michael@0: /*
michael@0: *******************************************************************************
michael@0: *
michael@0: *   Copyright (C) 2002-2011, International Business Machines
michael@0: *   Corporation and others.  All Rights Reserved.
michael@0: *
michael@0: *******************************************************************************
michael@0: *   file name:  uset.cpp
michael@0: *   encoding:   US-ASCII
michael@0: *   tab size:   8 (not used)
michael@0: *   indentation:4
michael@0: *
michael@0: *   created on: 2002mar07
michael@0: *   created by: Markus W. Scherer
michael@0: *
michael@0: *   There are functions to efficiently serialize a USet into an array of uint16_t
michael@0: *   and functions to use such a serialized form efficiently without
michael@0: *   instantiating a new USet.
michael@0: */
michael@0: 
michael@0: #include "unicode/utypes.h"
michael@0: #include "unicode/uobject.h"
michael@0: #include "unicode/uset.h"
michael@0: #include "unicode/uniset.h"
michael@0: #include "cmemory.h"
michael@0: #include "unicode/ustring.h"
michael@0: #include "unicode/parsepos.h"
michael@0: 
michael@0: U_NAMESPACE_USE
michael@0: 
michael@0: U_CAPI USet* U_EXPORT2
michael@0: uset_openEmpty() {
michael@0:     return (USet*) new UnicodeSet();
michael@0: }
michael@0: 
michael@0: U_CAPI USet* U_EXPORT2
michael@0: uset_open(UChar32 start, UChar32 end) {
michael@0:     return (USet*) new UnicodeSet(start, end);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_close(USet* set) {
michael@0:     delete (UnicodeSet*) set;
michael@0: }
michael@0: 
michael@0: U_CAPI USet * U_EXPORT2
michael@0: uset_clone(const USet *set) {
michael@0:     return (USet*) (((UnicodeSet*) set)->UnicodeSet::clone());
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_isFrozen(const USet *set) {
michael@0:     return ((UnicodeSet*) set)->UnicodeSet::isFrozen();
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_freeze(USet *set) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::freeze();
michael@0: }
michael@0: 
michael@0: U_CAPI USet * U_EXPORT2
michael@0: uset_cloneAsThawed(const USet *set) {
michael@0:     return (USet*) (((UnicodeSet*) set)->UnicodeSet::cloneAsThawed());
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_set(USet* set,
michael@0:      UChar32 start, UChar32 end) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::set(start, end);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_addAll(USet* set, const USet *additionalSet) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::addAll(*((const UnicodeSet*)additionalSet));
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_add(USet* set, UChar32 c) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::add(c);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_addRange(USet* set, UChar32 start, UChar32 end) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::add(start, end);    
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_addString(USet* set, const UChar* str, int32_t strLen) {
michael@0:     // UnicodeString handles -1 for strLen
michael@0:     UnicodeString s(strLen<0, str, strLen);
michael@0:     ((UnicodeSet*) set)->UnicodeSet::add(s);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_addAllCodePoints(USet* set, const UChar *str, int32_t strLen) {
michael@0:     // UnicodeString handles -1 for strLen
michael@0:     UnicodeString s(str, strLen);
michael@0:     ((UnicodeSet*) set)->UnicodeSet::addAll(s);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_remove(USet* set, UChar32 c) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::remove(c);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_removeRange(USet* set, UChar32 start, UChar32 end) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::remove(start, end);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_removeString(USet* set, const UChar* str, int32_t strLen) {
michael@0:     UnicodeString s(strLen==-1, str, strLen);
michael@0:     ((UnicodeSet*) set)->UnicodeSet::remove(s);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_removeAll(USet* set, const USet* remove) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::removeAll(*(const UnicodeSet*)remove);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_retain(USet* set, UChar32 start, UChar32 end) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::retain(start, end);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_retainAll(USet* set, const USet* retain) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::retainAll(*(const UnicodeSet*)retain);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_compact(USet* set) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::compact();
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_complement(USet* set) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::complement();
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_complementAll(USet* set, const USet* complement) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::complementAll(*(const UnicodeSet*)complement);
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_clear(USet* set) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::clear();
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_removeAllStrings(USet* set) {
michael@0:     ((UnicodeSet*) set)->UnicodeSet::removeAllStrings();
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_isEmpty(const USet* set) {
michael@0:     return ((const UnicodeSet*) set)->UnicodeSet::isEmpty();
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_contains(const USet* set, UChar32 c) {
michael@0:     return ((const UnicodeSet*) set)->UnicodeSet::contains(c);
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_containsRange(const USet* set, UChar32 start, UChar32 end) {
michael@0:     return ((const UnicodeSet*) set)->UnicodeSet::contains(start, end);
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_containsString(const USet* set, const UChar* str, int32_t strLen) {
michael@0:     UnicodeString s(strLen==-1, str, strLen);
michael@0:     return ((const UnicodeSet*) set)->UnicodeSet::contains(s);
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_containsAll(const USet* set1, const USet* set2) {
michael@0:     return ((const UnicodeSet*) set1)->UnicodeSet::containsAll(* (const UnicodeSet*) set2);
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_containsAllCodePoints(const USet* set, const UChar *str, int32_t strLen) {
michael@0:     // Create a string alias, since nothing is being added to the set.
