icuSources/common/patternprops.h

   1 // © 2016 and later: Unicode, Inc. and others.
   2 // License & terms of use: http://www.unicode.org/copyright.html
   3 /*
   4 *******************************************************************************
   5 *   Copyright (C) 2011, International Business Machines
   6 *   Corporation and others.  All Rights Reserved.
   7 *******************************************************************************
   8 *   file name:  patternprops.h
   9 *   encoding:   UTF-8
  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/unistr.h"
  21 #include "unicode/utypes.h"
  22
  23 U_NAMESPACE_BEGIN
  24
  25 /**
  26  * Implements the immutable Unicode properties Pattern_Syntax and Pattern_White_Space.
  27  * Hardcodes these properties, does not load data, does not depend on other ICU classes.
  28  * <p>
  29  * Note: Both properties include ASCII as well as non-ASCII, non-Latin-1 code points,
  30  * and both properties only include BMP code points (no supplementary ones).
  31  * Pattern_Syntax includes some unassigned code points.
  32  * <p>
  33  * [:Pattern_White_Space:] =
  34  *   [\u0009-\u000D\ \u0085\u200E\u200F\u2028\u2029]
  35  * <p>
  36  * [:Pattern_Syntax:] =
  37  *   [!-/\:-@\[-\^`\{-~\u00A1-\u00A7\u00A9\u00AB\u00AC\u00AE
  38  *    \u00B0\u00B1\u00B6\u00BB\u00BF\u00D7\u00F7
  39  *    \u2010-\u2027\u2030-\u203E\u2041-\u2053\u2055-\u205E
  40  *    \u2190-\u245F\u2500-\u2775\u2794-\u2BFF\u2E00-\u2E7F
  41  *    \u3001-\u3003\u3008-\u3020\u3030\uFD3E\uFD3F\uFE45\uFE46]
  42  * @author mscherer
  43  */
  44 class U_COMMON_API PatternProps {
  45 public:
  46     /**
  47      * @return TRUE if c is a Pattern_Syntax code point.
  48      */
  49     static UBool isSyntax(UChar32 c);
  50
  51     /**
  52      * @return TRUE if c is a Pattern_Syntax or Pattern_White_Space code point.
  53      */
  54     static UBool isSyntaxOrWhiteSpace(UChar32 c);
  55
  56     /**
  57      * @return TRUE if c is a Pattern_White_Space character.
  58      */
  59     static UBool isWhiteSpace(UChar32 c);
  60
  61     /**
  62      * Skips over Pattern_White_Space starting at s.
  63      * @return The smallest pointer at or after s with a non-white space character.
  64      */
  65     static const UChar *skipWhiteSpace(const UChar *s, int32_t length);
  66
  67     /**
  68      * Skips over Pattern_White_Space starting at index start in s.
  69      * @return The smallest index at or after start with a non-white space character.
  70      */
  71     static int32_t skipWhiteSpace(const UnicodeString &s, int32_t start);
  72
  73     /**
  74      * @return s except with leading and trailing Pattern_White_Space removed and length adjusted.
  75      */
  76     static const UChar *trimWhiteSpace(const UChar *s, int32_t &length);
  77
  78     /**
  79      * Tests whether the string contains a "pattern identifier", that is,
  80      * whether it contains only non-Pattern_White_Space, non-Pattern_Syntax characters.
  81      * @return TRUE if there are no Pattern_White_Space or Pattern_Syntax characters in s.
  82      */
  83     static UBool isIdentifier(const UChar *s, int32_t length);
  84
  85     /**
  86      * Skips over a "pattern identifier" starting at index s.
  87      * @return The smallest pointer at or after s with
  88      *         a Pattern_White_Space or Pattern_Syntax character.
  89      */
  90     static const UChar *skipIdentifier(const UChar *s, int32_t length);
  91
  92 private:
  93     PatternProps();  // no constructor: all static methods
  94 };
  95
  96 U_NAMESPACE_END
  97
  98 #endif  // __PATTERNPROPS_H__