U_CAPI int32_t U_EXPORT2
u_charDigitValue(UChar32 c) {
uint32_t props;
int32_t value;
GET_PROPS(c, props);
value=(int32_t)GET_NUMERIC_TYPE_VALUE(props)-UPROPS_NTV_DECIMAL_START;
if(value<=9) {
return value;
} else {
return -1;
}
}
U_CAPI double U_EXPORT2
u_getNumericValue(UChar32 c) {
uint32_t props;
int32_t ntv;
GET_PROPS(c, props);
ntv=(int32_t)GET_NUMERIC_TYPE_VALUE(props);
if(ntv==UPROPS_NTV_NONE) {
return U_NO_NUMERIC_VALUE;
} else if(ntv<UPROPS_NTV_DIGIT_START) {
/* decimal digit */
return ntv-UPROPS_NTV_DECIMAL_START;
} else if(ntv<UPROPS_NTV_NUMERIC_START) {
/* other digit */
return ntv-UPROPS_NTV_DIGIT_START;
} else if(ntv<UPROPS_NTV_FRACTION_START) {
/* small integer */
return ntv-UPROPS_NTV_NUMERIC_START;
} else if(ntv<UPROPS_NTV_LARGE_START) {
/* fraction */
int32_t numerator=(ntv>>4)-12;
int32_t denominator=(ntv&0xf)+1;
return (double)numerator/denominator;
} else if(ntv<UPROPS_NTV_BASE60_START) {
U_CAPI int32_t U_EXPORT2
u_getIntPropertyValue(UChar32 c, UProperty which) {
UErrorCode errorCode;
if(which<UCHAR_BINARY_START) {
return 0; /* undefined */
} else if(which<UCHAR_BINARY_LIMIT) {
return (int32_t)u_hasBinaryProperty(c, which);
} else if(which<UCHAR_INT_START) {
return 0; /* undefined */
} else if(which<UCHAR_INT_LIMIT) {
switch(which) {
case UCHAR_BIDI_CLASS:
return (int32_t)u_charDirection(c);
case UCHAR_BLOCK:
return (int32_t)ublock_getCode(c);
#if !UCONFIG_NO_NORMALIZATION
case UCHAR_CANONICAL_COMBINING_CLASS:
return u_getCombiningClass(c);
#endif
case UCHAR_DECOMPOSITION_TYPE:
return (int32_t)(u_getUnicodeProperties(c, 2)&UPROPS_DT_MASK);
case UCHAR_EAST_ASIAN_WIDTH:
return (int32_t)(u_getUnicodeProperties(c, 0)&UPROPS_EA_MASK)>>UPROPS_EA_SHIFT;
case UCHAR_GENERAL_CATEGORY:
return (int32_t)u_charType(c);
case UCHAR_JOINING_GROUP:
return ubidi_getJoiningGroup(GET_BIDI_PROPS(), c);
case UCHAR_JOINING_TYPE:
return ubidi_getJoiningType(GET_BIDI_PROPS(), c);
case UCHAR_LINE_BREAK:
return (int32_t)(u_getUnicodeProperties(c, UPROPS_LB_VWORD)&UPROPS_LB_MASK)>>UPROPS_LB_SHIFT;
case UCHAR_NUMERIC_TYPE: {
int32_t ntv=(int32_t)GET_NUMERIC_TYPE_VALUE(u_getUnicodeProperties(c, -1));
return UPROPS_NTV_GET_TYPE(ntv);
}
case UCHAR_SCRIPT:
errorCode=U_ZERO_ERROR;
return (int32_t)uscript_getScript(c, &errorCode);
case UCHAR_HANGUL_SYLLABLE_TYPE: {
/* see comments on gcbToHst[] above */
int32_t gcb=(int32_t)(u_getUnicodeProperties(c, 2)&UPROPS_GCB_MASK)>>UPROPS_GCB_SHIFT;
if(gcb<LENGTHOF(gcbToHst)) {
return gcbToHst[gcb];
} else {
return U_HST_NOT_APPLICABLE;
}
}
#if !UCONFIG_NO_NORMALIZATION
case UCHAR_NFD_QUICK_CHECK:
case UCHAR_NFKD_QUICK_CHECK:
case UCHAR_NFC_QUICK_CHECK:
case UCHAR_NFKC_QUICK_CHECK:
return (int32_t)unorm_getQuickCheck(c, (UNormalizationMode)(which-UCHAR_NFD_QUICK_CHECK+UNORM_NFD));
case UCHAR_LEAD_CANONICAL_COMBINING_CLASS:
return getFCD16(c)>>8;
case UCHAR_TRAIL_CANONICAL_COMBINING_CLASS:
return getFCD16(c)&0xff;
#endif
case UCHAR_GRAPHEME_CLUSTER_BREAK:
return (int32_t)(u_getUnicodeProperties(c, 2)&UPROPS_GCB_MASK)>>UPROPS_GCB_SHIFT;
case UCHAR_SENTENCE_BREAK:
return (int32_t)(u_getUnicodeProperties(c, 2)&UPROPS_SB_MASK)>>UPROPS_SB_SHIFT;
case UCHAR_WORD_BREAK:
return (int32_t)(u_getUnicodeProperties(c, 2)&UPROPS_WB_MASK)>>UPROPS_WB_SHIFT;
default:
return 0; /* undefined */
}
} else if(which==UCHAR_GENERAL_CATEGORY_MASK) {
static void U_CALLCONV
numericLineFn(void *context,
char *fields[][2], int32_t fieldCount,
UErrorCode *pErrorCode) {
Props newProps={ 0 };
char *s, *numberLimit;
uint32_t start, end, value, oldProps32;
char c;
UBool isFraction;
/* get the code point range */
u_parseCodePointRange(fields[0][0], &start, &end, pErrorCode);
if(U_FAILURE(*pErrorCode)) {
fprintf(stderr, "genprops: syntax error in DerivedNumericValues.txt field 0 at %s\n", fields[0][0]);
exit(*pErrorCode);
}
/*
* Ignore the
* # @missing: 0000..10FFFF; NaN
* line from Unicode 5.1's DerivedNumericValues.txt:
* The following code cannot parse "NaN", and we don't want to overwrite
* the numeric values for all characters after reading most
* from UnicodeData.txt already.
