1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/propsvec.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,176 @@
1.4 +/*
1.5 +*******************************************************************************
1.6 +*
1.7 +* Copyright (C) 2002-2010, International Business Machines
1.8 +* Corporation and others. All Rights Reserved.
1.9 +*
1.10 +*******************************************************************************
1.11 +* file name: propsvec.h
1.12 +* encoding: US-ASCII
1.13 +* tab size: 8 (not used)
1.14 +* indentation:4
1.15 +*
1.16 +* created on: 2002feb22
1.17 +* created by: Markus W. Scherer
1.18 +*
1.19 +* Store bits (Unicode character properties) in bit set vectors.
1.20 +*/
1.21 +
1.22 +#ifndef __UPROPSVEC_H__
1.23 +#define __UPROPSVEC_H__
1.24 +
1.25 +#include "unicode/utypes.h"
1.26 +#include "utrie.h"
1.27 +#include "utrie2.h"
1.28 +
1.29 +U_CDECL_BEGIN
1.30 +
1.31 +/**
1.32 + * Unicode Properties Vectors associated with code point ranges.
1.33 + *
1.34 + * Rows of uint32_t integers in a contiguous array store
1.35 + * the range limits and the properties vectors.
1.36 + *
1.37 + * Logically, each row has a certain number of uint32_t values,
1.38 + * which is set via the upvec_open() "columns" parameter.
1.39 + *
1.40 + * Internally, two additional columns are stored.
1.41 + * In each internal row,
1.42 + * row[0] contains the start code point and
1.43 + * row[1] contains the limit code point,
1.44 + * which is the start of the next range.
1.45 + *
1.46 + * Initially, there is only one "normal" row for
1.47 + * range [0..0x110000[ with values 0.
1.48 + * There are additional rows for special purposes, see UPVEC_FIRST_SPECIAL_CP.
1.49 + *
1.50 + * It would be possible to store only one range boundary per row,
1.51 + * but self-contained rows allow to later sort them by contents.
1.52 + */
1.53 +struct UPropsVectors;
1.54 +typedef struct UPropsVectors UPropsVectors;
1.55 +
1.56 +/*
1.57 + * Special pseudo code points for storing the initialValue and the errorValue,
1.58 + * which are used to initialize a UTrie2 or similar.
1.59 + */
1.60 +#define UPVEC_FIRST_SPECIAL_CP 0x110000
1.61 +#define UPVEC_INITIAL_VALUE_CP 0x110000
1.62 +#define UPVEC_ERROR_VALUE_CP 0x110001
1.63 +#define UPVEC_MAX_CP 0x110001
1.64 +
1.65 +/*
1.66 + * Special pseudo code point used in upvec_compact() signalling the end of
1.67 + * delivering special values and the beginning of delivering real ones.
1.68 + * Stable value, unlike UPVEC_MAX_CP which might grow over time.
1.69 + */
1.70 +#define UPVEC_START_REAL_VALUES_CP 0x200000
1.71 +
1.72 +/*
1.73 + * Open a UPropsVectors object.
1.74 + * @param columns Number of value integers (uint32_t) per row.
1.75 + */
1.76 +U_CAPI UPropsVectors * U_EXPORT2
1.77 +upvec_open(int32_t columns, UErrorCode *pErrorCode);
1.78 +
1.79 +U_CAPI void U_EXPORT2
1.80 +upvec_close(UPropsVectors *pv);
1.81 +
1.82 +/*
1.83 + * In rows for code points [start..end], select the column,
1.84 + * reset the mask bits and set the value bits (ANDed with the mask).
1.85 + *
1.86 + * Will set U_NO_WRITE_PERMISSION if called after upvec_compact().
1.87 + */
1.88 +U_CAPI void U_EXPORT2
1.89 +upvec_setValue(UPropsVectors *pv,
1.90 + UChar32 start, UChar32 end,
1.91 + int32_t column,
1.92 + uint32_t value, uint32_t mask,
1.93 + UErrorCode *pErrorCode);
1.94 +
1.95 +/*
1.96 + * Logically const but must not be used on the same pv concurrently!
1.97 + * Always returns 0 if called after upvec_compact().