michael@0:     UnicodeString s(strLen==-1, str, strLen);
michael@0:     return ((const UnicodeSet*) set)->UnicodeSet::containsAll(s);
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_containsNone(const USet* set1, const USet* set2) {
michael@0:     return ((const UnicodeSet*) set1)->UnicodeSet::containsNone(* (const UnicodeSet*) set2);
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_containsSome(const USet* set1, const USet* set2) {
michael@0:     return ((const UnicodeSet*) set1)->UnicodeSet::containsSome(* (const UnicodeSet*) set2);
michael@0: }
michael@0: 
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uset_span(const USet *set, const UChar *s, int32_t length, USetSpanCondition spanCondition) {
michael@0:     return ((UnicodeSet*) set)->UnicodeSet::span(s, length, spanCondition);
michael@0: }
michael@0: 
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uset_spanBack(const USet *set, const UChar *s, int32_t length, USetSpanCondition spanCondition) {
michael@0:     return ((UnicodeSet*) set)->UnicodeSet::spanBack(s, length, spanCondition);
michael@0: }
michael@0: 
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uset_spanUTF8(const USet *set, const char *s, int32_t length, USetSpanCondition spanCondition) {
michael@0:     return ((UnicodeSet*) set)->UnicodeSet::spanUTF8(s, length, spanCondition);
michael@0: }
michael@0: 
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uset_spanBackUTF8(const USet *set, const char *s, int32_t length, USetSpanCondition spanCondition) {
michael@0:     return ((UnicodeSet*) set)->UnicodeSet::spanBackUTF8(s, length, spanCondition);
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_equals(const USet* set1, const USet* set2) {
michael@0:     return *(const UnicodeSet*)set1 == *(const UnicodeSet*)set2;
michael@0: }
michael@0: 
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uset_indexOf(const USet* set, UChar32 c) {
michael@0:     return ((UnicodeSet*) set)->UnicodeSet::indexOf(c);
michael@0: }
michael@0: 
michael@0: U_CAPI UChar32 U_EXPORT2
michael@0: uset_charAt(const USet* set, int32_t index) {
michael@0:     return ((UnicodeSet*) set)->UnicodeSet::charAt(index);
michael@0: }
michael@0: 
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uset_size(const USet* set) {
michael@0:     return ((const UnicodeSet*) set)->UnicodeSet::size();
michael@0: }
michael@0: 
michael@0: U_NAMESPACE_BEGIN
michael@0: /**
michael@0:  * This class only exists to provide access to the UnicodeSet private
michael@0:  * USet support API.  Declaring a class a friend is more portable than
michael@0:  * trying to declare extern "C" functions as friends.
