ICU-1501 Ported back the Upper/Lower/TitlecaseTransliterator changes from Java.

X-SVN-Rev: 7192

icu4c/source/i18n/titletrn.cpp

@@ -13,6 +13,9 @@
 #include "unicode/uniset.h"
 #include "mutex.h"
 #include "ucln_in.h"
+#include "unicode/ustring.h"
+#include "ustr_imp.h"
+#include "cpputils.h"
 
 U_NAMESPACE_BEGIN
 
@@ -33,8 +36,11 @@ static UnicodeSet* SKIP = NULL;
  */
 static UnicodeSet* CASED = NULL;
 
-TitlecaseTransliterator::TitlecaseTransliterator(UnicodeFilter* adoptedFilter) :
-    Transliterator(_ID, adoptedFilter) {
+TitlecaseTransliterator::TitlecaseTransliterator(const Locale& theLoc) :
+    Transliterator(_ID, 0),
+    loc(theLoc), 
+    buffer(0) {
+    buffer = new UChar[u_getMaxCaseExpansion()];
     // Need to look back 2 characters in the case of "can't"
     setMaximumContextLength(2);
 }
@@ -42,13 +48,20 @@ TitlecaseTransliterator::TitlecaseTransliterator(UnicodeFilter* adoptedFilter) :
 /**
  * Destructor.
  */
-TitlecaseTransliterator::~TitlecaseTransliterator() {}
+TitlecaseTransliterator::~TitlecaseTransliterator() {
+    delete [] buffer;
+}
 
 /**
  * Copy constructor.
  */
 TitlecaseTransliterator::TitlecaseTransliterator(const TitlecaseTransliterator& o) :
-    Transliterator(o) {}
+    Transliterator(o),
+    loc(o.loc),
+    buffer(0) {
+    buffer = new UChar[u_getMaxCaseExpansion()];    
+    uprv_arrayCopy(o.buffer, 0, this->buffer, 0, u_getMaxCaseExpansion());
+}
 
 /**
  * Assignment operator.
@@ -56,6 +69,8 @@ TitlecaseTransliterator::TitlecaseTransliterator(const TitlecaseTransliterator&
 TitlecaseTransliterator& TitlecaseTransliterator::operator=(
                              const TitlecaseTransliterator& o) {
     Transliterator::operator=(o);
+    loc = o.loc;
+    uprv_arrayCopy(o.buffer, 0, this->buffer, 0, u_getMaxCaseExpansion());
     return *this;
 }
 
@@ -103,22 +118,58 @@ void TitlecaseTransliterator::handleTransliterate(
     // Convert things after a CASED character toLower; things
     // after a non-CASED, non-SKIP character toTitle.  SKIP
     // characters are copied directly and do not change the mode.
-    UnicodeString str("A", "");
-    for (start=offsets.start; start<offsets.limit; ++start) {
-        UChar c = text.charAt(start);
-        if (SKIP->contains(c)) {
-            continue;
-        }
-        UChar d = (UChar) (doTitle ? u_totitle(c)
-                                   : u_tolower(c));
-        if (c != d) {
-            str.setCharAt(0, d);
-            text.handleReplaceBetween(start, start+1, str);
-        }
-        doTitle = !CASED->contains(c);
+    int32_t textPos = offsets.start;
+    if (textPos >= offsets.limit) return;
+
+    // get string for context
+    // TODO: add convenience method to do this, since we do it all over
+
+    int32_t loop = 0;
+    UnicodeString original;
+    /* UChar *original = new UChar[offsets.contextLimit - offsets.contextStart+1]; */// get whole context
+    /* Extract the characters from Replaceable */
+    for (loop = offsets.contextStart; loop < offsets.contextLimit; loop++) {
+        original.append(text.charAt(loop));
     }
-    
-    offsets.start = start;
+    // Walk through original string
+    // If there is a case change, modify corresponding position in replaceable
+
+    int32_t i = textPos - offsets.contextStart;
+    int32_t limit = offsets.limit - offsets.contextStart;
+    UChar32 cp;
+    int32_t oldLen;
+    int32_t newLen;
+
+    for (; i < limit; ) {
+        UErrorCode status = U_ZERO_ERROR;
+        int32_t s = i;
+        buffer[0] = original.charAt(s);
+
+        UTF_GET_CHAR(original.getBuffer(), 0, i, original.length(), cp);
+        oldLen = UTF_CHAR_LENGTH(cp);
+        i += oldLen;
+        if (!SKIP->contains(cp)) {
+            if (doTitle) {
+                newLen = u_internalTitleCase(cp, buffer, u_getMaxCaseExpansion(), loc.getName());
+            } else {
+                u_strToLower(buffer, u_getMaxCaseExpansion(), original.getBuffer()+s, i-s, loc.getName(), &status);
+                newLen = buffer[0] == original.charAt(s) ? -1 : u_strlen(buffer);
+            }
+            doTitle = !CASED->contains(cp);
+            if (newLen >= 0) {
+                UnicodeString temp(buffer, newLen);
+                text.handleReplaceBetween(textPos, textPos + oldLen, temp);
+                if (newLen != oldLen) {
+                    textPos += newLen;
+                    offsets.limit += newLen - oldLen;
+                    offsets.contextLimit += newLen - oldLen;
+                    continue;
+                }
+            }
+        }
+        textPos += oldLen;
+    }
+    offsets.start = offsets.limit;
 }
 
