Update to ICU 76104.3 (#80)

* Update to icu76 (original)

* do not build with darwin platform flag to avoid entering the path that  requires darwin internal SDK

* Fix Linux build

* Reapply renaming expansion like #63

* Fix windows build: Use PI defined in the source in case M_PI isn't defined.

* Add the same flag that we added to Package.swift to cmakefile

* Update to c++17

* Fix cmake flag

* Reapply WASI fix from PR #35

* fix WASI build

---------

Co-authored-by: Yuta Saito <[email protected]>

Author: itingliu

CMakeLists.txt

@@ -19,7 +19,7 @@ project(SwiftFoundationICU
 
 option(BUILD_SHARED_LIBS "build shared libraries" ON)
 
-set(CMAKE_CXX_STANDARD 14)
+set(CMAKE_CXX_STANDARD 17)
 set(CMAKE_CXX_STANDARD_REQUIRED ON)
 
 include(CheckLinkerFlag)
@@ -52,7 +52,9 @@ add_compile_definitions(
     $<$<COMPILE_LANGUAGE:C,CXX>:U_IO_IMPLEMENTATION>
     $<$<COMPILE_LANGUAGE:C,CXX>:ICU_DATA_DIR="/usr/share/icu">
     $<$<COMPILE_LANGUAGE:C,CXX>:USE_PACKAGE_DATA=1>
-    $<$<COMPILE_LANGUAGE:C,CXX>:APPLE_ICU_CHANGES=1>)
+    $<$<COMPILE_LANGUAGE:C,CXX>:APPLE_ICU_CHANGES=1>
+    $<$<COMPILE_LANGUAGE:C,CXX>:UCHAR_TYPE=char16_t>
+    $<$<COMPILE_LANGUAGE:C,CXX>:U_PLATFORM_IS_DARWIN_BASED=0>)
 
 # Linux specific settings
 if(CMAKE_SYSTEM_NAME STREQUAL Linux)

Package.swift

@@ -63,6 +63,8 @@ var buildSettings: [CXXSetting] = [
   .define("ICU_DATA_DIR", to: "\"/usr/share/icu/\""),
   .define("USE_PACKAGE_DATA", to: "1"),
   .define("APPLE_ICU_CHANGES", to: "1"),
+  .define("UCHAR_TYPE", to: "char16_t"),
+  .define("U_PLATFORM_IS_DARWIN_BASED", to: "0"),
 
   .headerSearchPath("common"),
   .headerSearchPath("io"),
@@ -101,5 +103,5 @@ let package = Package(
       linkerSettings: linkerSettings
     )
   ],
-  cxxLanguageStandard: .cxx14
+  cxxLanguageStandard: .cxx17
 )

icuSources/common/appendable.cpp

@@ -25,7 +25,7 @@ Appendable::~Appendable() {}
 UBool
 Appendable::appendCodePoint(UChar32 c) {
     if(c<=0xffff) {
-        return appendCodeUnit((char16_t)c);
+        return appendCodeUnit(static_cast<char16_t>(c));
     } else {
         return appendCodeUnit(U16_LEAD(c)) && appendCodeUnit(U16_TRAIL(c));
     }

icuSources/common/bmpset.cpp

@@ -75,7 +75,7 @@ static void set32x64Bits(uint32_t table[64], int32_t start, int32_t limit) {
     int32_t trail=start&0x3f;  // Named for UTF-8 2-byte trail byte with lower 6 bits.
 
     // Set one bit indicating an all-one block.
-    uint32_t bits=(uint32_t)1<<lead;
+    uint32_t bits = static_cast<uint32_t>(1) << lead;
     if((start+1)==limit) {  // Single-character shortcut.
         table[trail]|=bits;
         return;
@@ -100,9 +100,9 @@ static void set32x64Bits(uint32_t table[64], int32_t start, int32_t limit) {
             ++lead;
         }
         if(lead<limitLead) {
-            bits=~(((unsigned)1<<lead)-1);
+            bits = ~((static_cast<unsigned>(1) << lead) - 1);
             if(limitLead<0x20) {
-                bits&=((unsigned)1<<limitLead)-1;
+                bits &= (static_cast<unsigned>(1) << limitLead) - 1;
             }
             for(trail=0; trail<64; ++trail) {
                 table[trail]|=bits;
@@ -111,7 +111,7 @@ static void set32x64Bits(uint32_t table[64], int32_t start, int32_t limit) {
         // limit<=0x800. If limit==0x800 then limitLead=32 and limitTrail=0.
         // In that case, bits=1<<limitLead is undefined but the bits value
         // is not used because trail<limitTrail is already false.
-        bits=(uint32_t)1<<((limitLead == 0x20) ? (limitLead - 1) : limitLead);
+        bits = static_cast<uint32_t>(1) << ((limitLead == 0x20) ? (limitLead - 1) : limitLead);
         for(trail=0; trail<limitTrail; ++trail) {
             table[trail]|=bits;
         }
@@ -290,22 +290,22 @@ int32_t BMPSet::findCodePoint(UChar32 c, int32_t lo, int32_t hi) const {
 
