ICU-1611 add and test u_enumCharTypes

X-SVN-Rev: 7434

icu4c/source/common/uchar.c

@@ -423,6 +423,39 @@ u_charType(UChar32 c) {
     return (int8_t)GET_CATEGORY(props);
 }
 
+/* Enumerate all code points with their general categories. */
+struct _EnumTypeCallback {
+    UCharEnumTypeRange *enumRange;
+    const void *context;
+};
+
+static uint32_t U_CALLCONV
+_enumTypeValue(const void *context, uint32_t value) {
+    /* access the general category from the 32-bit properties, and those from the 16-bit trie value */
+    return GET_CATEGORY(props32Table[value]);
+}
+
+static UBool U_CALLCONV
+_enumTypeRange(const void *context, UChar32 start, UChar32 limit, uint32_t value) {
+    /* just cast the value to UCharCategory */
+    return ((struct _EnumTypeCallback *)context)->
+        enumRange(((struct _EnumTypeCallback *)context)->context,
+                  start, limit, (UCharCategory)value);
+}
+
+U_CAPI void U_EXPORT2
+u_enumCharTypes(UCharEnumTypeRange *enumRange, const void *context) {
+    struct _EnumTypeCallback callback;
+
+    if(enumRange==NULL || !HAVE_DATA) {
+        return;
+    }
+
+    callback.enumRange=enumRange;
+    callback.context=context;
+    utrie_enum(&propsTrie, _enumTypeValue, _enumTypeRange, &callback);
+}
+
 /* Checks if ch is a lower case letter.*/
 U_CAPI UBool U_EXPORT2
 u_islower(UChar32 c) {

icu4c/source/common/unicode/uchar.h

@@ -24,6 +24,9 @@
 #define UCHAR_H
 
 #include "unicode/utypes.h"
+
+U_CDECL_BEGIN
+
 /*==========================================================================*/
 /* Unicode version number                                                   */
 /*==========================================================================*/
@@ -995,6 +998,48 @@ u_charCellWidth(UChar32 c);
 U_CAPI int8_t U_EXPORT2
 u_charType(UChar32 c);
 
+/**
+ * Callback from u_enumCharTypes(), is called for each contiguous range
+ * of code points c (where start<=c<limit)
+ * with the same Unicode general category ("character type").
+ *
+ * The callback function can stop the enumeration by returning FALSE.
+ *
+ * @param context an opaque pointer, as passed into utrie_enum()
+ * @param start the first code point in a contiguous range with value
+ * @param limit one past the last code point in a contiguous range with value
+ * @param type the general category for all code points in [start..limit[
+ * @return FALSE to stop the enumeration
+ *
+ * @draft ICU 2.1
+ * @see UCharCategory
+ * @see u_enumCharTypes
+ */
+typedef UBool U_CALLCONV
+UCharEnumTypeRange(const void *context, UChar32 start, UChar32 limit, UCharCategory type);
+
+/**
+ * Enumerate efficiently all code points with their Unicode general categories.
+ *
+ * This is useful for building data structures (e.g., UnicodeSet's),
+ * for enumerating all assigned code points (type!=U_UNASSIGNED), etc.
+ *
+ * For each contiguous range of code points with a given general category ("character type"),
+ * the UCharEnumTypeRange function is called.
+ * Adjacent ranges have different types.
+ * The Unicode Standard guarantees that the numeric value of the type is 0..31.
+ *
+ * @param enumRange a pointer to a function that is called for each contiguous range
+ *                  of code points with the same general category
+ * @param context an opaque pointer that is passed on to the callback function
+ *
+ * @draft ICU 2.1
+ * @see UCharCategory
+ * @see UCharEnumTypeRange
+ */
+U_CAPI void U_EXPORT2
+u_enumCharTypes(UCharEnumTypeRange *enumRange, const void *context);
+
 /**
  * Returns the combining class of the code point as specified in UnicodeData.txt.
  *
@@ -1083,8 +1128,6 @@ u_charFromName(UCharNameChoice nameChoice,
                const char *name,
                UErrorCode *pErrorCode);
 
-U_CDECL_BEGIN
-
 /**
  * Type of a callback function for u_enumCharNames() that gets called
  * for each Unicode character with the code point value and
@@ -1107,8 +1150,6 @@ typedef UBool UEnumCharNamesFn(void *context,
                                const char *name,
                                UTextOffset length);
 
-U_CDECL_END
-
 /**
  * Enumerate all assigned Unicode characters between the start and limit
  * code points (start inclusive, limit exclusive) and call a function
@@ -1429,5 +1470,7 @@ u_getUnicodeVersion(UVersionInfo info);
 /** @deprecated  Use the enum UCharBlock instead. Remove after Aug,2002*/
 typedef UBlockCode UCharScript;
 
+U_CDECL_END
+
 #endif /*_UCHAR*/
 /*eof*/

icu4c/source/test/cintltst/cucdtst.c

@@ -573,6 +573,41 @@ unicodeDataLineFn(void *context,
     }
 }
 
+static UBool U_CALLCONV
+enumTypeRange(const void *context, UChar32 start, UChar32 limit, UCharCategory type) {
+    static UChar32 test[][2]={
+        0x41, U_UPPERCASE_LETTER,
+        0x308, U_NON_SPACING_MARK,
+        0xfffe, U_GENERAL_OTHER_TYPES,
+        0xe0041, U_FORMAT_CHAR,
+        0xeffff, U_UNASSIGNED
+    };
+    int i, count;
+
+    if(0!=uprv_strcmp((const char *)context, "a1")) {
+        log_err("error: u_enumCharTypes() passes on an incorrect context pointer\n");
+    }
+
+    count=sizeof(test)/sizeof(test[0]);
+    for(i=0; i<count; ++i) {
+        if(start<=test[i][0] && test[i][0]<limit) {
+            if(type!=(UCharCategory)test[i][1]) {
+                log_err("error: u_enumCharTypes() has range [U+%04lx, U+%04lx[ with %ld instead of U+%04lx with %ld\n",
+                        start, limit, (long)type, test[i][0], test[i][1]);
+            }
+            /* stop at the range that includes the last test code point */
+            return i==(count-1) ? FALSE : TRUE;
+        }
+    }
+
+    if(start>test[count-1][0]) {
+        log_err("error: u_enumCharTypes() has range [U+%04lx, U+%04lx[ with %ld after it should have stopped\n",
+                start, limit, (long)type);
+        return FALSE;
+    }
+    return TRUE;
+}
+
 /* tests for several properties */
 static void TestUnicodeData()
 {
@@ -658,6 +693,9 @@ static void TestUnicodeData()
             ++c;
         }
     }
+
+    /* test u_enumCharTypes() */
+    u_enumCharTypes(enumTypeRange, "a1");
 }
 
 /*internal functions ----*/