michael@0:  */
michael@0: class USetAccess /* not : public UObject because all methods are static */ {
michael@0: public:
michael@0:     /* Try to have the compiler inline these*/
michael@0:     inline static int32_t getStringCount(const UnicodeSet& set) {
michael@0:         return set.getStringCount();
michael@0:     }
michael@0:     inline static const UnicodeString* getString(const UnicodeSet& set,
michael@0:                                                  int32_t i) {
michael@0:         return set.getString(i);
michael@0:     }
michael@0: private:
michael@0:     /* do not instantiate*/
michael@0:     USetAccess();
michael@0: };
michael@0: U_NAMESPACE_END
michael@0: 
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uset_getItemCount(const USet* uset) {
michael@0:     const UnicodeSet& set = *(const UnicodeSet*)uset;
michael@0:     return set.getRangeCount() + USetAccess::getStringCount(set);
michael@0: }
michael@0: 
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uset_getItem(const USet* uset, int32_t itemIndex,
michael@0:              UChar32* start, UChar32* end,
michael@0:              UChar* str, int32_t strCapacity,
michael@0:              UErrorCode* ec) {
michael@0:     if (U_FAILURE(*ec)) return 0;
michael@0:     const UnicodeSet& set = *(const UnicodeSet*)uset;
michael@0:     int32_t rangeCount;
michael@0: 
michael@0:     if (itemIndex < 0) {
michael@0:         *ec = U_ILLEGAL_ARGUMENT_ERROR;
michael@0:         return -1;
michael@0:     } else if (itemIndex < (rangeCount = set.getRangeCount())) {
michael@0:         *start = set.getRangeStart(itemIndex);
michael@0:         *end = set.getRangeEnd(itemIndex);
michael@0:         return 0;
michael@0:     } else {
michael@0:         itemIndex -= rangeCount;
michael@0:         if (itemIndex < USetAccess::getStringCount(set)) {
michael@0:             const UnicodeString* s = USetAccess::getString(set, itemIndex);
michael@0:             return s->extract(str, strCapacity, *ec);
michael@0:         } else {
michael@0:             *ec = U_INDEX_OUTOFBOUNDS_ERROR;
michael@0:             return -1;
michael@0:         }
michael@0:     }
michael@0: }
michael@0: 
michael@0: //U_CAPI int32_t U_EXPORT2
michael@0: //uset_getRangeCount(const USet* set) {
michael@0: //    return ((const UnicodeSet*) set)->getRangeCount();
michael@0: //}
michael@0: //
michael@0: //U_CAPI UBool U_EXPORT2
michael@0: //uset_getRange(const USet* set, int32_t rangeIndex,
michael@0: //              UChar32* pStart, UChar32* pEnd) {
michael@0: //    if ((uint32_t) rangeIndex >= (uint32_t) uset_getRangeCount(set)) {
michael@0: //        return FALSE;
michael@0: //    }
michael@0: //    const UnicodeSet* us = (const UnicodeSet*) set;
michael@0: //    *pStart = us->getRangeStart(rangeIndex);
michael@0: //    *pEnd = us->getRangeEnd(rangeIndex);
michael@0: //    return TRUE;
michael@0: //}
michael@0: 
michael@0: /*
michael@0:  * Serialize a USet into 16-bit units.
michael@0:  * Store BMP code points as themselves with one 16-bit unit each.
michael@0:  *
michael@0:  * Important: the code points in the array are in ascending order,
michael@0:  * therefore all BMP code points precede all supplementary code points.
michael@0:  *
michael@0:  * Store each supplementary code point in 2 16-bit units,
michael@0:  * simply with higher-then-lower 16-bit halfs.
michael@0:  *
michael@0:  * Precede the entire list with the length.
michael@0:  * If there are supplementary code points, then set bit 15 in the length
michael@0:  * and add the bmpLength between it and the array.
michael@0:  *
michael@0:  * In other words:
michael@0:  * - all BMP:            (length=bmpLength) BMP, .., BMP
michael@0:  * - some supplementary: (length|0x8000) (bmpLength<length) BMP, .., BMP, supp-high, supp-low, ..
