|
1 /* |
|
2 ********************************************************************** |
|
3 * Copyright (C) 2001-2011 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/normalizer2.h" |
|
17 #include "unicode/ucol.h" |
|
18 #include "unicode/ucoleitr.h" |
|
19 #include "unicode/ubrk.h" |
|
20 |
|
21 #define INITIAL_ARRAY_SIZE_ 256 |
|
22 #define MAX_TABLE_SIZE_ 257 |
|
23 |
|
24 struct USearch { |
|
25 // required since collation element iterator does not have a getText API |
|
26 const UChar *text; |
|
27 int32_t textLength; // exact length |
|
28 UBool isOverlap; |
|
29 UBool isCanonicalMatch; |
|
30 int16_t elementComparisonType; |
|
31 UBreakIterator *internalBreakIter; //internal character breakiterator |
|
32 UBreakIterator *breakIter; |
|
33 // value USEARCH_DONE is the default value |
|
34 // if we are not at the start of the text or the end of the text, |
|
35 // depending on the iteration direction and matchedIndex is USEARCH_DONE |
|
36 // it means that we can't find any more matches in that particular direction |
|
37 int32_t matchedIndex; |
|
38 int32_t matchedLength; |
|
39 UBool isForwardSearching; |
|
40 UBool reset; |
|
41 }; |
|
42 |
|
43 struct UPattern { |
|
44 const UChar *text; |
|
45 int32_t textLength; // exact length |
|
46 // length required for backwards ce comparison |
|
47 int32_t CELength; |
|
48 int32_t *CE; |
|
49 int32_t CEBuffer[INITIAL_ARRAY_SIZE_]; |
|
50 int32_t PCELength; |
|
51 int64_t *PCE; |
|
52 int64_t PCEBuffer[INITIAL_ARRAY_SIZE_]; |
|
53 UBool hasPrefixAccents; |
|
54 UBool hasSuffixAccents; |
|
55 int16_t defaultShiftSize; |
|
56 int16_t shift[MAX_TABLE_SIZE_]; |
|
57 int16_t backShift[MAX_TABLE_SIZE_]; |
|
58 }; |
|
59 |
|
60 struct UStringSearch { |
|
61 struct USearch *search; |
|
62 struct UPattern pattern; |
|
63 const UCollator *collator; |
|
64 const icu::Normalizer2 *nfd; |
|
65 // positions within the collation element iterator is used to determine |
|
66 // if we are at the start of the text. |
|
67 UCollationElements *textIter; |
|
68 // utility collation element, used throughout program for temporary |
|
69 // iteration. |
|
70 UCollationElements *utilIter; |
|
71 UBool ownCollator; |
|
72 UCollationStrength strength; |
|
73 uint32_t ceMask; |
|
74 uint32_t variableTop; |
|
75 UBool toShift; |
|
76 UChar canonicalPrefixAccents[INITIAL_ARRAY_SIZE_]; |
|
77 UChar canonicalSuffixAccents[INITIAL_ARRAY_SIZE_]; |
|
78 }; |
|
79 |
|
80 /** |
|
81 * Exact matches without checking for the ends for extra accents. |
|
82 * The match after the position within the collation element iterator is to be |
|
83 * found. |
|
84 * After a match is found the offset in the collation element iterator will be |
|
85 * shifted to the start of the match. |
|
86 * Implementation note: |
|
87 * For tertiary we can't use the collator->tertiaryMask, that is a |
|
88 * preprocessed mask that takes into account case options. since we are only |
|
89 * concerned with exact matches, we don't need that. |
|
90 * Alternate handling - since only the 16 most significant digits is only used, |
|
91 * we can safely do a compare without masking if the ce is a variable, we mask |
|
92 * and get only the primary values no shifting to quartenary is required since |
|
93 * all primary values less than variabletop will need to be masked off anyway. |
|
94 * If the end character is composite and the pattern ce does not match the text |
|
95 * ce, we skip it until we find a match in the end composite character or when |
|
96 * it has passed the character. This is so that we can match pattern "a" with |
|
97 * the text "\u00e6" |
|
98 * @param strsrch string search data |
|
99 * @param status error status if any |
|
100 * @return TRUE if an exact match is found, FALSE otherwise |
|
101 */ |
|
102 U_CFUNC |
|
103 UBool usearch_handleNextExact(UStringSearch *strsrch, UErrorCode *status); |
|
104 |
|
105 /** |
|
106 * Canonical matches. |
|
107 * According to the definition, matches found here will include the whole span |
|
108 * of beginning and ending accents if it overlaps that region. |
|
109 * @param strsrch string search data |
|
110 * @param status error status if any |
|
111 * @return TRUE if a canonical match is found, FALSE otherwise |
|
112 */ |
|
113 U_CFUNC |
|
114 UBool usearch_handleNextCanonical(UStringSearch *strsrch, UErrorCode *status); |
|
115 |
|
116 /** |
|
117 * Gets the previous match. |
|
118 * Comments follows from handleNextExact |
|
119 * @param strsrch string search data |
|
120 * @param status error status if any |
|
121 * @return True if a exact math is found, FALSE otherwise. |
|
122 */ |
|
123 U_CFUNC |
|
124 UBool usearch_handlePreviousExact(UStringSearch *strsrch, UErrorCode *status); |
|
125 |
|
126 /** |
|
127 * Canonical matches. |
|
128 * According to the definition, matches found here will include the whole span |
|
129 * of beginning and ending accents if it overlaps that region. |
|
130 * @param strsrch string search data |
|
131 * @param status error status if any |
|
132 * @return TRUE if a canonical match is found, FALSE otherwise |
|
133 */ |
|
134 U_CFUNC |
|
135 UBool usearch_handlePreviousCanonical(UStringSearch *strsrch, |
|
136 UErrorCode *status); |
|
137 |
|
138 #endif /* #if !UCONFIG_NO_COLLATION */ |
|
139 |
|
140 #endif |