 UBool
 BMPSet::contains(UChar32 c) const {
-    if((uint32_t)c<=0xff) {
-        return (UBool)latin1Contains[c];
-    } else if((uint32_t)c<=0x7ff) {
-        return (UBool)((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))!=0);
-    } else if((uint32_t)c<0xd800 || (c>=0xe000 && c<=0xffff)) {
+    if (static_cast<uint32_t>(c) <= 0xff) {
+        return latin1Contains[c];
+    } else if (static_cast<uint32_t>(c) <= 0x7ff) {
+        return (table7FF[c & 0x3f] & (static_cast<uint32_t>(1) << (c >> 6))) != 0;
+    } else if (static_cast<uint32_t>(c) < 0xd800 || (c >= 0xe000 && c <= 0xffff)) {
         int lead=c>>12;
         uint32_t twoBits=(bmpBlockBits[(c>>6)&0x3f]>>lead)&0x10001;
         if(twoBits<=1) {
             // All 64 code points with the same bits 15..6
             // are either in the set or not.
-            return (UBool)twoBits;
+            return twoBits;
         } else {
             // Look up the code point in its 4k block of code points.
             return containsSlow(c, list4kStarts[lead], list4kStarts[lead+1]);
         }
-    } else if((uint32_t)c<=0x10ffff) {
+    } else if (static_cast<uint32_t>(c) <= 0x10ffff) {
         // surrogate or supplementary code point
         return containsSlow(c, list4kStarts[0xd], list4kStarts[0x11]);
     } else {
@@ -332,7 +332,7 @@ BMPSet::span(const char16_t *s, const char16_t *limit, USetSpanCondition spanCon
                     break;
                 }
             } else if(c<=0x7ff) {
-                if((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))==0) {
+                if ((table7FF[c & 0x3f] & (static_cast<uint32_t>(1) << (c >> 6))) == 0) {
                     break;
                 }
             } else if(c<0xd800 || c>=0xe000) {
@@ -372,7 +372,7 @@ BMPSet::span(const char16_t *s, const char16_t *limit, USetSpanCondition spanCon
                     break;
                 }
             } else if(c<=0x7ff) {
-                if((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))!=0) {
+                if ((table7FF[c & 0x3f] & (static_cast<uint32_t>(1) << (c >> 6))) != 0) {
                     break;
                 }
             } else if(c<0xd800 || c>=0xe000) {
@@ -421,7 +421,7 @@ BMPSet::spanBack(const char16_t *s, const char16_t *limit, USetSpanCondition spa
                     break;
                 }
             } else if(c<=0x7ff) {
-                if((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))==0) {
+                if ((table7FF[c & 0x3f] & (static_cast<uint32_t>(1) << (c >> 6))) == 0) {
                     break;
                 }
             } else if(c<0xd800 || c>=0xe000) {
@@ -464,7 +464,7 @@ BMPSet::spanBack(const char16_t *s, const char16_t *limit, USetSpanCondition spa
                     break;
                 }
             } else if(c<=0x7ff) {
-                if((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))!=0) {
+                if ((table7FF[c & 0x3f] & (static_cast<uint32_t>(1) << (c >> 6))) != 0) {
                     break;
                 }
             } else if(c<0xd800 || c>=0xe000) {
@@ -527,7 +527,7 @@ BMPSet::spanUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanConditi
                 b=*s;
             } while(U8_IS_SINGLE(b));
         }
-        length=(int32_t)(limit-s);
+        length = static_cast<int32_t>(limit - s);
     }
 
