com.ibm.icu.impl.coll.CollationDataBuilder Maven / Gradle / Ivy
Show all versions of icu4j Show documentation
// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
*******************************************************************************
* Copyright (C) 2012-2015, International Business Machines
* Corporation and others. All Rights Reserved.
*******************************************************************************
* CollationDataBuilder.java, ported from collationdatabuilder.h/.cpp
*
* C++ version created on: 2012apr01
* created by: Markus W. Scherer
*/
package com.ibm.icu.impl.coll;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Iterator;
import com.ibm.icu.impl.Norm2AllModes;
import com.ibm.icu.impl.Normalizer2Impl;
import com.ibm.icu.impl.Normalizer2Impl.Hangul;
import com.ibm.icu.impl.Trie2;
import com.ibm.icu.impl.Trie2Writable;
import com.ibm.icu.lang.UCharacter;
import com.ibm.icu.text.UnicodeSet;
import com.ibm.icu.text.UnicodeSetIterator;
import com.ibm.icu.util.CharsTrie;
import com.ibm.icu.util.CharsTrieBuilder;
import com.ibm.icu.util.StringTrieBuilder;
/**
* Low-level CollationData builder.
* Takes (character, CE) pairs and builds them into runtime data structures.
* Supports characters with context prefixes and contraction suffixes.
*/
final class CollationDataBuilder { // not final in C++
/**
* Collation element modifier. Interface class for a modifier
* that changes a tailoring builder's temporary CEs to final CEs.
* Called for every non-special CE32 and every expansion CE.
*/
interface CEModifier {
/** Returns a new CE to replace the non-special input CE32, or else Collation.NO_CE. */
long modifyCE32(int ce32);
/** Returns a new CE to replace the input CE, or else Collation.NO_CE. */
long modifyCE(long ce);
}
CollationDataBuilder() {
nfcImpl = Norm2AllModes.getNFCInstance().impl;
base = null;
baseSettings = null;
trie = null;
ce32s = new UVector32();
ce64s = new UVector64();
conditionalCE32s = new ArrayList();
modified = false;
fastLatinEnabled = false;
fastLatinBuilder = null;
collIter = null;
// Reserve the first CE32 for U+0000.
ce32s.addElement(0);
}
void initForTailoring(CollationData b) {
if(trie != null) {
throw new IllegalStateException("attempt to reuse a CollationDataBuilder");
}
if(b == null) {
throw new IllegalArgumentException("null CollationData");
}
base = b;
// For a tailoring, the default is to fall back to the base.
trie = new Trie2Writable(Collation.FALLBACK_CE32, Collation.FFFD_CE32);
// Set the Latin-1 letters block so that it is allocated first in the data array,
// to try to improve locality of reference when sorting Latin-1 text.
// Do not use utrie2_setRange32() since that will not actually allocate blocks
// that are filled with the default value.
// ASCII (0..7F) is already preallocated anyway.
for(int c = 0xc0; c <= 0xff; ++c) {
trie.set(c, Collation.FALLBACK_CE32);
}
// Hangul syllables are not tailorable (except via tailoring Jamos).
// Always set the Hangul tag to help performance.
// Do this here, rather than in buildMappings(),
// so that we see the HANGUL_TAG in various assertions.
int hangulCE32 = Collation.makeCE32FromTagAndIndex(Collation.HANGUL_TAG, 0);
trie.setRange(Hangul.HANGUL_BASE, Hangul.HANGUL_END, hangulCE32, true);
// Copy the set contents but don't copy/clone the set as a whole because
// that would copy the isFrozen state too.
unsafeBackwardSet.addAll(b.unsafeBackwardSet);
}
boolean isCompressibleLeadByte(int b) {
return base.isCompressibleLeadByte(b);
}
boolean isCompressiblePrimary(long p) {
return isCompressibleLeadByte((int)p >>> 24);
}
/**
* @return true if this builder has mappings (e.g., add() has been called)
*/
boolean hasMappings() { return modified; }
/**
* @return true if c has CEs in this builder
*/
boolean isAssigned(int c) {
return Collation.isAssignedCE32(trie.get(c));
}
void add(CharSequence prefix, CharSequence s, long ces[], int cesLength) {
int ce32 = encodeCEs(ces, cesLength);
addCE32(prefix, s, ce32);
}
/**
* Encodes the ces as either the returned ce32 by itself,
* or by storing an expansion, with the returned ce32 referring to that.
*
* add(p, s, ces, cesLength) = addCE32(p, s, encodeCEs(ces, cesLength))
*/
int encodeCEs(long ces[], int cesLength) {
if(cesLength < 0 || cesLength > Collation.MAX_EXPANSION_LENGTH) {
throw new IllegalArgumentException("mapping to too many CEs");
}
if(!isMutable()) {
throw new IllegalStateException("attempt to add mappings after build()");
}
if(cesLength == 0) {
// Convenience: We cannot map to nothing, but we can map to a completely ignorable CE.
// Do this here so that callers need not do it.
return encodeOneCEAsCE32(0);
} else if(cesLength == 1) {
return encodeOneCE(ces[0]);
} else if(cesLength == 2) {
// Try to encode two CEs as one CE32.
long ce0 = ces[0];
long ce1 = ces[1];
long p0 = ce0 >>> 32;
if((ce0 & 0xffffffffff00ffL) == Collation.COMMON_SECONDARY_CE &&
(ce1 & 0xffffffff00ffffffL) == Collation.COMMON_TERTIARY_CE &&
p0 != 0) {
// Latin mini expansion
return
(int)p0 |
(((int)ce0 & 0xff00) << 8) |
(((int)ce1 >> 16) & 0xff00) |
Collation.SPECIAL_CE32_LOW_BYTE |
Collation.LATIN_EXPANSION_TAG;
}
}
// Try to encode two or more CEs as CE32s.
int[] newCE32s = new int[Collation.MAX_EXPANSION_LENGTH]; // TODO: instance field?
for(int i = 0;; ++i) {
if(i == cesLength) {
return encodeExpansion32(newCE32s, 0, cesLength);
}
int ce32 = encodeOneCEAsCE32(ces[i]);
if(ce32 == Collation.NO_CE32) { break; }
newCE32s[i] = ce32;
}
return encodeExpansion(ces, 0, cesLength);
}
void addCE32(CharSequence prefix, CharSequence s, int ce32) {
if(s.length() == 0) {
throw new IllegalArgumentException("mapping from empty string");
}
if(!isMutable()) {
throw new IllegalStateException("attempt to add mappings after build()");
}
int c = Character.codePointAt(s, 0);
int cLength = Character.charCount(c);
int oldCE32 = trie.get(c);
boolean hasContext = prefix.length() != 0|| s.length() > cLength;
if(oldCE32 == Collation.FALLBACK_CE32) {
// First tailoring for c.