michael@0:  */
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uset_serialize(const USet* set, uint16_t* dest, int32_t destCapacity, UErrorCode* ec) {
michael@0:     if (ec==NULL || U_FAILURE(*ec)) {
michael@0:         return 0;
michael@0:     }
michael@0: 
michael@0:     return ((const UnicodeSet*) set)->UnicodeSet::serialize(dest, destCapacity,* ec);
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_getSerializedSet(USerializedSet* fillSet, const uint16_t* src, int32_t srcLength) {
michael@0:     int32_t length;
michael@0: 
michael@0:     if(fillSet==NULL) {
michael@0:         return FALSE;
michael@0:     }
michael@0:     if(src==NULL || srcLength<=0) {
michael@0:         fillSet->length=fillSet->bmpLength=0;
michael@0:         return FALSE;
michael@0:     }
michael@0: 
michael@0:     length=*src++;
michael@0:     if(length&0x8000) {
michael@0:         /* there are supplementary values */
michael@0:         length&=0x7fff;
michael@0:         if(srcLength<(2+length)) {
michael@0:             fillSet->length=fillSet->bmpLength=0;
michael@0:             return FALSE;
michael@0:         }
michael@0:         fillSet->bmpLength=*src++;
michael@0:     } else {
michael@0:         /* only BMP values */
michael@0:         if(srcLength<(1+length)) {
michael@0:             fillSet->length=fillSet->bmpLength=0;
michael@0:             return FALSE;
michael@0:         }
michael@0:         fillSet->bmpLength=length;
michael@0:     }
michael@0:     fillSet->array=src;
michael@0:     fillSet->length=length;
michael@0:     return TRUE;
michael@0: }
michael@0: 
michael@0: U_CAPI void U_EXPORT2
michael@0: uset_setSerializedToOne(USerializedSet* fillSet, UChar32 c) {
michael@0:     if(fillSet==NULL || (uint32_t)c>0x10ffff) {
michael@0:         return;
michael@0:     }
michael@0: 
michael@0:     fillSet->array=fillSet->staticArray;
michael@0:     if(c<0xffff) {
michael@0:         fillSet->bmpLength=fillSet->length=2;
michael@0:         fillSet->staticArray[0]=(uint16_t)c;
michael@0:         fillSet->staticArray[1]=(uint16_t)c+1;
michael@0:     } else if(c==0xffff) {
michael@0:         fillSet->bmpLength=1;
michael@0:         fillSet->length=3;
michael@0:         fillSet->staticArray[0]=0xffff;
michael@0:         fillSet->staticArray[1]=1;
michael@0:         fillSet->staticArray[2]=0;
michael@0:     } else if(c<0x10ffff) {
michael@0:         fillSet->bmpLength=0;
michael@0:         fillSet->length=4;
michael@0:         fillSet->staticArray[0]=(uint16_t)(c>>16);
michael@0:         fillSet->staticArray[1]=(uint16_t)c;
michael@0:         ++c;
michael@0:         fillSet->staticArray[2]=(uint16_t)(c>>16);
michael@0:         fillSet->staticArray[3]=(uint16_t)c;
michael@0:     } else /* c==0x10ffff */ {
michael@0:         fillSet->bmpLength=0;
michael@0:         fillSet->length=2;
michael@0:         fillSet->staticArray[0]=0x10;
michael@0:         fillSet->staticArray[1]=0xffff;
michael@0:     }
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_serializedContains(const USerializedSet* set, UChar32 c) {
michael@0:     const uint16_t* array;
michael@0: 
michael@0:     if(set==NULL || (uint32_t)c>0x10ffff) {
michael@0:         return FALSE;
michael@0:     }
michael@0: 
michael@0:     array=set->array;
michael@0:     if(c<=0xffff) {
michael@0:         /* find c in the BMP part */
michael@0:         int32_t lo = 0;
michael@0:         int32_t hi = set->bmpLength-1;
michael@0:         if (c < array[0]) {
michael@0:             hi = 0;
michael@0:         } else if (c < array[hi]) {
michael@0:             for(;;) {
michael@0:                 int32_t i = (lo + hi) >> 1;
michael@0:                 if (i == lo) {
michael@0:                     break;  // Done!
michael@0:                 } else if (c < array[i]) {
michael@0:                     hi = i;
michael@0:                 } else {
michael@0:                     lo = i;
michael@0:                 }
michael@0:             }
michael@0:         } else {
michael@0:             hi += 1;
michael@0:         }
michael@0:         return (UBool)(hi&1);
michael@0:     } else {
michael@0:         /* find c in the supplementary part */
michael@0:         uint16_t high=(uint16_t)(c>>16), low=(uint16_t)c;
michael@0:         int32_t base = set->bmpLength;
michael@0:         int32_t lo = 0;
michael@0:         int32_t hi = set->length - 2 - base;
michael@0:         if (high < array[base] || (high==array[base] && low<array[base+1])) {
michael@0:             hi = 0;
michael@0:         } else if (high < array[base+hi] || (high==array[base+hi] && low<array[base+hi+1])) {
michael@0:             for (;;) {
michael@0:                 int32_t i = ((lo + hi) >> 1) & ~1;  // Guarantee even result
michael@0:                 int32_t iabs = i + base;
michael@0:                 if (i == lo) {
michael@0:                     break;  // Done!
