ICU-57132.0.1.tar.gz

[apple/icu.git] / icuSources / common / brkiter.cpp

diff --git a/icuSources/common/brkiter.cpp b/icuSources/common/brkiter.cpp
index 9e6ee4d545fb06e63bfbc5bba1122643103e89ca..065e6df9556d5fc84bab0a0c4b609424336bfe13 100644 (file)
--- a/icuSources/common/brkiter.cpp
+++ b/icuSources/common/brkiter.cpp
@@ -1,10 +1,10 @@
 /*
 *******************************************************************************
 /*
 *******************************************************************************

-* Copyright (C) 1997-2008, International Business Machines Corporation and    *
-* others. All Rights Reserved.                                                *
+* Copyright (C) 1997-2016, International Business Machines Corporation and
+* others. All Rights Reserved.

 *******************************************************************************
 *
 *******************************************************************************
 *

-* File TXTBDRY.CPP
+* File brkiter.cpp

 *
 * Modification History:
 *
 *
 * Modification History:
 *


@@ -27,6 +27,7 @@
 #include "unicode/udata.h"
 #include "unicode/ures.h"
 #include "unicode/ustring.h"
 #include "unicode/udata.h"
 #include "unicode/ures.h"
 #include "unicode/ustring.h"

+#include "unicode/filteredbrk.h"

 #include "ucln_cmn.h"
 #include "cstring.h"
 #include "umutex.h"
 #include "ucln_cmn.h"
 #include "cstring.h"
 #include "umutex.h"


@@ -35,6 +36,8 @@
 #include "uresimp.h"
 #include "uassert.h"
 #include "ubrkimpl.h"
 #include "uresimp.h"
 #include "uassert.h"
 #include "ubrkimpl.h"

+#include "charstr.h"
+#include "unicode/filteredbrk.h"

 
 // *****************************************************************************
 // class BreakIterator
 
 // *****************************************************************************
 // class BreakIterator


@@ -52,7 +55,7 @@ BreakIterator::buildInstance(const Locale& loc, const char *type, int32_t kind,
 {
     char fnbuff[256];
     char ext[4]={'\0'};
 {
     char fnbuff[256];
     char ext[4]={'\0'};

-    char actualLocale[ULOC_FULLNAME_CAPACITY];
+    CharString actualLocale;

     int32_t size;
     const UChar* brkfname = NULL;
     UResourceBundle brkRulesStack;
     int32_t size;
     const UChar* brkfname = NULL;
     UResourceBundle brkRulesStack;


@@ -68,13 +71,7 @@ BreakIterator::buildInstance(const Locale& loc, const char *type, int32_t kind,
     ures_initStackObject(brkName);
 
     // Get the locale
     ures_initStackObject(brkName);
 
     // Get the locale

-    UResourceBundle *b = ures_open(U_ICUDATA_BRKITR, loc.getName(), &status);
-    /* this is a hack for now. Should be fixed when the data is fetched from
-        brk_index.txt */
-    if(status==U_USING_DEFAULT_WARNING){
-        status=U_ZERO_ERROR;
-        ures_openFillIn(b, U_ICUDATA_BRKITR, "", &status);
-    }
+    UResourceBundle *b = ures_openNoDefault(U_ICUDATA_BRKITR, loc.getName(), &status);

 
     // Get the "boundaries" array.
     if (U_SUCCESS(status)) {
 
     // Get the "boundaries" array.
     if (U_SUCCESS(status)) {


@@ -93,14 +90,12 @@ BreakIterator::buildInstance(const Locale& loc, const char *type, int32_t kind,
 
         // Use the string if we found it
         if (U_SUCCESS(status) && brkfname) {
 
         // Use the string if we found it
         if (U_SUCCESS(status) && brkfname) {

-            uprv_strncpy(actualLocale,
-                ures_getLocale(brkName, &status),
-                sizeof(actualLocale)/sizeof(actualLocale[0]));
+            actualLocale.append(ures_getLocaleInternal(brkName, &status), -1, status);

 
             UChar* extStart=u_strchr(brkfname, 0x002e);
             int len = 0;
             if(extStart!=NULL){
 