// If c has contextual base mappings or if we add a contextual mapping,
// then copy the base mappings.
// Otherwise we just override the base mapping.
int baseCE32 = base.getFinalCE32(base.getCE32(c));
if(hasContext || Collation.ce32HasContext(baseCE32)) {
oldCE32 = copyFromBaseCE32(c, baseCE32, true);
trie.set(c, oldCE32);
}
}
if(!hasContext) {
// No prefix, no contraction.
if(!isBuilderContextCE32(oldCE32)) {
trie.set(c, ce32);
} else {
ConditionalCE32 cond = getConditionalCE32ForCE32(oldCE32);
cond.builtCE32 = Collation.NO_CE32;
cond.ce32 = ce32;
}
} else {
ConditionalCE32 cond;
if(!isBuilderContextCE32(oldCE32)) {
// Replace the simple oldCE32 with a builder context CE32
// pointing to a new ConditionalCE32 list head.
int index = addConditionalCE32("\0", oldCE32);
int contextCE32 = makeBuilderContextCE32(index);
trie.set(c, contextCE32);
contextChars.add(c);
cond = getConditionalCE32(index);
} else {
cond = getConditionalCE32ForCE32(oldCE32);
cond.builtCE32 = Collation.NO_CE32;
}
CharSequence suffix = s.subSequence(cLength, s.length());
String context = new StringBuilder().append((char)prefix.length()).
append(prefix).append(suffix).toString();
unsafeBackwardSet.addAll(suffix);
for(;;) {
// invariant: context > cond.context
int next = cond.next;
if(next < 0) {
// Append a new ConditionalCE32 after cond.
int index = addConditionalCE32(context, ce32);
cond.next = index;
break;
}
ConditionalCE32 nextCond = getConditionalCE32(next);
int cmp = context.compareTo(nextCond.context);
if(cmp < 0) {
// Insert a new ConditionalCE32 between cond and nextCond.
int index = addConditionalCE32(context, ce32);
cond.next = index;
getConditionalCE32(index).next = next;
break;
} else if(cmp == 0) {
// Same context as before, overwrite its ce32.
nextCond.ce32 = ce32;
break;
}
cond = nextCond;
}
}
modified = true;
}
/**
* Copies all mappings from the src builder, with modifications.
* This builder here must not be built yet, and should be empty.
*/
void copyFrom(CollationDataBuilder src, CEModifier modifier) {
if(!isMutable()) {
throw new IllegalStateException("attempt to copyFrom() after build()");
}
CopyHelper helper = new CopyHelper(src, this, modifier);
Iterator trieIterator = src.trie.iterator();
Trie2.Range range;
while(trieIterator.hasNext() && !(range = trieIterator.next()).leadSurrogate) {
enumRangeForCopy(range.startCodePoint, range.endCodePoint, range.value, helper);
}
// Update the contextChars and the unsafeBackwardSet while copying,
// in case a character had conditional mappings in the source builder
// and they were removed later.
modified |= src.modified;
}
void optimize(UnicodeSet set) {
if(set.isEmpty()) { return; }
UnicodeSetIterator iter = new UnicodeSetIterator(set);
while(iter.next() && iter.codepoint != UnicodeSetIterator.IS_STRING) {
int c = iter.codepoint;
int ce32 = trie.get(c);
if(ce32 == Collation.FALLBACK_CE32) {
ce32 = base.getFinalCE32(base.getCE32(c));
ce32 = copyFromBaseCE32(c, ce32, true);
trie.set(c, ce32);
}
}
modified = true;
}
void suppressContractions(UnicodeSet set) {
if(set.isEmpty()) { return; }
UnicodeSetIterator iter = new UnicodeSetIterator(set);
while(iter.next() && iter.codepoint != UnicodeSetIterator.IS_STRING) {
int c = iter.codepoint;
int ce32 = trie.get(c);
if(ce32 == Collation.FALLBACK_CE32) {
ce32 = base.getFinalCE32(base.getCE32(c));
if(Collation.ce32HasContext(ce32)) {
ce32 = copyFromBaseCE32(c, ce32, false /* without context */);
trie.set(c, ce32);
}
} else if(isBuilderContextCE32(ce32)) {
ce32 = getConditionalCE32ForCE32(ce32).ce32;
// Simply abandon the list of ConditionalCE32.
// The caller will copy this builder in the end,
// eliminating unreachable data.
trie.set(c, ce32);
contextChars.remove(c);
}
}
modified = true;
}
void enableFastLatin() { fastLatinEnabled = true; }
void build(CollationData data) {
buildMappings(data);
if(base != null) {
data.numericPrimary = base.numericPrimary;
data.compressibleBytes = base.compressibleBytes;
data.numScripts = base.numScripts;
data.scriptsIndex = base.scriptsIndex;
data.scriptStarts = base.scriptStarts;
}
buildFastLatinTable(data);
}
/**
* Looks up CEs for s and appends them to the ces array.
* Does not handle normalization: s should be in FCD form.
*
* Does not write completely ignorable CEs.
* Does not write beyond Collation.MAX_EXPANSION_LENGTH.
*
* @return incremented cesLength
*/
int getCEs(CharSequence s, long ces[], int cesLength) {
return getCEs(s, 0, ces, cesLength);
}
int getCEs(CharSequence prefix, CharSequence s, long ces[], int cesLength) {
int prefixLength = prefix.length();
if(prefixLength == 0) {
return getCEs(s, 0, ces, cesLength);
} else {
return getCEs(new StringBuilder(prefix).append(s), prefixLength, ces, cesLength);
}
}
/**
* Build-time context and CE32 for a code point.
* If a code point has contextual mappings, then the default (no-context) mapping
* and all conditional mappings are stored in a singly-linked list
* of ConditionalCE32, sorted by context strings.
*
* Context strings sort by prefix length, then by prefix, then by contraction suffix.
* Context strings must be unique and in ascending order.