michael@0:                 } else if (high < array[iabs] || (high==array[iabs] && low<array[iabs+1])) {
michael@0:                     hi = i;
michael@0:                 } else {
michael@0:                     lo = i;
michael@0:                 }
michael@0:             }
michael@0:         } else {
michael@0:             hi += 2;
michael@0:         }
michael@0:         /* count pairs of 16-bit units even per BMP and check if the number of pairs is odd */
michael@0:         return (UBool)(((hi+(base<<1))&2)!=0);
michael@0:     }
michael@0: }
michael@0: 
michael@0: U_CAPI int32_t U_EXPORT2
michael@0: uset_getSerializedRangeCount(const USerializedSet* set) {
michael@0:     if(set==NULL) {
michael@0:         return 0;
michael@0:     }
michael@0: 
michael@0:     return (set->bmpLength+(set->length-set->bmpLength)/2+1)/2;
michael@0: }
michael@0: 
michael@0: U_CAPI UBool U_EXPORT2
michael@0: uset_getSerializedRange(const USerializedSet* set, int32_t rangeIndex,
michael@0:                         UChar32* pStart, UChar32* pEnd) {
michael@0:     const uint16_t* array;
michael@0:     int32_t bmpLength, length;
michael@0: 
michael@0:     if(set==NULL || rangeIndex<0 || pStart==NULL || pEnd==NULL) {
michael@0:         return FALSE;
michael@0:     }
michael@0: 
michael@0:     array=set->array;
michael@0:     length=set->length;
michael@0:     bmpLength=set->bmpLength;
michael@0: 
michael@0:     rangeIndex*=2; /* address start/limit pairs */
michael@0:     if(rangeIndex<bmpLength) {
michael@0:         *pStart=array[rangeIndex++];
michael@0:         if(rangeIndex<bmpLength) {
michael@0:             *pEnd=array[rangeIndex]-1;
michael@0:         } else if(rangeIndex<length) {
michael@0:             *pEnd=((((int32_t)array[rangeIndex])<<16)|array[rangeIndex+1])-1;
michael@0:         } else {
michael@0:             *pEnd=0x10ffff;
michael@0:         }
michael@0:         return TRUE;
michael@0:     } else {
michael@0:         rangeIndex-=bmpLength;
michael@0:         rangeIndex*=2; /* address pairs of pairs of units */
michael@0:         length-=bmpLength;
michael@0:         if(rangeIndex<length) {
michael@0:             array+=bmpLength;
michael@0:             *pStart=(((int32_t)array[rangeIndex])<<16)|array[rangeIndex+1];
michael@0:             rangeIndex+=2;
michael@0:             if(rangeIndex<length) {
michael@0:                 *pEnd=((((int32_t)array[rangeIndex])<<16)|array[rangeIndex+1])-1;
michael@0:             } else {
michael@0:                 *pEnd=0x10ffff;
michael@0:             }
michael@0:             return TRUE;
michael@0:         } else {
michael@0:             return FALSE;
michael@0:         }
michael@0:     }
michael@0: }
michael@0: 
michael@0: // TODO The old, internal uset.c had an efficient uset_containsOne function.
michael@0: // Returned the one and only code point, or else -1 or something.
michael@0: // Consider adding such a function to both C and C++ UnicodeSet/uset.
michael@0: // See tools/gennorm/store.c for usage, now usetContainsOne there.
michael@0: 
michael@0: // TODO Investigate incorporating this code into UnicodeSet to improve
michael@0: // efficiency.
