ICU 59.1  59.1
unorm2.h
Go to the documentation of this file.
1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
3 /*
4 *******************************************************************************
5 *
6 * Copyright (C) 2009-2015, International Business Machines
7 * Corporation and others. All Rights Reserved.
8 *
9 *******************************************************************************
10 * file name: unorm2.h
11 * encoding: UTF-8
12 * tab size: 8 (not used)
13 * indentation:4
14 *
15 * created on: 2009dec15
16 * created by: Markus W. Scherer
17 */
18 
19 #ifndef __UNORM2_H__
20 #define __UNORM2_H__
21 
33 #include "unicode/utypes.h"
34 #include "unicode/localpointer.h"
35 #include "unicode/uset.h"
36 
44 typedef enum {
87 
113 
118 struct UNormalizer2;
119 typedef struct UNormalizer2 UNormalizer2;
121 #if !UCONFIG_NO_NORMALIZATION
122 
134 U_STABLE const UNormalizer2 * U_EXPORT2
135 unorm2_getNFCInstance(UErrorCode *pErrorCode);
136 
148 U_STABLE const UNormalizer2 * U_EXPORT2
149 unorm2_getNFDInstance(UErrorCode *pErrorCode);
150 
162 U_STABLE const UNormalizer2 * U_EXPORT2
164 
176 U_STABLE const UNormalizer2 * U_EXPORT2
178 
190 U_STABLE const UNormalizer2 * U_EXPORT2
192 
214 U_STABLE const UNormalizer2 * U_EXPORT2
215 unorm2_getInstance(const char *packageName,
216  const char *name,
217  UNormalization2Mode mode,
218  UErrorCode *pErrorCode);
219 
235 U_STABLE UNormalizer2 * U_EXPORT2
236 unorm2_openFiltered(const UNormalizer2 *norm2, const USet *filterSet, UErrorCode *pErrorCode);
237 
244 U_STABLE void U_EXPORT2
245 unorm2_close(UNormalizer2 *norm2);
246 
247 #if U_SHOW_CPLUSPLUS_API
248 
250 
261 
263 
264 #endif
265 
282 U_STABLE int32_t U_EXPORT2
283 unorm2_normalize(const UNormalizer2 *norm2,
284  const UChar *src, int32_t length,
285  UChar *dest, int32_t capacity,
286  UErrorCode *pErrorCode);
305 U_STABLE int32_t U_EXPORT2
307  UChar *first, int32_t firstLength, int32_t firstCapacity,
308  const UChar *second, int32_t secondLength,
309  UErrorCode *pErrorCode);
328 U_STABLE int32_t U_EXPORT2
329 unorm2_append(const UNormalizer2 *norm2,
330  UChar *first, int32_t firstLength, int32_t firstCapacity,
331  const UChar *second, int32_t secondLength,
332  UErrorCode *pErrorCode);
333 
353 U_STABLE int32_t U_EXPORT2
355  UChar32 c, UChar *decomposition, int32_t capacity,
356  UErrorCode *pErrorCode);
357 
387 U_STABLE int32_t U_EXPORT2
389  UChar32 c, UChar *decomposition, int32_t capacity,
390  UErrorCode *pErrorCode);
391 
407 U_STABLE UChar32 U_EXPORT2
408 unorm2_composePair(const UNormalizer2 *norm2, UChar32 a, UChar32 b);
409 
419 U_STABLE uint8_t U_EXPORT2
421 
438 U_STABLE UBool U_EXPORT2
439 unorm2_isNormalized(const UNormalizer2 *norm2,
440  const UChar *s, int32_t length,
441  UErrorCode *pErrorCode);
442 
461 unorm2_quickCheck(const UNormalizer2 *norm2,
462  const UChar *s, int32_t length,
463  UErrorCode *pErrorCode);
464 
489 U_STABLE int32_t U_EXPORT2
491  const UChar *s, int32_t length,
492  UErrorCode *pErrorCode);
493 
503 U_STABLE UBool U_EXPORT2
505 
515 U_STABLE UBool U_EXPORT2
517 
526 U_STABLE UBool U_EXPORT2
527 unorm2_isInert(const UNormalizer2 *norm2, UChar32 c);
528 
534 #define UNORM_INPUT_IS_FCD 0x20000
535 
541 #define U_COMPARE_IGNORE_CASE 0x10000
542 
543 #ifndef U_COMPARE_CODE_POINT_ORDER
544 /* see also unistr.h and ustring.h */
550 #define U_COMPARE_CODE_POINT_ORDER 0x8000
551 #endif
552 
619 U_STABLE int32_t U_EXPORT2
620 unorm_compare(const UChar *s1, int32_t length1,
621  const UChar *s2, int32_t length2,
622  uint32_t options,
623  UErrorCode *pErrorCode);
624 
625 #endif /* !UCONFIG_NO_NORMALIZATION */
626 #endif /* __UNORM2_H__ */
const UNormalizer2 * unorm2_getNFDInstance(UErrorCode *pErrorCode)
Returns a UNormalizer2 instance for Unicode NFD normalization.
"Fast C or D" form.
Definition: unorm2.h:75
The input string is not in the normalization form.
Definition: unorm2.h:98
int32_t unorm_compare(const UChar *s1, int32_t length1, const UChar *s2, int32_t length2, uint32_t options, UErrorCode *pErrorCode)
Compares two strings for canonical equivalence.
const UNormalizer2 * unorm2_getInstance(const char *packageName, const char *name, UNormalization2Mode mode, UErrorCode *pErrorCode)
Returns a UNormalizer2 instance which uses the specified data file (packageName/name similar to ucnv_...
int32_t unorm2_getRawDecomposition(const UNormalizer2 *norm2, UChar32 c, UChar *decomposition, int32_t capacity, UErrorCode *pErrorCode)
Gets the raw decomposition mapping of c.