     if(spanCondition!=USET_SPAN_NOT_CONTAINED) {
@@ -547,7 +547,7 @@ BMPSet::spanUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanConditi
      * the truncated sequence.
      */
     b=*(limit-1);
-    if((int8_t)b<0) {
+    if (static_cast<int8_t>(b) < 0) {
         // b>=0x80: lead or trail byte
         if(b<0xc0) {
             // single trail byte, check for preceding 3- or 4-byte lead byte
@@ -602,15 +602,15 @@ BMPSet::spanUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanConditi
         if(b>=0xe0) {
             if(b<0xf0) {
                 if( /* handle U+0000..U+FFFF inline */
-                    (t1=(uint8_t)(s[0]-0x80)) <= 0x3f &&
-                    (t2=(uint8_t)(s[1]-0x80)) <= 0x3f
+                    (t1 = static_cast<uint8_t>(s[0] - 0x80)) <= 0x3f &&
+                    (t2 = static_cast<uint8_t>(s[1] - 0x80)) <= 0x3f
                 ) {
                     b&=0xf;
                     uint32_t twoBits=(bmpBlockBits[t1]>>b)&0x10001;
                     if(twoBits<=1) {
                         // All 64 code points with this lead byte and middle trail byte
                         // are either in the set or not.
-                        if(twoBits!=(uint32_t)spanCondition) {
+                        if (twoBits != static_cast<uint32_t>(spanCondition)) {
                             return s-1;
                         }
                     } else {
@@ -624,12 +624,12 @@ BMPSet::spanUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanConditi
                     continue;
                 }
             } else if( /* handle U+10000..U+10FFFF inline */
-                (t1=(uint8_t)(s[0]-0x80)) <= 0x3f &&
-                (t2=(uint8_t)(s[1]-0x80)) <= 0x3f &&
-                (t3=(uint8_t)(s[2]-0x80)) <= 0x3f
+                (t1 = static_cast<uint8_t>(s[0] - 0x80)) <= 0x3f &&
+                (t2 = static_cast<uint8_t>(s[1] - 0x80)) <= 0x3f &&
+                (t3 = static_cast<uint8_t>(s[2] - 0x80)) <= 0x3f
             ) {
                 // Give an illegal sequence the same value as the result of contains(FFFD).
-                UChar32 c=((UChar32)(b-0xf0)<<18)|((UChar32)t1<<12)|(t2<<6)|t3;
+                UChar32 c = (static_cast<UChar32>(b - 0xf0) << 18) | (static_cast<UChar32>(t1) << 12) | (t2 << 6) | t3;
                 if( (   (0x10000<=c && c<=0x10ffff) ?
                             containsSlow(c, list4kStarts[0x10], list4kStarts[0x11]) :
                             containsFFFD
@@ -643,9 +643,9 @@ BMPSet::spanUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanConditi
         } else {
             if( /* handle U+0000..U+07FF inline */
                 b>=0xc0 &&
-                (t1=(uint8_t)(*s-0x80)) <= 0x3f
+                (t1 = static_cast<uint8_t>(*s - 0x80)) <= 0x3f
             ) {
-                if((USetSpanCondition)((table7FF[t1]&((uint32_t)1<<(b&0x1f)))!=0) != spanCondition) {
+                if (static_cast<USetSpanCondition>((table7FF[t1] & (static_cast<uint32_t>(1) << (b & 0x1f))) != 0) != spanCondition) {
                     return s-1;
                 }
                 ++s;
@@ -711,7 +711,7 @@ BMPSet::spanBackUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCon
         c=utf8_prevCharSafeBody(s, 0, &length, b, -3);
         // c is a valid code point, not ASCII, not a surrogate
         if(c<=0x7ff) {
-            if((USetSpanCondition)((table7FF[c&0x3f]&((uint32_t)1<<(c>>6)))!=0) != spanCondition) {
+            if (static_cast<USetSpanCondition>((table7FF[c & 0x3f] & (static_cast<uint32_t>(1) << (c >> 6))) != 0) != spanCondition) {
                 return prev+1;
             }
         } else if(c<=0xffff) {
@@ -720,7 +720,7 @@ BMPSet::spanBackUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCon
             if(twoBits<=1) {
                 // All 64 code points with the same bits 15..6
                 // are either in the set or not.
-                if(twoBits!=(uint32_t)spanCondition) {
+                if (twoBits != static_cast<uint32_t>(spanCondition)) {
                     return prev+1;
                 }
             } else {

icuSources/common/bmpset.h

@@ -156,7 +156,7 @@ class BMPSet : public UMemory {
 };
 
 inline UBool BMPSet::containsSlow(UChar32 c, int32_t lo, int32_t hi) const {
-    return (UBool)(findCodePoint(c, lo, hi) & 1);
+    return findCodePoint(c, lo, hi) & 1;
 }
 
