x1536,0x1536,0x1563,0x1563,0x1563,0x1563,
0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,
0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1563,0x1575,0x1575,0x1575,0x1575,
@ -3135,52 +3135,52 @@ static const uint16_t propsVectorsTrie_index[29236]={
0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,
0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a1,0x16a7,0x16a7,0x16a7,0x16a7,
0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,
0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x1707,0x1707,0x1707,0x1707,
0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,
0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1707,0x1743,0x1743,0x1743,0x1743,
0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,
0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1749,0x1746,
0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,0x1743,
0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,
0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,0x174c,
0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x16a7,0x170a,0x170a,0x170a,0x170a,
0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,
0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x170a,0x1746,0x1746,0x1746,0x1746,
0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,
0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x174c,0x1749,
0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,0x1746,
0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,
0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,0x174f,
0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,
0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,0x1761,
0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,
0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,0x1752,
0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,
0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,0x1764,
0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,
0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,
0x1767,0x1767,0x1767,0x176a,0x176a,0x176a,0x176a,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,0x1767,
0x1767,0x1767,0x1767,0x1767,0x1767,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x1767,0x176a,0x176a,
0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,
0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,
0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,
0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,0x1782,
0x17c7,0x17c7,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,
0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,0x17c4,
0x176a,0x176a,0x176a,0x176d,0x176d,0x176d,0x176d,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,0x176a,
0x176a,0x176a,0x176a,0x176a,0x176a,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176a,0x176d,0x176d,
0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,
0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,0x176d,
0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,
0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,0x1785,
0x17ca,0x17ca,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,
0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,
0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,0x17c7,
0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,
0x1815,0x1815,0x1815,0x1815,0x1815,0x1812,0x1812,0x1812,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,0x17fd,
0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,
0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,0x1815,
0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,
0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,0x183c,
0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,
0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,0x17ca,
0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,
0x1818,0x1818,0x1818,0x1818,0x1818,0x1815,0x1815,0x1815,0x1800,0x1800,0x1800,0x1800,0x1800,0x1800,0x1800,0x1800,
0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,
0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,0x1818,
0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,
0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,0x183f,
0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,
0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,0x1896,
0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,
0x18b7,0x18b7,0x18b7,0x18a2,0x18a8,0x18a5,0x18a5,0x18a5,0x18a5,0x18b4,0x18ba,0x18a5,0x18a5,0x18a5,0x18a5,0x18b1,
0x18b7,0x18a5,0x18a5,0x18a5,0x18a5,0x18a5,0x18a5,0x18b4,0x18b4,0x18a5,0x18a5,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,
0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,0x18b7,
0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,
0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,0x18c9,
0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,
0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,0x18d8,
0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,
0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,0x18de,
0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,
0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,0x1842,
0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,
0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,0x1899,
0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,
0x18ba,0x18ba,0x18ba,0x18a5,0x18ab,0x18a8,0x18a8,0x18a8,0x18a8,0x18b7,0x18bd,0x18a8,0x18a8,0x18a8,0x18a8,0x18b4,
0x18ba,0x18a8,0x18a8,0x18a8,0x18a8,0x18a8,0x18a8,0x18b7,0x18b7,0x18a8,0x18a8,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,
0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,0x18ba,
0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,
0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,0x18cc,
0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,
0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,0x18db,
0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,
0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,0x18e1,
0,0,0,0
};
@ -3199,7 +3199,7 @@ static const UTrie2 propsVectorsTrie={
NULL, 0, FALSE, FALSE, 0, NULL
};
static const uint32_t propsVectors[6372]={
static const uint32_t propsVectors[6375]={
0x67,0,0,0x67,0,0xe00000,0x67,0x80000,0x20,0x867,0,0,0xa67,0,0,0xb67,
0,0,0xc67,0,0,0xd67,0,0,0xe67,0,0,0x1067,0,0,0x1167,0,
0,0x1267,0,0,0x1367,0,0,0x1467,0,0,0x1567,0,0,0x1667,0,0,
@ -3564,44 +3564,44 @@ static const uint32_t propsVectors[6372]={
0x6800000,0x1329800,0x7000fb9e,0x6800100,0x962540,0x7000fb9e,0x6800100,0x962541,0x7000fb9e,0x7c00100,0x230400,0x7000fc92,0x4000000,0x200000,0x7000fc92,0x6800000,
0x1329800,0x7000fc92,0x7c00100,0x220400,0x7000fc92,0x7c00100,0x230400,0x7000fc92,0x7c00100,0x250400,0x700acd00,0x4000000,0x30e00000,0x700acd00,0x4000000,0xb28045a0,
0x700ace00,0x4000000,0x30e00000,0x700acf00,0x4000000,0x30e00000,0x700acf00,0x4000000,0xb28045a0,0x7040dfc0,0x4000000,0x200000,0x7040f7c4,0x80000,0x918820,0x7080af7b,
0x2802400,0x962460,0x7080dfc0,0x2802400,0x962460,0x70c0e4c2,0x2802400,0x962460,0x70c0e4c2,0x6800100,0x962540,0x8000120f,0x7c00100,0x230400,0x80001524,0x7c00100,
0x230400,0x8000171a,0x7c00100,0x230400,0x80002006,0x7c00100,0x220400,0x80002006,0x7c00100,0x250400,0x80002a00,0x4000000,0x1500000,0x80002d00,0x4000000,0x200000,
0x80005208,0x2802400,0x962460,0x80005c00,0x4000000,0x200000,0x80007300,0x24000000,0x200000,0x80009519,0x7c00100,0x220400,0x80009519,0x7c00100,0x230400,0x80009519,
0x7c00100,0x250400,0x80009865,0x7c00100,0x230400,0x8000a008,0x2802100,0x962460,0x8000b30a,0x4000000,0x500000,0x8000b30a,0x7c00100,0x230400,0x8000cd00,0x4000000,
0xe00000,0x8000d202,0x2802500,0x962460,0x8000d202,0x7c00100,0x230400,0x8000d68d,0x4000000,0x200000,0x8000d997,0x2802400,0x962460,0x8000d997,0x4000000,0x200000,
0x8000d997,0x4000000,0x400000,0x8000d997,0x4000000,0x500000,0x8000d997,0x7c00100,0x230400,0x8000d997,0xc000010,0x448000,0x8000e489,0x2802100,0x962460,0x8000e489,
0x7c00100,0x230400,0x8000e719,0x7c00100,0x220400,0x8000f8a6,0x2802100,0x962460,0x8000f8a6,0x7c00100,0x230400,0x8000f8a6,0xc000010,0x448000,0x8000fda1,0x2802100,
0x1862460,0x8000fda1,0x2806400,0x1862460,0x8000fda1,0x4000000,0x1800000,0x8000fda1,0x6800000,0x1329800,0x8000fda1,0x6800100,0x1862540,0x8000fda1,0x7c00100,0x1830000,
0x8000fda1,0xc000010,0x448000,0x8000fe9c,0x7c00100,0x230400,0x8000fe9c,0x7c00100,0x830400,0x8000fe9c,0x7c00100,0x1430400,0x8000ff06,0x7c00100,0x220400,0x80010165,
0x7c00100,0x230400,0x800102a2,0x4000000,0x200000,0x800102a2,0x7c00100,0x230400,0x800103a4,0x7c00100,0x230400,0x800103a4,0xc000010,0x448000,0x8001044c,0x4000000,
0x200000,0x8001044c,0x7c00100,0x220400,0x8001044c,0x7c00100,0x250400,0x80010670,0x2802000,0x962460,0x80010670,0x4000000,0x200000,0x80010670,0x4000010,0x400000,
0x80010670,0xc000010,0x448000,0x800a4711,0x7c40300,0xe30000,0x800acd00,0x4000000,0x30e00000,0x800acd00,0x4000000,0x7a904de0,0x800ace00,0x4000000,0x30e00000,0x800acf00,
0x4000000,0x30e00000,0x800b0011,0x7c40300,0xe30000,0x800b0500,0x4000000,0x30e00000,0x800b0500,0x4000000,0xb28045a0,0x90001615,0x7c00100,0x230400,0x9000171a,0x4000000,
0x200000,0x9000171a,0x7c00100,0x230400,0x90003000,0x24000000,0x200000,0x90007f0e,0x4000000,0x200000,0x90008301,0x2802000,0x962460,0x90008e00,0x24000000,0x400000,
0x90009519,0x7c00100,0x250400,0x9000a16f,0x2802100,0x962460,0x9000d200,0x80000000,0x218960,0x9000d202,0x2802000,0x962460,0x9000d202,0x2802100,0x962460,0x9000d202,
0x7c00100,0x230400,0x9000e59d,0x2802100,0x962460,0x900107a7,0x2802100,0x962460,0x900107a7,0x2802400,0x962460,0x900107a7,0x2802c00,0x962460,0x900107a7,0x4000000,
0x1400000,0x900107a7,0x6800000,0x1329800,0x900107a7,0x7c00100,0x220400,0x900107a7,0x7c00100,0x250400,0x900108a8,0x2802100,0x962460,0x900108a8,0x2806400,0x962460,
0x900108a8,0x4000000,0x200000,0x900108a8,0x4000000,0x400000,0x900108a8,0x4000010,0x400000,0x900108a8,0x6800000,0x1329800,0x900108a8,0x6800100,0x962540,0x900108a8,
0x7c00100,0x230400,0x900108a8,0xc000010,0x448000,0x90010908,0x7c00100,0x220400,0x90010a38,0x2802100,0x962460,0x90010ca9,0x2802100,0x962460,0x90010ca9,0x4000000,
0x500000,0x90010ca9,0x4000010,0xb00000,0x90010ca9,0x6800100,0x962540,0x90010ca9,0x7c00100,0x230400,0x90010d1b,0x4000000,0x500000,0x90010eaa,0x2802100,0x962460,
0x90010eaa,0x2802400,0x962460,0x90010eaa,0x2806400,0x962460,0x90010eaa,0x4000000,0x200000,0x90010eaa,0x4000000,0x400000,0x90010eaa,0x4000010,0x400000,0x90010eaa,
0x6800000,0x1329800,0x90010eaa,0x6800100,0x962540,0x90010eaa,0x7c00100,0x230400,0x90010eaa,0xc000010,0x448000,0x90010fab,0x7c00100,0x220400,0x90010fab,0x7c00100,
0x250400,0x9002c300,0x4000000,0x100000,0x900ac400,0x4000000,0xe0000d,0x900acd00,0x4000000,0x30e00000,0x900acd00,0x4000000,0xb28045a0,0x900acf00,0x4000000,0x30e00000,
0x900b0500,0x4000000,0xe00000,0x900b0500,0x4000000,0x30e00000,0x900b0500,0x4000000,0xb28045a0,0x900b0b9a,0x7c00900,0x1230400,0x900b109a,0x7c00300,0xe30000,0x900b119a,
0x7c00300,0xe30000,0x90408e06,0x24000000,0x400000,0xa0001004,0x4000000,0x200000,0xa0001004,0x7c00100,0x230400,0xa000120f,0x2802100,0x962460,0xa000120f,0x2802400,
0x962460,0xa000171a,0x2802100,0x962460,0xa000171a,0x2806400,0x962460,0xa0002a00,0x4000000,0x1600000,0xa0003000,0x24000000,0x200000,0xa000581e,0x7c00100,0x230400,
0xa0007300,0x24000000,0x200000,0xa0008301,0x2802400,0x962460,0xa0008e00,0x24000000,0x400000,0xa000cf00,0x4000000,0xe00000,0xa0010500,0x4000000,0x200000,0xa00114af,
0x2802100,0x962460,0xa00114af,0x2802400,0x962460,0xa00114af,0x2806400,0x962460,0xa00114af,0x6800000,0x1329800,0xa00114af,0x7c00100,0x230400,0xa00114af,0x7c00100,
0x230560,0xa00116b0,0x2802100,0x962460,0xa00116b0,0x2802800,0x962460,0xa00116b0,0x2806400,0x962460,0xa00116b0,0x4000000,0x400000,0xa00116b0,0x4000000,0x500000,
0xa00116b0,0x4000010,0x400000,0xa00116b0,0x6800100,0x962540,0xa00116b0,0x7c00100,0x230400,0xa00116b0,0x7c00100,0x230560,0xa00116b0,0xc000010,0x448000,0xa0011722,
0x7c00100,0x230400,0xa00118b1,0x2802000,0x962460,0xa00118b1,0x2802100,0x962460,0xa00118b1,0x2806400,0x962460,0xa00118b1,0x4000000,0x200000,0xa00118b1,0x4000000,
0x400000,0xa00118b1,0x4000000,0x500000,0xa00118b1,0x6800100,0x962540,0xa00118b1,0x7c00100,0x230400,0xa00118b1,0x7c00100,0x230560,0xa00118b1,0xc000010,0x448000,
0xa00a4005,0x7c00100,0xe30400,0xa00a4711,0x7c40300,0xe30000,0xa00ac400,0x4000000,0xe00000,0xa00acb14,0x7c00100,0xe30000,0xa00acf00,0x4000000,0x30e00000,0xa00b0500,
0x4000000,0x30e00000,0xa00b0500,0x4000000,0xb28045a0,0xa00b0b96,0x7c00900,0x1230400,0xa00b1211,0x7c40300,0xe30000,0xa00b1314,0x7c00100,0xe30000,0xa00b1596,0x7c00300,
0xe30000,0xa040af86,0x6800400,0x962540};
0x2802400,0x962460,0x7080dfc0,0x2802400,0x962460,0x70c0e4c2,0x2802100,0x962460,0x70c0e4c2,0x2802400,0x962460,0x70c0e4c2,0x6800100,0x962540,0x8000120f,0x7c00100,
0x230400,0x80001524,0x7c00100,0x230400,0x8000171a,0x7c00100,0x230400,0x80002006,0x7c00100,0x220400,0x80002006,0x7c00100,0x250400,0x80002a00,0x4000000,0x1500000,
0x80002d00,0x4000000,0x200000,0x80005208,0x2802400,0x962460,0x80005c00,0x4000000,0x200000,0x80007300,0x24000000,0x200000,0x80009519,0x7c00100,0x220400,0x80009519,
0x7c00100,0x230400,0x80009519,0x7c00100,0x250400,0x80009865,0x7c00100,0x230400,0x8000a008,0x2802100,0x962460,0x8000b30a,0x4000000,0x500000,0x8000b30a,0x7c00100,
0x230400,0x8000cd00,0x4000000,0xe00000,0x8000d202,0x2802500,0x962460,0x8000d202,0x7c00100,0x230400,0x8000d68d,0x4000000,0x200000,0x8000d997,0x2802400,0x962460,
0x8000d997,0x4000000,0x200000,0x8000d997,0x4000000,0x400000,0x8000d997,0x4000000,0x500000,0x8000d997,0x7c00100,0x230400,0x8000d997,0xc000010,0x448000,0x8000e489,
0x2802100,0x962460,0x8000e489,0x7c00100,0x230400,0x8000e719,0x7c00100,0x220400,0x8000f8a6,0x2802100,0x962460,0x8000f8a6,0x7c00100,0x230400,0x8000f8a6,0xc000010,
0x448000,0x8000fda1,0x2802100,0x1862460,0x8000fda1,0x2806400,0x1862460,0x8000fda1,0x4000000,0x1800000,0x8000fda1,0x6800000,0x1329800,0x8000fda1,0x6800100,0x1862540,
0x8000fda1,0x7c00100,0x1830000,0x8000fda1,0xc000010,0x448000,0x8000fe9c,0x7c00100,0x230400,0x8000fe9c,0x7c00100,0x830400,0x8000fe9c,0x7c00100,0x1430400,0x8000ff06,
0x7c00100,0x220400,0x80010165,0x7c00100,0x230400,0x800102a2,0x4000000,0x200000,0x800102a2,0x7c00100,0x230400,0x800103a4,0x7c00100,0x230400,0x800103a4,0xc000010,
0x448000,0x8001044c,0x4000000,0x200000,0x8001044c,0x7c00100,0x220400,0x8001044c,0x7c00100,0x250400,0x80010670,0x2802000,0x962460,0x80010670,0x4000000,0x200000,
0x80010670,0x4000010,0x400000,0x80010670,0xc000010,0x448000,0x800a4711,0x7c40300,0xe30000,0x800acd00,0x4000000,0x30e00000,0x800acd00,0x4000000,0x7a904de0,0x800ace00,
0x4000000,0x30e00000,0x800acf00,0x4000000,0x30e00000,0x800b0011,0x7c40300,0xe30000,0x800b0500,0x4000000,0x30e00000,0x800b0500,0x4000000,0xb28045a0,0x90001615,0x7c00100,
0x230400,0x9000171a,0x4000000,0x200000,0x9000171a,0x7c00100,0x230400,0x90003000,0x24000000,0x200000,0x90007f0e,0x4000000,0x200000,0x90008301,0x2802000,0x962460,
0x90008e00,0x24000000,0x400000,0x90009519,0x7c00100,0x250400,0x9000a16f,0x2802100,0x962460,0x9000d200,0x80000000,0x218960,0x9000d202,0x2802000,0x962460,0x9000d202,
0x2802100,0x962460,0x9000d202,0x7c00100,0x230400,0x9000e59d,0x2802100,0x962460,0x900107a7,0x2802100,0x962460,0x900107a7,0x2802400,0x962460,0x900107a7,0x2802c00,
0x962460,0x900107a7,0x4000000,0x1400000,0x900107a7,0x6800000,0x1329800,0x900107a7,0x7c00100,0x220400,0x900107a7,0x7c00100,0x250400,0x900108a8,0x2802100,0x962460,
0x900108a8,0x2806400,0x962460,0x900108a8,0x4000000,0x200000,0x900108a8,0x4000000,0x400000,0x900108a8,0x4000010,0x400000,0x900108a8,0x6800000,0x1329800,0x900108a8,
0x6800100,0x962540,0x900108a8,0x7c00100,0x230400,0x900108a8,0xc000010,0x448000,0x90010908,0x7c00100,0x220400,0x90010a38,0x2802100,0x962460,0x90010ca9,0x2802100,
0x962460,0x90010ca9,0x4000000,0x500000,0x90010ca9,0x4000010,0xb00000,0x90010ca9,0x6800100,0x962540,0x90010ca9,0x7c00100,0x230400,0x90010d1b,0x4000000,0x500000,
0x90010eaa,0x2802100,0x962460,0x90010eaa,0x2802400,0x962460,0x90010eaa,0x2806400,0x962460,0x90010eaa,0x4000000,0x200000,0x90010eaa,0x4000000,0x400000,0x90010eaa,
0x4000010,0x400000,0x90010eaa,0x6800000,0x1329800,0x90010eaa,0x6800100,0x962540,0x90010eaa,0x7c00100,0x230400,0x90010eaa,0xc000010,0x448000,0x90010fab,0x7c00100,
0x220400,0x90010fab,0x7c00100,0x250400,0x9002c300,0x4000000,0x100000,0x900ac400,0x4000000,0xe0000d,0x900acd00,0x4000000,0x30e00000,0x900acd00,0x4000000,0xb28045a0,
0x900acf00,0x4000000,0x30e00000,0x900b0500,0x4000000,0xe00000,0x900b0500,0x4000000,0x30e00000,0x900b0500,0x4000000,0xb28045a0,0x900b0b9a,0x7c00900,0x1230400,0x900b109a,
0x7c00300,0xe30000,0x900b119a,0x7c00300,0xe30000,0x90408e06,0x24000000,0x400000,0xa0001004,0x4000000,0x200000,0xa0001004,0x7c00100,0x230400,0xa000120f,0x2802100,
0x962460,0xa000120f,0x2802400,0x962460,0xa000171a,0x2802100,0x962460,0xa000171a,0x2806400,0x962460,0xa0002a00,0x4000000,0x1600000,0xa0003000,0x24000000,0x200000,
0xa000581e,0x7c00100,0x230400,0xa0007300,0x24000000,0x200000,0xa0008301,0x2802400,0x962460,0xa0008e00,0x24000000,0x400000,0xa000cf00,0x4000000,0xe00000,0xa0010500,
0x4000000,0x200000,0xa00114af,0x2802100,0x962460,0xa00114af,0x2802400,0x962460,0xa00114af,0x2806400,0x962460,0xa00114af,0x6800000,0x1329800,0xa00114af,0x7c00100,
0x230400,0xa00114af,0x7c00100,0x230560,0xa00116b0,0x2802100,0x962460,0xa00116b0,0x2802800,0x962460,0xa00116b0,0x2806400,0x962460,0xa00116b0,0x4000000,0x400000,
0xa00116b0,0x4000000,0x500000,0xa00116b0,0x4000010,0x400000,0xa00116b0,0x6800100,0x962540,0xa00116b0,0x7c00100,0x230400,0xa00116b0,0x7c00100,0x230560,0xa00116b0,
0xc000010,0x448000,0xa0011722,0x7c00100,0x230400,0xa00118b1,0x2802000,0x962460,0xa00118b1,0x2802100,0x962460,0xa00118b1,0x2806400,0x962460,0xa00118b1,0x4000000,
0x200000,0xa00118b1,0x4000000,0x400000,0xa00118b1,0x4000000,0x500000,0xa00118b1,0x6800100,0x962540,0xa00118b1,0x7c00100,0x230400,0xa00118b1,0x7c00100,0x230560,
0xa00118b1,0xc000010,0x448000,0xa00a4005,0x7c00100,0xe30400,0xa00a4711,0x7c40300,0xe30000,0xa00ac400,0x4000000,0xe00000,0xa00acb14,0x7c00100,0xe30000,0xa00acf00,
0x4000000,0x30e00000,0xa00b0500,0x4000000,0x30e00000,0xa00b0500,0x4000000,0xb28045a0,0xa00b0b96,0x7c00900,0x1230400,0xa00b1211,0x7c40300,0xe30000,0xa00b1314,0x7c00100,
0xe30000,0xa00b1596,0x7c00300,0xe30000,0xa040af86,0x6800400,0x962540};
static const int32_t countPropsVectors=6372;
static const int32_t propsVectorsColumns=3;
static const int32_t countPropsVectors=6375;
static const int32_t propsVectorsColumns=3; /* The code assumes that this value is > 0*/
static const uint16_t scriptExtensions[198]={
0x800e,0x8019,8,0x8059,8,2,8,0x8038,8,6,8,0x8019,3,0x800c,2,0x22,
0x8025,2,0xe,2,0x22,0x54,0x79,0x7b,0x80a7,2,0x8022,2,0x8025,2,0x1b,4,
@ -3617,6 +3617,6 @@ static const uint16_t scriptExtensions[198]={
0xa,0xae,0xa,0x8023,0xa,0xb2,0x19,0x1c,0x804f,0x37,0x804e,0x2f,0x31,0x8053,0x2f,0x8031,
2,0x8007,0x89,0x69,0x8087,0};
static const int32_t indexes[UPROPS_INDEX_COUNT]={0x2962,0x2962,0x2962,0x2962,0x6280,3,0x7b64,0x7bc7,0x7bc7,0x7bc7,0xb18b1,0x2a75631,0,0,0,0};
static const int32_t indexes[UPROPS_INDEX_COUNT]={0x2962,0x2962,0x2962,0x2962,0x6280,3,0x7b67,0x7bca,0x7bca,0x7bca,0xb18b1,0x2a75631,0,0,0,0};
#endif // INCLUDED_FROM_UCHAR_C

