1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/i18n/usrchimp.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,140 @@
1.4 +/*
1.5 +**********************************************************************
1.6 +* Copyright (C) 2001-2011 IBM and others. All rights reserved.
1.7 +**********************************************************************
1.8 +* Date Name Description
1.9 +* 08/13/2001 synwee Creation.
1.10 +**********************************************************************
1.11 +*/
1.12 +#ifndef USRCHIMP_H
1.13 +#define USRCHIMP_H
1.14 +
1.15 +#include "unicode/utypes.h"
1.16 +
1.17 +#if !UCONFIG_NO_COLLATION
1.18 +
1.19 +#include "unicode/normalizer2.h"
1.20 +#include "unicode/ucol.h"
1.21 +#include "unicode/ucoleitr.h"
1.22 +#include "unicode/ubrk.h"
1.23 +
1.24 +#define INITIAL_ARRAY_SIZE_ 256
1.25 +#define MAX_TABLE_SIZE_ 257
1.26 +
1.27 +struct USearch {
1.28 + // required since collation element iterator does not have a getText API
1.29 + const UChar *text;
1.30 + int32_t textLength; // exact length
1.31 + UBool isOverlap;
1.32 + UBool isCanonicalMatch;
1.33 + int16_t elementComparisonType;
1.34 + UBreakIterator *internalBreakIter; //internal character breakiterator
1.35 + UBreakIterator *breakIter;
1.36 + // value USEARCH_DONE is the default value
1.37 + // if we are not at the start of the text or the end of the text,
1.38 + // depending on the iteration direction and matchedIndex is USEARCH_DONE
1.39 + // it means that we can't find any more matches in that particular direction
1.40 + int32_t matchedIndex;
1.41 + int32_t matchedLength;
1.42 + UBool isForwardSearching;
1.43 + UBool reset;
1.44 +};
1.45 +
1.46 +struct UPattern {
1.47 + const UChar *text;
1.48 + int32_t textLength; // exact length
1.49 + // length required for backwards ce comparison
1.50 + int32_t CELength;
1.51 + int32_t *CE;
1.52 + int32_t CEBuffer[INITIAL_ARRAY_SIZE_];
1.53 + int32_t PCELength;
1.54 + int64_t *PCE;
1.55 + int64_t PCEBuffer[INITIAL_ARRAY_SIZE_];
1.56 + UBool hasPrefixAccents;
1.57 + UBool hasSuffixAccents;
1.58 + int16_t defaultShiftSize;
1.59 + int16_t shift[MAX_TABLE_SIZE_];
1.60 + int16_t backShift[MAX_TABLE_SIZE_];
1.61 +};
1.62 +
1.63 +struct UStringSearch {
1.64 + struct USearch *search;
1.65 + struct UPattern pattern;
1.66 + const UCollator *collator;
1.67 + const icu::Normalizer2 *nfd;
1.68 + // positions within the collation element iterator is used to determine
1.69 + // if we are at the start of the text.
1.70 + UCollationElements *textIter;
1.71 + // utility collation element, used throughout program for temporary
1.72 + // iteration.
1.73 + UCollationElements *utilIter;
1.74 + UBool ownCollator;
1.75 + UCollationStrength strength;
1.76 + uint32_t ceMask;
1.77 + uint32_t variableTop;
1.78 + UBool toShift;
1.79 + UChar canonicalPrefixAccents[INITIAL_ARRAY_SIZE_];
1.80 + UChar canonicalSuffixAccents[INITIAL_ARRAY_SIZE_];
1.81 +};
1.82 +
1.83 +/**
1.84 +* Exact matches without checking for the ends for extra accents.
1.85 +* The match after the position within the collation element iterator is to be
1.86 +* found.
1.87 +* After a match is found the offset in the collation element iterator will be
1.88 +* shifted to the start of the match.
1.89 +* Implementation note:
1.90 +* For tertiary we can't use the collator->tertiaryMask, that is a
1.91 +* preprocessed mask that takes into account case options. since we are only
1.92 +* concerned with exact matches, we don't need that.
1.93 +* Alternate handling - since only the 16 most significant digits is only used,
1.94 +* we can safely do a compare without masking if the ce is a variable, we mask
1.95 +* and get only the primary values no shifting to quartenary is required since
1.96 +* all primary values less than variabletop will need to be masked off anyway.
1.97 +* If the end character is composite and the pattern ce does not match the text
1.98 +* ce, we skip it until we find a match in the end composite character or when
1.99 +* it has passed the character. This is so that we can match pattern "a" with
1.100 +* the text "\u00e6"
1.101 +* @param strsrch string search data
1.102 +* @param status error status if any
1.103 +* @return TRUE if an exact match is found, FALSE otherwise
1.104 +*/
1.105 +U_CFUNC
1.106 +UBool usearch_handleNextExact(UStringSearch *strsrch, UErrorCode *status);
1.107 +
1.108 +/**
1.109 +* Canonical matches.
1.110 +* According to the definition, matches found here will include the whole span
1.111 +* of beginning and ending accents if it overlaps that region.
1.112 +* @param strsrch string search data
1.113 +* @param status error status if any
1.114 +* @return TRUE if a canonical match is found, FALSE otherwise
1.115 +*/
1.116 +U_CFUNC
1.117 +UBool usearch_handleNextCanonical(UStringSearch *strsrch, UErrorCode *status);
1.118 +
1.119 +/**
1.120 +* Gets the previous match.
1.121 +* Comments follows from handleNextExact
1.122 +* @param strsrch string search data
1.123 +* @param status error status if any
1.124 +* @return True if a exact math is found, FALSE otherwise.
1.125 +*/
1.126 +U_CFUNC
1.127 +UBool usearch_handlePreviousExact(UStringSearch *strsrch, UErrorCode *status);
1.128 +
1.129 +/**
1.130 +* Canonical matches.
1.131 +* According to the definition, matches found here will include the whole span
1.132 +* of beginning and ending accents if it overlaps that region.
1.133 +* @param strsrch string search data
1.134 +* @param status error status if any
1.135 +* @return TRUE if a canonical match is found, FALSE otherwise
1.136 +*/
1.137 +U_CFUNC
1.138 +UBool usearch_handlePreviousCanonical(UStringSearch *strsrch,
1.139 + UErrorCode *status);
1.140 +
1.141 +#endif /* #if !UCONFIG_NO_COLLATION */
1.142 +
1.143 +#endif
