icuSources/i18n/usrchimp.h

   1 /*
   2 **********************************************************************
   3 *   Copyright (C) 2001-2004 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     *breakIter;
  30     // value USEARCH_DONE is the default value
  31     // if we are not at the start of the text or the end of the text,
  32     // depending on the iteration direction and matchedIndex is USEARCH_DONE
  33     // it means that we can find any more matches in that particular direction
  34           int32_t         matchedIndex;
  35           int32_t             matchedLength;
  36           UBool               isForwardSearching;
  37           UBool               reset;
  38 };
  39
  40 struct UPattern {
  41     const UChar              *text;
  42           int32_t             textLength; // exact length
  43           // length required for backwards ce comparison
  44           int32_t             CELength;
  45           int32_t            *CE;
  46           int32_t             CEBuffer[INITIAL_ARRAY_SIZE_];
  47           UBool               hasPrefixAccents;
  48           UBool               hasSuffixAccents;
  49           int16_t             defaultShiftSize;
  50           int16_t             shift[MAX_TABLE_SIZE_];
  51           int16_t             backShift[MAX_TABLE_SIZE_];
  52 };
  53
  54 struct UStringSearch {
  55     struct USearch            *search;
  56     struct UPattern            pattern;
  57     const  UCollator          *collator;
  58     // positions within the collation element iterator is used to determine
  59     // if we are at the start of the text.
  60            UCollationElements *textIter;
  61     // utility collation element, used throughout program for temporary
  62     // iteration.
  63            UCollationElements *utilIter;
  64            UBool               ownCollator;
  65            UCollationStrength  strength;
  66            uint32_t            ceMask;
  67            uint32_t            variableTop;
  68            UBool               toShift;
  69            UChar               canonicalPrefixAccents[INITIAL_ARRAY_SIZE_];
  70            UChar               canonicalSuffixAccents[INITIAL_ARRAY_SIZE_];
  71 };
  72
  73 /**
  74 * Exact matches without checking for the ends for extra accents.
  75 * The match after the position within the collation element iterator is to be
  76 * found.
  77 * After a match is found the offset in the collation element iterator will be
  78 * shifted to the start of the match.
  79 * Implementation note:
  80 * For tertiary we can't use the collator->tertiaryMask, that is a
  81 * preprocessed mask that takes into account case options. since we are only
  82 * concerned with exact matches, we don't need that.
  83 * Alternate handling - since only the 16 most significant digits is only used,
  84 * we can safely do a compare without masking if the ce is a variable, we mask
  85 * and get only the primary values no shifting to quartenary is required since
  86 * all primary values less than variabletop will need to be masked off anyway.
  87 * If the end character is composite and the pattern ce does not match the text
  88 * ce, we skip it until we find a match in the end composite character or when
  89 * it has passed the character. This is so that we can match pattern "a" with
  90 * the text "\u00e6"
  91 * @param strsrch string search data
  92 * @param status error status if any
  93 * @return TRUE if an exact match is found, FALSE otherwise
  94 */
  95 U_CFUNC
  96 UBool usearch_handleNextExact(UStringSearch *strsrch, UErrorCode *status);
  97
  98 /**
  99 * Canonical matches.
 100 * According to the definition, matches found here will include the whole span
 101 * of beginning and ending accents if it overlaps that region.
 102 * @param strsrch string search data
 103 * @param status error status if any
 104 * @return TRUE if a canonical match is found, FALSE otherwise
 105 */
 106 U_CFUNC
 107 UBool usearch_handleNextCanonical(UStringSearch *strsrch, UErrorCode *status);
 108
 109 /**
 110 * Gets the previous match.
 111 * Comments follows from handleNextExact
 112 * @param strsrch string search data
 113 * @param status error status if any
 114 * @return True if a exact math is found, FALSE otherwise.
 115 */
 116 U_CFUNC
 117 UBool usearch_handlePreviousExact(UStringSearch *strsrch, UErrorCode *status);
 118
 119 /**
 120 * Canonical matches.
 121 * According to the definition, matches found here will include the whole span
 122 * of beginning and ending accents if it overlaps that region.
 123 * @param strsrch string search data
 124 * @param status error status if any
 125 * @return TRUE if a canonical match is found, FALSE otherwise
 126 */
 127 U_CFUNC
 128 UBool usearch_handlePreviousCanonical(UStringSearch *strsrch,
 129                                       UErrorCode    *status);
 130
 131 #endif /* #if !UCONFIG_NO_COLLATION */
 132
 133 #endif