2
icu4c/source/common/ucharstriebuilder.cpp
View file

@ -287,7 +287,7 @@ UCharsTrieBuilder::indexOfElementWithNextUnit(int32_t i, int32_t unitIndex, UCha
UCharsTrieBuilder::UCTLinearMatchNode::UCTLinearMatchNode(const UChar *units, int32_t len, Node *nextNode)
: LinearMatchNode(len, nextNode), s(units) {
hash=hash*37+ustr_hashUCharsN(units, len);
hash=hash*37u+ustr_hashUCharsN(units, len);
}
UBool

2
icu4c/source/common/ucnv_ct.cpp
View file

@ -519,7 +519,7 @@ UConverter_toUnicode_CompoundText_OFFSETS(UConverterToUnicodeArgs *args,
currentState = tmpState;
}
sourceOffset = uprv_strlen((char*)escSeqCompoundText[currentState]) - args->converter->toULength;
sourceOffset = static_cast<int32_t>(uprv_strlen((char*)escSeqCompoundText[currentState]) - args->converter->toULength);
mySource += sourceOffset;

8
icu4c/source/common/ucnv_lmb.cpp
View file

@ -966,26 +966,26 @@ _LMBCSFromUnicode(UConverterFromUnicodeArgs* args,
if(extraInfo->localeConverterIndex < ULMBCS_DOUBLEOPTGROUP_START)
{
bytes_written = LMBCSConversionWorker (extraInfo,
bytes_written = (int32_t)LMBCSConversionWorker (extraInfo,
ULMBCS_GRP_L1, pLMBCS, &uniChar,
&lastConverterIndex, groups_tried);
if(!bytes_written)
{
bytes_written = LMBCSConversionWorker (extraInfo,
bytes_written = (int32_t)LMBCSConversionWorker (extraInfo,
ULMBCS_GRP_EXCEPT, pLMBCS, &uniChar,
&lastConverterIndex, groups_tried);
}
if(!bytes_written)
{
bytes_written = LMBCSConversionWorker (extraInfo,
bytes_written = (int32_t)LMBCSConversionWorker (extraInfo,
extraInfo->localeConverterIndex, pLMBCS, &uniChar,
&lastConverterIndex, groups_tried);
}
}
else
{
bytes_written = LMBCSConversionWorker (extraInfo,
bytes_written = (int32_t)LMBCSConversionWorker (extraInfo,
extraInfo->localeConverterIndex, pLMBCS, &uniChar,
&lastConverterIndex, groups_tried);
}

4
icu4c/source/common/ucnvmbcs.cpp
View file

@ -5037,7 +5037,7 @@ ucnv_SBCSFromUTF8(UConverterFromUnicodeArgs *pFromUArgs,
uint8_t b, t1, t2;
uint32_t asciiRoundtrips;
uint16_t value, minValue;
uint16_t value, minValue = 0;
UBool hasSupplementary;
/* set up the local pointers */
@ -5344,7 +5344,7 @@ ucnv_DBCSFromUTF8(UConverterFromUnicodeArgs *pFromUArgs,
uint32_t stage2Entry;
uint32_t asciiRoundtrips;
uint16_t value;
uint16_t value = 0;
UBool hasSupplementary;
/* set up the local pointers */

8
icu4c/source/common/udata.cpp
View file

@ -206,6 +206,8 @@ setCommonICUData(UDataMemory *pData, /* The new common data. Belongs to ca
return didUpdate;
}
#if U_PLATFORM_HAS_WINUWP_API == 0
static UBool
setCommonICUDataPointer(const void *pData, UBool /*warn*/, UErrorCode *pErrorCode) {
UDataMemory tData;
@ -215,6 +217,8 @@ setCommonICUDataPointer(const void *pData, UBool /*warn*/, UErrorCode *pErrorCod
return setCommonICUData(&tData, FALSE, pErrorCode);
}
#endif
static const char *
findBasename(const char *path) {
const char *basename=uprv_strrchr(path, U_FILE_SEP_CHAR);
@ -982,7 +986,7 @@ static UDataMemory *doLoadFromIndividualFiles(const char *pkgName,
/* init path iterator for individual files */
UDataPathIterator iter(dataPath, pkgName, path, tocEntryPathSuffix, FALSE, pErrorCode);
while((pathBuffer = iter.next(pErrorCode)))
while((pathBuffer = iter.next(pErrorCode)) != NULL)
{
#ifdef UDATA_DEBUG
fprintf(stderr, "UDATA: trying individual file %s\n", pathBuffer);
@ -1165,7 +1169,7 @@ doOpenChoice(const char *path, const char *type, const char *name,
if(uprv_strchr(path,U_FILE_ALT_SEP_CHAR) != NULL) {
altSepPath.append(path, *pErrorCode);
char *p;
while((p=uprv_strchr(altSepPath.data(), U_FILE_ALT_SEP_CHAR))) {
while ((p = uprv_strchr(altSepPath.data(), U_FILE_ALT_SEP_CHAR)) != NULL) {
*p = U_FILE_SEP_CHAR;
}
#if defined (UDATA_DEBUG)

2
icu4c/source/common/uhash.cpp
View file

@ -844,7 +844,7 @@ uhash_hashUChars(const UHashTok key) {
U_CAPI int32_t U_EXPORT2
uhash_hashChars(const UHashTok key) {
const char *s = (const char *)key.pointer;
return s == NULL ? 0 : ustr_hashCharsN(s, uprv_strlen(s));
return s == NULL ? 0 : static_cast<int32_t>(ustr_hashCharsN(s, uprv_strlen(s)));
}
U_CAPI int32_t U_EXPORT2

4
icu4c/source/common/uinvchar.cpp
View file

@ -573,7 +573,7 @@ uprv_aestrncpy(uint8_t *dst, const uint8_t *src, int32_t n)
uint8_t *orig_dst = dst;
if(n==-1) {
n = uprv_strlen((const char*)src)+1; /* copy NUL */
n = static_cast<int32_t>(uprv_strlen((const char*)src)+1); /* copy NUL */
}
/* copy non-null */
while(*src && n>0) {
@ -594,7 +594,7 @@ uprv_eastrncpy(uint8_t *dst, const uint8_t *src, int32_t n)
uint8_t *orig_dst = dst;
if(n==-1) {
n = uprv_strlen((const char*)src)+1; /* copy NUL */
n = static_cast<int32_t>(uprv_strlen((const char*)src)+1); /* copy NUL */
}
/* copy non-null */
while(*src && n>0) {

2
icu4c/source/common/ulist.cpp
View file

@ -252,7 +252,7 @@ U_CAPI const char * U_EXPORT2 ulist_next_keyword_value(UEnumeration *en, int32_t
s = (const char *)ulist_getNext((UList *)(en->context));
if (s != NULL && resultLength != NULL) {
*resultLength = uprv_strlen(s);
*resultLength = static_cast<int32_t>(uprv_strlen(s));
}
return s;
}

10
icu4c/source/common/uloc.cpp
View file

@ -529,14 +529,16 @@ static const VariantMap VARIANT_MAP[] = {
#define _hasBCP47Extension(id) (id && uprv_strstr(id, "@") == NULL && getShortestSubtagLength(localeID) == 1)
/* Converts the BCP47 id to Unicode id. Does nothing to id if conversion fails */
#define _ConvertBCP47(finalID, id, buffer, length,err) \
if (uloc_forLanguageTag(id, buffer, length, NULL, err) <= 0 || U_FAILURE(*err)) { \
if (uloc_forLanguageTag(id, buffer, length, NULL, err) <= 0 || \
U_FAILURE(*err) || *err == U_STRING_NOT_TERMINATED_WARNING) { \
finalID=id; \
if (*err == U_STRING_NOT_TERMINATED_WARNING) { *err = U_BUFFER_OVERFLOW_ERROR; } \
} else { \
finalID=buffer; \
}
/* Gets the size of the shortest subtag in the given localeID. */
static int32_t getShortestSubtagLength(const char *localeID) {
int32_t localeIDLength = uprv_strlen(localeID);
int32_t localeIDLength = static_cast<int32_t>(uprv_strlen(localeID));
int32_t length = localeIDLength;
int32_t tmpLength = 0;
int32_t i;
@ -2486,7 +2488,7 @@ uloc_acceptLanguage(char *result, int32_t resultAvailable,
#if defined(ULOC_DEBUG)
fprintf(stderr,"%02d: %s\n", i, acceptList[i]);
#endif
while((l=uenum_next(availableLocales, NULL, status))) {
while((l=uenum_next(availableLocales, NULL, status)) != NULL) {
#if defined(ULOC_DEBUG)
fprintf(stderr," %s\n", l);
#endif
@ -2526,7 +2528,7 @@ uloc_acceptLanguage(char *result, int32_t resultAvailable,
#if defined(ULOC_DEBUG)
fprintf(stderr,"Try: [%s]", fallbackList[i]);
#endif
while((l=uenum_next(availableLocales, NULL, status))) {
while((l=uenum_next(availableLocales, NULL, status)) != NULL) {
#if defined(ULOC_DEBUG)
fprintf(stderr," %s\n", l);
#endif