*/
private static final class ConditionalCE32 {
ConditionalCE32(String ct, int ce) {
context = ct;
ce32 = ce;
defaultCE32 = Collation.NO_CE32;
builtCE32 = Collation.NO_CE32;
next = -1;
}
boolean hasContext() { return context.length() > 1; }
int prefixLength() { return context.charAt(0); }
/**
* "\0" for the first entry for any code point, with its default CE32.
*
* Otherwise one unit with the length of the prefix string,
* then the prefix string, then the contraction suffix.
*/
String context;
/**
* CE32 for the code point and its context.
* Can be special (e.g., for an expansion) but not contextual (prefix or contraction tag).
*/
int ce32;
/**
* Default CE32 for all contexts with this same prefix.
* Initially NO_CE32. Set only while building runtime data structures,
* and only on one of the nodes of a sub-list with the same prefix.
*/
int defaultCE32;
/**
* CE32 for the built contexts.
* When fetching CEs from the builder, the contexts are built into their runtime form
* so that the normal collation implementation can process them.
* The result is cached in the list head. It is reset when the contexts are modified.
*/
int builtCE32;
/**
* Index of the next ConditionalCE32.
* Negative for the end of the list.
*/
int next;
}
protected int getCE32FromOffsetCE32(boolean fromBase, int c, int ce32) {
int i = Collation.indexFromCE32(ce32);
long dataCE = fromBase ? base.ces[i] : ce64s.elementAti(i);
long p = Collation.getThreeBytePrimaryForOffsetData(c, dataCE);
return Collation.makeLongPrimaryCE32(p);
}
protected int addCE(long ce) {
int length = ce64s.size();
for(int i = 0; i < length; ++i) {
if(ce == ce64s.elementAti(i)) { return i; }
}
ce64s.addElement(ce);
return length;
}
protected int addCE32(int ce32) {
int length = ce32s.size();
for(int i = 0; i < length; ++i) {
if(ce32 == ce32s.elementAti(i)) { return i; }
}
ce32s.addElement(ce32);
return length;
}
protected int addConditionalCE32(String context, int ce32) {
assert(context.length() != 0);
int index = conditionalCE32s.size();
if(index > Collation.MAX_INDEX) {
throw new IndexOutOfBoundsException("too many context-sensitive mappings");
// BufferOverflowException is a better fit
// but cannot be constructed with a message string.
}
ConditionalCE32 cond = new ConditionalCE32(context, ce32);
conditionalCE32s.add(cond);
return index;
}
protected ConditionalCE32 getConditionalCE32(int index) {
return conditionalCE32s.get(index);
}
protected ConditionalCE32 getConditionalCE32ForCE32(int ce32) {
return getConditionalCE32(Collation.indexFromCE32(ce32));
}
protected static int makeBuilderContextCE32(int index) {
return Collation.makeCE32FromTagAndIndex(Collation.BUILDER_DATA_TAG, index);
}
protected static boolean isBuilderContextCE32(int ce32) {
return Collation.hasCE32Tag(ce32, Collation.BUILDER_DATA_TAG);
}
protected static int encodeOneCEAsCE32(long ce) {
long p = ce >>> 32;
int lower32 = (int)ce;
int t = lower32 & 0xffff;
assert((t & 0xc000) != 0xc000); // Impossible case bits 11 mark special CE32s.
if((ce & 0xffff00ff00ffL) == 0) {
// normal form ppppsstt
return (int)p | (lower32 >>> 16) | (t >> 8);
} else if((ce & 0xffffffffffL) == Collation.COMMON_SEC_AND_TER_CE) {
// long-primary form ppppppC1
return Collation.makeLongPrimaryCE32(p);
} else if(p == 0 && (t & 0xff) == 0) {
// long-secondary form ssssttC2
return Collation.makeLongSecondaryCE32(lower32);
}
return Collation.NO_CE32;
}
protected int encodeOneCE(long ce) {
// Try to encode one CE as one CE32.
int ce32 = encodeOneCEAsCE32(ce);
if(ce32 != Collation.NO_CE32) { return ce32; }
int index = addCE(ce);
if(index > Collation.MAX_INDEX) {
throw new IndexOutOfBoundsException("too many mappings");
// BufferOverflowException is a better fit
// but cannot be constructed with a message string.
}
return Collation.makeCE32FromTagIndexAndLength(Collation.EXPANSION_TAG, index, 1);
}
protected int encodeExpansion(long ces[], int start, int length) {
// See if this sequence of CEs has already been stored.
long first = ces[start];
int ce64sMax = ce64s.size() - length;
for(int i = 0; i <= ce64sMax; ++i) {
if(first == ce64s.elementAti(i)) {
if(i > Collation.MAX_INDEX) {
throw new IndexOutOfBoundsException("too many mappings");
// BufferOverflowException is a better fit
// but cannot be constructed with a message string.
}
for(int j = 1;; ++j) {
if(j == length) {
return Collation.makeCE32FromTagIndexAndLength(
Collation.EXPANSION_TAG, i, length);
}
if(ce64s.elementAti(i + j) != ces[start + j]) { break; }
}
}
}
// Store the new sequence.
int i = ce64s.size();
if(i > Collation.MAX_INDEX) {
throw new IndexOutOfBoundsException("too many mappings");
// BufferOverflowException is a better fit
// but cannot be constructed with a message string.
}
for(int j = 0; j < length; ++j) {
ce64s.addElement(ces[start + j]);
}
return Collation.makeCE32FromTagIndexAndLength(Collation.EXPANSION_TAG, i, length);
}
protected int encodeExpansion32(int newCE32s[], int start, int length) {
// See if this sequence of CE32s has already been stored.
int first = newCE32s[start];
int ce32sMax = ce32s.size() - length;
for(int i = 0; i <= ce32sMax; ++i) {
if(first == ce32s.elementAti(i)) {
if(i > Collation.MAX_INDEX) {
throw new IndexOutOfBoundsException("too many mappings");
// BufferOverflowException is a better fit
// but cannot be constructed with a message string.
}
for(int j = 1;; ++j) {
if(j == length) {
return Collation.makeCE32FromTagIndexAndLength(
Collation.EXPANSION32_TAG, i, length);
}
if(ce32s.elementAti(i + j) != newCE32s[start + j]) { break; }
}
}
}
// Store the new sequence.
int i = ce32s.size();
if(i > Collation.MAX_INDEX) {
throw new IndexOutOfBoundsException("too many mappings");
// BufferOverflowException is a better fit
// but cannot be constructed with a message string.