Map, and reorder canonically.
Definition: unorm2.h:62
UBool unorm2_isInert(const UNormalizer2 *norm2, UChar32 c)
Tests if the character is normalization-inert.
The input string is in the normalization form.
Definition: unorm2.h:103
void unorm2_close(UNormalizer2 *norm2)
Closes a UNormalizer2 instance from unorm2_openFiltered().
UBool unorm2_isNormalized(const UNormalizer2 *norm2, const UChar *s, int32_t length, UErrorCode *pErrorCode)
Tests if the string is normalized.
const UNormalizer2 * unorm2_getNFCInstance(UErrorCode *pErrorCode)
Returns a UNormalizer2 instance for Unicode NFC normalization.
The input string may or may not be in the normalization form.
Definition: unorm2.h:111
Compose only contiguously.
Definition: unorm2.h:85
int32_t unorm2_getDecomposition(const UNormalizer2 *norm2, UChar32 c, UChar *decomposition, int32_t capacity, UErrorCode *pErrorCode)
Gets the decomposition mapping of c.
UNormalizer2 * unorm2_openFiltered(const UNormalizer2 *norm2, const USet *filterSet, UErrorCode *pErrorCode)
Constructs a filtered normalizer wrapping any UNormalizer2 instance and a filter set.
C API: Unicode Set.
UNormalizationCheckResult unorm2_quickCheck(const UNormalizer2 *norm2, const UChar *s, int32_t length, UErrorCode *pErrorCode)
Tests if the string is normalized.
#define U_NAMESPACE_BEGIN
This is used to begin a declaration of a public ICU C++ API.
Definition: uversion.h:131
int32_t unorm2_normalize(const UNormalizer2 *norm2, const UChar *src, int32_t length, UChar *dest, int32_t capacity, UErrorCode *pErrorCode)
Writes the normalized form of the source string to the destination string (replacing its contents) an...
#define U_DEFINE_LOCAL_OPEN_POINTER(LocalPointerClassName, Type, closeFunction)
"Smart pointer" definition macro, deletes objects via the closeFunction.
Definition: localpointer.h:496
C++ API: "Smart pointers" for use with and in ICU4C C++ code.
struct UNormalizer2 UNormalizer2
C typedef for struct UNormalizer2.
Definition: unorm2.h:119
int32_t UChar32
Define UChar32 as a type for single Unicode code points.
Definition: umachine.h:396
const UNormalizer2 * unorm2_getNFKDInstance(UErrorCode *pErrorCode)
Returns a UNormalizer2 instance for Unicode NFKD normalization.
UChar32 unorm2_composePair(const UNormalizer2 *norm2, UChar32 a, UChar32 b)
Performs pairwise composition of a & b and returns the composite if there is one. ...
const UNormalizer2 * unorm2_getNFKCCasefoldInstance(UErrorCode *pErrorCode)
Returns a UNormalizer2 instance for Unicode NFKC_Casefold normalization.
uint16_t UChar
The base type for UTF-16 code units and pointers.
Definition: umachine.h:349
#define U_NAMESPACE_END
This is used to end a declaration of a public ICU C++ API.
Definition: uversion.h:132
struct USet USet
Definition: ucnv.h:69
UNormalization2Mode
Constants for normalization modes.
Definition: unorm2.h:44
UErrorCode
Error code to replace exception handling, so that the code is compatible with all C++ compilers...
Definition: utypes.h:396
uint8_t unorm2_getCombiningClass(const UNormalizer2 *norm2, UChar32 c)
Gets the combining class of c.
int32_t unorm2_normalizeSecondAndAppend(const UNormalizer2 *norm2, UChar *first, int32_t firstLength, int32_t firstCapacity, const UChar *second, int32_t secondLength, UErrorCode *pErrorCode)
Appends the normalized form of the second string to the first string (merging them at the boundary) a...
UBool unorm2_hasBoundaryAfter(const UNormalizer2 *norm2, UChar32 c)
Tests if the character always has a normalization boundary after it, regardless of context...
int32_t unorm2_spanQuickCheckYes(const UNormalizer2 *norm2, const UChar *s, int32_t length, UErrorCode *pErrorCode)
Returns the end of the normalized substring of the input string.
Basic definitions for ICU, for both C and C++ APIs.
UBool unorm2_hasBoundaryBefore(const UNormalizer2 *norm2, UChar32 c)
Tests if the character always has a normalization boundary before it, regardless of context...
int32_t unorm2_append(const UNormalizer2 *norm2, UChar *first, int32_t firstLength, int32_t firstCapacity, const UChar *second, int32_t secondLength, UErrorCode *pErrorCode)
Appends the second string to the first string (merging them at the boundary) and returns the length o...
Decomposition followed by composition.
Definition: unorm2.h:53
"Smart pointer" class, closes a UNormalizer2 via unorm2_close().
const UNormalizer2 * unorm2_getNFKCInstance(UErrorCode *pErrorCode)
Returns a UNormalizer2 instance for Unicode NFKC normalization.
UNormalizationCheckResult
Result values for normalization quick check functions.
Definition: unorm2.h:93
#define U_STABLE
This is used to declare a function as a stable public ICU C API.
Definition: umachine.h:111
int8_t UBool
The ICU boolean type.
Definition: umachine.h:236