ICU-1052 redesign of engine to support supplemental characters

X-SVN-Rev: 5340

icu4c/source/i18n/unicode/translit.h

@@ -559,7 +559,7 @@ public:
      * @draft
      */
     virtual void transliterate(Replaceable& text, UTransPosition& index,
-                               UChar insertion,
+                               UChar32 insertion,
                                UErrorCode& status) const;
 
     /**

icu4c/source/i18n/unicode/unifilt.h

@@ -8,7 +8,7 @@
 #ifndef UNIFILT_H
 #define UNIFILT_H
 
-#include "unicode/utypes.h"
+#include "unicode/unimatch.h"
 
 /**
  * <code>UnicodeFilter</code> defines a protocol for selecting a
@@ -20,7 +20,7 @@
  * @see UnicodeFilterLogic
  * @draft
  */
-class U_I18N_API UnicodeFilter {
+class U_I18N_API UnicodeFilter : public UnicodeMatcher {
 
 public:
     /**
@@ -35,7 +35,7 @@ public:
      * <b><tt>false</tt></b>.
      * @draft
      */
-    virtual UBool contains(UChar c) const = 0;
+    virtual UBool contains(UChar32 c) const = 0;
 
     /**
      * Returns a copy of this object.  All UnicodeFilter objects have
@@ -45,6 +45,14 @@ public:
      */
     virtual UnicodeFilter* clone() const = 0;
 
+    /**
+     * Implement UnicodeMatcher API.
+     */
+    virtual UMatchDegree matches(const Replaceable& text,
+                                 int32_t& offset,
+                                 int32_t limit,
+                                 UBool incremental) const;
+
 protected:
 
     UnicodeFilter();

icu4c/source/i18n/unicode/unimatch.h

@@ -0,0 +1,115 @@
+/*
+* Copyright (C) 2001, International Business Machines Corporation and others. All Rights Reserved.
+**********************************************************************
+*   Date        Name        Description
+*   07/18/01    aliu        Creation.
+**********************************************************************
+*/
+#ifndef UNIMATCH_H
+#define UNIMATCH_H
+
+#include "unicode/utypes.h"
+
+class Replaceable;
+
+/**
+ * Constants returned by <code>UnicodeMatcher::matches()</code>
+ * indicating the degree of match.
+ */
+enum UMatchDegree {
+    /**
+     * Constant returned by <code>matches()</code> indicating a
+     * mismatch between the text and this matcher.  The text contains
+     * a character which does not match, or the text does not contain
+     * all desired characters for a non-incremental match.
+     */
+    U_MISMATCH,
+    
+    /**
+     * Constant returned by <code>matches()</code> indicating a
+     * partial match between the text and this matcher.  This value is
+     * only returned for incremental match operations.  All characters
+     * of the text match, but more characters are required for a
+     * complete match.  Alternatively, for variable-length matchers,
+     * all characters of the text match, and if more characters were
+     * supplied at limit, they might also match.
+     */
+    U_PARTIAL_MATCH,
+    
+    /**
+     * Constant returned by <code>matches()</code> indicating a
+     * complete match between the text and this matcher.  For an
+     * incremental variable-length match, this value is returned if
+     * the given text matches, and it is known that additional
+     * characters would not alter the extent of the match.
+     */
+    U_MATCH
+};
+
+/**
+ * <code>UnicodeMatcher</code> defines a protocol for objects that can
+ * match a range of characters in a Replaceable string.
+ */
+class U_I18N_API UnicodeMatcher {
+
+public:
+
+    /**
+     * Destructor
+     */
+    virtual ~UnicodeMatcher();
+
+    /**
+     * Return a UMatchDegree value indicating the degree of match for
+     * the given text at the given offset.  Zero, one, or more
+     * characters may be matched.
+     *
+     * Matching in the forward direction is indicated by limit >
+     * offset.  Characters from offset forwards to limit-1 will be
+     * considered for matching.
+     * 
+     * Matching in the reverse direction is indicated by limit <
+     * offset.  Characters from offset backwards to limit+1 will be
+     * considered for matching.
+     *
+     * If limit == offset then the only match possible is a zero
+     * character match (which subclasses may implement if desired).
+     *
+     * As a side effect, advance the offset parameter to the limit of
+     * the matched substring.  In the forward direction, this will be
+     * the index of the last matched character plus one.  In the
+     * reverse direction, this will be the index of the last matched
+     * character minus one.
+     *
+     * @param text the text to be matched
+     * @param offset on input, the index into text at which to begin
+     * matching.  On output, the limit of the matched text.  The
+     * number of matched characters is the output value of offset
+     * minus the input value.
+     * @param limit the limit index of text to be matched.  Greater
+     * than offset for a forward direction match, less than offset for
+     * a backward direction match.  The last character to be
+     * considered for matching will be text.charAt(limit-1) in the
+     * forward direction or text.charAt(limit+1) in the backward
+     * direction.
+     * @param incremental if TRUE, then assume further characters may
+     * be inserted at limit and check for partial matching.  Otherwise
+     * assume the text as given is complete.
+     * @return a match degree value indicating a full match, a partial
+     * match, or a mismatch.  If incremental is FALSE then
+     * U_PARTIAL_MATCH should never be returned.
+     */
+    virtual UMatchDegree matches(const Replaceable& text,
+                                 int32_t& offset,
+                                 int32_t limit,
+                                 UBool incremental) const = 0;
+
+protected:
+
+    UnicodeMatcher();
+};
+
+inline UnicodeMatcher::UnicodeMatcher() {}
+inline UnicodeMatcher::~UnicodeMatcher() {}
+
+#endif