}
for(int j = 0; j < length; ++j) {
ce32s.addElement(newCE32s[start + j]);
}
return Collation.makeCE32FromTagIndexAndLength(Collation.EXPANSION32_TAG, i, length);
}
protected int copyFromBaseCE32(int c, int ce32, boolean withContext) {
if(!Collation.isSpecialCE32(ce32)) { return ce32; }
switch(Collation.tagFromCE32(ce32)) {
case Collation.LONG_PRIMARY_TAG:
case Collation.LONG_SECONDARY_TAG:
case Collation.LATIN_EXPANSION_TAG:
// copy as is
break;
case Collation.EXPANSION32_TAG: {
int index = Collation.indexFromCE32(ce32);
int length = Collation.lengthFromCE32(ce32);
ce32 = encodeExpansion32(base.ce32s, index, length);
break;
}
case Collation.EXPANSION_TAG: {
int index = Collation.indexFromCE32(ce32);
int length = Collation.lengthFromCE32(ce32);
ce32 = encodeExpansion(base.ces, index, length);
break;
}
case Collation.PREFIX_TAG: {
// Flatten prefixes and nested suffixes (contractions)
// into a linear list of ConditionalCE32.
int trieIndex = Collation.indexFromCE32(ce32);
ce32 = base.getCE32FromContexts(trieIndex); // Default if no prefix match.
if(!withContext) {
return copyFromBaseCE32(c, ce32, false);
}
ConditionalCE32 head = new ConditionalCE32("", 0);
StringBuilder context = new StringBuilder("\0");
int index;
if(Collation.isContractionCE32(ce32)) {
index = copyContractionsFromBaseCE32(context, c, ce32, head);
} else {
ce32 = copyFromBaseCE32(c, ce32, true);
head.next = index = addConditionalCE32(context.toString(), ce32);
}
ConditionalCE32 cond = getConditionalCE32(index); // the last ConditionalCE32 so far
CharsTrie.Iterator prefixes = CharsTrie.iterator(base.contexts, trieIndex + 2, 0);
while(prefixes.hasNext()) {
CharsTrie.Entry entry = prefixes.next();
context.setLength(0);
context.append(entry.chars).reverse().insert(0, (char)entry.chars.length());
ce32 = entry.value;
if(Collation.isContractionCE32(ce32)) {
index = copyContractionsFromBaseCE32(context, c, ce32, cond);
} else {
ce32 = copyFromBaseCE32(c, ce32, true);
cond.next = index = addConditionalCE32(context.toString(), ce32);
}
cond = getConditionalCE32(index);
}
ce32 = makeBuilderContextCE32(head.next);
contextChars.add(c);
break;
}
case Collation.CONTRACTION_TAG: {
if(!withContext) {
int index = Collation.indexFromCE32(ce32);
ce32 = base.getCE32FromContexts(index); // Default if no suffix match.
return copyFromBaseCE32(c, ce32, false);
}
ConditionalCE32 head = new ConditionalCE32("", 0);
StringBuilder context = new StringBuilder("\0");
copyContractionsFromBaseCE32(context, c, ce32, head);
ce32 = makeBuilderContextCE32(head.next);
contextChars.add(c);
break;
}
case Collation.HANGUL_TAG:
throw new UnsupportedOperationException("We forbid tailoring of Hangul syllables.");
case Collation.OFFSET_TAG:
ce32 = getCE32FromOffsetCE32(true, c, ce32);
break;
case Collation.IMPLICIT_TAG:
ce32 = encodeOneCE(Collation.unassignedCEFromCodePoint(c));
break;
default:
throw new AssertionError("copyFromBaseCE32(c, ce32, withContext) " +
"requires ce32 == base.getFinalCE32(ce32)");
}
return ce32;
}
/**
* Copies base contractions to a list of ConditionalCE32.
* Sets cond.next to the index of the first new item
* and returns the index of the last new item.
*/
protected int copyContractionsFromBaseCE32(StringBuilder context, int c, int ce32,
ConditionalCE32 cond) {
int trieIndex = Collation.indexFromCE32(ce32);
int index;
if((ce32 & Collation.CONTRACT_SINGLE_CP_NO_MATCH) != 0) {
// No match on the single code point.
// We are underneath a prefix, and the default mapping is just
// a fallback to the mappings for a shorter prefix.
assert(context.length() > 1);
index = -1;
} else {
ce32 = base.getCE32FromContexts(trieIndex); // Default if no suffix match.
assert(!Collation.isContractionCE32(ce32));
ce32 = copyFromBaseCE32(c, ce32, true);
cond.next = index = addConditionalCE32(context.toString(), ce32);
cond = getConditionalCE32(index);
}
int suffixStart = context.length();
CharsTrie.Iterator suffixes = CharsTrie.iterator(base.contexts, trieIndex + 2, 0);
while(suffixes.hasNext()) {
CharsTrie.Entry entry = suffixes.next();
context.append(entry.chars);
ce32 = copyFromBaseCE32(c, entry.value, true);
cond.next = index = addConditionalCE32(context.toString(), ce32);
// No need to update the unsafeBackwardSet because the tailoring set
// is already a copy of the base set.
cond = getConditionalCE32(index);
context.setLength(suffixStart);
}
assert(index >= 0);
return index;
}
private static final class CopyHelper {
CopyHelper(CollationDataBuilder s, CollationDataBuilder d,
CollationDataBuilder.CEModifier m) {
src = s;
dest = d;
modifier = m;
}
void copyRangeCE32(int start, int end, int ce32) {
ce32 = copyCE32(ce32);
dest.trie.setRange(start, end, ce32, true);
if(CollationDataBuilder.isBuilderContextCE32(ce32)) {
dest.contextChars.add(start, end);
}
}
int copyCE32(int ce32) {
if(!Collation.isSpecialCE32(ce32)) {
long ce = modifier.modifyCE32(ce32);
if(ce != Collation.NO_CE) {
ce32 = dest.encodeOneCE(ce);
}
} else {
int tag = Collation.tagFromCE32(ce32);
if(tag == Collation.EXPANSION32_TAG) {
int[] srcCE32s = src.ce32s.getBuffer();
int srcIndex = Collation.indexFromCE32(ce32);
int length = Collation.lengthFromCE32(ce32);
// Inspect the source CE32s. Just copy them if none are modified.
