int ICUTransService::compareNIString(const XMLCh* const comp1
, const XMLCh* const comp2
, const unsigned int maxChars)
{
if (maxChars > 0)
{
// Note that this function has somewhat broken semantics, as it's
// possible for two strings of different lengths to compare as equal
// in a case-insensitive manner, since one character could be
// represented as a surrogate pair.
size_t i = 0;
size_t j = 0;
for(;;)
{
UChar32 ch1;
UChar32 ch2;
U16_NEXT_UNSAFE(comp1, i, ch1);
U16_NEXT_UNSAFE(comp2, j, ch2);
const UChar32 folded1 =
u_foldCase(ch1, U_FOLD_CASE_DEFAULT);
const UChar32 folded2 =
u_foldCase(ch2, U_FOLD_CASE_DEFAULT);
if (folded1 != folded2)
{
return folded1 - folded2;
}
else if (i == maxChars)
{
// If we're at the end of both strings, return 0.
// Otherwise, we've run out of characters in the
// left string, so return -1.
return j == maxChars ? 0 : -1;
}
else if (j == maxChars)
{
// We've run out of characters in the right string,
// but not the left, so return 1.
return 1;
}
}
}
return 0;
}
U_CFUNC UBool U_EXPORT2
ucase_addStringCaseClosure(const UCaseProps *csp, const UChar *s, int32_t length, const USetAdder *sa) {
const UChar *unfold, *p;
int32_t i, start, limit, result, unfoldRows, unfoldRowWidth, unfoldStringWidth;
if(csp->unfold==NULL || s==NULL) {
return FALSE; /* no reverse case folding data, or no string */
}
if(length<=1) {
/* the string is too short to find any match */
/*
* more precise would be:
* if(!u_strHasMoreChar32Than(s, length, 1))
* but this does not make much practical difference because
* a single supplementary code point would just not be found
*/
return FALSE;
}
unfold=csp->unfold;
unfoldRows=unfold[UCASE_UNFOLD_ROWS];
unfoldRowWidth=unfold[UCASE_UNFOLD_ROW_WIDTH];
unfoldStringWidth=unfold[UCASE_UNFOLD_STRING_WIDTH];
unfold+=unfoldRowWidth;
if(length>unfoldStringWidth) {
/* the string is too long to find any match */
return FALSE;
}
/* do a binary search for the string */
start=0;
limit=unfoldRows;
while(start<limit) {
i=(start+limit)/2;
p=unfold+(i*unfoldRowWidth);
result=strcmpMax(s, length, p, unfoldStringWidth);
if(result==0) {
/* found the string: add each code point, and its case closure */
UChar32 c;
for(i=unfoldStringWidth; i<unfoldRowWidth && p[i]!=0;) {
U16_NEXT_UNSAFE(p, i, c);
sa->add(sa->set, c);
ucase_addCaseClosure(csp, c, sa);
}
return TRUE;
} else if(result<0) {
limit=i;
} else /* result>0 */ {
start=i+1;
}
}
return FALSE; /* string not found */
}
// ---------------------------------------------------------------------------
// ICUTransService: The virtual transcoding service API
// ---------------------------------------------------------------------------
int ICUTransService::compareIString(const XMLCh* const comp1
, const XMLCh* const comp2)
{
size_t i = 0;
size_t j = 0;
for(;;)
{
UChar32 ch1;
UChar32 ch2;
U16_NEXT_UNSAFE(comp1, i, ch1);
U16_NEXT_UNSAFE(comp2, j, ch2);
const UChar32 folded1 =
u_foldCase(ch1, U_FOLD_CASE_DEFAULT);
const UChar32 folded2 =
u_foldCase(ch2, U_FOLD_CASE_DEFAULT);
if (folded1 !=
folded2)
{
return folded1 - folded2;
}
else if (ch1 == 0)
{
// If ch1 is 0, the ch2 must also be
// 0. Otherwise, the previous if
// would have failed.
break;
}
}
return 0;
}
static void
doCaseConvert(
XMLCh* convertString,
FunctionType caseFunction)
{
// Note the semantics of this function are broken, since it's
// possible that changing the case of a string could increase
// its length, but there's no way to handle such a situation.
const unsigned int len =
XMLString::stringLen(convertString);
size_t readPos = 0;
size_t writePos = 0;
while(readPos < len)
{
UChar32 original;
// Get the next Unicode code point.
