intl/icu/source/common/unorm_it.h@6474c204b198 (annotated)
intl/icu/source/common/unorm_it.h
Wed, 31 Dec 2014 06:09:35 +0100
- author
- Michael Schloh von Bennewitz <michael@schloh.com>
- date
- Wed, 31 Dec 2014 06:09:35 +0100
- changeset 0
- 6474c204b198
- permissions
- -rw-r--r--
Cloned upstream origin tor-browser at tor-browser-31.3.0esr-4.5-1-build1
revision ID fc1c9ff7c1b2defdbc039f12214767608f46423f for hacking purpose.
| michael@0 | 1 | /* |
| michael@0 | 2 | ******************************************************************************* |
| michael@0 | 3 | * |
| michael@0 | 4 | * Copyright (C) 2003, International Business Machines |
| michael@0 | 5 | * Corporation and others. All Rights Reserved. |
| michael@0 | 6 | * |
| michael@0 | 7 | ******************************************************************************* |
| michael@0 | 8 | * file name: unorm_it.h |
| michael@0 | 9 | * encoding: US-ASCII |
| michael@0 | 10 | * tab size: 8 (not used) |
| michael@0 | 11 | * indentation:4 |
| michael@0 | 12 | * |
| michael@0 | 13 | * created on: 2003jan21 |
| michael@0 | 14 | * created by: Markus W. Scherer |
| michael@0 | 15 | */ |
| michael@0 | 16 | |
| michael@0 | 17 | #ifndef __UNORM_IT_H__ |
| michael@0 | 18 | #define __UNORM_IT_H__ |
| michael@0 | 19 | |
| michael@0 | 20 | #include "unicode/utypes.h" |
| michael@0 | 21 | |
| michael@0 | 22 | #if !UCONFIG_NO_COLLATION && !UCONFIG_NO_NORMALIZATION |
| michael@0 | 23 | |
| michael@0 | 24 | #include "unicode/uiter.h" |
| michael@0 | 25 | #include "unicode/unorm.h" |
| michael@0 | 26 | |
| michael@0 | 27 | /** |
| michael@0 | 28 | * Normalizing UCharIterator wrapper. |
| michael@0 | 29 | * This internal API basically duplicates the functionality of the C++ Normalizer |
| michael@0 | 30 | * but |
| michael@0 | 31 | * - it actually implements a character iterator (UCharIterator) |
| michael@0 | 32 | * with few restrictions (see unorm_setIter()) |
| michael@0 | 33 | * - it supports UCharIterator getState()/setState() |
| michael@0 | 34 | * - it uses lower-level APIs and buffers more text and states, |
| michael@0 | 35 | * hopefully resulting in higher performance |
| michael@0 | 36 | * |
| michael@0 | 37 | * Usage example: |
| michael@0 | 38 | * \code |
| michael@0 | 39 | * function(UCharIterator *srcIter) { |
| michael@0 | 40 | * UNormIterator *uni; |
| michael@0 | 41 | * UCharIterator *iter; |
| michael@0 | 42 | * UErrorCode errorCode; |
| michael@0 | 43 | * |
| michael@0 | 44 | * errorCode=U_ZERO_ERROR; |
| michael@0 | 45 | * uni=unorm_openIter(&errorCode); |
| michael@0 | 46 | * if(U_FAILURE(errorCode)) { |
| michael@0 | 47 | * // report error |
| michael@0 | 48 | * return; |
| michael@0 | 49 | * } |
| michael@0 | 50 | * |
| michael@0 | 51 | * iter=unorm_setIter(uni, srcIter, UNORM_FCD, &errorCode); |
| michael@0 | 52 | * if(U_FAILURE(errorCode)) { |
| michael@0 | 53 | * // report error |
| michael@0 | 54 | * } else { |
| michael@0 | 55 | * // use iter to iterate over the canonically ordered |
| michael@0 | 56 | * // version of srcIter's text |
| michael@0 | 57 | * uint32_t state; |
| michael@0 | 58 | * |
| michael@0 | 59 | * ... |
| michael@0 | 60 | * |
| michael@0 | 61 | * state=uiter_getState(iter); |
| michael@0 | 62 | * if(state!=UITER_NO_STATE) { |
| michael@0 | 63 | * // use valid state, store it, use iter some more |
| michael@0 | 64 | * ... |
| michael@0 | 65 | * |
| michael@0 | 66 | * // later restore iter to the saved state: |
| michael@0 | 67 | * uiter_setState(iter, state, &errorCode); |
| michael@0 | 68 | * |
| michael@0 | 69 | * ... |
| michael@0 | 70 | * } |
| michael@0 | 71 | * |
| michael@0 | 72 | * ... |
| michael@0 | 73 | * } |
| michael@0 | 74 | * unorm_closeIter(uni); |
| michael@0 | 75 | * } |
| michael@0 | 76 | * \endcode |
| michael@0 | 77 | * |
| michael@0 | 78 | * See also the ICU test suites. |
