com.ibm.icu.impl.PatternProps Maven / Gradle / Ivy

Show more of this group Show more artifacts with this name
Show all versions of virtdata-lib-realer Show documentation
With inspiration from other libraries
There is a newer version: 2.12.15
// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html#License
/*
*******************************************************************************
*   Copyright (C) 2011, International Business Machines
*   Corporation and others.  All Rights Reserved.
*******************************************************************************
*   created on: 2011feb25
*   created by: Markus W. Scherer
*/

package com.ibm.icu.impl;

/**
 * Implements the immutable Unicode properties Pattern_Syntax and Pattern_White_Space.
 * Hardcodes these properties, does not load data, does not depend on other ICU classes.
 * 
 * Note: Both properties include ASCII as well as non-ASCII, non-Latin-1 code points,
 * and both properties only include BMP code points (no supplementary ones).
 * Pattern_Syntax includes some unassigned code points.
 * 

 * [:Pattern_White_Space:] =
 *   [\u0009-\u000D\ \u0085\u200E\u200F\u2028\u2029]
 * 

 * [:Pattern_Syntax:] =
 *   [!-/\:-@\[-\^`\{-~\u00A1-\u00A7\u00A9\u00AB\u00AC\u00AE
 *    \u00B0\u00B1\u00B6\u00BB\u00BF\u00D7\u00F7
 *    \u2010-\u2027\u2030-\u203E\u2041-\u2053\u2055-\u205E
 *    \u2190-\u245F\u2500-\u2775\u2794-\u2BFF\u2E00-\u2E7F
 *    \u3001-\u3003\u3008-\u3020\u3030\uFD3E\uFD3F\uFE45\uFE46]
 * @author mscherer
 */
public final class PatternProps {
    /**
     * @return true if c is a Pattern_Syntax code point.
     */
    public static boolean isSyntax(int c) {
        if(c<0) {
            return false;
        } else if(c<=0xff) {
            return latin1[c]==3;
        } else if(c<0x2010) {
            return false;
        } else if(c<=0x3030) {
            int bits=syntax2000[index2000[(c-0x2000)>>5]];
            return ((bits>>(c&0x1f))&1)!=0;
        } else if(0xfd3e<=c && c<=0xfe46) {
            return c<=0xfd3f || 0xfe45<=c;
        } else {
            return false;
        }
    }

    /**
     * @return true if c is a Pattern_Syntax or Pattern_White_Space code point.
     */
    public static boolean isSyntaxOrWhiteSpace(int c) {
        if(c<0) {
            return false;
        } else if(c<=0xff) {
            return latin1[c]!=0;
        } else if(c<0x200e) {
            return false;
        } else if(c<=0x3030) {
            int bits=syntaxOrWhiteSpace2000[index2000[(c-0x2000)>>5]];
            return ((bits>>(c&0x1f))&1)!=0;
        } else if(0xfd3e<=c && c<=0xfe46) {
            return c<=0xfd3f || 0xfe45<=c;
        } else {
            return false;
        }
    }

    /**
     * @return true if c is a Pattern_White_Space character.
     */
    public static boolean isWhiteSpace(int c) {
        if(c<0) {
            return false;
        } else if(c<=0xff) {
            return latin1[c]==5;
        } else if(0x200e<=c && c<=0x2029) {
            return c<=0x200f || 0x2028<=c;
        } else {
            return false;
        }
    }

    /**
     * Skips over Pattern_White_Space starting at index i of the CharSequence.
     * @return The smallest index at or after i with a non-white space character.
     */
    public static int skipWhiteSpace(CharSequence s, int i) {
        while(i=limit) {
            return false;
        }
        do {
            if(isSyntaxOrWhiteSpace(s.charAt(start++))) {
                return false;
            }
        } while(start