6
icu4c/source/common/uloc_tag.cpp
View file

@ -1022,7 +1022,7 @@ _appendKeywordsToLanguageTag(const char* localeID, char* appendAt, int32_t capac
no known mapping. This implementation normalizes the
the value to lower case
*/
int32_t bcpValueLen = uprv_strlen(bcpValue);
int32_t bcpValueLen = static_cast<int32_t>(uprv_strlen(bcpValue));
if (bcpValueLen < extBufCapacity) {
uprv_strcpy(pExtBuf, bcpValue);
T_CString_toLowerCase(pExtBuf);
@ -1288,7 +1288,7 @@ _appendLDMLExtensionAsKeywords(const char* ldmlext, ExtensionListEntry** appendT
bufIdx++;
}
len = uprv_strlen(attr->attribute);
len = static_cast<int32_t>(uprv_strlen(attr->attribute));
uprv_memcpy(buf + bufIdx, attr->attribute, len);
bufIdx += len;
@ -1841,7 +1841,7 @@ ultag_parse(const char* tag, int32_t tagLen, int32_t* parsedLen, UErrorCode* sta
int32_t newTagLength;
grandfatheredLen = tagLen; /* back up for output parsedLen */
newTagLength = uprv_strlen(GRANDFATHERED[i+1]);
newTagLength = static_cast<int32_t>(uprv_strlen(GRANDFATHERED[i+1]));
if (tagLen < newTagLength) {
uprv_free(tagBuf);
tagBuf = (char*)uprv_malloc(newTagLength + 1);

10
icu4c/source/common/umapfile.cpp
View file

@ -102,10 +102,7 @@
{
HANDLE map;
HANDLE file;
SECURITY_ATTRIBUTES mappingAttributes;
SECURITY_ATTRIBUTES *mappingAttributesPtr = NULL;
SECURITY_DESCRIPTOR securityDesc;
UDataMemory_init(pData); /* Clear the output struct. */
/* open the input file */
@ -143,6 +140,11 @@
This is required for multiuser systems on Windows 2000 SP4 and beyond */
// TODO: UWP does not have this function and I do not think it is required?
#if U_PLATFORM_HAS_WINUWP_API == 0
SECURITY_ATTRIBUTES mappingAttributes;
SECURITY_ATTRIBUTES *mappingAttributesPtr = NULL;
SECURITY_DESCRIPTOR securityDesc;
if (InitializeSecurityDescriptor(&securityDesc, SECURITY_DESCRIPTOR_REVISION)) {
/* give the security descriptor a Null Dacl done using the "TRUE, (PACL)NULL" here */
if (SetSecurityDescriptorDacl(&securityDesc, TRUE, (PACL)NULL, FALSE)) {

2
icu4c/source/common/umutex.cpp
View file

@ -132,7 +132,7 @@ umtx_condBroadcast(UConditionVar *condition) {
}
U_CAPI void U_EXPORT2
umtx_condSignal(UConditionVar *condition) {
umtx_condSignal(UConditionVar * /* condition */) {
// Function not implemented. There is no immediate requirement from ICU to have it.
// Once ICU drops support for Windows XP and Server 2003, ICU Condition Variables will be
// changed to be thin wrappers on native Windows CONDITION_VARIABLEs, and this function

24
icu4c/source/common/unicode/brkiter.h
View file

@ -250,7 +250,7 @@ public:
virtual int32_t next(void) = 0;
/**
* Return character index of the current interator position within the text.
* Return character index of the current iterator position within the text.
* @return The boundary most recently returned.
* @stable ICU 2.0
*/
@ -277,7 +277,7 @@ public:
virtual int32_t preceding(int32_t offset) = 0;
/**
* Return true if the specfied position is a boundary position.
* Return true if the specified position is a boundary position.
* As a side effect, the current position of the iterator is set
* to the first boundary position at or following the specified offset.
* @param offset the offset to check.
@ -331,7 +331,7 @@ public:
* @param fillInVec an array to be filled in with the status values.
* @param capacity the length of the supplied vector. A length of zero causes
* the function to return the number of status values, in the
* normal way, without attemtping to store any values.
* normal way, without attempting to store any values.
* @param status receives error codes.
* @return The number of rule status values from rules that determined
* the most recent boundary returned by the break iterator.
@ -469,7 +469,7 @@ public:
static const Locale* U_EXPORT2 getAvailableLocales(int32_t& count);
/**
* Get name of the object for the desired Locale, in the desired langauge.
* Get name of the object for the desired Locale, in the desired language.
* @param objectLocale must be from getAvailableLocales.
* @param displayLocale specifies the desired locale for output.
* @param name the fill-in parameter of the return value
@ -482,7 +482,7 @@ public:
UnicodeString& name);
/**
* Get name of the object for the desired Locale, in the langauge of the
* Get name of the object for the desired Locale, in the language of the
* default locale.
* @param objectLocale must be from getMatchingLocales
* @param name the fill-in parameter of the return value
@ -629,10 +629,12 @@ protected:
/** @internal */
BreakIterator();
/** @internal */
BreakIterator (const BreakIterator &other) : UObject(other) {}
BreakIterator (const BreakIterator &other);
#ifndef U_HIDE_INTERNAL_API
/** @internal */
BreakIterator (const Locale& valid, const Locale& actual);
BreakIterator (const Locale& valid, const Locale &actual);
/** @internal. Assignment Operator, used by RuleBasedBreakIterator. */
BreakIterator &operator = (const BreakIterator &other);
#endif /* U_HIDE_INTERNAL_API */
private:
@ -640,12 +642,6 @@ private:
/** @internal */
char actualLocale[ULOC_FULLNAME_CAPACITY];
char validLocale[ULOC_FULLNAME_CAPACITY];
/**
* The assignment operator has no real implementation.
* It's provided to make the compiler happy. Do not call.
*/
BreakIterator& operator=(const BreakIterator&);
};
#ifndef U_HIDE_DEPRECATED_API
@ -661,5 +657,5 @@ U_NAMESPACE_END
#endif /* #if !UCONFIG_NO_BREAK_ITERATION */
#endif // _BRKITER
#endif // BRKITER_H
//eof

33
icu4c/source/common/unicode/bytestream.h
View file

@ -126,8 +126,8 @@ public:
virtual void Flush();
private:
ByteSink(const ByteSink &); // copy constructor not implemented
ByteSink &operator=(const ByteSink &); // assignment operator not implemented
ByteSink(const ByteSink &) = delete;
ByteSink &operator=(const ByteSink &) = delete;
};
// -------------------------------------------------------------
@ -217,9 +217,10 @@ private:
int32_t size_;
int32_t appended_;
UBool overflowed_;
CheckedArrayByteSink(); ///< default constructor not implemented
CheckedArrayByteSink(const CheckedArrayByteSink &); ///< copy constructor not implemented
CheckedArrayByteSink &operator=(const CheckedArrayByteSink &); ///< assignment operator not implemented
CheckedArrayByteSink() = delete;
CheckedArrayByteSink(const CheckedArrayByteSink &) = delete;
CheckedArrayByteSink &operator=(const CheckedArrayByteSink &) = delete;
};
/**
@ -236,6 +237,21 @@ class StringByteSink : public ByteSink {
* @stable ICU 4.2
*/
StringByteSink(StringClass* dest) : dest_(dest) { }
#ifndef U_HIDE_DRAFT_API
/**
* Constructs a ByteSink that reserves append capacity and will append bytes to the dest string.
*
* @param dest pointer to string object to append to
* @param initialAppendCapacity capacity beyond dest->length() to be reserve()d
* @draft ICU 60
*/
StringByteSink(StringClass* dest, int32_t initialAppendCapacity) : dest_(dest) {
if (initialAppendCapacity > 0 &&
(uint32_t)initialAppendCapacity > (dest->capacity() - dest->length())) {
dest->reserve(dest->length() + initialAppendCapacity);
}
}
#endif // U_HIDE_DRAFT_API
/**
* Append "bytes[0,n-1]" to this.
* @param data the pointer to the bytes
@ -245,9 +261,10 @@ class StringByteSink : public ByteSink {
virtual void Append(const char* data, int32_t n) { dest_->append(data, n); }
private:
StringClass* dest_;
StringByteSink(); ///< default constructor not implemented
StringByteSink(const StringByteSink &); ///< copy constructor not implemented
StringByteSink &operator=(const StringByteSink &); ///< assignment operator not implemented
StringByteSink() = delete;
StringByteSink(const StringByteSink &) = delete;
StringByteSink &operator=(const StringByteSink &) = delete;
};
U_NAMESPACE_END

26
icu4c/source/common/unicode/casemap.h
View file

@ -36,7 +36,7 @@ public:
* The source string and the destination buffer must not overlap.
*
* @param locale The locale ID. ("" = root locale, NULL = default locale.)
* @param options Options bit set, usually 0. See UCASEMAP_OMIT_UNCHANGED_TEXT.
* @param options Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT.
* @param src The original string.
* @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
* @param dest A buffer for the result string. The result will be NUL-terminated if
@ -71,7 +71,7 @@ public:
* The source string and the destination buffer must not overlap.
*
* @param locale The locale ID. ("" = root locale, NULL = default locale.)
* @param options Options bit set, usually 0. See UCASEMAP_OMIT_UNCHANGED_TEXT.
* @param options Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT.
* @param src The original string.
* @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
* @param dest A buffer for the result string. The result will be NUL-terminated if
@ -112,8 +112,10 @@ public:
* all others. (This can be modified with options bits.)
*
* @param locale The locale ID. ("" = root locale, NULL = default locale.)
* @param options Options bit set, usually 0. See UCASEMAP_OMIT_UNCHANGED_TEXT,
* U_TITLECASE_NO_LOWERCASE, U_TITLECASE_NO_BREAK_ADJUSTMENT.
* @param options Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT,
* U_TITLECASE_NO_LOWERCASE,
* U_TITLECASE_NO_BREAK_ADJUSTMENT, U_TITLECASE_ADJUST_TO_CASED,
* U_TITLECASE_WHOLE_STRING, U_TITLECASE_SENTENCES.
* @param iter A break iterator to find the first characters of words that are to be titlecased.
* It is set to the source string (setText())
* and used one or more times for iteration (first() and next()).
@ -159,7 +161,7 @@ public:
* The result may be longer or shorter than the original.
* The source string and the destination buffer must not overlap.
*
* @param options Options bit set, usually 0. See UCASEMAP_OMIT_UNCHANGED_TEXT,
* @param options Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT,
* U_FOLD_CASE_DEFAULT, U_FOLD_CASE_EXCLUDE_SPECIAL_I.
* @param src The original string.
* @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
@ -195,7 +197,7 @@ public:
* The source string and the destination buffer must not overlap.
*
* @param locale The locale ID. ("" = root locale, NULL = default locale.)
* @param options Options bit set, usually 0. See UCASEMAP_OMIT_UNCHANGED_TEXT.
* @param options Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT.
* @param src The original string.
* @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
* @param dest A buffer for the result string. The result will be NUL-terminated if
@ -230,7 +232,7 @@ public:
* The source string and the destination buffer must not overlap.
*
* @param locale The locale ID. ("" = root locale, NULL = default locale.)
* @param options Options bit set, usually 0. See UCASEMAP_OMIT_UNCHANGED_TEXT.
* @param options Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT.
* @param src The original string.
* @param srcLength The length of the original string. If -1, then src must be NUL-terminated.
* @param dest A buffer for the result string. The result will be NUL-terminated if
@ -271,10 +273,12 @@ public:
* all others. (This can be modified with options bits.)
*
* @param locale The locale ID. ("" = root locale, NULL = default locale.)
* @param options Options bit set, usually 0. See UCASEMAP_OMIT_UNCHANGED_TEXT,
* U_TITLECASE_NO_LOWERCASE, U_TITLECASE_NO_BREAK_ADJUSTMENT.
* @param options Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT,
* U_TITLECASE_NO_LOWERCASE,
* U_TITLECASE_NO_BREAK_ADJUSTMENT, U_TITLECASE_ADJUST_TO_CASED,
* U_TITLECASE_WHOLE_STRING, U_TITLECASE_SENTENCES.
* @param iter A break iterator to find the first characters of words that are to be titlecased.
* It is set to the source string (setText())
* It is set to the source string (setUText())
* and used one or more times for iteration (first() and next()).
* If NULL, then a word break iterator for the locale is used
* (or something equivalent).
@ -317,7 +321,7 @@ public:
* The result may be longer or shorter than the original.
* The source string and the destination buffer must not overlap.
*
* @param options Options bit set, usually 0. See UCASEMAP_OMIT_UNCHANGED_TEXT,
* @param options Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT,
* U_FOLD_CASE_DEFAULT, U_FOLD_CASE_EXCLUDE_SPECIAL_I.
* @param src The original string.
* @param srcLength The length of the original string. If -1, then src must be NUL-terminated.

52
icu4c/source/common/unicode/char16ptr.h
View file

@ -95,45 +95,45 @@ private:
return reinterpret_cast<char16_t *>(t);
}
char16_t *p;
char16_t *p_;
#else
union {
char16_t *cp;
uint16_t *up;
wchar_t *wp;
} u;
} u_;
#endif
};
#ifdef U_ALIASING_BARRIER
Char16Ptr::Char16Ptr(char16_t *p) : p(p) {}
Char16Ptr::Char16Ptr(char16_t *p) : p_(p) {}
#if !U_CHAR16_IS_TYPEDEF
Char16Ptr::Char16Ptr(uint16_t *p) : p(cast(p)) {}
Char16Ptr::Char16Ptr(uint16_t *p) : p_(cast(p)) {}
#endif
#if U_SIZEOF_WCHAR_T==2
Char16Ptr::Char16Ptr(wchar_t *p) : p(cast(p)) {}
Char16Ptr::Char16Ptr(wchar_t *p) : p_(cast(p)) {}
#endif
Char16Ptr::Char16Ptr(std::nullptr_t p) : p(p) {}
Char16Ptr::Char16Ptr(std::nullptr_t p) : p_(p) {}
Char16Ptr::~Char16Ptr() {
U_ALIASING_BARRIER(p);
U_ALIASING_BARRIER(p_);
}
char16_t *Char16Ptr::get() const { return p; }
char16_t *Char16Ptr::get() const { return p_; }
#else
Char16Ptr::Char16Ptr(char16_t *p) { u.cp = p; }
Char16Ptr::Char16Ptr(char16_t *p) { u_.cp = p; }
#if !U_CHAR16_IS_TYPEDEF
Char16Ptr::Char16Ptr(uint16_t *p) { u.up = p; }
Char16Ptr::Char16Ptr(uint16_t *p) { u_.up = p; }
#endif
#if U_SIZEOF_WCHAR_T==2
Char16Ptr::Char16Ptr(wchar_t *p) { u.wp = p; }
Char16Ptr::Char16Ptr(wchar_t *p) { u_.wp = p; }
#endif
Char16Ptr::Char16Ptr(std::nullptr_t p) { u.cp = p; }
Char16Ptr::Char16Ptr(std::nullptr_t p) { u_.cp = p; }
Char16Ptr::~Char16Ptr() {}
char16_t *Char16Ptr::get() const { return u.cp; }
char16_t *Char16Ptr::get() const { return u_.cp; }
#endif
@ -203,45 +203,45 @@ private:
return reinterpret_cast<const char16_t *>(t);
}
const char16_t *p;
const char16_t *p_;
#else
union {
const char16_t *cp;
const uint16_t *up;
const wchar_t *wp;
} u;
} u_;
#endif
};
#ifdef U_ALIASING_BARRIER
ConstChar16Ptr::ConstChar16Ptr(const char16_t *p) : p(p) {}
ConstChar16Ptr::ConstChar16Ptr(const char16_t *p) : p_(p) {}
#if !U_CHAR16_IS_TYPEDEF
ConstChar16Ptr::ConstChar16Ptr(const uint16_t *p) : p(cast(p)) {}
ConstChar16Ptr::ConstChar16Ptr(const uint16_t *p) : p_(cast(p)) {}
#endif
#if U_SIZEOF_WCHAR_T==2
ConstChar16Ptr::ConstChar16Ptr(const wchar_t *p) : p(cast(p)) {}
ConstChar16Ptr::ConstChar16Ptr(const wchar_t *p) : p_(cast(p)) {}
#endif
ConstChar16Ptr::ConstChar16Ptr(const std::nullptr_t p) : p(p) {}
ConstChar16Ptr::ConstChar16Ptr(const std::nullptr_t p) : p_(p) {}
ConstChar16Ptr::~ConstChar16Ptr() {
U_ALIASING_BARRIER(p);
U_ALIASING_BARRIER(p_);
}
const char16_t *ConstChar16Ptr::get() const { return p; }
const char16_t *ConstChar16Ptr::get() const { return p_; }
#else
ConstChar16Ptr::ConstChar16Ptr(const char16_t *p) { u.cp = p; }
ConstChar16Ptr::ConstChar16Ptr(const char16_t *p) { u_.cp = p; }
#if !U_CHAR16_IS_TYPEDEF
ConstChar16Ptr::ConstChar16Ptr(const uint16_t *p) { u.up = p; }
ConstChar16Ptr::ConstChar16Ptr(const uint16_t *p) { u_.up = p; }
#endif
#if U_SIZEOF_WCHAR_T==2
ConstChar16Ptr::ConstChar16Ptr(const wchar_t *p) { u.wp = p; }
ConstChar16Ptr::ConstChar16Ptr(const wchar_t *p) { u_.wp = p; }
#endif
ConstChar16Ptr::ConstChar16Ptr(const std::nullptr_t p) { u.cp = p; }
ConstChar16Ptr::ConstChar16Ptr(const std::nullptr_t p) { u_.cp = p; }
ConstChar16Ptr::~ConstChar16Ptr() {}
const char16_t *ConstChar16Ptr::get() const { return u.cp; }
const char16_t *ConstChar16Ptr::get() const { return u_.cp; }
#endif