 /**

icu4c/source/i18n/titletrn.h

@@ -11,6 +11,7 @@
 #define TITLETRN_H
 
 #include "unicode/translit.h"
+#include "unicode/locid.h"
 
 U_NAMESPACE_BEGIN
 
@@ -33,7 +34,7 @@ class U_I18N_API TitlecaseTransliterator : public Transliterator {
     /**
      * Constructs a transliterator.
      */
-    TitlecaseTransliterator(UnicodeFilter* adoptedFilter = 0);
+    TitlecaseTransliterator(const Locale& loc = Locale::getDefault());
 
     /**
      * Destructor.
@@ -70,6 +71,9 @@ class U_I18N_API TitlecaseTransliterator : public Transliterator {
      * CALL.
      */
     static void cleanup();
+ private:
+    Locale loc;
+    UChar* buffer;
 };
 
 U_NAMESPACE_END

icu4c/source/i18n/tolowtrn.cpp

@@ -9,6 +9,9 @@
 */
 
 #include "tolowtrn.h"
+#include "unicode/ustring.h"
+#include "ustr_imp.h"
+#include "cpputils.h"
 #include "unicode/uchar.h"
 
 U_NAMESPACE_BEGIN
@@ -18,31 +21,35 @@ const char LowercaseTransliterator::_ID[] = "Any-Lower";
 /**
  * Constructs a transliterator.
  */
-LowercaseTransliterator::LowercaseTransliterator(const Locale& theLoc,
-                                                   UnicodeFilter* adoptedFilter) :
-    TransformTransliterator(_ID, adoptedFilter),
-    loc(theLoc) {
+LowercaseTransliterator::LowercaseTransliterator(const Locale& theLoc) : Transliterator(_ID, 0),
+    loc(theLoc) , buffer(0) {
+    buffer = new UChar[u_getMaxCaseExpansion()];
 }
 
 /**
  * Destructor.
  */
-LowercaseTransliterator::~LowercaseTransliterator() {}
+LowercaseTransliterator::~LowercaseTransliterator() {
+    delete [] buffer;
+}
 
 /**
  * Copy constructor.
  */
 LowercaseTransliterator::LowercaseTransliterator(const LowercaseTransliterator& o) :
-    TransformTransliterator(o),
-    loc(o.loc) {}
+    Transliterator(o),
+    loc(o.loc), buffer(0) {
+    buffer = new UChar[u_getMaxCaseExpansion()];
+}
 
 /**
  * Assignment operator.
  */
 LowercaseTransliterator& LowercaseTransliterator::operator=(
                              const LowercaseTransliterator& o) {
-    TransformTransliterator::operator=(o);
+    Transliterator::operator=(o);
     loc = o.loc;
+    uprv_arrayCopy((const UChar*)o.buffer, 0, this->buffer, 0, u_getMaxCaseExpansion());
     return *this;
 }
 
@@ -54,18 +61,58 @@ Transliterator* LowercaseTransliterator::clone(void) const {
 }
 
 /**
- * TransformTransliterator framework method.
+ * Implements {@link Transliterator#handleTransliterate}.
  */
-UBool LowercaseTransliterator::hasTransform(UChar32 c) const {
-    return c != u_tolower(c);
-}
+void LowercaseTransliterator::handleTransliterate(Replaceable& text,
+                                 UTransPosition& offsets, 
+                                 UBool isIncremental) const
+{
+    int32_t textPos = offsets.start;
+    int32_t loop;
+    if (textPos >= offsets.limit) return;
 