             UChar* extStart=u_strchr(brkfname, 0x002e);
             int len = 0;
             if(extStart!=NULL){

-                len = extStart-brkfname;
+                len = (int)(extStart-brkfname);

                 u_UCharsToChars(extStart+1, ext, sizeof(ext)); // nul terminates the buff
                 u_UCharsToChars(brkfname, fnbuff, len);
             }
                 u_UCharsToChars(extStart+1, ext, sizeof(ext)); // nul terminates the buff
                 u_UCharsToChars(brkfname, fnbuff, len);
             }


@@ -123,7 +118,8 @@ BreakIterator::buildInstance(const Locale& loc, const char *type, int32_t kind,
     // If there is a result, set the valid locale and actual locale, and the kind
     if (U_SUCCESS(status) && result != NULL) {
         U_LOCALE_BASED(locBased, *(BreakIterator*)result);
     // If there is a result, set the valid locale and actual locale, and the kind
     if (U_SUCCESS(status) && result != NULL) {
         U_LOCALE_BASED(locBased, *(BreakIterator*)result);

-        locBased.setLocaleIDs(ures_getLocaleByType(b, ULOC_VALID_LOCALE, &status), actualLocale);
+        locBased.setLocaleIDs(ures_getLocaleByType(b, ULOC_VALID_LOCALE, &status), 
+                              actualLocale.data());

         result->setBreakType(kind);
     }
 
         result->setBreakType(kind);
     }
 


@@ -196,25 +192,6 @@ BreakIterator::getAvailableLocales(int32_t& count)
     return Locale::getAvailableLocales(count);
 }
 
     return Locale::getAvailableLocales(count);
 }
 

-// -------------------------------------
-// Gets the objectLocale display name in the default locale language.
-UnicodeString& U_EXPORT2
-BreakIterator::getDisplayName(const Locale& objectLocale,
-                             UnicodeString& name)
-{
-    return objectLocale.getDisplayName(name);
-}
-
-// -------------------------------------
-// Gets the objectLocale display name in the displayLocale language.
-UnicodeString& U_EXPORT2
-BreakIterator::getDisplayName(const Locale& objectLocale,
-                             const Locale& displayLocale,
-                             UnicodeString& name)
-{
-    return objectLocale.getDisplayName(displayLocale, name);
-}
-

 // ------------------------------------------
 //
 // Default constructor and destructor
 // ------------------------------------------
 //
 // Default constructor and destructor


@@ -222,8 +199,8 @@ BreakIterator::getDisplayName(const Locale& objectLocale,
 //-------------------------------------------
 
 BreakIterator::BreakIterator()
 //-------------------------------------------
 
 BreakIterator::BreakIterator()

+: fKeepAll(FALSE)

 {
 {

-    fBufferClone = FALSE;

     *validLocale = *actualLocale = 0;
 }
 
     *validLocale = *actualLocale = 0;
 }
 


@@ -241,12 +218,16 @@ BreakIterator::~BreakIterator()
 // -------------------------------------
 
 class ICUBreakIteratorFactory : public ICUResourceBundleFactory {
 // -------------------------------------
 
 class ICUBreakIteratorFactory : public ICUResourceBundleFactory {

+public:
+    virtual ~ICUBreakIteratorFactory();

 protected:
     virtual UObject* handleCreate(const Locale& loc, int32_t kind, const ICUService* /*service*/, UErrorCode& status) const {
         return BreakIterator::makeInstance(loc, kind, status);
     }
 };
 
 protected:
     virtual UObject* handleCreate(const Locale& loc, int32_t kind, const ICUService* /*service*/, UErrorCode& status) const {
         return BreakIterator::makeInstance(loc, kind, status);
     }
 };
 

+ICUBreakIteratorFactory::~ICUBreakIteratorFactory() {}
+

 // -------------------------------------
 
 class ICUBreakIteratorService : public ICULocaleService {
 // -------------------------------------
 
 class ICUBreakIteratorService : public ICULocaleService {


@@ -258,6 +239,8 @@ public:
         registerFactory(new ICUBreakIteratorFactory(), status);
     }
 
         registerFactory(new ICUBreakIteratorFactory(), status);
     }
 

+    virtual ~ICUBreakIteratorService();
+

     virtual UObject* cloneInstance(UObject* instance) const {
         return ((BreakIterator*)instance)->clone();
     }
     virtual UObject* cloneInstance(UObject* instance) const {
         return ((BreakIterator*)instance)->clone();
     }