195
icu4c/source/common/unicode/edits.h
View file

@ -36,19 +36,61 @@ public:
* @draft ICU 59
*/
Edits() :
array(stackArray), capacity(STACK_CAPACITY), length(0), delta(0),
errorCode(U_ZERO_ERROR) {}
array(stackArray), capacity(STACK_CAPACITY), length(0), delta(0), numChanges(0),
errorCode_(U_ZERO_ERROR) {}
/**
* Copy constructor.
* @param other source edits
* @draft ICU 60
*/
Edits(const Edits &other) :
array(stackArray), capacity(STACK_CAPACITY), length(other.length),
delta(other.delta), numChanges(other.numChanges),
errorCode_(other.errorCode_) {
copyArray(other);
}
/**
* Move constructor, might leave src empty.
* This object will have the same contents that the source object had.
* @param src source edits
* @draft ICU 60
*/
Edits(Edits &&src) U_NOEXCEPT :
array(stackArray), capacity(STACK_CAPACITY), length(src.length),
delta(src.delta), numChanges(src.numChanges),
errorCode_(src.errorCode_) {
moveArray(src);
}
/**
* Destructor.
* @draft ICU 59
*/
~Edits();
/**
* Assignment operator.
* @param other source edits
* @return *this
* @draft ICU 60
*/
Edits &operator=(const Edits &other);
/**
* Move assignment operator, might leave src empty.
* This object will have the same contents that the source object had.
* The behavior is undefined if *this and src are the same object.
* @param src source edits
* @return *this
* @draft ICU 60
*/
Edits &operator=(Edits &&src) U_NOEXCEPT;
/**
* Resets the data but may not release memory.
* @draft ICU 59
*/
void reset();
void reset() U_NOEXCEPT;
/**
* Adds a record for an unchanged segment of text.
@ -66,6 +108,9 @@ public:
* Sets the UErrorCode if an error occurred while recording edits.
* Preserves older error codes in the outErrorCode.
* Normally called from inside ICU string transformation functions, not user code.
* @param outErrorCode Set to an error code if it does not contain one already
* and an error occurred while recording edits.
* Otherwise unchanged.
* @return TRUE if U_FAILURE(outErrorCode)
* @draft ICU 59
*/
@ -81,7 +126,13 @@ public:
* @return TRUE if there are any change edits
* @draft ICU 59
*/
UBool hasChanges() const;
UBool hasChanges() const { return numChanges != 0; }
/**
* @return the number of change edits
* @draft ICU 60
*/
int32_t numberOfChanges() const { return numChanges; }
/**
* Access to the list of edits.
@ -90,6 +141,15 @@ public:
* @draft ICU 59
*/
struct U_COMMON_API Iterator U_FINAL : public UMemory {
/**
* Default constructor, empty iterator.
* @draft ICU 60
*/
Iterator() :
array(nullptr), index(0), length(0),
remaining(0), onlyChanges_(FALSE), coarse(FALSE),
dir(0), changed(FALSE), oldLength_(0), newLength_(0),
srcIndex(0), replIndex(0), destIndex(0) {}
/**
* Copy constructor.
* @draft ICU 59
@ -103,6 +163,9 @@ public:
/**
* Advances to the next edit.
* @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
* or else the function returns immediately. Check for U_FAILURE()
* on output or use with function chaining. (See User Guide for details.)
* @return TRUE if there is another edit
* @draft ICU 59
*/
@ -121,10 +184,86 @@ public:
* if the source index is out of bounds for the source string.
*
* @param i source index
* @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
* or else the function returns immediately. Check for U_FAILURE()
* on output or use with function chaining. (See User Guide for details.)
* @return TRUE if the edit for the source index was found
* @draft ICU 59
*/
UBool findSourceIndex(int32_t i, UErrorCode &errorCode);
UBool findSourceIndex(int32_t i, UErrorCode &errorCode) {
return findIndex(i, TRUE, errorCode) == 0;
}
/**
* Finds the edit that contains the destination index.
* The destination index may be found in a non-change
* even if normal iteration would skip non-changes.
* Normal iteration can continue from a found edit.
*
* The iterator state before this search logically does not matter.
* (It may affect the performance of the search.)
*
* The iterator state after this search is undefined
* if the source index is out of bounds for the source string.
*
* @param i destination index
* @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
* or else the function returns immediately. Check for U_FAILURE()
* on output or use with function chaining. (See User Guide for details.)
* @return TRUE if the edit for the destination index was found
* @draft ICU 60
*/
UBool findDestinationIndex(int32_t i, UErrorCode &errorCode) {
return findIndex(i, FALSE, errorCode) == 0;
}
/**
* Returns the destination index corresponding to the given source index.
* If the source index is inside a change edit (not at its start),
* then the destination index at the end of that edit is returned,
* since there is no information about index mapping inside a change edit.
*
* (This means that indexes to the start and middle of an edit,
* for example around a grapheme cluster, are mapped to indexes
* encompassing the entire edit.
* The alternative, mapping an interior index to the start,
* would map such an interval to an empty one.)
*
* This operation will usually but not always modify this object.
* The iterator state after this search is undefined.
*
* @param i source index
* @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
* or else the function returns immediately. Check for U_FAILURE()
* on output or use with function chaining. (See User Guide for details.)
* @return destination index; undefined if i is not 0..string length
* @draft ICU 60
*/
int32_t destinationIndexFromSourceIndex(int32_t i, UErrorCode &errorCode);
/**
* Returns the source index corresponding to the given destination index.
* If the destination index is inside a change edit (not at its start),
* then the source index at the end of that edit is returned,
* since there is no information about index mapping inside a change edit.
*
* (This means that indexes to the start and middle of an edit,
* for example around a grapheme cluster, are mapped to indexes
* encompassing the entire edit.
* The alternative, mapping an interior index to the start,
* would map such an interval to an empty one.)
*
* This operation will usually but not always modify this object.
* The iterator state after this search is undefined.
*
* @param i destination index
* @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
* or else the function returns immediately. Check for U_FAILURE()
* on output or use with function chaining. (See User Guide for details.)
* @return source index; undefined if i is not 0..string length
* @draft ICU 60
*/
int32_t sourceIndexFromDestinationIndex(int32_t i, UErrorCode &errorCode);
/**
* @return TRUE if this edit replaces oldLength() units with newLength() different ones.
@ -167,15 +306,22 @@ public:
Iterator(const uint16_t *a, int32_t len, UBool oc, UBool crs);
int32_t readLength(int32_t head);
void updateIndexes();
void updateNextIndexes();
void updatePreviousIndexes();
UBool noNext();
UBool next(UBool onlyChanges, UErrorCode &errorCode);
UBool previous(UErrorCode &errorCode);
/** @return -1: error or i<0; 0: found; 1: i>=string length */
int32_t findIndex(int32_t i, UBool findSource, UErrorCode &errorCode);
const uint16_t *array;
int32_t index, length;
// 0 if we are not within compressed equal-length changes.
// Otherwise the number of remaining changes, including the current one.
int32_t remaining;
UBool onlyChanges_, coarse;
int8_t dir; // iteration direction: back(<0), initial(0), forward(>0)
UBool changed;
int32_t oldLength_, newLength_;
int32_t srcIndex, replIndex, destIndex;
@ -219,9 +365,39 @@ public:
return Iterator(array, length, FALSE, FALSE);
}
/**
* Merges the two input Edits and appends the result to this object.
*
* Consider two string transformations (for example, normalization and case mapping)
* where each records Edits in addition to writing an output string.<br>
* Edits ab reflect how substrings of input string a
* map to substrings of intermediate string b.<br>
* Edits bc reflect how substrings of intermediate string b
* map to substrings of output string c.<br>
* This function merges ab and bc such that the additional edits
* recorded in this object reflect how substrings of input string a
* map to substrings of output string c.
*
* If unrelated Edits are passed in where the output string of the first
* has a different length than the input string of the second,
* then a U_ILLEGAL_ARGUMENT_ERROR is reported.
*
* @param ab reflects how substrings of input string a
* map to substrings of intermediate string b.
* @param bc reflects how substrings of intermediate string b
* map to substrings of output string c.
* @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
* or else the function returns immediately. Check for U_FAILURE()
* on output or use with function chaining. (See User Guide for details.)
* @return *this, with the merged edits appended
* @draft ICU 60
*/
Edits &mergeAndAppend(const Edits &ab, const Edits &bc, UErrorCode &errorCode);
private:
Edits(const Edits &) = delete;
Edits &operator=(const Edits &) = delete;
void releaseArray() U_NOEXCEPT;
Edits &copyArray(const Edits &other);
Edits &moveArray(Edits &src) U_NOEXCEPT;
void setLastUnit(int32_t last) { array[length - 1] = (uint16_t)last; }
int32_t lastUnit() const { return length > 0 ? array[length - 1] : 0xffff; }
@ -234,7 +410,8 @@ private:
int32_t capacity;
int32_t length;
int32_t delta;
UErrorCode errorCode;
int32_t numChanges;
UErrorCode errorCode_;
uint16_t stackArray[STACK_CAPACITY];
};

2
icu4c/source/common/unicode/locid.h
View file

@ -88,7 +88,7 @@ class UnicodeString;
* <P>
* The third constructor requires a third argument--the <STRONG>Variant.</STRONG>
* The Variant codes are vendor and browser-specific.
* For example, use REVISED for a langauge's revised script orthography, and POSIX for POSIX.
* For example, use REVISED for a language's revised script orthography, and POSIX for POSIX.
* Where there are two variants, separate them with an underscore, and
* put the most important one first. For
* example, a Traditional Spanish collation might be referenced, with