U16_NEXT_UNSAFE(convertString, readPos, original);
// Convert the code point
const UChar32 converted = caseFunction(original);
// OK, now here's where it gets ugly.
if (!U_IS_BMP(converted) && U_IS_BMP(original) &&
readPos - writePos == 1)
{
// We do not have room to convert the
// character without overwriting the next
// character, so we will just stop.
break;
}
else
{
U16_APPEND_UNSAFE(convertString, writePos, converted);
}
}
convertString[writePos] = 0;
}
static void TestNextPrevChar(){
static UChar input[]={0x0061, 0xd800, 0xdc00, 0xdbff, 0xdfff, 0x0062, 0xd841, 0xd7ff, 0xd841, 0xdc41, 0xdc00, 0x0000};
static UChar32 result[]={
/*next_unsafe next_safe_ns next_safe_s prev_unsafe prev_safe_ns prev_safe_s*/
0x0061, 0x0061, 0x0061, 0x0000, 0x0000, 0x0000,
0x10000, 0x10000, 0x10000, 0x120400, 0xdc00, UTF_ERROR_VALUE,
0xdc00, 0xdc00, UTF_ERROR_VALUE, 0x20441, 0x20441, 0x20441,
0x10ffff, 0x10ffff, 0x10ffff, 0xd841, 0xd841, UTF_ERROR_VALUE,
0xdfff, 0xdfff, UTF_ERROR_VALUE, 0xd7ff, 0xd7ff, 0xd7ff,
0x0062, 0x0062, 0x0062, 0xd841, 0xd841, UTF_ERROR_VALUE,
0x1ffff, 0xd841, UTF_ERROR_VALUE, 0x0062, 0x0062, 0x0062,
0xd7ff, 0xd7ff, 0xd7ff, 0x10ffff, 0x10ffff, 0x10ffff,
0x20441, 0x20441, 0x20441, 0xdbff, 0xdbff, UTF_ERROR_VALUE,
0xdc41, 0xdc41, UTF_ERROR_VALUE, 0x10000, 0x10000, 0x10000,
0xdc00, 0xdc00, UTF_ERROR_VALUE, 0xd800, 0xd800, UTF_ERROR_VALUE,
0x0000, 0x0000, 0x0000, 0x0061, 0x0061, 0x0061
};
static uint16_t movedOffset[]={
/*next_unsafe next_safe_ns next_safe_s prev_unsafe prev_safe_ns prev_safe_s*/
1, 1, 1, 11, 11, 11,
3, 3, 3, 9, 10 , 10,
3, 3, 3, 8, 8, 8,
5, 5, 4, 8, 8, 8,
5, 5, 5, 7, 7, 7,
6, 6, 6, 6, 6, 6,
8, 7, 7, 5, 5, 5,
8, 8, 8, 3, 3, 3,
10, 10, 10, 3, 3, 3,
10, 10, 10, 1, 1, 1,
11, 11, 11, 1, 1, 1,
12, 12, 12, 0, 0, 0,
};
UChar32 c=0x0000;
uint16_t i=0;
uint16_t offset=0, setOffset=0;
for(offset=0; offset<sizeof(input)/U_SIZEOF_UCHAR; offset++){
setOffset=offset;
UTF16_NEXT_CHAR_UNSAFE(input, setOffset, c);
if(setOffset != movedOffset[i]){
log_err("ERROR: UTF16_NEXT_CHAR_UNSAFE failed to move the offset correctly at %d\n ExpectedOffset:%d Got %d\n",
offset, movedOffset[i], setOffset);
}
if(c != result[i]){
log_err("ERROR: UTF16_NEXT_CHAR_UNSAFE failed for offset=%ld. Expected:%lx Got:%lx\n", offset, result[i], c);
}
setOffset=offset;
U16_NEXT_UNSAFE(input, setOffset, c);
if(setOffset != movedOffset[i]){
log_err("ERROR: U16_NEXT_CHAR_UNSAFE failed to move the offset correctly at %d\n ExpectedOffset:%d Got %d\n",
offset, movedOffset[i], setOffset);
}
if(c != result[i]){
log_err("ERROR: U16_NEXT_CHAR_UNSAFE failed for offset=%ld. Expected:%lx Got:%lx\n", offset, result[i], c);
}
setOffset=offset;
UTF16_NEXT_CHAR_SAFE(input, setOffset, sizeof(input)/U_SIZEOF_UCHAR, c, FALSE);
if(setOffset != movedOffset[i+1]){
log_err("ERROR: UTF16_NEXT_CHAR_SAFE failed to move the offset correctly at %d\n ExpectedOffset:%d Got %d\n",
offset, movedOffset[i+1], setOffset);
}
if(c != result[i+1]){
log_err("ERROR: UTF16_NEXT_CHAR_SAFE failed for input=%ld. Expected:%lx Got:%lx\n", offset, result[i+1], c);
}