*/
if(start==0 && end==0x10ffff) {
return;
}
/* check if the numeric value is a fraction (this code does not handle any) */
isFraction=FALSE;
s=uprv_strchr(fields[1][0], '.');
if(s!=NULL) {
numberLimit=s+1;
while('0'<=(c=*numberLimit++) && c<='9') {
if(c!='0') {
isFraction=TRUE;
break;
}
}
}
if(isFraction) {
value=0;
} else {
/* parse numeric value */
s=(char *)u_skipWhitespace(fields[1][0]);
/* try large, single-significant-digit numbers, may otherwise overflow strtoul() */
if('1'<=s[0] && s[0]<='9' && s[1]=='0' && s[2]=='0') {
/* large integers are encoded in a special way, see store.c */
uint8_t exp=0;
value=s[0]-'0';
numberLimit=s;
while(*(++numberLimit)=='0') {
++exp;
}
newProps.exponent=exp;
} else {
/* normal number parsing */
value=(uint32_t)uprv_strtoul(s, &numberLimit, 10);
}
if(numberLimit<=s || (*numberLimit!='.' && u_skipWhitespace(numberLimit)!=fields[1][1]) || value>=0x80000000) {
fprintf(stderr, "genprops: syntax error in DerivedNumericValues.txt field 1 at %s\n", fields[0][0]);
exit(U_PARSE_ERROR);
}
}
/*
* Unicode 4.0.1 removes the third column that used to list the numeric type.
* Assume that either the data is the same as in UnicodeData.txt,
* or else that the numeric type is "numeric".
* This should work because we only expect to add numeric values for
* Han characters; for those, UnicodeData.txt lists only ranges without
* specific properties for single characters.
*/
/* set the new numeric value */
newProps.code=start;
newProps.numericValue=(int32_t)value; /* newly parsed numeric value */
/* the exponent may have been set above */
for(; start<=end; ++start) {
uint32_t newProps32;
int32_t oldNtv;
oldProps32=getProps(start);
oldNtv=(int32_t)GET_NUMERIC_TYPE_VALUE(oldProps32);
if(isFraction) {
if(UPROPS_NTV_FRACTION_START<=oldNtv && oldNtv<UPROPS_NTV_LARGE_START) {
/* this code point was already listed with its numeric value in UnicodeData.txt */
continue;
} else {
fprintf(stderr, "genprops: not prepared for new fractions in DerivedNumericValues.txt field 1 at %s\n", fields[1][0]);
exit(U_PARSE_ERROR);
}
}
/*
* For simplicity, and because we only expect to set numeric values for Han characters,
* for now we only allow to set these values for Lo characters.
*/
if(oldNtv==UPROPS_NTV_NONE && GET_CATEGORY(oldProps32)!=U_OTHER_LETTER) {
fprintf(stderr, "genprops error: new numeric value for a character other than Lo in DerivedNumericValues.txt at %s\n", fields[0][0]);
exit(U_PARSE_ERROR);
}
/* verify that we do not change an existing value (fractions were excluded above) */
if(oldNtv!=UPROPS_NTV_NONE) {
/* the code point already has a value stored */
newProps.numericType=UPROPS_NTV_GET_TYPE(oldNtv);
newProps32=makeProps(&newProps);
if(oldNtv!=GET_NUMERIC_TYPE_VALUE(newProps32)) {
fprintf(stderr, "genprops error: new numeric value differs from old one for U+%04lx\n", (long)start);
exit(U_PARSE_ERROR);
}
/* same value, continue */
} else {
/* the code point is getting a new numeric value */
newProps.numericType=(uint8_t)U_NT_NUMERIC; /* assumed numeric type, see Unicode 4.0.1 comment */
newProps32=makeProps(&newProps);
if(beVerbose) {
printf("adding U+%04x numeric type %d encoded-numeric-type-value 0x%03x from %s\n",
(int)start, U_NT_NUMERIC, (int)GET_NUMERIC_TYPE_VALUE(newProps32), fields[0][0]);
}
addProps(start, newProps32|GET_CATEGORY(oldProps32));
}
}
}