138
icu4c/source/common/unicode/normalizer2.h
View file

@ -28,12 +28,15 @@
#if !UCONFIG_NO_NORMALIZATION
#include "unicode/stringpiece.h"
#include "unicode/uniset.h"
#include "unicode/unistr.h"
#include "unicode/unorm2.h"
U_NAMESPACE_BEGIN
class ByteSink;
/**
* Unicode normalization functionality for standard Unicode normalization or
* for using custom mapping tables.
@ -215,6 +218,34 @@ public:
normalize(const UnicodeString &src,
UnicodeString &dest,
UErrorCode &errorCode) const = 0;
/**
* Normalizes a UTF-8 string and optionally records how source substrings
* relate to changed and unchanged result substrings.
*
* Currently implemented completely only for "compose" modes,
* such as for NFC, NFKC, and NFKC_Casefold
* (UNORM2_COMPOSE and UNORM2_COMPOSE_CONTIGUOUS).
* Otherwise currently converts to & from UTF-16 and does not support edits.
*
* @param options Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT.
* @param src Source UTF-8 string.
* @param sink A ByteSink to which the normalized UTF-8 result string is written.
* sink.Flush() is called at the end.
* @param edits Records edits for index mapping, working with styled text,
* and getting only changes (if any).
* The Edits contents is undefined if any error occurs.
* This function calls edits->reset() first. edits can be nullptr.
* @param errorCode Standard ICU error code. Its input value must
* pass the U_SUCCESS() test, or else the function returns
* immediately. Check for U_FAILURE() on output or use with
* function chaining. (See User Guide for details.)
* @draft ICU 60
*/
virtual void
normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
Edits *edits, UErrorCode &errorCode) const;
/**
* Appends the normalized form of the second string to the first string
* (merging them at the boundary) and returns the first string.
@ -340,6 +371,30 @@ public:
*/
virtual UBool
isNormalized(const UnicodeString &s, UErrorCode &errorCode) const = 0;
/**
* Tests if the UTF-8 string is normalized.
* Internally, in cases where the quickCheck() method would return "maybe"
* (which is only possible for the two COMPOSE modes) this method
* resolves to "yes" or "no" to provide a definitive result,
* at the cost of doing more work in those cases.
*
* This works for all normalization modes,
* but it is currently optimized for UTF-8 only for "compose" modes,
* such as for NFC, NFKC, and NFKC_Casefold
* (UNORM2_COMPOSE and UNORM2_COMPOSE_CONTIGUOUS).
* For other modes it currently converts to UTF-16 and calls isNormalized().
*
* @param s UTF-8 input string
* @param errorCode Standard ICU error code. Its input value must
* pass the U_SUCCESS() test, or else the function returns
* immediately. Check for U_FAILURE() on output or use with
* function chaining. (See User Guide for details.)
* @return TRUE if s is normalized
* @draft ICU 60
*/
virtual UBool
isNormalizedUTF8(StringPiece s, UErrorCode &errorCode) const;
/**
* Tests if the string is normalized.
@ -479,7 +534,35 @@ public:
virtual UnicodeString &
normalize(const UnicodeString &src,
UnicodeString &dest,
UErrorCode &errorCode) const;
UErrorCode &errorCode) const override;
/**
* Normalizes a UTF-8 string and optionally records how source substrings
* relate to changed and unchanged result substrings.
*
* Currently implemented completely only for "compose" modes,
* such as for NFC, NFKC, and NFKC_Casefold
* (UNORM2_COMPOSE and UNORM2_COMPOSE_CONTIGUOUS).
* Otherwise currently converts to & from UTF-16 and does not support edits.
*
* @param options Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT.
* @param src Source UTF-8 string.
* @param sink A ByteSink to which the normalized UTF-8 result string is written.
* sink.Flush() is called at the end.
* @param edits Records edits for index mapping, working with styled text,
* and getting only changes (if any).
* The Edits contents is undefined if any error occurs.
* This function calls edits->reset() first. edits can be nullptr.
* @param errorCode Standard ICU error code. Its input value must
* pass the U_SUCCESS() test, or else the function returns
* immediately. Check for U_FAILURE() on output or use with
* function chaining. (See User Guide for details.)
* @draft ICU 60
*/
virtual void
normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
Edits *edits, UErrorCode &errorCode) const override;
/**
* Appends the normalized form of the second string to the first string
* (merging them at the boundary) and returns the first string.
@ -497,7 +580,7 @@ public:
virtual UnicodeString &
normalizeSecondAndAppend(UnicodeString &first,
const UnicodeString &second,
UErrorCode &errorCode) const;
UErrorCode &errorCode) const override;
/**
* Appends the second string to the first string
* (merging them at the boundary) and returns the first string.
@ -515,7 +598,7 @@ public:
virtual UnicodeString &
append(UnicodeString &first,
const UnicodeString &second,
UErrorCode &errorCode) const;
UErrorCode &errorCode) const override;
/**
* Gets the decomposition mapping of c.
@ -529,7 +612,7 @@ public:
* @stable ICU 4.6
*/
virtual UBool
getDecomposition(UChar32 c, UnicodeString &decomposition) const;
getDecomposition(UChar32 c, UnicodeString &decomposition) const override;
/**
* Gets the raw decomposition mapping of c.
@ -543,7 +626,7 @@ public:
* @stable ICU 49
*/
virtual UBool
getRawDecomposition(UChar32 c, UnicodeString &decomposition) const;
getRawDecomposition(UChar32 c, UnicodeString &decomposition) const override;
/**
* Performs pairwise composition of a & b and returns the composite if there is one.
@ -556,7 +639,7 @@ public:
* @stable ICU 49
*/
virtual UChar32
composePair(UChar32 a, UChar32 b) const;
composePair(UChar32 a, UChar32 b) const override;
/**
* Gets the combining class of c.
@ -567,7 +650,7 @@ public:
* @stable ICU 49
*/
virtual uint8_t
getCombiningClass(UChar32 c) const;
getCombiningClass(UChar32 c) const override;
/**
* Tests if the string is normalized.
@ -581,7 +664,30 @@ public:
* @stable ICU 4.4
*/
virtual UBool
isNormalized(const UnicodeString &s, UErrorCode &errorCode) const;
isNormalized(const UnicodeString &s, UErrorCode &errorCode) const override;
/**
* Tests if the UTF-8 string is normalized.
* Internally, in cases where the quickCheck() method would return "maybe"
* (which is only possible for the two COMPOSE modes) this method
* resolves to "yes" or "no" to provide a definitive result,
* at the cost of doing more work in those cases.
*
* This works for all normalization modes,
* but it is currently optimized for UTF-8 only for "compose" modes,
* such as for NFC, NFKC, and NFKC_Casefold
* (UNORM2_COMPOSE and UNORM2_COMPOSE_CONTIGUOUS).
* For other modes it currently converts to UTF-16 and calls isNormalized().
*
* @param s UTF-8 input string
* @param errorCode Standard ICU error code. Its input value must
* pass the U_SUCCESS() test, or else the function returns
* immediately. Check for U_FAILURE() on output or use with
* function chaining. (See User Guide for details.)
* @return TRUE if s is normalized
* @draft ICU 60
*/
virtual UBool
isNormalizedUTF8(StringPiece s, UErrorCode &errorCode) const override;
/**
* Tests if the string is normalized.
* For details see the Normalizer2 base class documentation.
@ -594,7 +700,7 @@ public:
* @stable ICU 4.4
*/
virtual UNormalizationCheckResult
quickCheck(const UnicodeString &s, UErrorCode &errorCode) const;
quickCheck(const UnicodeString &s, UErrorCode &errorCode) const override;
/**
* Returns the end of the normalized substring of the input string.
* For details see the Normalizer2 base class documentation.
@ -607,7 +713,7 @@ public:
* @stable ICU 4.4
*/
virtual int32_t
spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const;
spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const override;
/**
* Tests if the character always has a normalization boundary before it,
@ -617,7 +723,7 @@ public:
* @return TRUE if c has a normalization boundary before it
* @stable ICU 4.4
*/
virtual UBool hasBoundaryBefore(UChar32 c) const;
virtual UBool hasBoundaryBefore(UChar32 c) const override;
/**
* Tests if the character always has a normalization boundary after it,
@ -627,7 +733,7 @@ public:
* @return TRUE if c has a normalization boundary after it
* @stable ICU 4.4
*/
virtual UBool hasBoundaryAfter(UChar32 c) const;
virtual UBool hasBoundaryAfter(UChar32 c) const override;
/**
* Tests if the character is normalization-inert.
@ -636,7 +742,7 @@ public:
* @return TRUE if c is normalization-inert
* @stable ICU 4.4
*/
virtual UBool isInert(UChar32 c) const;
virtual UBool isInert(UChar32 c) const override;
private:
UnicodeString &
normalize(const UnicodeString &src,
@ -644,6 +750,12 @@ private:
USetSpanCondition spanCondition,
UErrorCode &errorCode) const;
void
normalizeUTF8(uint32_t options, const char *src, int32_t length,
ByteSink &sink, Edits *edits,
USetSpanCondition spanCondition,
UErrorCode &errorCode) const;
UnicodeString &
normalizeSecondAndAppend(UnicodeString &first,
const UnicodeString &second,

2
icu4c/source/common/unicode/rbbi.h
View file

@ -32,8 +32,6 @@
#include "unicode/uchriter.h"
struct UTrie;
U_NAMESPACE_BEGIN
/** @internal */

188
icu4c/source/common/unicode/stringoptions.h Normal file
View file

@ -0,0 +1,188 @@
// © 2017 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
// stringoptions.h
// created: 2017jun08 Markus W. Scherer
#ifndef __STRINGOPTIONS_H__
#define __STRINGOPTIONS_H__
#include "unicode/utypes.h"
/**
* \file
* \brief C API: Bit set option bit constants for various string and character processing functions.
*/
/**
* Option value for case folding: Use default mappings defined in CaseFolding.txt.
*
* @stable ICU 2.0
*/
#define U_FOLD_CASE_DEFAULT 0
/**
* Option value for case folding:
*
* Use the modified set of mappings provided in CaseFolding.txt to handle dotted I
* and dotless i appropriately for Turkic languages (tr, az).
*
* Before Unicode 3.2, CaseFolding.txt contains mappings marked with 'I' that
* are to be included for default mappings and
* excluded for the Turkic-specific mappings.
*
* Unicode 3.2 CaseFolding.txt instead contains mappings marked with 'T' that
* are to be excluded for default mappings and
* included for the Turkic-specific mappings.
*
* @stable ICU 2.0
*/
#define U_FOLD_CASE_EXCLUDE_SPECIAL_I 1
#ifndef U_HIDE_DRAFT_API
/**
* Titlecase the string as a whole rather than each word.
* (Titlecase only the character at index 0, possibly adjusted.)
* Option bits value for titlecasing APIs that take an options bit set.
*
* It is an error to specify multiple titlecasing iterator options together,
* including both an options bit and an explicit BreakIterator.
*
* @see U_TITLECASE_ADJUST_TO_CASED
* @draft ICU 60
*/
#define U_TITLECASE_WHOLE_STRING 0x20
/**
* Titlecase sentences rather than words.
* (Titlecase only the first character of each sentence, possibly adjusted.)
* Option bits value for titlecasing APIs that take an options bit set.
*
* It is an error to specify multiple titlecasing iterator options together,
* including both an options bit and an explicit BreakIterator.
*
* @see U_TITLECASE_ADJUST_TO_CASED
* @draft ICU 60
*/
#define U_TITLECASE_SENTENCES 0x40
#endif // U_HIDE_DRAFT_API
/**
* Do not lowercase non-initial parts of words when titlecasing.
* Option bit for titlecasing APIs that take an options bit set.
*
* By default, titlecasing will titlecase the character at each
* (possibly adjusted) BreakIterator index and
* lowercase all other characters up to the next iterator index.
* With this option, the other characters will not be modified.
*
* @see U_TITLECASE_ADJUST_TO_CASED
* @see UnicodeString::toTitle
* @see CaseMap::toTitle
* @see ucasemap_setOptions
* @see ucasemap_toTitle
* @see ucasemap_utf8ToTitle
* @stable ICU 3.8
*/
#define U_TITLECASE_NO_LOWERCASE 0x100
/**
* Do not adjust the titlecasing BreakIterator indexes;
* titlecase exactly the characters at breaks from the iterator.
* Option bit for titlecasing APIs that take an options bit set.
*
* By default, titlecasing will take each break iterator index,
* adjust it to the next relevant character (see U_TITLECASE_ADJUST_TO_CASED),
* and titlecase that one.
*
* Other characters are lowercased.
*
* It is an error to specify multiple titlecasing adjustment options together.
*
* @see U_TITLECASE_ADJUST_TO_CASED
* @see U_TITLECASE_NO_LOWERCASE
* @see UnicodeString::toTitle
* @see CaseMap::toTitle
* @see ucasemap_setOptions
* @see ucasemap_toTitle
* @see ucasemap_utf8ToTitle
* @stable ICU 3.8
*/
#define U_TITLECASE_NO_BREAK_ADJUSTMENT 0x200
#ifndef U_HIDE_DRAFT_API
/**
* Adjust each titlecasing BreakIterator index to the next cased character.
* (See the Unicode Standard, chapter 3, Default Case Conversion, R3 toTitlecase(X).)
* Option bit for titlecasing APIs that take an options bit set.
*
* This used to be the default index adjustment in ICU.
* Since ICU 60, the default index adjustment is to the next character that is
* a letter, number, symbol, or private use code point.
* (Uncased modifier letters are skipped.)
* The difference in behavior is small for word titlecasing,
* but the new adjustment is much better for whole-string and sentence titlecasing:
* It yields "49ers" and "«丰(abc)»" instead of "49Ers" and "«丰(Abc)»".
*
* It is an error to specify multiple titlecasing adjustment options together.
*
* @see U_TITLECASE_NO_BREAK_ADJUSTMENT
* @draft ICU 60
*/
#define U_TITLECASE_ADJUST_TO_CASED 0x400
/**
* Omit unchanged text when recording how source substrings
* relate to changed and unchanged result substrings.
* Used for example in some case-mapping and normalization functions.
*
* @see CaseMap
* @see Edits
* @see Normalizer2
* @draft ICU 60
*/
#define U_OMIT_UNCHANGED_TEXT 0x4000
#endif // U_HIDE_DRAFT_API
/**
* Option bit for u_strCaseCompare, u_strcasecmp, unorm_compare, etc:
* Compare strings in code point order instead of code unit order.
* @stable ICU 2.2
*/
#define U_COMPARE_CODE_POINT_ORDER 0x8000
/**
* Option bit for unorm_compare:
* Perform case-insensitive comparison.
* @stable ICU 2.2
*/
#define U_COMPARE_IGNORE_CASE 0x10000
/**
* Option bit for unorm_compare:
* Both input strings are assumed to fulfill FCD conditions.
* @stable ICU 2.2
*/
#define UNORM_INPUT_IS_FCD 0x20000
// Related definitions elsewhere.
// Options that are not meaningful in the same functions
// can share the same bits.
//
// Public:
// unicode/unorm.h #define UNORM_COMPARE_NORM_OPTIONS_SHIFT 20
//
// Internal: (may change or be removed)
// ucase.h #define _STRCASECMP_OPTIONS_MASK 0xffff
// ucase.h #define _FOLD_CASE_OPTIONS_MASK 7
// ucasemap_imp.h #define U_TITLECASE_ITERATOR_MASK 0xe0
// ucasemap_imp.h #define U_TITLECASE_ADJUSTMENT_MASK 0x600
// ustr_imp.h #define _STRNCMP_STYLE 0x1000
// ustr_imp.h #define U_EDITS_NO_RESET 0x2000
// unormcmp.cpp #define _COMPARE_EQUIV 0x80000
#endif // __STRINGOPTIONS_H__

18
icu4c/source/common/unicode/stringtriebuilder.h
View file

@ -256,7 +256,7 @@ protected:
/** @internal */
class FinalValueNode : public Node {
public:
FinalValueNode(int32_t v) : Node(0x111111*37+v), value(v) {}
FinalValueNode(int32_t v) : Node(0x111111u*37u+v), value(v) {}
virtual UBool operator==(const Node &other) const;
virtual void write(StringTrieBuilder &builder);
protected:
@ -276,7 +276,7 @@ protected:
void setValue(int32_t v) {
hasValue=TRUE;
value=v;
hash=hash*37+v;
hash=hash*37u+v;
}
protected:
UBool hasValue;
@ -290,7 +290,7 @@ protected:
class IntermediateValueNode : public ValueNode {
public:
IntermediateValueNode(int32_t v, Node *nextNode)
: ValueNode(0x222222*37+hashCode(nextNode)), next(nextNode) { setValue(v); }
: ValueNode(0x222222u*37u+hashCode(nextNode)), next(nextNode) { setValue(v); }
virtual UBool operator==(const Node &other) const;
virtual int32_t markRightEdgesFirst(int32_t edgeNumber);
virtual void write(StringTrieBuilder &builder);
@ -307,7 +307,7 @@ protected:
class LinearMatchNode : public ValueNode {
public:
LinearMatchNode(int32_t len, Node *nextNode)
: ValueNode((0x333333*37+len)*37+hashCode(nextNode)),
: ValueNode((0x333333u*37u+len)*37u+hashCode(nextNode)),
length(len), next(nextNode) {}
virtual UBool operator==(const Node &other) const;
virtual int32_t markRightEdgesFirst(int32_t edgeNumber);
@ -342,7 +342,7 @@ protected:
equal[length]=NULL;
values[length]=value;
++length;
hash=(hash*37+c)*37+value;
hash=(hash*37u+c)*37u+value;
}
// Adds a unit which leads to another match node.
void add(int32_t c, Node *node) {
@ -350,7 +350,7 @@ protected:
equal[length]=node;
values[length]=0;
++length;
hash=(hash*37+c)*37+hashCode(node);
hash=(hash*37u+c)*37u+hashCode(node);
}
protected:
Node *equal[kMaxBranchLinearSubNodeLength]; // NULL means "has final value".
@ -365,8 +365,8 @@ protected:
class SplitBranchNode : public BranchNode {
public:
SplitBranchNode(char16_t middleUnit, Node *lessThanNode, Node *greaterOrEqualNode)
: BranchNode(((0x555555*37+middleUnit)*37+
hashCode(lessThanNode))*37+hashCode(greaterOrEqualNode)),
: BranchNode(((0x555555u*37u+middleUnit)*37u+
hashCode(lessThanNode))*37u+hashCode(greaterOrEqualNode)),
unit(middleUnit), lessThan(lessThanNode), greaterOrEqual(greaterOrEqualNode) {}
virtual UBool operator==(const Node &other) const;
virtual int32_t markRightEdgesFirst(int32_t edgeNumber);
@ -382,7 +382,7 @@ protected:
class BranchHeadNode : public ValueNode {
public:
BranchHeadNode(int32_t len, Node *subNode)
: ValueNode((0x666666*37+len)*37+hashCode(subNode)),
: ValueNode((0x666666u*37u+len)*37u+hashCode(subNode)),
length(len), next(subNode) {}
virtual UBool operator==(const Node &other) const;
virtual int32_t markRightEdgesFirst(int32_t edgeNumber);

53
icu4c/source/common/unicode/ucasemap.h
View file

@ -23,6 +23,7 @@
#include "unicode/utypes.h"
#include "unicode/localpointer.h"
#include "unicode/stringoptions.h"
#include "unicode/ustring.h"
/**
@ -144,56 +145,6 @@ ucasemap_setLocale(UCaseMap *csm, const char *locale, UErrorCode *pErrorCode);
U_STABLE void U_EXPORT2
ucasemap_setOptions(UCaseMap *csm, uint32_t options, UErrorCode *pErrorCode);
/**
* Do not lowercase non-initial parts of words when titlecasing.
* Option bit for titlecasing APIs that take an options bit set.
*
* By default, titlecasing will titlecase the first cased character
* of a word and lowercase all other characters.
* With this option, the other characters will not be modified.
*
* @see ucasemap_setOptions
* @see ucasemap_toTitle
* @see ucasemap_utf8ToTitle
* @see UnicodeString::toTitle
* @stable ICU 3.8
*/
#define U_TITLECASE_NO_LOWERCASE 0x100
/**
* Do not adjust the titlecasing indexes from BreakIterator::next() indexes;
* titlecase exactly the characters at breaks from the iterator.
* Option bit for titlecasing APIs that take an options bit set.
*
* By default, titlecasing will take each break iterator index,
* adjust it by looking for the next cased character, and titlecase that one.
* Other characters are lowercased.
*
* This follows Unicode 4 & 5 section 3.13 Default Case Operations:
*
* R3 toTitlecase(X): Find the word boundaries based on Unicode Standard Annex
* #29, "Text Boundaries." Between each pair of word boundaries, find the first
* cased character F. If F exists, map F to default_title(F); then map each
* subsequent character C to default_lower(C).
*
* @see ucasemap_setOptions
* @see ucasemap_toTitle
* @see ucasemap_utf8ToTitle
* @see UnicodeString::toTitle
* @see U_TITLECASE_NO_LOWERCASE
* @stable ICU 3.8
*/
#define U_TITLECASE_NO_BREAK_ADJUSTMENT 0x200
/**
* Omit unchanged text when case-mapping with Edits.
*
* @see CaseMap
* @see Edits
* @draft ICU 59
*/
#define UCASEMAP_OMIT_UNCHANGED_TEXT 0x4000
#if !UCONFIG_NO_BREAK_ITERATION
/**
@ -251,7 +202,7 @@ ucasemap_setBreakIterator(UCaseMap *csm, UBreakIterator *iterToAdopt, UErrorCode
* The standard titlecase iterator for the root locale implements the
* algorithm of Unicode TR 21.
*
* This function uses only the setUText(), first(), next() and close() methods of the
* This function uses only the setText(), first() and next() methods of the
* provided break iterator.
*
* The result may be longer or shorter than the original.

