icuSources/i18n/usrchimp.h

   1 /*
   2 **********************************************************************
   3 *   Copyright (C) 2001-2008 IBM and others. All rights reserved.
   4 **********************************************************************
   5 *   Date        Name        Description
   6 *  08/13/2001   synwee      Creation.
   7 **********************************************************************
   8 */
   9 #ifndef USRCHIMP_H
  10 #define USRCHIMP_H
  11
  12 #include "unicode/utypes.h"
  13
  14 #if !UCONFIG_NO_COLLATION
  15
  16 #include "unicode/ucol.h"
  17 #include "unicode/ucoleitr.h"
  18 #include "unicode/ubrk.h"
  19
  20 #define INITIAL_ARRAY_SIZE_       256
  21 #define MAX_TABLE_SIZE_           257
  22
  23 struct USearch {
  24     // required since collation element iterator does not have a getText API
  25     const UChar              *text;
  26           int32_t             textLength; // exact length
  27           UBool               isOverlap;
  28           UBool               isCanonicalMatch;
  29           UBreakIterator         *internalBreakIter;  //internal character breakiterator
  30           UBreakIterator     *breakIter;
  31     // value USEARCH_DONE is the default value
  32     // if we are not at the start of the text or the end of the text,
  33     // depending on the iteration direction and matchedIndex is USEARCH_DONE
  34     // it means that we can't find any more matches in that particular direction
  35           int32_t             matchedIndex;
  36           int32_t             matchedLength;
  37           UBool               isForwardSearching;
  38           UBool               reset;
  39 };
  40
  41 struct UPattern {
  42     const UChar              *text;
  43           int32_t             textLength; // exact length
  44           // length required for backwards ce comparison
  45           int32_t             CELength;
  46           int32_t            *CE;
  47           int32_t             CEBuffer[INITIAL_ARRAY_SIZE_];
  48           int32_t             PCELength;
  49           int64_t            *PCE;
  50           int64_t             PCEBuffer[INITIAL_ARRAY_SIZE_];
  51           UBool               hasPrefixAccents;
  52           UBool               hasSuffixAccents;
  53           int16_t             defaultShiftSize;
  54           int16_t             shift[MAX_TABLE_SIZE_];
  55           int16_t             backShift[MAX_TABLE_SIZE_];
  56 };
  57
  58 struct UStringSearch {
  59     struct USearch            *search;
  60     struct UPattern            pattern;
  61     const  UCollator          *collator;
  62     // positions within the collation element iterator is used to determine
  63     // if we are at the start of the text.
  64            UCollationElements *textIter;
  65     // utility collation element, used throughout program for temporary
  66     // iteration.
  67            UCollationElements *utilIter;
  68            UBool               ownCollator;
  69            UCollationStrength  strength;
  70            uint32_t            ceMask;
  71            uint32_t            variableTop;
  72            UBool               toShift;
  73            UChar               canonicalPrefixAccents[INITIAL_ARRAY_SIZE_];
  74            UChar               canonicalSuffixAccents[INITIAL_ARRAY_SIZE_];
  75 };
  76
  77 /**
  78 * Exact matches without checking for the ends for extra accents.
  79 * The match after the position within the collation element iterator is to be
  80 * found.
  81 * After a match is found the offset in the collation element iterator will be
  82 * shifted to the start of the match.
  83 * Implementation note:
  84 * For tertiary we can't use the collator->tertiaryMask, that is a
  85 * preprocessed mask that takes into account case options. since we are only
  86 * concerned with exact matches, we don't need that.
  87 * Alternate handling - since only the 16 most significant digits is only used,
  88 * we can safely do a compare without masking if the ce is a variable, we mask
  89 * and get only the primary values no shifting to quartenary is required since
  90 * all primary values less than variabletop will need to be masked off anyway.
  91 * If the end character is composite and the pattern ce does not match the text
  92 * ce, we skip it until we find a match in the end composite character or when
  93 * it has passed the character. This is so that we can match pattern "a" with
  94 * the text "\u00e6"
  95 * @param strsrch string search data
  96 * @param status error status if any
  97 * @return TRUE if an exact match is found, FALSE otherwise
  98 */
  99 U_CFUNC
 100 UBool usearch_handleNextExact(UStringSearch *strsrch, UErrorCode *status);
 101
 102 /**
 103 * Canonical matches.
 104 * According to the definition, matches found here will include the whole span
 105 * of beginning and ending accents if it overlaps that region.
 106 * @param strsrch string search data
 107 * @param status error status if any
 108 * @return TRUE if a canonical match is found, FALSE otherwise
 109 */
 110 U_CFUNC
 111 UBool usearch_handleNextCanonical(UStringSearch *strsrch, UErrorCode *status);
 112
 113 /**
 114 * Gets the previous match.
 115 * Comments follows from handleNextExact
 116 * @param strsrch string search data
 117 * @param status error status if any
 118 * @return True if a exact math is found, FALSE otherwise.
 119 */
 120 U_CFUNC
 121 UBool usearch_handlePreviousExact(UStringSearch *strsrch, UErrorCode *status);
 122
 123 /**
 124 * Canonical matches.
 125 * According to the definition, matches found here will include the whole span
 126 * of beginning and ending accents if it overlaps that region.
 127 * @param strsrch string search data
 128 * @param status error status if any
 129 * @return TRUE if a canonical match is found, FALSE otherwise
 130 */
 131 U_CFUNC
 132 UBool usearch_handlePreviousCanonical(UStringSearch *strsrch,
 133                                       UErrorCode    *status);
 134
 135 #endif /* #if !UCONFIG_NO_COLLATION */
 136
 137 #endif