// Otherwise copy to modifiedCEs, with modifications.
boolean isModified = false;
for(int i = 0; i < length; ++i) {
ce32 = srcCE32s[srcIndex + i];
long ce;
if(Collation.isSpecialCE32(ce32) ||
(ce = modifier.modifyCE32(ce32)) == Collation.NO_CE) {
if(isModified) {
modifiedCEs[i] = Collation.ceFromCE32(ce32);
}
} else {
if(!isModified) {
for(int j = 0; j < i; ++j) {
modifiedCEs[j] = Collation.ceFromCE32(srcCE32s[srcIndex + j]);
}
isModified = true;
}
modifiedCEs[i] = ce;
}
}
if(isModified) {
ce32 = dest.encodeCEs(modifiedCEs, length);
} else {
ce32 = dest.encodeExpansion32(srcCE32s, srcIndex, length);
}
} else if(tag == Collation.EXPANSION_TAG) {
long[] srcCEs = src.ce64s.getBuffer();
int srcIndex = Collation.indexFromCE32(ce32);
int length = Collation.lengthFromCE32(ce32);
// Inspect the source CEs. Just copy them if none are modified.
// Otherwise copy to modifiedCEs, with modifications.
boolean isModified = false;
for(int i = 0; i < length; ++i) {
long srcCE = srcCEs[srcIndex + i];
long ce = modifier.modifyCE(srcCE);
if(ce == Collation.NO_CE) {
if(isModified) {
modifiedCEs[i] = srcCE;
}
} else {
if(!isModified) {
for(int j = 0; j < i; ++j) {
modifiedCEs[j] = srcCEs[srcIndex + j];
}
isModified = true;
}
modifiedCEs[i] = ce;
}
}
if(isModified) {
ce32 = dest.encodeCEs(modifiedCEs, length);
} else {
ce32 = dest.encodeExpansion(srcCEs, srcIndex, length);
}
} else if(tag == Collation.BUILDER_DATA_TAG) {
// Copy the list of ConditionalCE32.
ConditionalCE32 cond = src.getConditionalCE32ForCE32(ce32);
assert(!cond.hasContext());
int destIndex = dest.addConditionalCE32(
cond.context, copyCE32(cond.ce32));
ce32 = CollationDataBuilder.makeBuilderContextCE32(destIndex);
while(cond.next >= 0) {
cond = src.getConditionalCE32(cond.next);
ConditionalCE32 prevDestCond = dest.getConditionalCE32(destIndex);
destIndex = dest.addConditionalCE32(
cond.context, copyCE32(cond.ce32));
int suffixStart = cond.prefixLength() + 1;
dest.unsafeBackwardSet.addAll(cond.context.substring(suffixStart));
prevDestCond.next = destIndex;
}
} else {
// Just copy long CEs and Latin mini expansions (and other expected values) as is,
// assuming that the modifier would not modify them.
assert(tag == Collation.LONG_PRIMARY_TAG ||
tag == Collation.LONG_SECONDARY_TAG ||
tag == Collation.LATIN_EXPANSION_TAG ||
tag == Collation.HANGUL_TAG);
}
}
return ce32;
}
CollationDataBuilder src;
CollationDataBuilder dest;
CollationDataBuilder.CEModifier modifier;
long[] modifiedCEs = new long[Collation.MAX_EXPANSION_LENGTH];
}
private static void
enumRangeForCopy(int start, int end, int value, CopyHelper helper) {
if(value != Collation.UNASSIGNED_CE32 && value != Collation.FALLBACK_CE32) {
helper.copyRangeCE32(start, end, value);
}
}
protected boolean getJamoCE32s(int jamoCE32s[]) {
boolean anyJamoAssigned = base == null; // always set jamoCE32s in the base data
boolean needToCopyFromBase = false;
for(int j = 0; j < CollationData.JAMO_CE32S_LENGTH; ++j) { // Count across Jamo types.
int jamo = jamoCpFromIndex(j);
boolean fromBase = false;
int ce32 = trie.get(jamo);
anyJamoAssigned |= Collation.isAssignedCE32(ce32);
// TODO: Try to prevent [optimize [Jamo]] from counting as anyJamoAssigned.
// (As of CLDR 24 [2013] the Korean tailoring does not optimize conjoining Jamo.)
if(ce32 == Collation.FALLBACK_CE32) {
fromBase = true;
ce32 = base.getCE32(jamo);
}
if(Collation.isSpecialCE32(ce32)) {
switch(Collation.tagFromCE32(ce32)) {
case Collation.LONG_PRIMARY_TAG:
case Collation.LONG_SECONDARY_TAG:
case Collation.LATIN_EXPANSION_TAG:
// Copy the ce32 as-is.
break;
case Collation.EXPANSION32_TAG:
case Collation.EXPANSION_TAG:
case Collation.PREFIX_TAG:
case Collation.CONTRACTION_TAG:
if(fromBase) {
// Defer copying until we know if anyJamoAssigned.
ce32 = Collation.FALLBACK_CE32;
needToCopyFromBase = true;
}
break;
case Collation.IMPLICIT_TAG:
// An unassigned Jamo should only occur in tests with incomplete bases.
assert(fromBase);
ce32 = Collation.FALLBACK_CE32;
needToCopyFromBase = true;
break;
case Collation.OFFSET_TAG:
ce32 = getCE32FromOffsetCE32(fromBase, jamo, ce32);
break;
case Collation.FALLBACK_TAG:
case Collation.RESERVED_TAG_3:
case Collation.BUILDER_DATA_TAG:
case Collation.DIGIT_TAG:
case Collation.U0000_TAG:
case Collation.HANGUL_TAG:
case Collation.LEAD_SURROGATE_TAG:
throw new AssertionError(String.format("unexpected special tag in ce32=0x%08x", ce32));
}
}
jamoCE32s[j] = ce32;
}
if(anyJamoAssigned && needToCopyFromBase) {
for(int j = 0; j < CollationData.JAMO_CE32S_LENGTH; ++j) {
if(jamoCE32s[j] == Collation.FALLBACK_CE32) {
int jamo = jamoCpFromIndex(j);
jamoCE32s[j] = copyFromBaseCE32(jamo, base.getCE32(jamo),
/*withContext=*/ true);
}
}
}
return anyJamoAssigned;
}
protected void setDigitTags() {
UnicodeSet digits = new UnicodeSet("[:Nd:]");
UnicodeSetIterator iter = new UnicodeSetIterator(digits);
while(iter.next()) {
assert(iter.codepoint != UnicodeSetIterator.IS_STRING);
int c = iter.codepoint;
int ce32 = trie.get(c);
if(ce32 != Collation.FALLBACK_CE32 && ce32 != Collation.UNASSIGNED_CE32) {
int index = addCE32(ce32);
if(index > Collation.MAX_INDEX) {
throw new IndexOutOfBoundsException("too many mappings");
// BufferOverflowException is a better fit
// but cannot be constructed with a message string.
