summaryrefslogtreecommitdiff
path: root/thirdparty/icu4c/common/locid.cpp
diff options
context:
space:
mode:
Diffstat (limited to 'thirdparty/icu4c/common/locid.cpp')
-rw-r--r--thirdparty/icu4c/common/locid.cpp106
1 files changed, 41 insertions, 65 deletions
diff --git a/thirdparty/icu4c/common/locid.cpp b/thirdparty/icu4c/common/locid.cpp
index 0d506293a9..e8859c7048 100644
--- a/thirdparty/icu4c/common/locid.cpp
+++ b/thirdparty/icu4c/common/locid.cpp
@@ -297,13 +297,12 @@ Locale::Locale( const char * newLanguage,
else
{
UErrorCode status = U_ZERO_ERROR;
- int32_t size = 0;
int32_t lsize = 0;
int32_t csize = 0;
int32_t vsize = 0;
int32_t ksize = 0;
- // Calculate the size of the resulting string.
+ // Check the sizes of the input strings.
// Language
if ( newLanguage != NULL )
@@ -313,7 +312,6 @@ Locale::Locale( const char * newLanguage,
setToBogus();
return;
}
- size = lsize;
}
CharString togo(newLanguage, lsize, status); // start with newLanguage
@@ -326,7 +324,6 @@ Locale::Locale( const char * newLanguage,
setToBogus();
return;
}
- size += csize;
}
// _Variant
@@ -350,21 +347,6 @@ Locale::Locale( const char * newLanguage,
}
}
- if( vsize > 0 )
- {
- size += vsize;
- }
-
- // Separator rules:
- if ( vsize > 0 )
- {
- size += 2; // at least: __v
- }
- else if ( csize > 0 )
- {
- size += 1; // at least: _v
- }
-
if ( newKeywords != NULL)
{
ksize = (int32_t)uprv_strlen(newKeywords);
@@ -372,11 +354,9 @@ Locale::Locale( const char * newLanguage,
setToBogus();
return;
}
- size += ksize + 1;
}
- // NOW we have the full locale string..
- // Now, copy it back.
+ // We've checked the input sizes, now build up the full locale string..
// newLanguage is already copied
@@ -469,14 +449,18 @@ Locale& Locale::operator=(Locale&& other) U_NOEXCEPT {
if ((baseName != fullName) && (baseName != fullNameBuffer)) uprv_free(baseName);
if (fullName != fullNameBuffer) uprv_free(fullName);
- if (other.fullName == other.fullNameBuffer) {
+ if (other.fullName == other.fullNameBuffer || other.baseName == other.fullNameBuffer) {
uprv_strcpy(fullNameBuffer, other.fullNameBuffer);
+ }
+ if (other.fullName == other.fullNameBuffer) {
fullName = fullNameBuffer;
} else {
fullName = other.fullName;
}
- if (other.baseName == other.fullName) {
+ if (other.baseName == other.fullNameBuffer) {
+ baseName = fullNameBuffer;
+ } else if (other.baseName == other.fullName) {
baseName = fullName;
} else {
baseName = other.baseName;
@@ -499,7 +483,7 @@ Locale::clone() const {
return new Locale(*this);
}
-UBool
+bool
Locale::operator==( const Locale& other) const
{
return (uprv_strcmp(other.fullName, fullName) == 0);
@@ -768,7 +752,7 @@ AliasDataBuilder::readLanguageAlias(
alias, strings, types, replacementIndexes, length,
#if U_DEBUG
[](const char* type) {
- // Assert the aliasFrom only contains the following possibilties
+ // Assert the aliasFrom only contains the following possibilities
// language_REGION_variant
// language_REGION
// language_variant
@@ -1227,7 +1211,7 @@ AliasReplacer::parseLanguageReplacement(
status = U_MEMORY_ALLOCATION_ERROR;
return;
}
- toBeFreed.addElement(str, status);
+ toBeFreed.addElementX(str, status);
char* data = str->data();
replacedLanguage = (const char*) data;
char* endOfField = uprv_strchr(data, '_');
@@ -1367,7 +1351,7 @@ AliasReplacer::replaceLanguage(
}
if (replacedExtensions != nullptr) {
// DO NOTHING
- // UTS35 does not specifiy what should we do if we have extensions in the