-/**
- * TransformTransliterator framework method.
- */
-void LowercaseTransliterator::transform(UnicodeString& s) const {
-    s.toLower(loc);
-}
+    // get string for context
+    // TODO: add convenience method to do this, since we do it all over
+    
+    UnicodeString original;
+    /*UChar *original = new UChar[offsets.contextLimit - offsets.contextStart+1];*/ // get whole context
+    /* Extract the characters from Replaceable */
+    for (loop = offsets.contextStart; loop < offsets.contextLimit; loop++) {
+        original.append(text.charAt(loop));
+    }
+    
+    // Walk through original string
+    // If there is a case change, modify corresponding position in replaceable
+    
+    int32_t i = textPos - offsets.contextStart;
+    int32_t limit = offsets.limit - offsets.contextStart;
+    UChar32 cp;
+    int32_t oldLen;
+    
+    for (; i < limit; ) { 
+        UErrorCode status = U_ZERO_ERROR;
+        int32_t s = i;
+        buffer[0] = original.charAt(s);
 
+        UTF_GET_CHAR(original.getBuffer(), 0, i, original.length(), cp);
+        oldLen = UTF_CHAR_LENGTH(cp);
+        i += oldLen;
+        u_strToLower(buffer, u_getMaxCaseExpansion(), original.getBuffer()+s, i-s, loc.getName(), &status);
+        /* Skip checking of status code here because the buffer should not have overflowed. */
+        if ( buffer[0] != original.charAt(s) ) {
+            int len = u_strlen(buffer);
+            UnicodeString temp(buffer);
+            text.handleReplaceBetween(textPos, textPos + oldLen, temp);
+            if (len != oldLen) {
+                textPos += len;
+                offsets.limit += len - oldLen;
+                offsets.contextLimit += len - oldLen;
+                continue;
+            }
+        }
+        textPos += oldLen;
+    }
+    offsets.start = offsets.limit;
+}
 U_NAMESPACE_END
 

icu4c/source/i18n/tolowtrn.h

@@ -20,15 +20,14 @@ U_NAMESPACE_BEGIN
  * case mapping.
  * @author Alan Liu
  */
-class U_I18N_API LowercaseTransliterator : public TransformTransliterator {
+class U_I18N_API LowercaseTransliterator : public Transliterator {
 
  public:
 
     /**
      * Constructs a transliterator.
      */
-    LowercaseTransliterator(const Locale& loc = Locale::getDefault(),
-                             UnicodeFilter* adoptedFilter = 0);
+    LowercaseTransliterator(const Locale& loc = Locale::getDefault());
 
     /**
      * Destructor.
@@ -53,19 +52,16 @@ class U_I18N_API LowercaseTransliterator : public TransformTransliterator {
  protected:
 
     /**
-     * TransformTransliterator framework method.
+     * Implements {@link Transliterator#handleTransliterate}.
      */
-    virtual UBool hasTransform(UChar32 c) const;
-
-    /**
-     * TransformTransliterator framework method.
-     */
-    virtual void transform(UnicodeString& s) const;
+    virtual void handleTransliterate(Replaceable& text,
+                                     UTransPosition& offsets, 
+                                     UBool isIncremental) const;
 
  private:
 
     Locale loc;
-
+    UChar* buffer;
     static const char _ID[];
 };
 

icu4c/source/i18n/toupptrn.cpp

@@ -9,6 +9,9 @@
 */
 
 #include "toupptrn.h"
+#include "unicode/ustring.h"
+#include "ustr_imp.h"
+#include "cpputils.h"
 #include "unicode/uchar.h"
 
 U_NAMESPACE_BEGIN
@@ -18,31 +21,39 @@ const char UppercaseTransliterator::_ID[] = "Any-Upper";
 /**
  * Constructs a transliterator.
  */
-UppercaseTransliterator::UppercaseTransliterator(const Locale& theLoc,
-                                                   UnicodeFilter* adoptedFilter) :
-    TransformTransliterator(_ID, adoptedFilter),
-    loc(theLoc) {
+UppercaseTransliterator::UppercaseTransliterator(const Locale& theLoc) :
+    Transliterator(_ID, 0),
+    loc(theLoc), 
+    buffer(0) {
+    buffer = new UChar[u_getMaxCaseExpansion()];
 }
 
 /**
  * Destructor.
  */
-UppercaseTransliterator::~UppercaseTransliterator() {}
+UppercaseTransliterator::~UppercaseTransliterator() {
+    delete [] buffer;
+}
 
 /**
  * Copy constructor.
  */
 UppercaseTransliterator::UppercaseTransliterator(const UppercaseTransliterator& o) :
-    TransformTransliterator(o),
-    loc(o.loc) {}
+    Transliterator(o),
+    loc(o.loc),
+    buffer(0){
+    buffer = new UChar[u_getMaxCaseExpansion()];
+    uprv_arrayCopy(o.buffer, 0, this->buffer, 0, u_getMaxCaseExpansion());
+}
 