 U_NAMESPACE_END

icuSources/common/brkeng.cpp

@@ -86,7 +86,7 @@ UnhandledEngine::findBreaks( UText *text,
     if (U_FAILURE(status)) return 0;
     utext_setNativeIndex(text, startPos);
     UChar32 c = utext_current32(text);
-    while((int32_t)utext_getNativeIndex(text) < endPos && fHandled->contains(c)) {
+    while (static_cast<int32_t>(utext_getNativeIndex(text)) < endPos && fHandled->contains(c)) {
         utext_next32(text);            // TODO:  recast loop to work with post-increment operations.
         c = utext_current32(text);
     }
@@ -114,13 +114,11 @@ UnhandledEngine::handleCharacter(UChar32 c) {
  */
 
 ICULanguageBreakFactory::ICULanguageBreakFactory(UErrorCode &/*status*/) {
-    fEngines = 0;
+    fEngines = nullptr;
 }
 
 ICULanguageBreakFactory::~ICULanguageBreakFactory() {
-    if (fEngines != 0) {
-        delete fEngines;
-    }
+    delete fEngines;
 }
 
 void ICULanguageBreakFactory::ensureEngines(UErrorCode& status) {
@@ -148,7 +146,7 @@ ICULanguageBreakFactory::getEngineFor(UChar32 c, const char* locale) {
     Mutex m(&gBreakEngineMutex);
     int32_t i = fEngines->size();
     while (--i >= 0) {
-        lbe = (const LanguageBreakEngine *)(fEngines->elementAt(i));
+        lbe = static_cast<const LanguageBreakEngine*>(fEngines->elementAt(i));
         if (lbe != nullptr && lbe->handles(c, locale)) {
             return lbe;
         }
@@ -261,7 +259,7 @@ ICULanguageBreakFactory::loadDictionaryMatcherFor(UScriptCode script) {
     CharString ext;
     const char16_t *extStart = u_memrchr(dictfname, 0x002e, dictnlength);  // last dot
     if (extStart != nullptr) {
-        int32_t len = (int32_t)(extStart - dictfname);
+        int32_t len = static_cast<int32_t>(extStart - dictfname);
         ext.appendInvariantChars(UnicodeString(false, extStart + 1, dictnlength - len - 1), status);
         dictnlength = len;
     }
@@ -271,18 +269,18 @@ ICULanguageBreakFactory::loadDictionaryMatcherFor(UScriptCode script) {
     UDataMemory *file = udata_open(U_ICUDATA_BRKITR, ext.data(), dictnbuf.data(), &status);
     if (U_SUCCESS(status)) {
         // build trie
-        const uint8_t *data = (const uint8_t *)udata_getMemory(file);
-        const int32_t *indexes = (const int32_t *)data;
+        const uint8_t* data = static_cast<const uint8_t*>(udata_getMemory(file));
+        const int32_t* indexes = reinterpret_cast<const int32_t*>(data);
         const int32_t offset = indexes[DictionaryData::IX_STRING_TRIE_OFFSET];
         const int32_t trieType = indexes[DictionaryData::IX_TRIE_TYPE] & DictionaryData::TRIE_TYPE_MASK;
         DictionaryMatcher *m = nullptr;
         if (trieType == DictionaryData::TRIE_TYPE_BYTES) {
             const int32_t transform = indexes[DictionaryData::IX_TRANSFORM];
-            const char *characters = (const char *)(data + offset);
+            const char* characters = reinterpret_cast<const char*>(data + offset);
             m = new BytesDictionaryMatcher(characters, transform, file);
         }
         else if (trieType == DictionaryData::TRIE_TYPE_UCHARS) {
-            const char16_t *characters = (const char16_t *)(data + offset);
+            const char16_t* characters = reinterpret_cast<const char16_t*>(data + offset);
             m = new UCharsDictionaryMatcher(characters, file);
         }
         if (m == nullptr) {
@@ -339,12 +337,12 @@ int32_t BreakEngineWrapper::findBreaks(
     //   extends towards the start or end of the text, depending on 'reverse'.
 