22
icu4c/source/common/unicode/uchar.h
View file

@ -26,6 +26,7 @@
#define UCHAR_H
#include "unicode/utypes.h"
#include "unicode/stringoptions.h"
U_CDECL_BEGIN
@ -3569,27 +3570,6 @@ u_toupper(UChar32 c);
U_STABLE UChar32 U_EXPORT2
u_totitle(UChar32 c);
/** Option value for case folding: use default mappings defined in CaseFolding.txt. @stable ICU 2.0 */
#define U_FOLD_CASE_DEFAULT 0
/**
* Option value for case folding:
*
* Use the modified set of mappings provided in CaseFolding.txt to handle dotted I
* and dotless i appropriately for Turkic languages (tr, az).
*
* Before Unicode 3.2, CaseFolding.txt contains mappings marked with 'I' that
* are to be included for default mappings and
* excluded for the Turkic-specific mappings.
*
* Unicode 3.2 CaseFolding.txt instead contains mappings marked with 'T' that
* are to be excluded for default mappings and
* included for the Turkic-specific mappings.
*
* @stable ICU 2.0
*/
#define U_FOLD_CASE_EXCLUDE_SPECIAL_I 1
/**
* The given character is mapped to its case folding equivalent according to
* UnicodeData.txt and CaseFolding.txt;

18
icu4c/source/common/unicode/unistr.h
View file

@ -38,16 +38,6 @@
struct UConverter; // unicode/ucnv.h
#ifndef U_COMPARE_CODE_POINT_ORDER
/* see also ustring.h and unorm.h */
/**
* Option bit for u_strCaseCompare, u_strcasecmp, unorm_compare, etc:
* Compare strings in code point order instead of code unit order.
* @stable ICU 2.2
*/
#define U_COMPARE_CODE_POINT_ORDER 0x8000
#endif
#ifndef USTRING_H
/**
* \ingroup ustring_ustrlen
@ -1730,7 +1720,7 @@ public:
*/
template<typename StringClass>
StringClass &toUTF8String(StringClass &result) const {
StringByteSink<StringClass> sbs(&result);
StringByteSink<StringClass> sbs(&result, length());
toUTF8(sbs);
return result;
}
@ -2785,11 +2775,11 @@ public:
* break iterator is opened.
* Otherwise the provided iterator is set to the string's text.
* @param locale The locale to consider.
* @param options Options bit set, usually 0. See U_TITLECASE_NO_LOWERCASE,
* U_TITLECASE_NO_BREAK_ADJUSTMENT, U_TITLECASE_ADJUST_TO_CASED,
* U_TITLECASE_WHOLE_STRING, U_TITLECASE_SENTENCES.
* @param options Options bit set, see ucasemap_open().
* @return A reference to this.
* @see U_TITLECASE_NO_LOWERCASE
* @see U_TITLECASE_NO_BREAK_ADJUSTMENT
* @see ucasemap_open
* @stable ICU 3.8
*/
UnicodeString &toTitle(BreakIterator *titleIter, const Locale &locale, uint32_t options);

25
icu4c/source/common/unicode/unorm2.h
View file

@ -32,6 +32,7 @@
#include "unicode/utypes.h"
#include "unicode/localpointer.h"
#include "unicode/stringoptions.h"
#include "unicode/uset.h"
/**
@ -526,30 +527,6 @@ unorm2_hasBoundaryAfter(const UNormalizer2 *norm2, UChar32 c);
U_STABLE UBool U_EXPORT2
unorm2_isInert(const UNormalizer2 *norm2, UChar32 c);
/**
* Option bit for unorm_compare:
* Both input strings are assumed to fulfill FCD conditions.
* @stable ICU 2.2
*/
#define UNORM_INPUT_IS_FCD 0x20000
/**
* Option bit for unorm_compare:
* Perform case-insensitive comparison.
* @stable ICU 2.2
*/
#define U_COMPARE_IGNORE_CASE 0x10000
#ifndef U_COMPARE_CODE_POINT_ORDER
/* see also unistr.h and ustring.h */
/**
* Option bit for u_strCaseCompare, u_strcasecmp, unorm_compare, etc:
* Compare strings in code point order instead of code unit order.
* @stable ICU 2.2
*/
#define U_COMPARE_CODE_POINT_ORDER 0x8000
#endif
/**
* Compares two strings for canonical equivalence.
* Further options include case-insensitive comparison and

1
icu4c/source/common/unicode/urename.h
View file

@ -1654,6 +1654,7 @@
#define ustr_hashICharsN U_ICU_ENTRY_POINT_RENAME(ustr_hashICharsN)
#define ustr_hashUCharsN U_ICU_ENTRY_POINT_RENAME(ustr_hashUCharsN)
#define ustrcase_getCaseLocale U_ICU_ENTRY_POINT_RENAME(ustrcase_getCaseLocale)
#define ustrcase_getTitleBreakIterator U_ICU_ENTRY_POINT_RENAME(ustrcase_getTitleBreakIterator)
#define ustrcase_internalFold U_ICU_ENTRY_POINT_RENAME(ustrcase_internalFold)
#define ustrcase_internalToLower U_ICU_ENTRY_POINT_RENAME(ustrcase_internalToLower)
#define ustrcase_internalToTitle U_ICU_ENTRY_POINT_RENAME(ustrcase_internalToTitle)

10
icu4c/source/common/unicode/ustring.h
View file

@ -497,16 +497,6 @@ u_strCompare(const UChar *s1, int32_t length1,
U_STABLE int32_t U_EXPORT2
u_strCompareIter(UCharIterator *iter1, UCharIterator *iter2, UBool codePointOrder);
#ifndef U_COMPARE_CODE_POINT_ORDER
/* see also unistr.h and unorm.h */
/**
* Option bit for u_strCaseCompare, u_strcasecmp, unorm_compare, etc:
* Compare strings in code point order instead of code unit order.
* @stable ICU 2.2
*/
#define U_COMPARE_CODE_POINT_ORDER 0x8000
#endif
/**
* Compare two strings case-insensitively using full case folding.
* This is equivalent to

2
icu4c/source/common/unicode/utext.h
View file

@ -768,7 +768,7 @@ utext_extract(UText *ut,
*/
#define UTEXT_SETNATIVEINDEX(ut, ix) \
{ int64_t __offset = (ix) - (ut)->chunkNativeStart; \
if (__offset>=0 && __offset<=(int64_t)(ut)->nativeIndexingLimit) { \
if (__offset>=0 && __offset<(int64_t)(ut)->nativeIndexingLimit && (ut)->chunkContents[__offset]<0xdc00) { \
(ut)->chunkOffset=(int32_t)__offset; \
} else { \
utext_setNativeIndex((ut), (ix)); } }

14
icu4c/source/common/unicode/uvernum.h
View file

@ -58,19 +58,19 @@
* This value will change in the subsequent releases of ICU
* @stable ICU 2.4
*/
#define U_ICU_VERSION_MAJOR_NUM 59
#define U_ICU_VERSION_MAJOR_NUM 60
/** The current ICU minor version as an integer.
* This value will change in the subsequent releases of ICU
* @stable ICU 2.6
*/
#define U_ICU_VERSION_MINOR_NUM 1
#define U_ICU_VERSION_MINOR_NUM 0
/** The current ICU patchlevel version as an integer.
* This value will change in the subsequent releases of ICU
* @stable ICU 2.4
*/
#define U_ICU_VERSION_PATCHLEVEL_NUM 0
#define U_ICU_VERSION_PATCHLEVEL_NUM 1
/** The current ICU build level version as an integer.
* This value is for use by ICU clients. It defaults to 0.
@ -84,7 +84,7 @@
* This value will change in the subsequent releases of ICU
* @stable ICU 2.6
*/
#define U_ICU_VERSION_SUFFIX _59
#define U_ICU_VERSION_SUFFIX _60
/**
* \def U_DEF2_ICU_ENTRY_POINT_RENAME
@ -119,14 +119,14 @@
* This value will change in the subsequent releases of ICU
* @stable ICU 2.4
*/
#define U_ICU_VERSION "59.1"
#define U_ICU_VERSION "60.0.1"
/** The current ICU library major/minor version as a string without dots, for library name suffixes.
* This value will change in the subsequent releases of ICU
* @stable ICU 2.6
*/
#if U_PLATFORM_HAS_WINUWP_API == 0
#define U_ICU_VERSION_SHORT "59"
#define U_ICU_VERSION_SHORT "60"
#else
// U_DISABLE_RENAMING does not impact dat file name
#define U_ICU_VERSION_SHORT
@ -136,7 +136,7 @@
/** Data version in ICU4C.
* @internal ICU 4.4 Internal Use Only
**/
#define U_ICU_DATA_VERSION "59.1"
#define U_ICU_DATA_VERSION "60.0.1"
#endif /* U_HIDE_INTERNAL_API */
/*===========================================================================

2
icu4c/source/common/unifiedcache.h
View file

@ -107,7 +107,7 @@ class CacheKey : public CacheKeyBase {
*/
virtual int32_t hashCode() const {
const char *s = typeid(T).name();
return ustr_hashCharsN(s, uprv_strlen(s));
return ustr_hashCharsN(s, static_cast<int32_t>(uprv_strlen(s)));
}
/**

21
icu4c/source/common/uniset_props.cpp
View file

@ -987,7 +987,7 @@ UnicodeSet::applyPropertyAlias(const UnicodeString& prop,
UProperty p;
int32_t v;
UBool mustNotBeEmpty = FALSE, invert = FALSE;
UBool invert = FALSE;
if (value.length() > 0) {
p = u_getPropertyEnum(pname.data());
@ -1009,14 +1009,15 @@ UnicodeSet::applyPropertyAlias(const UnicodeString& prop,
p == UCHAR_LEAD_CANONICAL_COMBINING_CLASS) {
char* end;
double value = uprv_strtod(vname.data(), &end);
v = (int32_t) value;
if (v != value || v < 0 || *end != 0) {
// non-integral or negative value, or trailing junk
// Anything between 0 and 255 is valid even if unused.
// Cast double->int only after range check.
// We catch NaN here because comparing it with both 0 and 255 will be false
// (as are all comparisons with NaN).
if (*end != 0 || !(0 <= value && value <= 255) ||
(v = (int32_t)value) != value) {
// non-integral value or outside 0..255, or trailing junk
FAIL(ec);
}
// If the resultant set is empty then the numeric value
// was invalid.
mustNotBeEmpty = TRUE;
} else {
FAIL(ec);
}
@ -1115,12 +1116,6 @@ UnicodeSet::applyPropertyAlias(const UnicodeString& prop,
complement();
}
if (U_SUCCESS(ec) && (mustNotBeEmpty && isEmpty())) {
// mustNotBeEmpty is set to true if an empty set indicates
// invalid input.
ec = U_ILLEGAL_ARGUMENT_ERROR;
}
if (isBogus() && U_SUCCESS(ec)) {
// We likely ran out of memory. AHHH!
ec = U_MEMORY_ALLOCATION_ERROR;

2
icu4c/source/common/unistr_case.cpp
View file

@ -162,7 +162,7 @@ UnicodeString::caseMap(int32_t caseLocale, uint32_t options, UCASEMAP_BREAK_ITER
iter->setText(oldString);
}
#endif
stringCaseMapper(caseLocale, options | UCASEMAP_OMIT_UNCHANGED_TEXT, UCASEMAP_BREAK_ITERATOR
stringCaseMapper(caseLocale, options | U_OMIT_UNCHANGED_TEXT, UCASEMAP_BREAK_ITERATOR
replacementChars, UPRV_LENGTHOF(replacementChars),
oldArray, oldLength, &edits, errorCode);
if (U_SUCCESS(errorCode)) {

31
icu4c/source/common/unistr_titlecase_brkiter.cpp
View file

@ -30,31 +30,26 @@
U_NAMESPACE_BEGIN
UnicodeString &
UnicodeString::toTitle(BreakIterator *titleIter) {
return toTitle(titleIter, Locale::getDefault(), 0);
UnicodeString::toTitle(BreakIterator *iter) {
return toTitle(iter, Locale::getDefault(), 0);
}
UnicodeString &
UnicodeString::toTitle(BreakIterator *titleIter, const Locale &locale) {
return toTitle(titleIter, locale, 0);
UnicodeString::toTitle(BreakIterator *iter, const Locale &locale) {
return toTitle(iter, locale, 0);
}
UnicodeString &
UnicodeString::toTitle(BreakIterator *titleIter, const Locale &locale, uint32_t options) {
BreakIterator *bi=titleIter;
if(bi==NULL) {
UErrorCode errorCode=U_ZERO_ERROR;
bi=BreakIterator::createWordInstance(locale, errorCode);
if(U_FAILURE(errorCode)) {
setToBogus();
return *this;
UnicodeString::toTitle(BreakIterator *iter, const Locale &locale, uint32_t options) {
LocalPointer<BreakIterator> ownedIter;
UErrorCode errorCode = U_ZERO_ERROR;
iter = ustrcase_getTitleBreakIterator(&locale, "", options, iter, ownedIter, errorCode);
if (iter == nullptr) {
setToBogus();
return *this;
}
}
caseMap(ustrcase_getCaseLocale(locale.getBaseName()), options, bi, ustrcase_internalToTitle);
if(titleIter==NULL) {
delete bi;
}
return *this;
caseMap(ustrcase_getCaseLocale(locale.getBaseName()), options, iter, ustrcase_internalToTitle);
return *this;
}
U_NAMESPACE_END

1
icu4c/source/common/uresbund.cpp
View file

@ -1083,6 +1083,7 @@ static UResourceBundle *init_resb_result(const ResourceData *rdata, Resource r,
pathBuf = (char *)uprv_malloc((uprv_strlen(keyPath)+1)*sizeof(char));
if(pathBuf == NULL) {
*status = U_MEMORY_ALLOCATION_ERROR;
ures_close(mainRes);
return NULL;
}
}

5
icu4c/source/common/ustr_imp.h
View file

@ -25,6 +25,11 @@
*/
#define _STRNCMP_STYLE 0x1000
/**
* Internal option for string transformation functions to not first reset the Edits object.
*/
#define U_EDITS_NO_RESET 0x2000
/**
* Compare two strings in code point order or code unit order.
* Works in strcmp style (both lengths -1),