michael@0: // ---
michael@0: // #define USET_GROW_DELTA 20
michael@0: // 
michael@0: // static int32_t
michael@0: // findChar(const UChar32* array, int32_t length, UChar32 c) {
michael@0: //     int32_t i;
michael@0: // 
michael@0: //     /* check the last range limit first for more efficient appending */
michael@0: //     if(length>0) {
michael@0: //         if(c>=array[length-1]) {
michael@0: //             return length;
michael@0: //         }
michael@0: // 
michael@0: //         /* do not check the last range limit again in the loop below */
michael@0: //         --length;
michael@0: //     }
michael@0: // 
michael@0: //     for(i=0; i<length && c>=array[i]; ++i) {}
michael@0: //     return i;
michael@0: // }
michael@0: // 
michael@0: // static UBool
michael@0: // addRemove(USet* set, UChar32 c, int32_t doRemove) {
michael@0: //     int32_t i, length, more;
michael@0: // 
michael@0: //     if(set==NULL || (uint32_t)c>0x10ffff) {
michael@0: //         return FALSE;
michael@0: //     }
michael@0: // 
michael@0: //     length=set->length;
michael@0: //     i=findChar(set->array, length, c);
michael@0: //     if((i&1)^doRemove) {
michael@0: //         /* c is already in the set */
michael@0: //         return TRUE;
michael@0: //     }
michael@0: // 
michael@0: //     /* how many more array items do we need? */
michael@0: //     if(i<length && (c+1)==set->array[i]) {
michael@0: //         /* c is just before the following range, extend that in-place by one */
michael@0: //         set->array[i]=c;
michael@0: //         if(i>0) {
michael@0: //             --i;
michael@0: //             if(c==set->array[i]) {
michael@0: //                 /* the previous range collapsed, remove it */
michael@0: //                 set->length=length-=2;
michael@0: //                 if(i<length) {
michael@0: //                     uprv_memmove(set->array+i, set->array+i+2, (length-i)*4);
michael@0: //                 }
michael@0: //             }
michael@0: //         }
michael@0: //         return TRUE;
michael@0: //     } else if(i>0 && c==set->array[i-1]) {
michael@0: //         /* c is just after the previous range, extend that in-place by one */
michael@0: //         if(++c<=0x10ffff) {
michael@0: //             set->array[i-1]=c;
michael@0: //             if(i<length && c==set->array[i]) {
michael@0: //                 /* the following range collapsed, remove it */
michael@0: //                 --i;
michael@0: //                 set->length=length-=2;
michael@0: //                 if(i<length) {
michael@0: //                     uprv_memmove(set->array+i, set->array+i+2, (length-i)*4);
michael@0: //                 }
michael@0: //             }
michael@0: //         } else {
michael@0: //             /* extend the previous range (had limit 0x10ffff) to the end of Unicode */
michael@0: //             set->length=i-1;
michael@0: //         }
michael@0: //         return TRUE;
michael@0: //     } else if(i==length && c==0x10ffff) {
michael@0: //         /* insert one range limit c */
michael@0: //         more=1;
michael@0: //     } else {
michael@0: //         /* insert two range limits c, c+1 */
michael@0: //         more=2;
michael@0: //     }
michael@0: // 
michael@0: //     /* insert <more> range limits */
michael@0: //     if(length+more>set->capacity) {
michael@0: //         /* reallocate */
michael@0: //         int32_t newCapacity=set->capacity+set->capacity/2+USET_GROW_DELTA;
michael@0: //         UChar32* newArray=(UChar32* )uprv_malloc(newCapacity*4);
michael@0: //         if(newArray==NULL) {
michael@0: //             return FALSE;
michael@0: //         }
michael@0: //         set->capacity=newCapacity;
michael@0: //         uprv_memcpy(newArray, set->array, length*4);
michael@0: // 
michael@0: //         if(set->array!=set->staticBuffer) {
michael@0: //             uprv_free(set->array);
michael@0: //         }
michael@0: //         set->array=newArray;
michael@0: //     }
michael@0: // 
michael@0: //     if(i<length) {
michael@0: //         uprv_memmove(set->array+i+more, set->array+i, (length-i)*4);
michael@0: //     }
michael@0: //     set->array[i]=c;
michael@0: //     if(more==2) {
michael@0: //         set->array[i+1]=c+1;
michael@0: //     }
michael@0: //     set->length+=more;
michael@0: // 
michael@0: //     return TRUE;
michael@0: // }
michael@0: // 
michael@0: // U_CAPI UBool U_EXPORT2
michael@0: // uset_add(USet* set, UChar32 c) {
michael@0: //     return addRemove(set, c, 0);
michael@0: // }
michael@0: // 
michael@0: // U_CAPI void U_EXPORT2
michael@0: // uset_remove(USet* set, UChar32 c) {
michael@0: //     addRemove(set, c, 1);
michael@0: // }