1// © 2016 and later: Unicode, Inc. and others.
2// License & terms of use: http://www.unicode.org/copyright.html
3/*
4*******************************************************************************
5* Copyright (C) 2013-2016, International Business Machines
6* Corporation and others. All Rights Reserved.
7*******************************************************************************
8* collationfastlatinbuilder.h
9*
10* created on: 2013aug09
11* created by: Markus W. Scherer
12*/
13
14#ifndef __COLLATIONFASTLATINBUILDER_H__
15#define __COLLATIONFASTLATINBUILDER_H__
16
17#include "unicode/utypes.h"
18
19#if !UCONFIG_NO_COLLATION
20
21#include "unicode/ucol.h"
22#include "unicode/unistr.h"
23#include "unicode/uobject.h"
24#include "collation.h"
25#include "collationfastlatin.h"
26#include "uvectr64.h"
27
28U_NAMESPACE_BEGIN
29
30struct CollationData;
31
32class U_I18N_API CollationFastLatinBuilder : public UObject {
33public:
34 CollationFastLatinBuilder(UErrorCode &errorCode);
35 ~CollationFastLatinBuilder();
36
37 UBool forData(const CollationData &data, UErrorCode &errorCode);
38
39 const uint16_t *getTable() const {
40 return reinterpret_cast<const uint16_t *>(result.getBuffer());
41 }
42 int32_t lengthOfTable() const { return result.length(); }
43
44private:
45 // space, punct, symbol, currency (not digit)
46 enum { NUM_SPECIAL_GROUPS = UCOL_REORDER_CODE_CURRENCY - UCOL_REORDER_CODE_FIRST + 1 };
47
48 UBool loadGroups(const CollationData &data, UErrorCode &errorCode);
49 UBool inSameGroup(uint32_t p, uint32_t q) const;
50
51 void resetCEs();
52 void getCEs(const CollationData &data, UErrorCode &errorCode);
53 UBool getCEsFromCE32(const CollationData &data, UChar32 c, uint32_t ce32,
54 UErrorCode &errorCode);
55 UBool getCEsFromContractionCE32(const CollationData &data, uint32_t ce32,
56 UErrorCode &errorCode);
57 void addContractionEntry(int32_t x, int64_t cce0, int64_t cce1, UErrorCode &errorCode);
58 void addUniqueCE(int64_t ce, UErrorCode &errorCode);
59 uint32_t getMiniCE(int64_t ce) const;
60 UBool encodeUniqueCEs(UErrorCode &errorCode);
61 UBool encodeCharCEs(UErrorCode &errorCode);
62 UBool encodeContractions(UErrorCode &errorCode);
63 uint32_t encodeTwoCEs(int64_t first, int64_t second) const;
64
65 static UBool isContractionCharCE(int64_t ce) {
66 return (uint32_t)(ce >> 32) == Collation::NO_CE_PRIMARY && ce != Collation::NO_CE;
67 }
68
69 static const uint32_t CONTRACTION_FLAG = 0x80000000;
70
71 // temporary "buffer"
72 int64_t ce0, ce1;
73
74 int64_t charCEs[CollationFastLatin::NUM_FAST_CHARS][2];
75
76 UVector64 contractionCEs;
77 UVector64 uniqueCEs;
78
79 /** One 16-bit mini CE per unique CE. */
80 uint16_t *miniCEs;
81
82 // These are constant for a given root collator.
83 uint32_t lastSpecialPrimaries[NUM_SPECIAL_GROUPS];
84 uint32_t firstDigitPrimary;
85 uint32_t firstLatinPrimary;
86 uint32_t lastLatinPrimary;
87 // This determines the first normal primary weight which is mapped to
88 // a short mini primary. It must be >=firstDigitPrimary.
89 uint32_t firstShortPrimary;
90
91 UBool shortPrimaryOverflow;
92
93 UnicodeString result;
94 int32_t headerLength;
95};
96
97U_NAMESPACE_END
98
99#endif // !UCONFIG_NO_COLLATION
100#endif // __COLLATIONFASTLATINBUILDER_H__
101