static void
addClosureMapping(UChar32 src, UChar32 dest) {
uint32_t value;
if(beVerbose) {
printf("add closure mapping U+%04lx->U+%04lx\n",
(unsigned long)src, (unsigned long)dest);
}
value=upvec_getValue(pv, src, 0);
if(value&UCASE_EXCEPTION) {
Props *p=excProps+(value>>UGENCASE_EXC_SHIFT);
int32_t i;
/* append dest to src's closure array */
for(i=0;; ++i) {
if(i==LENGTHOF(p->closure)) {
fprintf(stderr, "closure[] overflow for U+%04lx->U+%04lx\n",
(unsigned long)src, (unsigned long)dest);
exit(U_BUFFER_OVERFLOW_ERROR);
} else if(p->closure[i]==dest) {
break; /* do not store duplicates */
} else if(p->closure[i]==0) {
p->closure[i]=dest;
break;
}
}
} else {
/* Improve code coverage of UPropsVectors */
static void TestUPropsVector() {
UErrorCode errorCode = U_ILLEGAL_ARGUMENT_ERROR;
UPropsVectors *pv = upvec_open(100, &errorCode);
if (pv != NULL) {
log_err("Should have returned NULL if UErrorCode is an error.");
return;
}
errorCode = U_ZERO_ERROR;
pv = upvec_open(-1, &errorCode);
if (pv != NULL || U_SUCCESS(errorCode)) {
log_err("Should have returned NULL if column is less than 0.\n");
return;
}
errorCode = U_ZERO_ERROR;
pv = upvec_open(100, &errorCode);
if (pv == NULL || U_FAILURE(errorCode)) {
log_err("Unable to open UPropsVectors.\n");
return;
}
if (upvec_getValue(pv, 0, 1) != 0) {
log_err("upvec_getValue should return 0.\n");
}
if (upvec_getRow(pv, 0, NULL, NULL) == NULL) {
log_err("upvec_getRow should not return NULL.\n");
}
if (upvec_getArray(pv, NULL, NULL) != NULL) {
log_err("upvec_getArray should return NULL.\n");
}
upvec_close(pv);
}
extern void
setProps(Props *p) {
UErrorCode errorCode;
uint32_t value, oldValue;
int32_t delta;
UBool isCaseIgnorable;
/* get the non-UnicodeData.txt properties */
value=oldValue=upvec_getValue(pv, p->code, 0);
/* default: map to self */
delta=0;
if(p->gc==U_TITLECASE_LETTER) {
/* the Titlecase property is read late, from UnicodeData.txt */
value|=UCASE_TITLE;
}
if(p->upperCase!=0) {
/* uppercase mapping as delta if the character is lowercase */
if((value&UCASE_TYPE_MASK)==UCASE_LOWER) {
delta=p->upperCase-p->code;
} else {
value|=UCASE_EXCEPTION;
}
}
if(p->lowerCase!=0) {
/* lowercase mapping as delta if the character is uppercase or titlecase */
if((value&UCASE_TYPE_MASK)>=UCASE_UPPER) {
delta=p->lowerCase-p->code;
} else {
value|=UCASE_EXCEPTION;
}
}
if(p->upperCase!=p->titleCase) {
value|=UCASE_EXCEPTION;
}
if(p->closure[0]!=0) {
value|=UCASE_EXCEPTION;
}
if(p->specialCasing!=NULL) {
value|=UCASE_EXCEPTION;
}
if(p->caseFolding!=NULL) {
value|=UCASE_EXCEPTION;
}
if(delta<UCASE_MIN_DELTA || UCASE_MAX_DELTA<delta) {
value|=UCASE_EXCEPTION;
}
if(p->cc!=0) {
if(value&UCASE_DOT_MASK) {
fprintf(stderr, "gencase: a soft-dotted character has cc!=0\n");
exit(U_INTERNAL_PROGRAM_ERROR);
}
if(p->cc==230) {
value|=UCASE_ABOVE;
} else {
value|=UCASE_OTHER_ACCENT;
}
}
/* encode case-ignorable as delta==1 on uncased characters */
isCaseIgnorable=FALSE;
if((value&UCASE_TYPE_MASK)==UCASE_NONE) {
if(ucdVersion>=UNI_4_1) {
/*
* Unicode 4.1 and up: (D47a) Word_Break=MidLetter or Mn, Me, Cf, Lm, Sk
* Unicode 5.1 and up: Word_Break=(MidLetter or MidNumLet) or Mn, Me, Cf, Lm, Sk
* The UGENCASE_IS_MID_LETTER_SHIFT bit is set for both WB=MidLetter and WB=MidNumLet.
*/
if(
(U_MASK(p->gc)&(U_GC_MN_MASK|U_GC_ME_MASK|U_GC_CF_MASK|U_GC_LM_MASK|U_GC_SK_MASK))!=0 ||
(upvec_getValue(pv, p->code, 1)&U_MASK(UGENCASE_IS_MID_LETTER_SHIFT))!=0
) {
isCaseIgnorable=TRUE;
}
} else {
/* before Unicode 4.1: Mn, Me, Cf, Lm, Sk or 0027 or 00AD or 2019 */
if(
(U_MASK(p->gc)&(U_GC_MN_MASK|U_GC_ME_MASK|U_GC_CF_MASK|U_GC_LM_MASK|U_GC_SK_MASK))!=0 ||
p->code==0x27 || p->code==0xad || p->code==0x2019
) {
isCaseIgnorable=TRUE;
}
}
}
if(isCaseIgnorable && p->code!=0x307) {
/*
* We use one of the delta/exception bits, which works because we only
* store the case-ignorable flag for uncased characters.
* There is no delta for uncased characters (see checks above).
* If there is an exception for an uncased, case-ignorable character
* (although there should not be any case mappings if it's uncased)
* then we have a problem.
* There is one character which is case-ignorable but has an exception:
* U+0307 is uncased, Mn, has conditional special casing and
* is therefore handled in code instead.
*/
if(value&UCASE_EXCEPTION) {
fprintf(stderr, "gencase error: unable to encode case-ignorable for U+%04lx with exceptions\n",
(unsigned long)p->code);
exit(U_INTERNAL_PROGRAM_ERROR);
}
delta=1;
}
/* handle exceptions */
if(value&UCASE_EXCEPTION) {
/* simply store exceptions for later processing and encoding */
value|=(uint32_t)exceptionsCount<<UGENCASE_EXC_SHIFT;
uprv_memcpy(excProps+exceptionsCount, p, sizeof(*p));
if(++exceptionsCount==MAX_EXC_COUNT) {
fprintf(stderr, "gencase: too many exceptions\n");
exit(U_INDEX_OUTOFBOUNDS_ERROR);
}
} else {
/* store the simple case mapping delta */
value|=((uint32_t)delta<<UCASE_DELTA_SHIFT)&UCASE_DELTA_MASK;
}
errorCode=U_ZERO_ERROR;
if(value!=oldValue) {
upvec_setValue(pv, p->code, p->code, 0, value, 0xffffffff, &errorCode);
if(U_FAILURE(errorCode)) {
fprintf(stderr, "gencase error: unable to set case mapping values, code: %s\n",
u_errorName(errorCode));
exit(errorCode);
}
}
/* add the multi-character case folding to the "unfold" data */
if(p->caseFolding!=NULL) {
int32_t length=p->caseFolding->full[0];
if(length>1 && u_strHasMoreChar32Than(p->caseFolding->full+1, length, 1)) {
addUnfolding(p->code, p->caseFolding->full+1, length);
}
}
}