@@ -275,13 +258,17 @@ public:
     }
 };
 
     }
 };
 

+ICUBreakIteratorService::~ICUBreakIteratorService() {}
+

 // -------------------------------------
 
 // -------------------------------------
 

+// defined in ucln_cmn.h

 U_NAMESPACE_END
 
 U_NAMESPACE_END
 

-// defined in ucln_cmn.h
+static icu::UInitOnce gInitOnce;
+static icu::ICULocaleService* gService = NULL;
+

 
 

-static U_NAMESPACE_QUALIFIER ICULocaleService* gService = NULL;

 
 /**
  * Release all static memory held by breakiterator.
 
 /**
  * Release all static memory held by breakiterator.


@@ -293,40 +280,33 @@ static UBool U_CALLCONV breakiterator_cleanup(void) {
         delete gService;
         gService = NULL;
     }
         delete gService;
         gService = NULL;
     }

+    gInitOnce.reset();

 #endif
     return TRUE;
 }
 U_CDECL_END
 U_NAMESPACE_BEGIN
 
 #endif
     return TRUE;
 }
 U_CDECL_END
 U_NAMESPACE_BEGIN
 

+static void U_CALLCONV 
+initService(void) {
+    gService = new ICUBreakIteratorService();
+    ucln_common_registerCleanup(UCLN_COMMON_BREAKITERATOR, breakiterator_cleanup);
+}
+

 static ICULocaleService*
 getService(void)
 {
 static ICULocaleService*
 getService(void)
 {

-    UBool needsInit;
-    UMTX_CHECK(NULL, (UBool)(gService == NULL), needsInit);
-
-    if (needsInit) {
-        ICULocaleService  *tService = new ICUBreakIteratorService();
-        umtx_lock(NULL);
-        if (gService == NULL) {
-            gService = tService;
-            tService = NULL;
-            ucln_common_registerCleanup(UCLN_COMMON_BREAKITERATOR, breakiterator_cleanup);
-        }
-        umtx_unlock(NULL);
-        delete tService;
-    }
+    umtx_initOnce(gInitOnce, &initService);

     return gService;
 }
 
     return gService;
 }
 

+

 // -------------------------------------
 
 static inline UBool
 hasService(void)
 {
 // -------------------------------------
 
 static inline UBool
 hasService(void)
 {

-    UBool retVal;
-    UMTX_CHECK(NULL, gService != NULL, retVal);
-    return retVal;
+    return !gInitOnce.isReset() && getService() != NULL;

 }
 
 // -------------------------------------
 }
 
 // -------------------------------------


@@ -378,7 +358,6 @@ BreakIterator::createInstance(const Locale& loc, int32_t kind, UErrorCode& statu
         return NULL;
     }
 
         return NULL;
     }
 

-    u_init(&status);

 #if !UCONFIG_NO_SERVICE
     if (hasService()) {
         Locale actualLoc("");
 #if !UCONFIG_NO_SERVICE
     if (hasService()) {
         Locale actualLoc("");


@@ -407,6 +386,7 @@ BreakIterator::createInstance(const Locale& loc, int32_t kind, UErrorCode& statu
 }
 
 // -------------------------------------
 }
 
 // -------------------------------------

+enum { kKeyValueLenMax = 32 };

 
 BreakIterator*
 BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
 
 BreakIterator*
 BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)


@@ -415,6 +395,7 @@ BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
     if (U_FAILURE(status)) {
         return NULL;
     }
     if (U_FAILURE(status)) {
         return NULL;
     }

+    char lbType[kKeyValueLenMax];

 
     BreakIterator *result = NULL;
     switch (kind) {
 
     BreakIterator *result = NULL;
     switch (kind) {


@@ -425,10 +406,38 @@ BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
         result = BreakIterator::buildInstance(loc, "word", kind, status);
         break;
     case UBRK_LINE:
         result = BreakIterator::buildInstance(loc, "word", kind, status);
         break;
     case UBRK_LINE:

-        result = BreakIterator::buildInstance(loc, "line", kind, status);
+        uprv_strcpy(lbType, "line");
+        {
+            char lbKeyValue[kKeyValueLenMax] = {0};
+            UErrorCode kvStatus = U_ZERO_ERROR;
+            int32_t kLen = loc.getKeywordValue("lb", lbKeyValue, kKeyValueLenMax, kvStatus);
+            if (U_SUCCESS(kvStatus) && kLen > 0 && (uprv_strcmp(lbKeyValue,"strict")==0 || uprv_strcmp(lbKeyValue,"normal")==0 || uprv_strcmp(lbKeyValue,"loose")==0)) {
+                uprv_strcat(lbType, "_");
+                uprv_strcat(lbType, lbKeyValue);
+            }
+        }
+        result = BreakIterator::buildInstance(loc, lbType, kind, status);
+        if (U_SUCCESS(status) && result != NULL) {
+            char lwKeyValue[kKeyValueLenMax] = {0};
+            UErrorCode kvStatus = U_ZERO_ERROR;
+            int32_t kLen = loc.getKeywordValue("lw", lwKeyValue, kKeyValueLenMax, kvStatus);
+            result->setKeepAll(U_SUCCESS(kvStatus) && kLen > 0 && uprv_strcmp(lwKeyValue,"keepall")==0);
+        }

         break;
     case UBRK_SENTENCE:
         result = BreakIterator::buildInstance(loc, "sentence", kind, status);
         break;
     case UBRK_SENTENCE:
         result = BreakIterator::buildInstance(loc, "sentence", kind, status);

+        {
+            char ssKeyValue[kKeyValueLenMax] = {0};
+            UErrorCode kvStatus = U_ZERO_ERROR;
+            int32_t kLen = loc.getKeywordValue("ss", ssKeyValue, kKeyValueLenMax, kvStatus);
+            if (U_SUCCESS(kvStatus) && kLen > 0 && uprv_strcmp(ssKeyValue,"standard")==0) {
+                FilteredBreakIteratorBuilder* fbiBuilder = FilteredBreakIteratorBuilder::createInstance(loc, kvStatus);
+                if (U_SUCCESS(kvStatus)) {
+                    result = fbiBuilder->build(result, status);
+                    delete fbiBuilder;
+                }
+            }
+        }

         break;
     case UBRK_TITLE:
         result = BreakIterator::buildInstance(loc, "title", kind, status);
         break;
     case UBRK_TITLE:
         result = BreakIterator::buildInstance(loc, "title", kind, status);


@@ -456,6 +465,34 @@ BreakIterator::getLocaleID(ULocDataLocaleType type, UErrorCode& status) const {
     return locBased.getLocaleID(type, status);
 }
 
     return locBased.getLocaleID(type, status);
 }
 

+
+// This implementation of getRuleStatus is a do-nothing stub, here to
+// provide a default implementation for any derived BreakIterator classes that
+// do not implement it themselves.
+int32_t BreakIterator::getRuleStatus() const {
+    return 0;
+}
+
+// This implementation of getRuleStatusVec is a do-nothing stub, here to
+// provide a default implementation for any derived BreakIterator classes that
+// do not implement it themselves.
+int32_t BreakIterator::getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status) {
+    if (U_FAILURE(status)) {
+        return 0;
+    }
+    if (capacity < 1) {
+        status = U_BUFFER_OVERFLOW_ERROR;
+        return 1;
+    }
+    *fillInVec = 0;
+    return 1;
+}
+
+BreakIterator::BreakIterator (const Locale& valid, const Locale& actual) {
+  U_LOCALE_BASED(locBased, (*this));
+  locBased.setLocaleIDs(valid, actual);
+}
+

 U_NAMESPACE_END
 
 #endif /* #if !UCONFIG_NO_BREAK_ITERATION */
 U_NAMESPACE_END
 
 #endif /* #if !UCONFIG_NO_BREAK_ITERATION */