}
ce32 = Collation.makeCE32FromTagIndexAndLength(
Collation.DIGIT_TAG, index, UCharacter.digit(c)); // u_charDigitValue(c)
trie.set(c, ce32);
}
}
}
protected void setLeadSurrogates() {
for(char lead = 0xd800; lead < 0xdc00; ++lead) {
int leadValue = -1;
// utrie2_enumForLeadSurrogate(trie, lead, null, , &value);
Iterator trieIterator = trie.iteratorForLeadSurrogate(lead);
while(trieIterator.hasNext()) {
Trie2.Range range = trieIterator.next();
// The rest of this loop is equivalent to C++ enumRangeLeadValue().
int value = range.value;
if(value == Collation.UNASSIGNED_CE32) {
value = Collation.LEAD_ALL_UNASSIGNED;
} else if(value == Collation.FALLBACK_CE32) {
value = Collation.LEAD_ALL_FALLBACK;
} else {
leadValue = Collation.LEAD_MIXED;
break;
}
if(leadValue < 0) {
leadValue = value;
} else if(leadValue != value) {
leadValue = Collation.LEAD_MIXED;
break;
}
}
trie.setForLeadSurrogateCodeUnit(lead,
Collation.makeCE32FromTagAndIndex(Collation.LEAD_SURROGATE_TAG, 0) | leadValue);
}
}
protected void buildMappings(CollationData data) {
if(!isMutable()) {
throw new IllegalStateException("attempt to build() after build()");
}
buildContexts();
int[] jamoCE32s = new int[CollationData.JAMO_CE32S_LENGTH];
int jamoIndex = -1;
if(getJamoCE32s(jamoCE32s)) {
jamoIndex = ce32s.size();
for(int i = 0; i < CollationData.JAMO_CE32S_LENGTH; ++i) {
ce32s.addElement(jamoCE32s[i]);
}
// Small optimization: Use a bit in the Hangul ce32
// to indicate that none of the Jamo CE32s are isSpecialCE32()
// (as it should be in the root collator).
// It allows CollationIterator to avoid recursive function calls and per-Jamo tests.
// In order to still have good trie compression and keep this code simple,
// we only set this flag if a whole block of 588 Hangul syllables starting with
// a common leading consonant (Jamo L) has this property.
boolean isAnyJamoVTSpecial = false;
for(int i = Hangul.JAMO_L_COUNT; i < CollationData.JAMO_CE32S_LENGTH; ++i) {
if(Collation.isSpecialCE32(jamoCE32s[i])) {
isAnyJamoVTSpecial = true;
break;
}
}
int hangulCE32 = Collation.makeCE32FromTagAndIndex(Collation.HANGUL_TAG, 0);
int c = Hangul.HANGUL_BASE;
for(int i = 0; i < Hangul.JAMO_L_COUNT; ++i) { // iterate over the Jamo L
int ce32 = hangulCE32;
if(!isAnyJamoVTSpecial && !Collation.isSpecialCE32(jamoCE32s[i])) {
ce32 |= Collation.HANGUL_NO_SPECIAL_JAMO;
}
int limit = c + Hangul.JAMO_VT_COUNT;
trie.setRange(c, limit - 1, ce32, true);
c = limit;
}
} else {
// Copy the Hangul CE32s from the base in blocks per Jamo L,
// assuming that HANGUL_NO_SPECIAL_JAMO is set or not set for whole blocks.
for(int c = Hangul.HANGUL_BASE; c < Hangul.HANGUL_LIMIT;) {
int ce32 = base.getCE32(c);
assert(Collation.hasCE32Tag(ce32, Collation.HANGUL_TAG));
int limit = c + Hangul.JAMO_VT_COUNT;
trie.setRange(c, limit - 1, ce32, true);
c = limit;
}
}
setDigitTags();
setLeadSurrogates();
// For U+0000, move its normal ce32 into CE32s[0] and set U0000_TAG.
ce32s.setElementAt(trie.get(0), 0);
trie.set(0, Collation.makeCE32FromTagAndIndex(Collation.U0000_TAG, 0));
data.trie = trie.toTrie2_32();
// Mark each lead surrogate as "unsafe"
// if any of its 1024 associated supplementary code points is "unsafe".
int c = 0x10000;
for(char lead = 0xd800; lead < 0xdc00; ++lead, c += 0x400) {
if(unsafeBackwardSet.containsSome(c, c + 0x3ff)) {
unsafeBackwardSet.add(lead);
}
}
unsafeBackwardSet.freeze();
data.ce32s = ce32s.getBuffer();
data.ces = ce64s.getBuffer();
data.contexts = contexts.toString();
data.base = base;
if(jamoIndex >= 0) {
data.jamoCE32s = jamoCE32s; // C++: data.ce32s + jamoIndex
} else {
data.jamoCE32s = base.jamoCE32s;
}
data.unsafeBackwardSet = unsafeBackwardSet;
}
protected void clearContexts() {
contexts.setLength(0);
UnicodeSetIterator iter = new UnicodeSetIterator(contextChars);
while(iter.next()) {
assert(iter.codepoint != UnicodeSetIterator.IS_STRING);
int ce32 = trie.get(iter.codepoint);
assert(isBuilderContextCE32(ce32));
getConditionalCE32ForCE32(ce32).builtCE32 = Collation.NO_CE32;
}
}
protected void buildContexts() {
// Ignore abandoned lists and the cached builtCE32,
// and build all contexts from scratch.
contexts.setLength(0);
UnicodeSetIterator iter = new UnicodeSetIterator(contextChars);
while(iter.next()) {
assert(iter.codepoint != UnicodeSetIterator.IS_STRING);
int c = iter.codepoint;
int ce32 = trie.get(c);
if(!isBuilderContextCE32(ce32)) {
throw new AssertionError("Impossible: No context data for c in contextChars.");
}
ConditionalCE32 cond = getConditionalCE32ForCE32(ce32);
ce32 = buildContext(cond);
trie.set(c, ce32);
}
}
protected int buildContext(ConditionalCE32 head) {
// The list head must have no context.
assert(!head.hasContext());
// The list head must be followed by one or more nodes that all do have context.
assert(head.next >= 0);
CharsTrieBuilder prefixBuilder = new CharsTrieBuilder();
CharsTrieBuilder contractionBuilder = new CharsTrieBuilder();
for(ConditionalCE32 cond = head;; cond = getConditionalCE32(cond.next)) {
// After the list head, the prefix or suffix can be empty, but not both.
assert(cond == head || cond.hasContext());
int prefixLength = cond.prefixLength();
StringBuilder prefix = new StringBuilder().append(cond.context, 0, prefixLength + 1);
String prefixString = prefix.toString();
// Collect all contraction suffixes for one prefix.