201
icu4c/source/common/ustr_titlecase_brkiter.cpp
View file

@ -23,46 +23,153 @@
#include "unicode/brkiter.h"
#include "unicode/casemap.h"
#include "unicode/chariter.h"
#include "unicode/localpointer.h"
#include "unicode/ubrk.h"
#include "unicode/ucasemap.h"
#include "unicode/utext.h"
#include "cmemory.h"
#include "uassert.h"
#include "ucase.h"
#include "ucasemap_imp.h"
U_NAMESPACE_USE
U_NAMESPACE_BEGIN
/* functions available in the common library (for unistr_case.cpp) */
/**
* Whole-string BreakIterator.
* Titlecasing only calls setText(), first(), and next().
* We implement the rest only to satisfy the abstract interface.
*/
class WholeStringBreakIterator : public BreakIterator {
public:
WholeStringBreakIterator() : BreakIterator(), length(0) {}
~WholeStringBreakIterator() override;
UBool operator==(const BreakIterator&) const override;
BreakIterator *clone() const override;
static UClassID U_EXPORT2 getStaticClassID();
UClassID getDynamicClassID() const override;
CharacterIterator &getText() const override;
UText *getUText(UText *fillIn, UErrorCode &errorCode) const override;
void setText(const UnicodeString &text) override;
void setText(UText *text, UErrorCode &errorCode) override;
void adoptText(CharacterIterator* it) override;
int32_t first() override;
int32_t last() override;
int32_t previous() override;
int32_t next() override;
int32_t current() const override;
int32_t following(int32_t offset) override;
int32_t preceding(int32_t offset) override;
UBool isBoundary(int32_t offset) override;
int32_t next(int32_t n) override;
BreakIterator *createBufferClone(void *stackBuffer, int32_t &BufferSize,
UErrorCode &errorCode) override;
BreakIterator &refreshInputText(UText *input, UErrorCode &errorCode) override;
/* public API functions */
private:
int32_t length;
};
U_CAPI int32_t U_EXPORT2
u_strToTitle(UChar *dest, int32_t destCapacity,
const UChar *src, int32_t srcLength,
UBreakIterator *titleIter,
const char *locale,
UErrorCode *pErrorCode) {
LocalPointer<BreakIterator> ownedIter;
BreakIterator *iter;
if(titleIter!=NULL) {
iter=reinterpret_cast<BreakIterator *>(titleIter);
} else {
iter=BreakIterator::createWordInstance(Locale(locale), *pErrorCode);
ownedIter.adoptInstead(iter);
UOBJECT_DEFINE_RTTI_IMPLEMENTATION(WholeStringBreakIterator)
WholeStringBreakIterator::~WholeStringBreakIterator() {}
UBool WholeStringBreakIterator::operator==(const BreakIterator&) const { return FALSE; }
BreakIterator *WholeStringBreakIterator::clone() const { return nullptr; }
CharacterIterator &WholeStringBreakIterator::getText() const {
U_ASSERT(FALSE); // really should not be called
// Returns a null reference.
// Otherwise we would have to define a dummy CharacterIterator,
// and either have it as a field and const_cast it to a non-const reference,
// or have it via a pointer and return a reference to that.
CharacterIterator *none = nullptr;
return *none;
}
UText *WholeStringBreakIterator::getUText(UText * /*fillIn*/, UErrorCode &errorCode) const {
if (U_SUCCESS(errorCode)) {
errorCode = U_UNSUPPORTED_ERROR;
}
if(U_FAILURE(*pErrorCode)) {
return 0;
}
UnicodeString s(srcLength<0, src, srcLength);
iter->setText(s);
return ustrcase_mapWithOverlap(
ustrcase_getCaseLocale(locale), 0, iter,
dest, destCapacity,
src, srcLength,
ustrcase_internalToTitle, *pErrorCode);
return nullptr;
}
U_NAMESPACE_BEGIN
void WholeStringBreakIterator::setText(const UnicodeString &text) {
length = text.length();
}
void WholeStringBreakIterator::setText(UText *text, UErrorCode &errorCode) {
if (U_SUCCESS(errorCode)) {
int64_t length64 = utext_nativeLength(text);
if (length64 <= INT32_MAX) {
length = (int32_t)length64;
} else {
errorCode = U_INDEX_OUTOFBOUNDS_ERROR;
}
}
}
void WholeStringBreakIterator::adoptText(CharacterIterator* it) {
U_ASSERT(FALSE); // should not be called
length = it->getLength();
delete it;
}
int32_t WholeStringBreakIterator::first() { return 0; }
int32_t WholeStringBreakIterator::last() { return length; }
int32_t WholeStringBreakIterator::previous() { return 0; }
int32_t WholeStringBreakIterator::next() { return length; }
int32_t WholeStringBreakIterator::current() const { return 0; }
int32_t WholeStringBreakIterator::following(int32_t /*offset*/) { return length; }
int32_t WholeStringBreakIterator::preceding(int32_t /*offset*/) { return 0; }
UBool WholeStringBreakIterator::isBoundary(int32_t /*offset*/) { return FALSE; }
int32_t WholeStringBreakIterator::next(int32_t /*n*/) { return length; }
BreakIterator *WholeStringBreakIterator::createBufferClone(
void * /*stackBuffer*/, int32_t & /*BufferSize*/, UErrorCode &errorCode) {
if (U_SUCCESS(errorCode)) {
errorCode = U_UNSUPPORTED_ERROR;
}
return nullptr;
}
BreakIterator &WholeStringBreakIterator::refreshInputText(
UText * /*input*/, UErrorCode &errorCode) {
if (U_SUCCESS(errorCode)) {
errorCode = U_UNSUPPORTED_ERROR;
}
return *this;
}
U_CFUNC
BreakIterator *ustrcase_getTitleBreakIterator(
const Locale *locale, const char *locID, uint32_t options, BreakIterator *iter,
LocalPointer<BreakIterator> &ownedIter, UErrorCode &errorCode) {
if (U_FAILURE(errorCode)) { return nullptr; }
options &= U_TITLECASE_ITERATOR_MASK;
if (options != 0 && iter != nullptr) {
errorCode = U_ILLEGAL_ARGUMENT_ERROR;
return nullptr;
}
if (iter == nullptr) {
switch (options) {
case 0:
iter = BreakIterator::createWordInstance(
locale != nullptr ? *locale : Locale(locID), errorCode);
break;
case U_TITLECASE_WHOLE_STRING:
iter = new WholeStringBreakIterator();
if (iter == nullptr) {
errorCode = U_MEMORY_ALLOCATION_ERROR;
}
break;
case U_TITLECASE_SENTENCES:
iter = BreakIterator::createSentenceInstance(
locale != nullptr ? *locale : Locale(locID), errorCode);
break;
default:
errorCode = U_ILLEGAL_ARGUMENT_ERROR;
break;
}
ownedIter.adoptInstead(iter);
}
return iter;
}
int32_t CaseMap::toTitle(
const char *locale, uint32_t options, BreakIterator *iter,
@ -70,11 +177,8 @@ int32_t CaseMap::toTitle(
UChar *dest, int32_t destCapacity, Edits *edits,
UErrorCode &errorCode) {
LocalPointer<BreakIterator> ownedIter;
iter = ustrcase_getTitleBreakIterator(nullptr, locale, options, iter, ownedIter, errorCode);
if(iter==NULL) {
iter=BreakIterator::createWordInstance(Locale(locale), errorCode);
ownedIter.adoptInstead(iter);
}
if(U_FAILURE(errorCode)) {
return 0;
}
UnicodeString s(srcLength<0, src, srcLength);
@ -88,6 +192,30 @@ int32_t CaseMap::toTitle(
U_NAMESPACE_END
U_NAMESPACE_USE
U_CAPI int32_t U_EXPORT2
u_strToTitle(UChar *dest, int32_t destCapacity,
const UChar *src, int32_t srcLength,
UBreakIterator *titleIter,
const char *locale,
UErrorCode *pErrorCode) {
LocalPointer<BreakIterator> ownedIter;
BreakIterator *iter = ustrcase_getTitleBreakIterator(
nullptr, locale, 0, reinterpret_cast<BreakIterator *>(titleIter),
ownedIter, *pErrorCode);
if (iter == nullptr) {
return 0;
}
UnicodeString s(srcLength<0, src, srcLength);
iter->setText(s);
return ustrcase_mapWithOverlap(
ustrcase_getCaseLocale(locale), 0, iter,
dest, destCapacity,
src, srcLength,
ustrcase_internalToTitle, *pErrorCode);
}
U_CAPI int32_t U_EXPORT2
ucasemap_toTitle(UCaseMap *csm,
UChar *dest, int32_t destCapacity,
@ -97,10 +225,13 @@ ucasemap_toTitle(UCaseMap *csm,
return 0;
}
if (csm->iter == NULL) {
csm->iter = BreakIterator::createWordInstance(Locale(csm->locale), *pErrorCode);
}
if (U_FAILURE(*pErrorCode)) {
return 0;
LocalPointer<BreakIterator> ownedIter;
BreakIterator *iter = ustrcase_getTitleBreakIterator(
nullptr, csm->locale, csm->options, nullptr, ownedIter, *pErrorCode);
if (iter == nullptr) {
return 0;
}
csm->iter = ownedIter.orphan();
}
UnicodeString s(srcLength<0, src, srcLength);
csm->iter->setText(s);

55
icu4c/source/common/ustrcase.cpp
View file

@ -72,7 +72,7 @@ appendResult(UChar *dest, int32_t destIndex, int32_t destCapacity,
/* (not) original code point */
if(edits!=NULL) {
edits->addUnchanged(cpLength);
if(options & UCASEMAP_OMIT_UNCHANGED_TEXT) {
if(options & U_OMIT_UNCHANGED_TEXT) {
return destIndex;
}
}
@ -149,7 +149,7 @@ appendUnchanged(UChar *dest, int32_t destIndex, int32_t destCapacity,
if(length>0) {
if(edits!=NULL) {
edits->addUnchanged(length);
if(options & UCASEMAP_OMIT_UNCHANGED_TEXT) {
if(options & U_OMIT_UNCHANGED_TEXT) {
return destIndex;
}
}
@ -237,7 +237,7 @@ ustrcase_internalToTitle(int32_t caseLocale, uint32_t options, BreakIterator *it
const UChar *src, int32_t srcLength,
icu::Edits *edits,
UErrorCode &errorCode) {
if(U_FAILURE(errorCode)) {
if (!ustrcase_checkTitleAdjustmentOptions(options, errorCode)) {
return 0;
}
@ -264,45 +264,38 @@ ustrcase_internalToTitle(int32_t caseLocale, uint32_t options, BreakIterator *it
}
/*
* Unicode 4 & 5 section 3.13 Default Case Operations:
*
* R3 toTitlecase(X): Find the word boundaries based on Unicode Standard Annex
* #29, "Text Boundaries." Between each pair of word boundaries, find the first
* cased character F. If F exists, map F to default_title(F); then map each
* subsequent character C to default_lower(C).
*
* In this implementation, segment [prev..index[ into 3 parts:
* a) uncased characters (copy as-is) [prev..titleStart[
* b) first case letter (titlecase) [titleStart..titleLimit[
* Segment [prev..index[ into 3 parts:
* a) skipped characters (copy as-is) [prev..titleStart[
* b) first letter (titlecase) [titleStart..titleLimit[
* c) subsequent characters (lowercase) [titleLimit..index[
*/
if(prev<index) {
/* find and copy uncased characters [prev..titleStart[ */
// Find and copy skipped characters [prev..titleStart[
int32_t titleStart=prev;
int32_t titleLimit=prev;
UChar32 c;
U16_NEXT(src, titleLimit, index, c);
if((options&U_TITLECASE_NO_BREAK_ADJUSTMENT)==0 && UCASE_NONE==ucase_getType(c)) {
/* Adjust the titlecasing index (titleStart) to the next cased character. */
for(;;) {
if ((options&U_TITLECASE_NO_BREAK_ADJUSTMENT)==0) {
// Adjust the titlecasing index to the next cased character,
// or to the next letter/number/symbol/private use.
// Stop with titleStart<titleLimit<=index
// if there is a character to be titlecased,
// or else stop with titleStart==titleLimit==index.
UBool toCased = (options&U_TITLECASE_ADJUST_TO_CASED) != 0;
while (toCased ? UCASE_NONE==ucase_getType(c) : !ustrcase_isLNS(c)) {
titleStart=titleLimit;
if(titleLimit==index) {
/*
* only uncased characters in [prev..index[
* stop with titleStart==titleLimit==index
*/
break;
}
U16_NEXT(src, titleLimit, index, c);
if(UCASE_NONE!=ucase_getType(c)) {
break; /* cased letter at [titleStart..titleLimit[ */
}
}
destIndex=appendUnchanged(dest, destIndex, destCapacity,
src+prev, titleStart-prev, options, edits);
if(destIndex<0) {
errorCode=U_INDEX_OUTOFBOUNDS_ERROR;
return 0;
if (prev < titleStart) {
destIndex=appendUnchanged(dest, destIndex, destCapacity,
src+prev, titleStart-prev, options, edits);
if(destIndex<0) {
errorCode=U_INDEX_OUTOFBOUNDS_ERROR;
return 0;
}
}
}
@ -965,7 +958,7 @@ int32_t toUpper(uint32_t options,
edits->addUnchanged(oldLength);
}
// Write unchanged text?
change = (options & UCASEMAP_OMIT_UNCHANGED_TEXT) == 0;
change = (options & U_OMIT_UNCHANGED_TEXT) == 0;
}
}
@ -1110,7 +1103,7 @@ ustrcase_map(int32_t caseLocale, uint32_t options, UCASEMAP_BREAK_ITERATOR_PARAM
return 0;
}
if(edits!=NULL) {
if (edits != nullptr && (options & U_EDITS_NO_RESET) == 0) {
edits->reset();
}
destLength=stringCaseMapper(caseLocale, options, UCASEMAP_BREAK_ITERATOR

2
icu4c/source/common/ustrtrns.cpp
View file

@ -1312,7 +1312,7 @@ u_strFromJavaModifiedUTF8WithSub(
u_terminateUChars(dest, destCapacity, reqLength, pErrorCode);
return dest;
}
srcLength = uprv_strlen((const char *)pSrc);
srcLength = static_cast<int32_t>(uprv_strlen((const char *)pSrc));
}
/* Faster loop without ongoing checking for pSrcLimit and pDestLimit. */

4
icu4c/source/common/utrie2.cpp
View file

@ -746,7 +746,7 @@ uint16_t BackwardUTrie2StringIterator::previous16() {
codePointLimit=codePointStart;
if(start>=codePointStart) {
codePoint=U_SENTINEL;
return 0;
return trie->errorValue;
}
uint16_t result;
UTRIE2_U16_PREV16(trie, start, codePointStart, codePoint, result);
@ -757,7 +757,7 @@ uint16_t ForwardUTrie2StringIterator::next16() {
codePointStart=codePointLimit;
if(codePointLimit==limit) {
codePoint=U_SENTINEL;
return 0;
return trie->errorValue;
}
uint16_t result;
UTRIE2_U16_NEXT16(trie, codePointLimit, limit, codePoint, result);

7
icu4c/source/common/utrie2.h
View file

@ -980,11 +980,4 @@ utrie2_internalU8PrevIndex(const UTrie2 *trie, UChar32 c,
U_CDECL_END
/**
* Work around MSVC 2003 optimization bugs.
*/
#if defined (U_HAVE_MSVC_2003_OR_EARLIER)
#pragma optimize("", off)
#endif
#endif