+ // UTS35 does not specify what should we do if we have extensions in the
// replacement. Currently we know only the following 4 "BCP47 LegacyRules" have
// extensions in them languageAlias:
// i_default => en_x_i_default
@@ -1441,7 +1425,7 @@ AliasReplacer::replaceTerritory(UVector& toBeFreed, UErrorCode& status)
return false;
}
replacedRegion = item->data();
- toBeFreed.addElement(item.orphan(), status);
+ toBeFreed.addElementX(item.orphan(), status);
}
U_ASSERT(!same(region, replacedRegion));
region = replacedRegion;
@@ -1567,6 +1551,7 @@ AliasReplacer::replaceTransformedExtensions(
const char* tvalue = uprv_strchr(tkey, '-');
if (tvalue == nullptr) {
status = U_ILLEGAL_ARGUMENT_ERROR;
+ return false;
}
const char* nextTKey = ultag_getTKeyStart(tvalue);
if (nextTKey != nullptr) {
@@ -1578,13 +1563,8 @@ AliasReplacer::replaceTransformedExtensions(
}
tkey = nextTKey;
} while (tkey != nullptr);
- tfields.sort([](UElement e1, UElement e2) -> int8_t {
- // uprv_strcmp return int and in some platform, such as arm64-v8a,
- // it may return positive values > 127 which cause the casted value
- // of int8_t negative.
- int res = uprv_strcmp(
- (const char*)e1.pointer, (const char*)e2.pointer);
- return (res == 0) ? 0 : ((res > 0) ? 1 : -1);
+ tfields.sort([](UElement e1, UElement e2) -> int32_t {
+ return uprv_strcmp((const char*)e1.pointer, (const char*)e2.pointer);
}, status);
for (int32_t i = 0; i < tfields.size(); i++) {
if (output.length() > 0) {
@@ -1592,8 +1572,11 @@ AliasReplacer::replaceTransformedExtensions(
}
const char* tfield = (const char*) tfields.elementAt(i);
const char* tvalue = uprv_strchr(tfield, '-');
+ if (tvalue == nullptr) {
+ status = U_ILLEGAL_ARGUMENT_ERROR;
+ return false;
+ }
// Split the "tkey-tvalue" pair string so that we can canonicalize the tvalue.
- U_ASSERT(tvalue != nullptr);
*((char*)tvalue++) = '\0'; // NULL terminate tkey
output.append(tfield, status).append('-', status);
const char* bcpTValue = ulocimp_toBcpType(tfield, tvalue, nullptr, nullptr);
@@ -1623,13 +1606,8 @@ AliasReplacer::outputToString(
if (!notEmpty(script) && !notEmpty(region)) {
out.append(SEP_CHAR, status);
}
- variants.sort([](UElement e1, UElement e2) -> int8_t {
- // uprv_strcmp return int and in some platform, such as arm64-v8a,
- // it may return positive values > 127 which cause the casted value
- // of int8_t negative.
- int res = uprv_strcmp(
- (const char*)e1.pointer, (const char*)e2.pointer);
- return (res == 0) ? 0 : ((res > 0) ? 1 : -1);
+ variants.sort([](UElement e1, UElement e2) -> int32_t {
+ return uprv_strcmp((const char*)e1.pointer, (const char*)e2.pointer);
}, status);
int32_t variantsStart = out.length();
for (int32_t i = 0; i < variants.size(); i++) {
@@ -1681,21 +1659,16 @@ AliasReplacer::replace(const Locale& locale, CharString& out, UErrorCode& status
while ((end = uprv_strchr(start, SEP_CHAR)) != nullptr &&
U_SUCCESS(status)) {
*end = NULL_CHAR; // null terminate inside variantsBuff
- variants.addElement(start, status);
+ variants.addElementX(start, status);
start = end + 1;
}
- variants.addElement(start, status);
+ variants.addElementX(start, status);
}
if (U_FAILURE(status)) { return false; }
// Sort the variants
- variants.sort([](UElement e1, UElement e2) -> int8_t {
- // uprv_strcmp return int and in some platform, such as arm64-v8a,
- // it may return positive values > 127 which cause the casted value
- // of int8_t negative.
