ICU-62107.0.1.tar.gz

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

diff --git a/icuSources/common/brkiter.cpp b/icuSources/common/brkiter.cpp
index ff950636fbd5016338d287f8e73db07fe19f2741..f14c23a03d974ef19669ffdf638e26d0041d07f2 100644 (file)
--- a/icuSources/common/brkiter.cpp
+++ b/icuSources/common/brkiter.cpp
@@ -1,3 +1,5 @@
+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html

 /*
 *******************************************************************************
 * Copyright (C) 1997-2015, International Business Machines Corporation and
 /*
 *******************************************************************************
 * Copyright (C) 1997-2015, International Business Machines Corporation and


@@ -27,6 +29,7 @@
 #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"


@@ -36,7 +39,6 @@
 #include "uassert.h"
 #include "ubrkimpl.h"
 #include "charstr.h"
 #include "uassert.h"
 #include "ubrkimpl.h"
 #include "charstr.h"

-#include "unicode/filteredbrk.h"

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


@@ -50,7 +52,7 @@ U_NAMESPACE_BEGIN
 // -------------------------------------
 
 BreakIterator*
 // -------------------------------------
 
 BreakIterator*

-BreakIterator::buildInstance(const Locale& loc, const char *type, int32_t kind, UErrorCode &status)
+BreakIterator::buildInstance(const Locale& loc, const char *type, UErrorCode &status)

 {
     char fnbuff[256];
     char ext[4]={'\0'};
 {
     char fnbuff[256];
     char ext[4]={'\0'};


@@ -119,7 +121,6 @@ BreakIterator::buildInstance(const Locale& loc, const char *type, int32_t kind,
         U_LOCALE_BASED(locBased, *(BreakIterator*)result);
         locBased.setLocaleIDs(ures_getLocaleByType(b, ULOC_VALID_LOCALE, &status), 
                               actualLocale.data());
         U_LOCALE_BASED(locBased, *(BreakIterator*)result);
         locBased.setLocaleIDs(ures_getLocaleByType(b, ULOC_VALID_LOCALE, &status), 
                               actualLocale.data());

-        result->setBreakType(kind);

     }
 
     ures_close(b);
     }
 
     ures_close(b);


@@ -193,15 +194,29 @@ BreakIterator::getAvailableLocales(int32_t& count)
 
 // ------------------------------------------
 //
 
 // ------------------------------------------
 //

-// Default constructor and destructor
+// Constructors, destructor and assignment operator

 //
 //-------------------------------------------
 
 BreakIterator::BreakIterator()
 //
 //-------------------------------------------
 
 BreakIterator::BreakIterator()

+: fLineWordOpts(UBRK_LINEWORD_NORMAL)

 {
     *validLocale = *actualLocale = 0;
 }
 
 {
     *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()
 {
 }
 BreakIterator::~BreakIterator()
 {
 }


@@ -263,7 +278,7 @@ ICUBreakIteratorService::~ICUBreakIteratorService() {}
 // defined in ucln_cmn.h
 U_NAMESPACE_END
 
 // defined in ucln_cmn.h
 U_NAMESPACE_END
 

-static icu::UInitOnce gInitOnce;
+static icu::UInitOnce gInitOnceBrkiter;

 static icu::ICULocaleService* gService = NULL;
 
 
 static icu::ICULocaleService* gService = NULL;
 
 


@@ -278,7 +293,7 @@ static UBool U_CALLCONV breakiterator_cleanup(void) {
         delete gService;
         gService = NULL;
     }
         delete gService;
         gService = NULL;
     }

-    gInitOnce.reset();
+    gInitOnceBrkiter.reset();

 #endif
     return TRUE;
 }
 #endif
     return TRUE;
 }


@@ -294,7 +309,7 @@ initService(void) {
 static ICULocaleService*
 getService(void)
 {
 static ICULocaleService*
 getService(void)
 {

-    umtx_initOnce(gInitOnce, &initService);
+    umtx_initOnce(gInitOnceBrkiter, &initService);

     return gService;
 }
 
     return gService;
 }
 


@@ -304,7 +319,7 @@ getService(void)
 static inline UBool
 hasService(void)
 {
 static inline UBool
 hasService(void)
 {

-    return !gInitOnce.isReset() && getService() != NULL;
+    return !gInitOnceBrkiter.isReset() && getService() != NULL;

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


@@ -398,10 +413,10 @@ BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
     BreakIterator *result = NULL;
     switch (kind) {
     case UBRK_CHARACTER:
     BreakIterator *result = NULL;
     switch (kind) {
     case UBRK_CHARACTER:

-        result = BreakIterator::buildInstance(loc, "grapheme", kind, status);
+        result = BreakIterator::buildInstance(loc, "grapheme", status);

         break;
     case UBRK_WORD:
         break;
     case UBRK_WORD:

-        result = BreakIterator::buildInstance(loc, "word", kind, status);
+        result = BreakIterator::buildInstance(loc, "word", status);

         break;
     case UBRK_LINE:
         uprv_strcpy(lbType, "line");
         break;
     case UBRK_LINE:
         uprv_strcpy(lbType, "line");


@@ -414,10 +429,25 @@ BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
                 uprv_strcat(lbType, lbKeyValue);
             }
         }
                 uprv_strcat(lbType, lbKeyValue);
             }
         }

-        result = BreakIterator::buildInstance(loc, lbType, kind, status);
+        result = BreakIterator::buildInstance(loc, lbType, 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);
+            ULineWordOptions lineWordOpts = UBRK_LINEWORD_NORMAL;
+            if (U_SUCCESS(kvStatus) && kLen > 0) {
+                if (uprv_strcmp(lwKeyValue,"keepall")==0 || uprv_strcmp(lwKeyValue,"keep-all")==0) {
+                    lineWordOpts = UBRK_LINEWORD_KEEP_ALL;
+                } else if (uprv_strcmp(lwKeyValue,"keep-hangul")==0) {
+                    lineWordOpts = UBRK_LINEWORD_KEEP_HANGUL;
+                }
+            }
+            result->setLineWordOpts(lineWordOpts);
+        }

         break;
     case UBRK_SENTENCE:
         break;
     case UBRK_SENTENCE:

-        result = BreakIterator::buildInstance(loc, "sentence", kind, status);
+        result = BreakIterator::buildInstance(loc, "sentence", status);
+#if !UCONFIG_NO_FILTERED_BREAK_ITERATION

         {
             char ssKeyValue[kKeyValueLenMax] = {0};
             UErrorCode kvStatus = U_ZERO_ERROR;
         {
             char ssKeyValue[kKeyValueLenMax] = {0};
             UErrorCode kvStatus = U_ZERO_ERROR;


@@ -430,9 +460,10 @@ BreakIterator::makeInstance(const Locale& loc, int32_t kind, UErrorCode& status)
                 }
             }
         }
                 }
             }
         }

+#endif

         break;
     case UBRK_TITLE:
         break;
     case UBRK_TITLE:

-        result = BreakIterator::buildInstance(loc, "title", kind, status);
+        result = BreakIterator::buildInstance(loc, "title", status);

         break;
     default:
         status = U_ILLEGAL_ARGUMENT_ERROR;
         break;
     default:
         status = U_ILLEGAL_ARGUMENT_ERROR;