| michael@0 | 79 | * |
| michael@0 | 80 | * @internal |
| michael@0 | 81 | */ |
| michael@0 | 82 | struct UNormIterator; |
| michael@0 | 83 | typedef struct UNormIterator UNormIterator; |
| michael@0 | 84 | |
| michael@0 | 85 | /** |
| michael@0 | 86 | * Size of a stack buffer to hold a UNormIterator, see the stackMem parameter |
| michael@0 | 87 | * of unorm_openIter(). |
| michael@0 | 88 | * |
| michael@0 | 89 | * @internal |
| michael@0 | 90 | */ |
| michael@0 | 91 | #define UNORM_ITER_SIZE 1024 |
| michael@0 | 92 | |
| michael@0 | 93 | /** |
| michael@0 | 94 | * Open a normalizing iterator. Must be closed later. |
| michael@0 | 95 | * Use unorm_setIter(). |
| michael@0 | 96 | * |
| michael@0 | 97 | * @param stackMem Pointer to preallocated (stack-allocated) buffer to hold |
| michael@0 | 98 | * the UNormIterator if possible; can be NULL. |
| michael@0 | 99 | * @param stackMemSize Number of bytes at stackMem; can be 0, |
| michael@0 | 100 | * or should be >= UNORM_ITER_SIZE for a non-NULL stackMem. |
| michael@0 | 101 | * @param pErrorCode ICU error code |
| michael@0 | 102 | * @return an allocated and pre-initialized UNormIterator |
| michael@0 | 103 | * @internal |
| michael@0 | 104 | */ |
| michael@0 | 105 | U_CAPI UNormIterator * U_EXPORT2 |
| michael@0 | 106 | unorm_openIter(void *stackMem, int32_t stackMemSize, UErrorCode *pErrorCode); |
| michael@0 | 107 | |
| michael@0 | 108 | /** |
| michael@0 | 109 | * Close a normalizing iterator. |
| michael@0 | 110 | * |
| michael@0 | 111 | * @param uni UNormIterator from unorm_openIter() |
| michael@0 | 112 | * @internal |
| michael@0 | 113 | */ |
| michael@0 | 114 | U_CAPI void U_EXPORT2 |
| michael@0 | 115 | unorm_closeIter(UNormIterator *uni); |
| michael@0 | 116 | |
| michael@0 | 117 | /** |
| michael@0 | 118 | * Set a UCharIterator and a normalization mode for the normalizing iterator |
| michael@0 | 119 | * to wrap. The normalizing iterator will read from the character iterator, |
| michael@0 | 120 | * normalize the text, and in turn deliver it with its own wrapper UCharIterator |
| michael@0 | 121 | * interface which it returns. |
| michael@0 | 122 | * |
| michael@0 | 123 | * The source iterator remains at its current position through the unorm_setIter() |
| michael@0 | 124 | * call but will be used and moved as soon as the |
| michael@0 | 125 | * the returned normalizing iterator is. |
| michael@0 | 126 | * |
| michael@0 | 127 | * The returned interface pointer is valid for as long as the normalizing iterator |
| michael@0 | 128 | * is open and until another unorm_setIter() call is made on it. |
| michael@0 | 129 | * |
| michael@0 | 130 | * The normalizing iterator's UCharIterator interface has the following properties: |
| michael@0 | 131 | * - getIndex() and move() will almost always return UITER_UNKNOWN_INDEX |
| michael@0 | 132 | * - getState() will return UITER_NO_STATE for unknown states for positions |
| michael@0 | 133 | * that are not at normalization boundaries |
| michael@0 | 134 | * |
| michael@0 | 135 | * @param uni UNormIterator from unorm_openIter() |
| michael@0 | 136 | * @param iter The source text UCharIterator to be wrapped. It is aliases into the normalizing iterator. |
| michael@0 | 137 | * Must support getState() and setState(). |
| michael@0 | 138 | * @param mode The normalization mode. |
| michael@0 | 139 | * @param pErrorCode ICU error code |
| michael@0 | 140 | * @return an alias to the normalizing iterator's UCharIterator interface |
| michael@0 | 141 | * @internal |
| michael@0 | 142 | */ |
| michael@0 | 143 | U_CAPI UCharIterator * U_EXPORT2 |
| michael@0 | 144 | unorm_setIter(UNormIterator *uni, UCharIterator *iter, UNormalizationMode mode, UErrorCode *pErrorCode); |
| michael@0 | 145 | |
| michael@0 | 146 | #endif /* uconfig.h switches */ |
| michael@0 | 147 | |
| michael@0 | 148 | #endif |