1.98 + */
1.99 +U_CAPI uint32_t U_EXPORT2
1.100 +upvec_getValue(const UPropsVectors *pv, UChar32 c, int32_t column);
1.101 +
1.102 +/*
1.103 + * pRangeStart and pRangeEnd can be NULL.
1.104 + * @return NULL if rowIndex out of range and for illegal arguments,
1.105 + * or if called after upvec_compact()
1.106 + */
1.107 +U_CAPI uint32_t * U_EXPORT2
1.108 +upvec_getRow(const UPropsVectors *pv, int32_t rowIndex,
1.109 + UChar32 *pRangeStart, UChar32 *pRangeEnd);
1.110 +
1.111 +/*
1.112 + * Compact the vectors:
1.113 + * - modify the memory
1.114 + * - keep only unique vectors
1.115 + * - store them contiguously from the beginning of the memory
1.116 + * - for each (non-unique) row, call the handler function
1.117 + *
1.118 + * The handler's rowIndex is the index of the row in the compacted
1.119 + * memory block.
1.120 + * (Therefore, it starts at 0 increases in increments of the columns value.)
1.121 + *
1.122 + * In a first phase, only special values are delivered (each exactly once),
1.123 + * with start==end both equalling a special pseudo code point.
1.124 + * Then the handler is called once more with start==end==UPVEC_START_REAL_VALUES_CP
1.125 + * where rowIndex is the length of the compacted array,
1.126 + * and the row is arbitrary (but not NULL).
1.127 + * Then, in the second phase, the handler is called for each row of real values.
1.128 + */
1.129 +typedef void U_CALLCONV
1.130 +UPVecCompactHandler(void *context,
1.131 + UChar32 start, UChar32 end,
1.132 + int32_t rowIndex, uint32_t *row, int32_t columns,
1.133 + UErrorCode *pErrorCode);
1.134 +
1.135 +U_CAPI void U_EXPORT2
1.136 +upvec_compact(UPropsVectors *pv, UPVecCompactHandler *handler, void *context, UErrorCode *pErrorCode);
1.137 +
1.138 +/*
1.139 + * Get the vectors array after calling upvec_compact().
1.140 + * The caller must not modify nor release the returned array.
1.141 + * Returns NULL if called before upvec_compact().
1.142 + */
1.143 +U_CAPI const uint32_t * U_EXPORT2
1.144 +upvec_getArray(const UPropsVectors *pv, int32_t *pRows, int32_t *pColumns);
1.145 +
1.146 +/*
1.147 + * Get a clone of the vectors array after calling upvec_compact().
1.148 + * The caller owns the returned array and must uprv_free() it.
1.149 + * Returns NULL if called before upvec_compact().
1.150 + */
1.151 +U_CAPI uint32_t * U_EXPORT2
1.152 +upvec_cloneArray(const UPropsVectors *pv,
1.153 + int32_t *pRows, int32_t *pColumns, UErrorCode *pErrorCode);
1.154 +
1.155 +/*
1.156 + * Call upvec_compact(), create a 16-bit UTrie2 with indexes into the compacted
1.157 + * vectors array, and freeze the trie.
1.158 + */
1.159 +U_CAPI UTrie2 * U_EXPORT2
1.160 +upvec_compactToUTrie2WithRowIndexes(UPropsVectors *pv, UErrorCode *pErrorCode);
1.161 +
1.162 +struct UPVecToUTrie2Context {
1.163 + UTrie2 *trie;
1.164 + int32_t initialValue;
1.165 + int32_t errorValue;
1.166 + int32_t maxValue;
1.167 +};
1.168 +typedef struct UPVecToUTrie2Context UPVecToUTrie2Context;
1.169 +
1.170 +/* context=UPVecToUTrie2Context, creates the trie and stores the rowIndex values */
1.171 +U_CAPI void U_CALLCONV
1.172 +upvec_compactToUTrie2Handler(void *context,
1.173 + UChar32 start, UChar32 end,
1.174 + int32_t rowIndex, uint32_t *row, int32_t columns,
1.175 + UErrorCode *pErrorCode);
1.176 +
1.177 +U_CDECL_END
1.178 +
1.179 +#endif