setOffset=offset;
U16_NEXT(input, setOffset, sizeof(input)/U_SIZEOF_UCHAR, c);
if(setOffset != movedOffset[i+1]){
log_err("ERROR: U16_NEXT failed to move the offset correctly at %d\n ExpectedOffset:%d Got %d\n",
offset, movedOffset[i+1], setOffset);
}
if(c != result[i+1]){
log_err("ERROR: U16_NEXT failed for input=%ld. Expected:%lx Got:%lx\n", offset, result[i+1], c);
}
setOffset=offset;
UTF16_NEXT_CHAR_SAFE(input, setOffset, sizeof(input)/U_SIZEOF_UCHAR, c, TRUE);
if(setOffset != movedOffset[i+1]){
log_err("ERROR: UTF16_NEXT_CHAR_SAFE(strict) failed to move the offset correctly at %d\n ExpectedOffset:%d Got %d\n",
offset, movedOffset[i+2], setOffset);
}
if(c != result[i+2]){
log_err("ERROR: UTF16_NEXT_CHAR_SAFE(strict) failed for input=%ld. Expected:%lx Got:%lx\n", offset, result[i+2], c);
}
i=(uint16_t)(i+6);
}
i=0;
for(offset=(uint16_t)sizeof(input)/U_SIZEOF_UCHAR; offset > 0; --offset){
setOffset=offset;
UTF16_PREV_CHAR_UNSAFE(input, setOffset, c);
if(setOffset != movedOffset[i+3]){
log_err("ERROR: UTF16_PREV_CHAR_UNSAFE failed to move the offset correctly at %d\n ExpectedOffset:%d Got %d\n",
offset, movedOffset[i+3], setOffset);
}
if(c != result[i+3]){
log_err("ERROR: UTF16_PREV_CHAR_UNSAFE failed for offset=%ld. Expected:%lx Got:%lx\n", offset, result[i+3], c);
}
setOffset=offset;
U16_PREV_UNSAFE(input, setOffset, c);
if(setOffset != movedOffset[i+3]){
log_err("ERROR: U16_PREV_CHAR_UNSAFE failed to move the offset correctly at %d\n ExpectedOffset:%d Got %d\n",
offset, movedOffset[i+3], setOffset);
}
if(c != result[i+3]){
log_err("ERROR: U16_PREV_CHAR_UNSAFE failed for offset=%ld. Expected:%lx Got:%lx\n", offset, result[i+3], c);
}
setOffset=offset;
UTF16_PREV_CHAR_SAFE(input, 0, setOffset, c, FALSE);
if(setOffset != movedOffset[i+4]){
log_err("ERROR: UTF16_PREV_CHAR_SAFE failed to move the offset correctly at %d\n ExpectedOffset:%d Got %d\n",
offset, movedOffset[i+4], setOffset);
}
if(c != result[i+4]){
log_err("ERROR: UTF16_PREV_CHAR_SAFE failed for input=%ld. Expected:%lx Got:%lx\n", offset, result[i+4], c);
}
setOffset=offset;
U16_PREV(input, 0, setOffset, c);
if(setOffset != movedOffset[i+4]){
log_err("ERROR: U16_PREV failed to move the offset correctly at %d\n ExpectedOffset:%d Got %d\n",
offset, movedOffset[i+4], setOffset);
}
if(c != result[i+4]){
log_err("ERROR: U16_PREV failed for input=%ld. Expected:%lx Got:%lx\n", offset, result[i+4], c);
}
setOffset=offset;
UTF16_PREV_CHAR_SAFE(input, 0, setOffset, c, TRUE);
if(setOffset != movedOffset[i+5]){
log_err("ERROR: UTF16_PREV_CHAR_SAFE(strict) failed to move the offset correctly at %d\n ExpectedOffset:%d Got %d\n",
offset, movedOffset[i+5], setOffset);
}
if(c != result[i+5]){
log_err("ERROR: UTF16_PREV_CHAR_SAFE(strict) failed for input=%ld. Expected:%lx Got:%lx\n", offset, result[i+5], c);
}
i=(uint16_t)(i+6);
}
}
U_CFUNC void U_EXPORT2
ucase_addCaseClosure(const UCaseProps *csp, UChar32 c, const USetAdder *sa) {
uint16_t props;
/*
* Hardcode the case closure of i and its relatives and ignore the
* data file data for these characters.