- int res = uprv_strcmp(
- (const char*)e1.pointer, (const char*)e2.pointer);
- return (res == 0) ? 0 : ((res > 0) ? 1 : -1);
+ variants.sort([](UElement e1, UElement e2) -> int32_t {
+ return uprv_strcmp((const char*)e1.pointer, (const char*)e2.pointer);
}, status);
// A changed count to assert when loop too many times.
@@ -1737,7 +1710,7 @@ AliasReplacer::replace(const Locale& locale, CharString& out, UErrorCode& status
} // while(1)
if (U_FAILURE(status)) { return false; }
- // Nothing changed and we know the order of the vaiants are not change
+ // Nothing changed and we know the order of the variants are not change
// because we have no variant or only one.
const char* extensionsStr = locale_getKeywordsStart(locale.getName());
if (changed == 0 && variants.size() <= 1 && extensionsStr == nullptr) {
@@ -2445,7 +2418,7 @@ private:
public:
static UClassID U_EXPORT2 getStaticClassID(void) { return (UClassID)&fgClassID; }
- virtual UClassID getDynamicClassID(void) const { return getStaticClassID(); }
+ virtual UClassID getDynamicClassID(void) const override { return getStaticClassID(); }
public:
KeywordEnumeration(const char *keys, int32_t keywordLen, int32_t currentIndex, UErrorCode &status)
: keywords((char *)&fgClassID), current((char *)&fgClassID), length(0) {
@@ -2469,13 +2442,13 @@ public:
virtual ~KeywordEnumeration();
- virtual StringEnumeration * clone() const
+ virtual StringEnumeration * clone() const override
{
UErrorCode status = U_ZERO_ERROR;
return new KeywordEnumeration(keywords, length, (int32_t)(current - keywords), status);
}
- virtual int32_t count(UErrorCode &/*status*/) const {
+ virtual int32_t count(UErrorCode &/*status*/) const override {
char *kw = keywords;
int32_t result = 0;
while(*kw) {
@@ -2485,7 +2458,7 @@ public:
return result;
}
- virtual const char* next(int32_t* resultLength, UErrorCode& status) {
+ virtual const char* next(int32_t* resultLength, UErrorCode& status) override {
const char* result;
int32_t len;
if(U_SUCCESS(status) && *current != 0) {
@@ -2504,13 +2477,13 @@ public:
return result;
}
- virtual const UnicodeString* snext(UErrorCode& status) {
+ virtual const UnicodeString* snext(UErrorCode& status) override {
int32_t resultLength = 0;
const char *s = next(&resultLength, status);
return setChars(s, resultLength, status);
}
- virtual void reset(UErrorCode& /*status*/) {
+ virtual void reset(UErrorCode& /*status*/) override {
current = keywords;
}
};
@@ -2528,18 +2501,18 @@ public:
using KeywordEnumeration::KeywordEnumeration;
virtual ~UnicodeKeywordEnumeration();
- virtual const char* next(int32_t* resultLength, UErrorCode& status) {
+ virtual const char* next(int32_t* resultLength, UErrorCode& status) override {
const char* legacy_key = KeywordEnumeration::next(nullptr, status);
- if (U_SUCCESS(status) && legacy_key != nullptr) {
+ while (U_SUCCESS(status) && legacy_key != nullptr) {
const char* key = uloc_toUnicodeLocaleKey(legacy_key);
- if (key == nullptr) {
- status = U_ILLEGAL_ARGUMENT_ERROR;
- } else {
+ if (key != nullptr) {
if (resultLength != nullptr) {
*resultLength = static_cast<int32_t>(uprv_strlen(key));
}
return key;
}
+ // Not a Unicode keyword, could be a t, x or other, continue to look at the next one.
+ legacy_key = KeywordEnumeration::next(nullptr, status);
}
if (resultLength != nullptr) *resultLength = 0;
return nullptr;
@@ -2696,6 +2669,9 @@ Locale::setKeywordValue(const char* keywordName, const char* keywordValue, UErro
if (fullName != fullNameBuffer) {
// if full Name is already on the heap, need to free it.
uprv_free(fullName);
+ if (baseName == fullName) {
+ baseName = newFullName; // baseName should not point to freed memory.
+ }
}
fullName = newFullName;
status = U_ZERO_ERROR;
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-02 09:41:03 +0000