projects / apple / icu.git / blame
| Commit | Line | Data |
|---|---|---|
| f3c0d7a5 A |
1 | // © 2016 and later: Unicode, Inc. and others. |
| 2 | // License & terms of use: http://www.unicode.org/copyright.html | |
| 4388f060 A |
3 | /* |
| 4 | ******************************************************************************* | |
| 5 | * Copyright (C) 2011, International Business Machines | |
| 6 | * Corporation and others. All Rights Reserved. | |
| 7 | ******************************************************************************* | |
| 8 | * file name: patternprops.h | |
| f3c0d7a5 | 9 | * encoding: UTF-8 |
| 4388f060 A |
10 | * tab size: 8 (not used) |
| 11 | * indentation:4 | |
| 12 | * | |
| 13 | * created on: 2011mar13 | |
| 14 | * created by: Markus W. Scherer | |
| 15 | */ | |
| 16 | ||
| 17 | #ifndef __PATTERNPROPS_H__ | |
| 18 | #define __PATTERNPROPS_H__ | |
| 19 | ||
| 20 | #include "unicode/utypes.h" | |
| 21 | ||
| 22 | U_NAMESPACE_BEGIN | |
| 23 | ||
| 24 | /** | |
| 25 | * Implements the immutable Unicode properties Pattern_Syntax and Pattern_White_Space. | |
| 26 | * Hardcodes these properties, does not load data, does not depend on other ICU classes. | |
| 27 | * <p> | |
| 28 | * Note: Both properties include ASCII as well as non-ASCII, non-Latin-1 code points, | |
| 29 | * and both properties only include BMP code points (no supplementary ones). | |
| 30 | * Pattern_Syntax includes some unassigned code points. | |
| 31 | * <p> | |
| 32 | * [:Pattern_White_Space:] = | |
| 33 | * [\u0009-\u000D\ \u0085\u200E\u200F\u2028\u2029] | |
| 34 | * <p> | |
| 35 | * [:Pattern_Syntax:] = | |
| 36 | * [!-/\:-@\[-\^`\{-~\u00A1-\u00A7\u00A9\u00AB\u00AC\u00AE | |
| 37 | * \u00B0\u00B1\u00B6\u00BB\u00BF\u00D7\u00F7 | |
| 38 | * \u2010-\u2027\u2030-\u203E\u2041-\u2053\u2055-\u205E | |
| 39 | * \u2190-\u245F\u2500-\u2775\u2794-\u2BFF\u2E00-\u2E7F | |
| 40 | * \u3001-\u3003\u3008-\u3020\u3030\uFD3E\uFD3F\uFE45\uFE46] | |
| 41 | * @author mscherer | |
| 42 | */ | |
| 43 | class U_COMMON_API PatternProps { | |
| 44 | public: | |
| 45 | /** | |
| 46 | * @return TRUE if c is a Pattern_Syntax code point. | |
| 47 | */ | |
| 48 | static UBool isSyntax(UChar32 c); | |
| 49 | ||
| 50 | /** | |
| 51 | * @return TRUE if c is a Pattern_Syntax or Pattern_White_Space code point. | |
| 52 | */ | |
| 53 | static UBool isSyntaxOrWhiteSpace(UChar32 c); | |
| 54 | ||
| 55 | /** | |
| 56 | * @return TRUE if c is a Pattern_White_Space character. | |
| 57 | */ | |
| 58 | static UBool isWhiteSpace(UChar32 c); | |
| 59 | ||
| 60 | /** | |
| 61 | * Skips over Pattern_White_Space starting at s. | |
| 62 | * @return The smallest pointer at or after s with a non-white space character. | |
| 63 | */ | |
| 64 | static const UChar *skipWhiteSpace(const UChar *s, int32_t length); | |
| 65 | ||
| 66 | /** | |
| 67 | * @return s except with leading and trailing Pattern_White_Space removed and length adjusted. | |
| 68 | */ | |
| 69 | static const UChar *trimWhiteSpace(const UChar *s, int32_t &length); | |
| 70 | ||
| 71 | /** | |
| 72 | * Tests whether the string contains a "pattern identifier", that is, | |
| 73 | * whether it contains only non-Pattern_White_Space, non-Pattern_Syntax characters. | |
| 74 | * @return TRUE if there are no Pattern_White_Space or Pattern_Syntax characters in s. | |
| 75 | */ | |
| 76 | static UBool isIdentifier(const UChar *s, int32_t length); | |
| 77 | ||
| 78 | /** | |
| 79 | * Skips over a "pattern identifier" starting at index s. | |
| 80 | * @return The smallest pointer at or after s with | |
| 81 | * a Pattern_White_Space or Pattern_Syntax character. | |
| 82 | */ | |
| 83 | static const UChar *skipIdentifier(const UChar *s, int32_t length); | |
| 84 | ||
| 85 | private: | |
| 86 | PatternProps(); // no constructor: all static methods | |
| 87 | }; | |
| 88 | ||
| 89 | U_NAMESPACE_END | |
| 90 | ||
| 91 | #endif // __PATTERNPROPS_H__ |