ConditionalCE32 firstCond = cond;
ConditionalCE32 lastCond = cond;
while(cond.next >= 0 &&
(cond = getConditionalCE32(cond.next)).context.startsWith(prefixString)) {
lastCond = cond;
}
int ce32;
int suffixStart = prefixLength + 1; // == prefix.length()
if(lastCond.context.length() == suffixStart) {
// One prefix without contraction suffix.
assert(firstCond == lastCond);
ce32 = lastCond.ce32;
cond = lastCond;
} else {
// Build the contractions trie.
contractionBuilder.clear();
// Entry for an empty suffix, to be stored before the trie.
int emptySuffixCE32 = Collation.NO_CE32; // Will always be set to a real value.
int flags = 0;
if(firstCond.context.length() == suffixStart) {
// There is a mapping for the prefix and the single character c. (p|c)
// If no other suffix matches, then we return this value.
emptySuffixCE32 = firstCond.ce32;
cond = getConditionalCE32(firstCond.next);
} else {
// There is no mapping for the prefix and just the single character.
// (There is no p|c, only p|cd, p|ce etc.)
flags |= Collation.CONTRACT_SINGLE_CP_NO_MATCH;
// When the prefix matches but none of the prefix-specific suffixes,
// then we fall back to the mappings with the next-longest prefix,
// and ultimately to mappings with no prefix.
// Each fallback might be another set of contractions.
// For example, if there are mappings for ch, p|cd, p|ce, but not for p|c,
// then in text "pch" we find the ch contraction.
for(cond = head;; cond = getConditionalCE32(cond.next)) {
int length = cond.prefixLength();
if(length == prefixLength) { break; }
if(cond.defaultCE32 != Collation.NO_CE32 &&
(length==0 || prefixString.regionMatches(
prefix.length() - length, cond.context, 1, length)
/* C++: prefix.endsWith(cond.context, 1, length) */)) {
emptySuffixCE32 = cond.defaultCE32;
}
}
cond = firstCond;
}
// Optimization: Set a flag when
// the first character of every contraction suffix has lccc!=0.
// Short-circuits contraction matching when a normal letter follows.
flags |= Collation.CONTRACT_NEXT_CCC;
// Add all of the non-empty suffixes into the contraction trie.
for(;;) {
String suffix = cond.context.substring(suffixStart);
int fcd16 = nfcImpl.getFCD16(suffix.codePointAt(0));
if(fcd16 <= 0xff) {
flags &= ~Collation.CONTRACT_NEXT_CCC;
}
fcd16 = nfcImpl.getFCD16(suffix.codePointBefore(suffix.length()));
if(fcd16 > 0xff) {
// The last suffix character has lccc!=0, allowing for discontiguous contractions.
flags |= Collation.CONTRACT_TRAILING_CCC;
}
contractionBuilder.add(suffix, cond.ce32);
if(cond == lastCond) { break; }
cond = getConditionalCE32(cond.next);
}
int index = addContextTrie(emptySuffixCE32, contractionBuilder);
if(index > Collation.MAX_INDEX) {
throw new IndexOutOfBoundsException("too many context-sensitive mappings");
// BufferOverflowException is a better fit
// but cannot be constructed with a message string.
}
ce32 = Collation.makeCE32FromTagAndIndex(Collation.CONTRACTION_TAG, index) | flags;
}
assert(cond == lastCond);
firstCond.defaultCE32 = ce32;
if(prefixLength == 0) {
if(cond.next < 0) {
// No non-empty prefixes, only contractions.
return ce32;
}
} else {
prefix.delete(0, 1); // Remove the length unit.
prefix.reverse();
prefixBuilder.add(prefix, ce32);
if(cond.next < 0) { break; }
}
}
assert(head.defaultCE32 != Collation.NO_CE32);
int index = addContextTrie(head.defaultCE32, prefixBuilder);
if(index > Collation.MAX_INDEX) {
throw new IndexOutOfBoundsException("too many context-sensitive mappings");
// BufferOverflowException is a better fit
// but cannot be constructed with a message string.
}
return Collation.makeCE32FromTagAndIndex(Collation.PREFIX_TAG, index);
}
protected int addContextTrie(int defaultCE32, CharsTrieBuilder trieBuilder) {
StringBuilder context = new StringBuilder();
context.append((char)(defaultCE32 >> 16)).append((char)defaultCE32);
context.append(trieBuilder.buildCharSequence(StringTrieBuilder.Option.SMALL));
int index = contexts.indexOf(context.toString());
if(index < 0) {
index = contexts.length();
contexts.append(context);
}
return index;
}
protected void buildFastLatinTable(CollationData data) {
if(!fastLatinEnabled) { return; }
fastLatinBuilder = new CollationFastLatinBuilder();
if(fastLatinBuilder.forData(data)) {
char[] header = fastLatinBuilder.getHeader();
char[] table = fastLatinBuilder.getTable();
if(base != null &&
Arrays.equals(header, base.fastLatinTableHeader) &&
Arrays.equals(table, base.fastLatinTable)) {
// Same fast Latin table as in the base, use that one instead.
fastLatinBuilder = null;
header = base.fastLatinTableHeader;
table = base.fastLatinTable;
}
data.fastLatinTableHeader = header;
data.fastLatinTable = table;
} else {
fastLatinBuilder = null;
}
}
protected int getCEs(CharSequence s, int start, long ces[], int cesLength) {
if(collIter == null) {
collIter = new DataBuilderCollationIterator(this, new CollationData(nfcImpl));
if(collIter == null) { return 0; }
}
return collIter.fetchCEs(s, start, ces, cesLength);
}
protected static int jamoCpFromIndex(int i) {
// 0 <= i < CollationData.JAMO_CE32S_LENGTH = 19 + 21 + 27
if(i < Hangul.JAMO_L_COUNT) { return Hangul.JAMO_L_BASE + i; }
i -= Hangul.JAMO_L_COUNT;
if(i < Hangul.JAMO_V_COUNT) { return Hangul.JAMO_V_BASE + i; }
i -= Hangul.JAMO_V_COUNT;
// i < 27
return Hangul.JAMO_T_BASE + 1 + i;
}
/**
* Build-time collation element and character iterator.