     utext_setNativeIndex(text, startPos);
-    int32_t start = (int32_t)utext_getNativeIndex(text);
+    int32_t start = static_cast<int32_t>(utext_getNativeIndex(text));
     int32_t current;
     int32_t rangeStart;
     int32_t rangeEnd;
     UChar32 c = utext_current32(text);
-    while((current = (int32_t)utext_getNativeIndex(text)) < endPos && delegate->handles(c)) {
+    while ((current = static_cast<int32_t>(utext_getNativeIndex(text))) < endPos && delegate->handles(c)) {
         utext_next32(text);         // TODO:  recast loop for postincrement
         c = utext_current32(text);
     }

icuSources/common/brkiter.cpp

@@ -85,7 +85,7 @@ BreakIterator::buildInstance(const Locale& loc, const char *type, UErrorCode &st
         // Get the actual string
         brkfname = ures_getString(brkName, &size, &status);
         U_ASSERT((size_t)size<sizeof(fnbuff));
-        if ((size_t)size>=sizeof(fnbuff)) {
+        if (static_cast<size_t>(size) >= sizeof(fnbuff)) {
             size=0;
             if (U_SUCCESS(status)) {
                 status = U_BUFFER_OVERFLOW_ERROR;
@@ -99,7 +99,7 @@ BreakIterator::buildInstance(const Locale& loc, const char *type, UErrorCode &st
             char16_t* extStart=u_strchr(brkfname, 0x002e);
             int len = 0;
             if (extStart != nullptr){
-                len = (int)(extStart-brkfname);
+                len = static_cast<int>(extStart - brkfname);
                 u_UCharsToChars(extStart+1, ext, sizeof(ext)); // nul terminates the buff
                 u_UCharsToChars(brkfname, fnbuff, len);
             }
@@ -442,34 +442,19 @@ BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
             UTRACE_ENTRY(UTRACE_UBRK_CREATE_LINE);
             uprv_strcpy(lb_lw, "line");
             UErrorCode kvStatus = U_ZERO_ERROR;
-            CharString value;
-            CharStringByteSink valueSink(&value);
-            loc.getKeywordValue("lb", valueSink, kvStatus);
+            auto value = loc.getKeywordValue<CharString>("lb", kvStatus);
             if (U_SUCCESS(kvStatus) && (value == "strict" || value == "normal" || value == "loose")) {
                 uprv_strcat(lb_lw, "_");
                 uprv_strcat(lb_lw, value.data());
             }
-#if APPLE_ICU_CHANGES
-// rdar://36667210 Add ubrk_setLineWordOpts to programmatically set @lw options, add lw=keep-hangul support via keyword or function
-            value.clear();
-            kvStatus = U_ZERO_ERROR;
-            loc.getKeywordValue("lw", valueSink, kvStatus);
-            // lw=phrase is only supported in Japanese.
-            if (U_SUCCESS(kvStatus) && value == "phrase" && (uprv_strcmp(loc.getLanguage(), "ja") == 0 || uprv_strcmp(loc.getLanguage(), "ko") == 0)) {
-                uprv_strcat(lb_lw, "_");
-                uprv_strcat(lb_lw, value.data());
-            }
-#else
             // lw=phrase is only supported in Japanese.
             if (uprv_strcmp(loc.getLanguage(), "ja") == 0 || uprv_strcmp(loc.getLanguage(), "ko") == 0) {
-                value.clear();
-                loc.getKeywordValue("lw", valueSink, kvStatus);
+                value = loc.getKeywordValue<CharString>("lw", kvStatus);
                 if (U_SUCCESS(kvStatus) && value == "phrase") {
                     uprv_strcat(lb_lw, "_");
                     uprv_strcat(lb_lw, value.data());
                 }
             }
-#endif // APPLE_ICU_CHANGES
             result = BreakIterator::buildInstance(loc, lb_lw, status);
 #if APPLE_ICU_CHANGES
 // rdar://36667210 Add ubrk_setLineWordOpts to programmatically set @lw options, add lw=keep-hangul support via keyword or function
@@ -532,7 +517,7 @@ BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
 Locale
 BreakIterator::getLocale(ULocDataLocaleType type, UErrorCode& status) const {
     if (type == ULOC_REQUESTED_LOCALE) {
-        return Locale(requestLocale);
+        return {requestLocale};
     }
     U_LOCALE_BASED(locBased, *this);
     return locBased.getLocale(type, status);