8
icu4c/source/common/uts46.cpp
View file

@ -1415,7 +1415,7 @@ uidna_labelToASCII_UTF8(const UIDNA *idna,
if(!checkArgs(label, length, dest, capacity, pInfo, pErrorCode)) {
return 0;
}
StringPiece src(label, length<0 ? uprv_strlen(label) : length);
StringPiece src(label, length<0 ? static_cast<int32_t>(uprv_strlen(label)) : length);
CheckedArrayByteSink sink(dest, capacity);
IDNAInfo info;
reinterpret_cast<const IDNA *>(idna)->labelToASCII_UTF8(src, sink, info, *pErrorCode);
@ -1431,7 +1431,7 @@ uidna_labelToUnicodeUTF8(const UIDNA *idna,
if(!checkArgs(label, length, dest, capacity, pInfo, pErrorCode)) {
return 0;
}
StringPiece src(label, length<0 ? uprv_strlen(label) : length);
StringPiece src(label, length<0 ? static_cast<int32_t>(uprv_strlen(label)) : length);
CheckedArrayByteSink sink(dest, capacity);
IDNAInfo info;
reinterpret_cast<const IDNA *>(idna)->labelToUnicodeUTF8(src, sink, info, *pErrorCode);
@ -1447,7 +1447,7 @@ uidna_nameToASCII_UTF8(const UIDNA *idna,
if(!checkArgs(name, length, dest, capacity, pInfo, pErrorCode)) {
return 0;
}
StringPiece src(name, length<0 ? uprv_strlen(name) : length);
StringPiece src(name, length<0 ? static_cast<int32_t>(uprv_strlen(name)) : length);
CheckedArrayByteSink sink(dest, capacity);
IDNAInfo info;
reinterpret_cast<const IDNA *>(idna)->nameToASCII_UTF8(src, sink, info, *pErrorCode);
@ -1463,7 +1463,7 @@ uidna_nameToUnicodeUTF8(const UIDNA *idna,
if(!checkArgs(name, length, dest, capacity, pInfo, pErrorCode)) {
return 0;
}
StringPiece src(name, length<0 ? uprv_strlen(name) : length);
StringPiece src(name, length<0 ? static_cast<int32_t>(uprv_strlen(name)) : length);
CheckedArrayByteSink sink(dest, capacity);
IDNAInfo info;
reinterpret_cast<const IDNA *>(idna)->nameToUnicodeUTF8(src, sink, info, *pErrorCode);

4
icu4c/source/config/mh-mingw
View file

@ -28,6 +28,10 @@ else
STATICCPPFLAGS = -DU_STATIC_IMPLEMENTATION
endif
## ICU requires a minimum target of Windows 7, and MinGW does not set this by default.
## https://msdn.microsoft.com/en-us/library/aa383745.aspx
CPPFLAGS += -DWINVER=0x0601 -D_WIN32_WINNT=0x0601
## Flags for position independent code
SHAREDLIBCFLAGS =
SHAREDLIBCXXFLAGS =

4
icu4c/source/config/mh-mingw64
View file

@ -28,6 +28,10 @@ else
STATICCPPFLAGS = -DU_STATIC_IMPLEMENTATION
endif
## ICU requires a minimum target of Windows 7, and MinGW does not set this by default.
## https://msdn.microsoft.com/en-us/library/aa383745.aspx
CPPFLAGS += -DWINVER=0x0601 -D_WIN32_WINNT=0x0601
## Flags for position independent code
SHAREDLIBCFLAGS =
SHAREDLIBCXXFLAGS =

32
icu4c/source/configure vendored
View file

@ -1,6 +1,6 @@
#! /bin/sh
# Guess values for system-dependent variables and create Makefiles.
# Generated by GNU Autoconf 2.69 for ICU 59.1.
# Generated by GNU Autoconf 2.69 for ICU 60.0.1.
#
# Report bugs to <http://icu-project.org/bugs>.
#
@ -582,8 +582,8 @@ MAKEFLAGS=
# Identity of this package.
PACKAGE_NAME='ICU'
PACKAGE_TARNAME='International Components for Unicode'
PACKAGE_VERSION='59.1'
PACKAGE_STRING='ICU 59.1'
PACKAGE_VERSION='60.0.1'
PACKAGE_STRING='ICU 60.0.1'
PACKAGE_BUGREPORT='http://icu-project.org/bugs'
PACKAGE_URL='http://icu-project.org'
@ -752,6 +752,7 @@ infodir
docdir
oldincludedir
includedir
runstatedir
localstatedir
sharedstatedir
sysconfdir
@ -855,6 +856,7 @@ datadir='${datarootdir}'
sysconfdir='${prefix}/etc'
sharedstatedir='${prefix}/com'
localstatedir='${prefix}/var'
runstatedir='${localstatedir}/run'
includedir='${prefix}/include'
oldincludedir='/usr/include'
docdir='${datarootdir}/doc/${PACKAGE_TARNAME}'
@ -1107,6 +1109,15 @@ do
| -silent | --silent | --silen | --sile | --sil)
silent=yes ;;
-runstatedir | --runstatedir | --runstatedi | --runstated \
| --runstate | --runstat | --runsta | --runst | --runs \
| --run | --ru | --r)
ac_prev=runstatedir ;;
-runstatedir=* | --runstatedir=* | --runstatedi=* | --runstated=* \
| --runstate=* | --runstat=* | --runsta=* | --runst=* | --runs=* \
| --run=* | --ru=* | --r=*)
runstatedir=$ac_optarg ;;
-sbindir | --sbindir | --sbindi | --sbind | --sbin | --sbi | --sb)
ac_prev=sbindir ;;
-sbindir=* | --sbindir=* | --sbindi=* | --sbind=* | --sbin=* \
@ -1244,7 +1255,7 @@ fi
for ac_var in exec_prefix prefix bindir sbindir libexecdir datarootdir \
datadir sysconfdir sharedstatedir localstatedir includedir \
oldincludedir docdir infodir htmldir dvidir pdfdir psdir \
libdir localedir mandir
libdir localedir mandir runstatedir
do
eval ac_val=\$$ac_var
# Remove trailing slashes.
@ -1357,7 +1368,7 @@ if test "$ac_init_help" = "long"; then
# Omit some internal or obsolete options to make the list less imposing.
# This message is too long to be a string in the A/UX 3.1 sh.
cat <<_ACEOF
\`configure' configures ICU 59.1 to adapt to many kinds of systems.
\`configure' configures ICU 60.0.1 to adapt to many kinds of systems.
Usage: $0 [OPTION]... [VAR=VALUE]...
@ -1397,6 +1408,7 @@ Fine tuning of the installation directories:
--sysconfdir=DIR read-only single-machine data [PREFIX/etc]
--sharedstatedir=DIR modifiable architecture-independent data [PREFIX/com]
--localstatedir=DIR modifiable single-machine data [PREFIX/var]
--runstatedir=DIR modifiable per-process data [LOCALSTATEDIR/run]
--libdir=DIR object code libraries [EPREFIX/lib]
--includedir=DIR C header files [PREFIX/include]
--oldincludedir=DIR C header files for non-gcc [/usr/include]
@ -1423,7 +1435,7 @@ fi
if test -n "$ac_init_help"; then
case $ac_init_help in
short | recursive ) echo "Configuration of ICU 59.1:";;
short | recursive ) echo "Configuration of ICU 60.0.1:";;
esac
cat <<\_ACEOF
@ -1559,7 +1571,7 @@ fi
test -n "$ac_init_help" && exit $ac_status
if $ac_init_version; then
cat <<\_ACEOF
ICU configure 59.1
ICU configure 60.0.1
generated by GNU Autoconf 2.69
Copyright (C) 2012 Free Software Foundation, Inc.
@ -2251,7 +2263,7 @@ cat >config.log <<_ACEOF
This file contains any messages produced by compilers while
running configure, to aid debugging if configure makes a mistake.
It was created by ICU $as_me 59.1, which was
It was created by ICU $as_me 60.0.1, which was
generated by GNU Autoconf 2.69. Invocation command line was
$ $0 $@
@ -8385,7 +8397,7 @@ cat >>$CONFIG_STATUS <<\_ACEOF || ac_write_fail=1
# report actual input values of CONFIG_FILES etc. instead of their
# values after options handling.
ac_log="
This file was extended by ICU $as_me 59.1, which was
This file was extended by ICU $as_me 60.0.1, which was
generated by GNU Autoconf 2.69. Invocation command line was
CONFIG_FILES = $CONFIG_FILES
@ -8439,7 +8451,7 @@ _ACEOF
cat >>$CONFIG_STATUS <<_ACEOF || ac_write_fail=1
ac_cs_config="`$as_echo "$ac_configure_args" | sed 's/^ //; s/[\\""\`\$]/\\\\&/g'`"
ac_cs_version="\\
ICU config.status 59.1
ICU config.status 60.0.1
configured by $0, generated by GNU Autoconf 2.69,
with options \\"\$ac_cs_config\\"

2
icu4c/source/data/brkitr/brkfiles.mk
View file

@ -1,6 +1,6 @@
# © 2016 and later: Unicode, Inc. and others.
# License & terms of use: http://www.unicode.org/copyright.html#License
BRK_RES_CLDR_VERSION = 31.0.1
BRK_RES_CLDR_VERSION = 32
# A list of txt's to build
# Note:
#

4
icu4c/source/data/brkitr/rules/line_loose_cj.txt
View file

@ -496,8 +496,8 @@ $CAN_CM $CM* $QU; # QU x .
[$LB20NonBreaks-$CM] $CM* $BB; # BB x .
[^$CB] $CM* $BB; #
# LB21a
[^$CM $CB]? $CM* ($HY | $BA | $BAX) $CM* $HL;
# LB21a Don't break after Hebrew + Hyphen.
([^$CB] $CM*)? ($HY | $BA | $BAX) $CM* $HL;
# LB21b (reverse)
$HL $CM* $SY;

2
icu4c/source/data/coll/af.txt
View file

@ -5,7 +5,7 @@ af{
collations{
standard{
Sequence{"&N<<<ʼn"}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/am.txt
View file

@ -5,7 +5,7 @@ am{
collations{
standard{
Sequence{"[reorder Ethi]"}
Version{"31.0.1"}
Version{"32"}
}
}
}

4
icu4c/source/data/coll/ar.txt
View file

@ -9,7 +9,7 @@ ar{
"&ت<<ة<<<ﺔ<<<ﺓ"
"&ي<<ى<<<ﯨ<<<ﯩ<<<ﻰ<<<ﻯ<<<ﲐ<<<ﱝ"
}
Version{"31.0.1"}
Version{"32"}
}
standard{
Sequence{
@ -397,7 +397,7 @@ ar{
"‎&ۓ‎=ﮰ‎=ﮱ"
"‎&ۀ‎=ﮤ‎=ﮥ"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/as.txt
View file

@ -11,7 +11,7 @@ as{
"&[before 1]ত<ৎ=ত্\u200D"
"&হ<ক্ষ"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

4
icu4c/source/data/coll/az.txt
View file

@ -9,7 +9,7 @@ az{
"[import az-u-co-standard]"
"[reorder others]"
}
Version{"31.0.1"}
Version{"32"}
}
standard{
Sequence{
@ -26,7 +26,7 @@ az{
"&H<x<<<X"
"&Z<w<<<W"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/be.txt
View file

@ -9,7 +9,7 @@ be{
"&Е<ё<<<Ё"
"&у<ў<<<Ў"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/bg.txt
View file

@ -5,7 +5,7 @@ bg{
collations{
standard{
Sequence{"[reorder Cyrl]"}
Version{"31.0.1"}
Version{"32"}
}
}
}

4
icu4c/source/data/coll/bn.txt
View file

@ -9,7 +9,7 @@ bn{
"[reorder Beng Deva Guru Gujr Orya Taml Telu Knda Mlym Sinh]"
"&ঔ<ং<ঃ<ঁ"
}
Version{"31.0.1"}
Version{"32"}
}
traditional{
Sequence{
@ -629,7 +629,7 @@ bn{
"&যৌ<<<য়ৌ"
"&য্<<<য়্"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

4
icu4c/source/data/coll/bs.txt
View file

@ -5,11 +5,11 @@ bs{
collations{
search{
Sequence{"[import hr-u-co-search]"}
Version{"31.0.1"}
Version{"32"}
}
standard{
Sequence{"[import hr]"}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/bs_Cyrl.txt
View file

@ -5,7 +5,7 @@ bs_Cyrl{
collations{
standard{
Sequence{"[import sr]"}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/ca.txt
View file

@ -8,7 +8,7 @@ ca{
"[import und-u-co-search]"
"&L<ŀ=l·<<<Ŀ=L·"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/chr.txt
View file

@ -5,7 +5,7 @@ chr{
collations{
standard{
Sequence{"[reorder Cher]"}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/colfiles.mk
View file

@ -1,6 +1,6 @@
# © 2016 and later: Unicode, Inc. and others.
# License & terms of use: http://www.unicode.org/copyright.html#License
COLLATION_CLDR_VERSION = 31.0.1
COLLATION_CLDR_VERSION = 32
# A list of txt's to build
# Note:
#

2
icu4c/source/data/coll/cs.txt
View file

@ -11,7 +11,7 @@ cs{
"&S<š<<<Š"
"&Z<ž<<<Ž"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/cy.txt
View file

@ -14,7 +14,7 @@ cy{
"&R<rh<<<Rh<<<RH"
"&T<th<<<Th<<<TH"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

13
icu4c/source/data/coll/da.txt
View file

@ -1,7 +1,7 @@
// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html#License
da{
Version{"2.1.19.14"}
Version{"2.1.34.8"}
collations{
search{
Sequence{
@ -9,18 +9,19 @@ da{
"[import da-u-co-standard]"
"[caseFirst off]"
}
Version{"31.0.1"}
Version{"32"}
}
standard{
Sequence{
"[caseFirst upper]"
"&D<<đ<<<Đ<<ð<<<Ð"
"&t<<<þ/h"
"&T<<<Þ/H"
"&th<<<þ"
"&TH<<<Þ"
"&Y<<ü<<<Ü<<ű<<<Ű"
"&[before 1]ǀ<æ<<<Æ<<ä<<<Ä<ø<<<Ø<<ö<<<Ö<<ő<<<Ő<<œ<<<Œ<å<<<Å<<<aa<<<Aa<<<AA"
"&[before 1]ǀ<æ<<<Æ<<ä<<<Ä<ø<<<Ø<<ö<<<Ö<<ő<<<Ő<å<<<Å<<<aa<<<Aa<<<AA"
"&oe<<œ<<<Œ"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

4
icu4c/source/data/coll/de.txt
View file

@ -9,14 +9,14 @@ de{
"&OE<<ö<<<Ö"
"&UE<<ü<<<Ü"
}
Version{"31.0.1"}
Version{"32"}
}
search{
Sequence{
"[import und-u-co-search]"
"[import de-u-co-phonebk]"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/de_AT.txt
View file

@ -10,7 +10,7 @@ de_AT{
"&u<ü<<<Ü"
"&ss<ß<<<ẞ"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/dsb.txt
View file

@ -14,7 +14,7 @@ dsb{
"&S<š<<<Š<ś<<<Ś"
"&Z<ž<<<Ž<ź<<<Ź"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/ee.txt
View file

@ -17,7 +17,7 @@ ee{
"&T<ts<<<Ts<<<TS"
"&V<ʋ<<<Ʋ"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/el.txt
View file

@ -8,7 +8,7 @@ el{
"[normalization on]"
"[reorder Grek]"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/en_US_POSIX.txt
View file

@ -8,7 +8,7 @@ en_US_POSIX{
"&A<*'\u0020'-'/'<*0-'@'<*ABCDEFGHIJKLMNOPQRSTUVWXYZ<*'['-'`'<*abcdefghijklmnopqrstuvwxyz"
"<*'{'-'\u007F'"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

2
icu4c/source/data/coll/eo.txt
View file

@ -12,7 +12,7 @@ eo{
"&S<ŝ<<<Ŝ"
"&U<ŭ<<<Ŭ"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

6
icu4c/source/data/coll/es.txt
View file

@ -8,11 +8,11 @@ es{
"[import und-u-co-search]"
"&N<ñ<<<Ñ"
}
Version{"31.0.1"}
Version{"32"}
}
standard{
Sequence{"&N<ñ<<<Ñ"}
Version{"31.0.1"}
Version{"32"}
}
traditional{
Sequence{
@ -20,7 +20,7 @@ es{
"&C<ch<<<Ch<<<CH"
"&l<ll<<<Ll<<<LL"
}
Version{"31.0.1"}
Version{"32"}
}
}
}

Some files were not shown because too many files have changed in this diff Show more