* The Turkic dotless i and dotted I with their case mapping conditions
* and case folding option make the related characters behave specially.
* This code matches their closure behavior to their case folding behavior.
*/
switch(c) {
case 0x49:
/* regular i and I are in one equivalence class */
sa->add(sa->set, 0x69);
return;
case 0x69:
sa->add(sa->set, 0x49);
return;
case 0x130:
/* dotted I is in a class with <0069 0307> (for canonical equivalence with <0049 0307>) */
sa->addString(sa->set, iDot, 2);
return;
case 0x131:
/* dotless i is in a class by itself */
return;
default:
/* otherwise use the data file data */
break;
}
props=UTRIE2_GET16(&csp->trie, c);
if(!PROPS_HAS_EXCEPTION(props)) {
if(UCASE_GET_TYPE(props)!=UCASE_NONE) {
/* add the one simple case mapping, no matter what type it is */
int32_t delta=UCASE_GET_DELTA(props);
if(delta!=0) {
sa->add(sa->set, c+delta);
}
}
} else {
/*
* c has exceptions, so there may be multiple simple and/or
* full case mappings. Add them all.
*/
const uint16_t *pe0, *pe=GET_EXCEPTIONS(csp, props);
const UChar *closure;
uint16_t excWord=*pe++;
int32_t idx, closureLength, fullLength, length;
pe0=pe;
/* add all simple case mappings */
for(idx=UCASE_EXC_LOWER; idx<=UCASE_EXC_TITLE; ++idx) {
if(HAS_SLOT(excWord, idx)) {
pe=pe0;
GET_SLOT_VALUE(excWord, idx, pe, c);
sa->add(sa->set, c);
}
}
/* get the closure string pointer & length */
if(HAS_SLOT(excWord, UCASE_EXC_CLOSURE)) {
pe=pe0;
GET_SLOT_VALUE(excWord, UCASE_EXC_CLOSURE, pe, closureLength);
closureLength&=UCASE_CLOSURE_MAX_LENGTH; /* higher bits are reserved */
closure=(const UChar *)pe+1; /* behind this slot, unless there are full case mappings */
} else {
closureLength=0;
closure=NULL;
}
/* add the full case folding */
if(HAS_SLOT(excWord, UCASE_EXC_FULL_MAPPINGS)) {
pe=pe0;
GET_SLOT_VALUE(excWord, UCASE_EXC_FULL_MAPPINGS, pe, fullLength);
/* start of full case mapping strings */
++pe;
fullLength&=0xffff; /* bits 16 and higher are reserved */
/* skip the lowercase result string */
pe+=fullLength&UCASE_FULL_LOWER;
fullLength>>=4;
/* add the full case folding string */
length=fullLength&0xf;
if(length!=0) {
sa->addString(sa->set, (const UChar *)pe, length);
pe+=length;
}
/* skip the uppercase and titlecase strings */
fullLength>>=4;
pe+=fullLength&0xf;
fullLength>>=4;
pe+=fullLength;
closure=(const UChar *)pe; /* behind full case mappings */
}
/* add each code point in the closure string */
for(idx=0; idx<closureLength;) {
U16_NEXT_UNSAFE(closure, idx, c);
sa->add(sa->set, c);
}
}
}