* Uses the runtime CollationIterator for fetching CEs for a string
* but reads from the builder's unfinished data structures.
* In particular, this class reads from the unfinished trie
* and has to avoid CollationIterator.nextCE() and redirect other
* calls to data.getCE32() and data.getCE32FromSupplementary().
*
* We do this so that we need not implement the collation algorithm
* again for the builder and make it behave exactly like the runtime code.
* That would be more difficult to test and maintain than this indirection.
*
* Some CE32 tags (for example, the DIGIT_TAG) do not occur in the builder data,
* so the data accesses from those code paths need not be modified.
*
* This class iterates directly over whole code points
* so that the CollationIterator does not need the finished trie
* for handling the LEAD_SURROGATE_TAG.
*/
private static final class DataBuilderCollationIterator extends CollationIterator {
DataBuilderCollationIterator(CollationDataBuilder b, CollationData newData) {
super(newData, /*numeric=*/ false);
builder = b;
builderData = newData;
builderData.base = builder.base;
// Set all of the jamoCE32s[] to indirection CE32s.
for(int j = 0; j < CollationData.JAMO_CE32S_LENGTH; ++j) { // Count across Jamo types.
int jamo = CollationDataBuilder.jamoCpFromIndex(j);
jamoCE32s[j] = Collation.makeCE32FromTagAndIndex(Collation.BUILDER_DATA_TAG, jamo) |
CollationDataBuilder.IS_BUILDER_JAMO_CE32;
}
builderData.jamoCE32s = jamoCE32s;
}
int fetchCEs(CharSequence str, int start, long ces[], int cesLength) {
// Set the pointers each time, in case they changed due to reallocation.
builderData.ce32s = builder.ce32s.getBuffer();
builderData.ces = builder.ce64s.getBuffer();
builderData.contexts = builder.contexts.toString();
// Modified copy of CollationIterator.nextCE() and CollationIterator.nextCEFromCE32().
reset();
s = str;
pos = start;
while(pos < s.length()) {
// No need to keep all CEs in the iterator buffer.
clearCEs();
int c = Character.codePointAt(s, pos);
pos += Character.charCount(c);
int ce32 = builder.trie.get(c);
CollationData d;
if(ce32 == Collation.FALLBACK_CE32) {
d = builder.base;
ce32 = builder.base.getCE32(c);
} else {
d = builderData;
}
appendCEsFromCE32(d, c, ce32, /*forward=*/ true);
for(int i = 0; i < getCEsLength(); ++i) {
long ce = getCE(i);
if(ce != 0) {
if(cesLength < Collation.MAX_EXPANSION_LENGTH) {
ces[cesLength] = ce;
}
++cesLength;
}
}
}
return cesLength;
}
@Override
public void resetToOffset(int newOffset) {
reset();
pos = newOffset;
}
@Override
public int getOffset() {
return pos;
}
@Override
public int nextCodePoint() {
if(pos == s.length()) {
return Collation.SENTINEL_CP;
}
int c = Character.codePointAt(s, pos);
pos += Character.charCount(c);
return c;
}
@Override
public int previousCodePoint() {
if(pos == 0) {
return Collation.SENTINEL_CP;
}
int c = Character.codePointBefore(s, pos);
pos -= Character.charCount(c);
return c;
}
@Override
protected void forwardNumCodePoints(int num) {
pos = Character.offsetByCodePoints(s, pos, num);
}
@Override
protected void backwardNumCodePoints(int num) {
pos = Character.offsetByCodePoints(s, pos, -num);
}
@Override
protected int getDataCE32(int c) {
return builder.trie.get(c);
}
@Override
protected int getCE32FromBuilderData(int ce32) {
assert(Collation.hasCE32Tag(ce32, Collation.BUILDER_DATA_TAG));
if((ce32 & CollationDataBuilder.IS_BUILDER_JAMO_CE32) != 0) {
int jamo = Collation.indexFromCE32(ce32);
return builder.trie.get(jamo);
} else {
ConditionalCE32 cond = builder.getConditionalCE32ForCE32(ce32);
if(cond.builtCE32 == Collation.NO_CE32) {
// Build the context-sensitive mappings into their runtime form and cache the result.
try {
cond.builtCE32 = builder.buildContext(cond);
} catch(IndexOutOfBoundsException e) {
builder.clearContexts();
cond.builtCE32 = builder.buildContext(cond);
}
builderData.contexts = builder.contexts.toString();
}
return cond.builtCE32;
}
}
protected final CollationDataBuilder builder;
protected final CollationData builderData;
protected final int[] jamoCE32s = new int[CollationData.JAMO_CE32S_LENGTH];
protected CharSequence s;
protected int pos;
}
protected final boolean isMutable() {
// C++ tests !(trie == NULL || utrie2_isFrozen(trie))
// but Java Trie2Writable does not have an observable isFrozen() state.
return trie != null && unsafeBackwardSet != null && !unsafeBackwardSet.isFrozen();
}
/** @see Collation.BUILDER_DATA_TAG */
private static final int IS_BUILDER_JAMO_CE32 = 0x100;
protected Normalizer2Impl nfcImpl;
protected CollationData base;
protected CollationSettings baseSettings;
protected Trie2Writable trie;
protected UVector32 ce32s;
protected UVector64 ce64s;
protected ArrayList conditionalCE32s; // vector of ConditionalCE32
// Characters that have context (prefixes or contraction suffixes).
protected UnicodeSet contextChars = new UnicodeSet();
// Serialized UCharsTrie structures for finalized contexts.
protected StringBuilder contexts = new StringBuilder();
protected UnicodeSet unsafeBackwardSet = new UnicodeSet();
protected boolean modified;
protected boolean fastLatinEnabled;
protected CollationFastLatinBuilder fastLatinBuilder;
protected DataBuilderCollationIterator collIter;
}