Check in the pristine copy of ICU 4.6...

icu46/source/i18n/usrchimp.h

+/*
+**********************************************************************
+*   Copyright (C) 2001-2010 IBM and others. All rights reserved.
+**********************************************************************
+*   Date        Name        Description
+*  08/13/2001   synwee      Creation.
+**********************************************************************
+*/
+#ifndef USRCHIMP_H
+#define USRCHIMP_H
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_COLLATION
+
+#include "unicode/normalizer2.h"
+#include "unicode/ucol.h"
+#include "unicode/ucoleitr.h"
+#include "unicode/ubrk.h"
+
+#define INITIAL_ARRAY_SIZE_       256
+#define MAX_TABLE_SIZE_           257
+
+struct USearch {
+    // required since collation element iterator does not have a getText API
+    const UChar              *text;
+          int32_t             textLength; // exact length
+          UBool               isOverlap;
+          UBool               isCanonicalMatch;
+          int16_t             elementComparisonType;
+          UBreakIterator     *internalBreakIter;  //internal character breakiterator
+          UBreakIterator     *breakIter;
+    // value USEARCH_DONE is the default value
+    // if we are not at the start of the text or the end of the text, 
+    // depending on the iteration direction and matchedIndex is USEARCH_DONE 
+    // it means that we can't find any more matches in that particular direction
+          int32_t             matchedIndex; 
+          int32_t             matchedLength;
+          UBool               isForwardSearching;
+          UBool               reset;
+};
+
+struct UPattern {
+    const UChar              *text;
+          int32_t             textLength; // exact length
+          // length required for backwards ce comparison
+          int32_t             CELength; 
+          int32_t            *CE;
+          int32_t             CEBuffer[INITIAL_ARRAY_SIZE_];
+          int32_t             PCELength;
+          int64_t            *PCE;
+          int64_t             PCEBuffer[INITIAL_ARRAY_SIZE_];
+          UBool               hasPrefixAccents;
+          UBool               hasSuffixAccents;
+          int16_t             defaultShiftSize;
+          int16_t             shift[MAX_TABLE_SIZE_];
+          int16_t             backShift[MAX_TABLE_SIZE_];
+};
+
+struct UStringSearch {
+    struct USearch            *search;
+    struct UPattern            pattern;
+    const  UCollator          *collator;
+    const  U_NAMESPACE_QUALIFIER Normalizer2 *nfd;
+    // positions within the collation element iterator is used to determine
+    // if we are at the start of the text.
+           UCollationElements *textIter;
+    // utility collation element, used throughout program for temporary 
+    // iteration.
+           UCollationElements *utilIter;
+           UBool               ownCollator;
+           UCollationStrength  strength;
+           uint32_t            ceMask;
+           uint32_t            variableTop;
+           UBool               toShift;
+           UChar               canonicalPrefixAccents[INITIAL_ARRAY_SIZE_];
+           UChar               canonicalSuffixAccents[INITIAL_ARRAY_SIZE_];
+};
+
+/**
+* Exact matches without checking for the ends for extra accents.
+* The match after the position within the collation element iterator is to be
+* found. 
+* After a match is found the offset in the collation element iterator will be
+* shifted to the start of the match.
+* Implementation note: 
+* For tertiary we can't use the collator->tertiaryMask, that is a 
+* preprocessed mask that takes into account case options. since we are only 
+* concerned with exact matches, we don't need that.
+* Alternate handling - since only the 16 most significant digits is only used, 
+* we can safely do a compare without masking if the ce is a variable, we mask 
+* and get only the primary values no shifting to quartenary is required since 
+* all primary values less than variabletop will need to be masked off anyway.
+* If the end character is composite and the pattern ce does not match the text 
+* ce, we skip it until we find a match in the end composite character or when 
+* it has passed the character. This is so that we can match pattern "a" with
+* the text "\u00e6" 
+* @param strsrch string search data
+* @param status error status if any
+* @return TRUE if an exact match is found, FALSE otherwise
+*/
+U_CFUNC
+UBool usearch_handleNextExact(UStringSearch *strsrch, UErrorCode *status);
+
+/**
+* Canonical matches.
+* According to the definition, matches found here will include the whole span 
+* of beginning and ending accents if it overlaps that region.
+* @param strsrch string search data
+* @param status error status if any
+* @return TRUE if a canonical match is found, FALSE otherwise
+*/
+U_CFUNC
+UBool usearch_handleNextCanonical(UStringSearch *strsrch, UErrorCode *status);
+
+/**
+* Gets the previous match.
+* Comments follows from handleNextExact
+* @param strsrch string search data
+* @param status error status if any
+* @return True if a exact math is found, FALSE otherwise.
+*/
+U_CFUNC
+UBool usearch_handlePreviousExact(UStringSearch *strsrch, UErrorCode *status);
+
+/**
+* Canonical matches.
+* According to the definition, matches found here will include the whole span 
+* of beginning and ending accents if it overlaps that region.
+* @param strsrch string search data
+* @param status error status if any
+* @return TRUE if a canonical match is found, FALSE otherwise
+*/
+U_CFUNC
+UBool usearch_handlePreviousCanonical(UStringSearch *strsrch, 
+                                      UErrorCode    *status);
+
+#endif /* #if !UCONFIG_NO_COLLATION */
+
+#endif