 /**
  * Assignment operator.
  */
 UppercaseTransliterator& UppercaseTransliterator::operator=(
                              const UppercaseTransliterator& o) {
-    TransformTransliterator::operator=(o);
+    Transliterator::operator=(o);
     loc = o.loc;
+    uprv_arrayCopy(o.buffer, 0, this->buffer, 0, u_getMaxCaseExpansion());
     return *this;
 }
 
@@ -54,17 +65,57 @@ Transliterator* UppercaseTransliterator::clone(void) const {
 }
 
 /**
- * TransformTransliterator framework method.
+ * Implements {@link Transliterator#handleTransliterate}.
  */
-UBool UppercaseTransliterator::hasTransform(UChar32 c) const {
-    return c != u_toupper(c);
-}
+void UppercaseTransliterator::handleTransliterate(Replaceable& text,
+                                 UTransPosition& offsets, 
+                                 UBool isIncremental) const {
+    int32_t textPos = offsets.start;
+    int32_t loop = 0;
+    if (textPos >= offsets.limit) return;
 
-/**
- * TransformTransliterator framework method.
- */
-void UppercaseTransliterator::transform(UnicodeString& s) const {
-    s.toUpper(loc);
+    // get string for context
+    // TODO: add convenience method to do this, since we do it all over
+    
+    UnicodeString original;
+    /* UChar *original = new UChar[offsets.contextLimit - offsets.contextStart+1]; */// get whole context
+    /* Extract the characters from Replaceable */
+    for (loop = offsets.contextStart; loop < offsets.contextLimit; loop++) {
+        original.append(text.charAt(loop));
+    }
+            
+    // Walk through original string
+    // If there is a case change, modify corresponding position in replaceable
+    
+    int32_t i = textPos - offsets.contextStart;
+    int32_t limit = offsets.limit - offsets.contextStart;
+    UChar32 cp;
+    int32_t oldLen;
+    
+    for (; i < limit; ) {
+        UErrorCode status = U_ZERO_ERROR;
+        int32_t s = i;
+        buffer[0] = original.charAt(s);
+
+        UTF_GET_CHAR(original.getBuffer(), 0, i, original.length(), cp);
+        oldLen = UTF_CHAR_LENGTH(cp);
+        i += oldLen;
+        u_strToUpper(buffer, u_getMaxCaseExpansion(), original.getBuffer()+s, i-s, loc.getName(), &status);
+        /* Skip checking of status code here because the buffer should not have overflowed. */
+        if (buffer[0] != original.charAt(s)) {
+            int len = u_strlen(buffer);
+            UnicodeString temp(buffer);
+            text.handleReplaceBetween(textPos, textPos + oldLen, temp);
+            if (len != oldLen) {
+                textPos += len;
+                offsets.limit += len - oldLen;
+                offsets.contextLimit += len - oldLen;
+                continue;
+            }
+        }
+        textPos += oldLen;
+    }
+    offsets.start = offsets.limit;
 }
 
 U_NAMESPACE_END

icu4c/source/i18n/toupptrn.h

@@ -20,15 +20,14 @@ U_NAMESPACE_BEGIN
  * case mapping.
  * @author Alan Liu
  */
-class U_I18N_API UppercaseTransliterator : public TransformTransliterator {
+class U_I18N_API UppercaseTransliterator : public Transliterator {
 
  public:
 
     /**
      * Constructs a transliterator.
      */
-    UppercaseTransliterator(const Locale& loc = Locale::getDefault(),
-                             UnicodeFilter* adoptedFilter = 0);
+    UppercaseTransliterator(const Locale& loc = Locale::getDefault());
 
     /**
      * Destructor.
@@ -52,20 +51,18 @@ class U_I18N_API UppercaseTransliterator : public TransformTransliterator {
 
  protected:
 
-    /**
-     * TransformTransliterator framework method.
-     */
-    virtual UBool hasTransform(UChar32 c) const;
 
     /**
-     * TransformTransliterator framework method.
+     * Implements {@link Transliterator#handleTransliterate}.
      */
-    virtual void transform(UnicodeString& s) const;
+    virtual void handleTransliterate(Replaceable& text,
+                                     UTransPosition& offsets, 
+                                     UBool isIncremental) const;
 
  private:
 
     Locale loc;
-
+    UChar* buffer;
     static const char _ID[];
 };