icu4c/source/i18n/unicode/uniset.h

@@ -458,14 +458,13 @@ public:
     virtual UBool contains(UChar32 c) const;
 
     /**
-     * Implement UnicodeFilter:
-     * Returns <tt>true</tt> if this set contains the specified char.
-     *
-     * @return <tt>true</tt> if this set contains the specified char.
-     * @draft
+     * Implement UnicodeMatcher::matches()
      */
-    virtual UBool contains(UChar c) const;
-
+    UMatchDegree matches(const Replaceable& text,
+                         int32_t& offset,
+                         int32_t limit,
+                         UBool incremental) const;
+    
     /**
      * Adds the specified range to this set if it is not already
      * present.  If this set already contains the specified range,
@@ -761,12 +760,6 @@ private:
     // Implementation: Utility methods
     //----------------------------------------------------------------
 
-    /**
-     * Returns the character after the given position, or '\uFFFE' if
-     * there is none.
-     */
-    static UChar charAfter(const UnicodeString& str, int32_t i);
-
     void ensureCapacity(int32_t newLen);
 
     void ensureBufferCapacity(int32_t newLen);

icu4c/source/i18n/unifilt.cpp

@@ -0,0 +1,36 @@
+/*
+* Copyright (C) 2001, International Business Machines Corporation and others. All Rights Reserved.
+**********************************************************************
+*   Date        Name        Description
+*   07/18/01    aliu        Creation.
+**********************************************************************
+*/
+
+#include "unicode/unifilt.h"
+#include "unicode/rep.h"
+#include "rbt_rule.h"
+
+/**
+ * Default implementation of UnicodeMatcher::matches() for Unicode
+ * filters.  Matches a single 16-bit code unit at offset.
+ */
+UMatchDegree UnicodeFilter::matches(const Replaceable& text,
+                                    int32_t& offset,
+                                    int32_t limit,
+                                    UBool incremental) const {
+    UChar32 c;
+    if (offset < limit &&
+        contains(c = text.char32At(offset))) {
+        offset += UTF_CHAR_LENGTH(c);
+        return U_MATCH;
+    }
+    if (offset > limit &&
+        contains(c = text.charAt(offset))) {
+        offset -= UTF_CHAR_LENGTH(c);
+        return U_MATCH;
+    }
+    if (incremental && offset == limit) {
+        return U_PARTIAL_MATCH;
+    }
+    return U_MISMATCH;
+}