sal/textenc/unichars.c

   1 /*************************************************************************
   2  *
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * Copyright 2000, 2010 Oracle and/or its affiliates.
   6  *
   7  * OpenOffice.org - a multi-platform office productivity suite
   8  *
   9  * This file is part of OpenOffice.org.
  10  *
  11  * OpenOffice.org is free software: you can redistribute it and/or modify
  12  * it under the terms of the GNU Lesser General Public License version 3
  13  * only, as published by the Free Software Foundation.
  14  *
  15  * OpenOffice.org is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18  * GNU Lesser General Public License version 3 for more details
  19  * (a copy is included in the LICENSE file that accompanied this code).
  20  *
  21  * You should have received a copy of the GNU Lesser General Public License
  22  * version 3 along with OpenOffice.org.  If not, see
  23  * <http://www.openoffice.org/license.html>
  24  * for a copy of the LGPLv3 License.
  25  *
  26  ************************************************************************/
  27
  28 #include "unichars.h"
  29 #include "osl/diagnose.h"
  30 #include "sal/types.h"
  31
  32 int ImplIsNoncharacter(sal_uInt32 nUtf32)
  33 {
  34     /* All code points that are noncharacters, as of Unicode 3.1.1. */
  35     return (nUtf32 >= 0xFDD0 && nUtf32 <= 0xFDEF)
  36            || (nUtf32 & 0xFFFF) >= 0xFFFE
  37            || nUtf32 > 0x10FFFF;
  38 }
  39
  40 int ImplIsControlOrFormat(sal_uInt32 nUtf32)
  41 {
  42     /* All code points of <http://www.unicode.org/Public/UNIDATA/
  43        UnicodeData.txt>, Version 3.1.1, that have a General Category of Cc
  44        (Other, Control) or Cf (Other, Format).
  45      */
  46     return nUtf32 <= 0x001F
  47            || (nUtf32 >= 0x007F && nUtf32 <= 0x009F)
  48            || nUtf32 == 0x070F /* SYRIAC ABBREVIATION MARK */
  49            || nUtf32 == 0x180B /* MONGOLIAN FREE VARIATION SELECTOR ONE */
  50            || nUtf32 == 0x180C /* MONGOLIAN FREE VARIATION SELECTOR TWO */
  51            || nUtf32 == 0x180D /* MONGOLIAN FREE VARIATION SELECTOR THREE */
  52            || nUtf32 == 0x180E /* MONGOLIAN VOWEL SEPARATOR */
  53            || nUtf32 == 0x200C /* ZERO WIDTH NON-JOINER */
  54            || nUtf32 == 0x200D /* ZERO WIDTH JOINER */
  55            || nUtf32 == 0x200E /* LEFT-TO-RIGHT MARK */
  56            || nUtf32 == 0x200F /* RIGHT-TO-LEFT MARK */
  57            || nUtf32 == 0x202A /* LEFT-TO-RIGHT EMBEDDING */
  58            || nUtf32 == 0x202B /* RIGHT-TO-LEFT EMBEDDING */
  59            || nUtf32 == 0x202C /* POP DIRECTIONAL FORMATTING */
  60            || nUtf32 == 0x202D /* LEFT-TO-RIGHT OVERRIDE */
  61            || nUtf32 == 0x202E /* RIGHT-TO-LEFT OVERRIDE */
  62            || nUtf32 == 0x206A /* INHIBIT SYMMETRIC SWAPPING */
  63            || nUtf32 == 0x206B /* ACTIVATE SYMMETRIC SWAPPING */
  64            || nUtf32 == 0x206C /* INHIBIT ARABIC FORM SHAPING */
  65            || nUtf32 == 0x206D /* ACTIVATE ARABIC FORM SHAPING */
  66            || nUtf32 == 0x206E /* NATIONAL DIGIT SHAPES */
  67            || nUtf32 == 0x206F /* NOMINAL DIGIT SHAPES */
  68            || nUtf32 == 0xFEFF /* ZERO WIDTH NO-BREAK SPACE */
  69            || nUtf32 == 0xFFF9 /* INTERLINEAR ANNOTATION ANCHOR */
  70            || nUtf32 == 0xFFFA /* INTERLINEAR ANNOTATION SEPARATOR */
  71            || nUtf32 == 0xFFFB /* INTERLINEAR ANNOTATION TERMINATOR */
  72            || nUtf32 == 0x1D173 /* MUSICAL SYMBOL BEGIN BEAM */
  73            || nUtf32 == 0x1D174 /* MUSICAL SYMBOL END BEAM */
  74            || nUtf32 == 0x1D175 /* MUSICAL SYMBOL BEGIN TIE */
  75            || nUtf32 == 0x1D176 /* MUSICAL SYMBOL END TIE */
  76            || nUtf32 == 0x1D177 /* MUSICAL SYMBOL BEGIN SLUR */
  77            || nUtf32 == 0x1D178 /* MUSICAL SYMBOL END SLUR */
  78            || nUtf32 == 0x1D179 /* MUSICAL SYMBOL BEGIN PHRASE */
  79            || nUtf32 == 0x1D17A /* MUSICAL SYMBOL END PHRASE */
  80            || nUtf32 == 0xE0001 /* LANGUAGE TAG */
  81            || (nUtf32 >= 0xE0020 && nUtf32 <= 0xE007F);
  82 }
  83
  84 int ImplIsHighSurrogate(sal_uInt32 nUtf32)
  85 {
  86     /* All code points that are high-surrogates, as of Unicode 3.1.1. */
  87     return nUtf32 >= 0xD800 && nUtf32 <= 0xDBFF;
  88 }
  89
  90 int ImplIsLowSurrogate(sal_uInt32 nUtf32)
  91 {
  92     /* All code points that are low-surrogates, as of Unicode 3.1.1. */
  93     return nUtf32 >= 0xDC00 && nUtf32 <= 0xDFFF;
  94 }
  95
  96 int ImplIsPrivateUse(sal_uInt32 nUtf32)
  97 {
  98     /* All code points of <http://www.unicode.org/Public/UNIDATA/
  99        UnicodeData.txt>, Version 3.1.1, that have a General Category of Co
 100        (Other, Private Use).
 101      */
 102     return (nUtf32 >= 0xE000 && nUtf32 <= 0xF8FF)
 103            || (nUtf32 >= 0xF0000 && nUtf32 <= 0xFFFFD)
 104            || (nUtf32 >= 0x100000 && nUtf32 <= 0x10FFFD);
 105 }
 106
 107 int ImplIsZeroWidth(sal_uInt32 nUtf32)
 108 {
 109     /* All code points of <http://www.unicode.org/Public/UNIDATA/
 110        UnicodeData.txt>, Version 3.1.1, that have "ZERO WIDTH" in their
 111        Character name.
 112      */
 113     return nUtf32 == 0x200B /* ZERO WIDTH SPACE */
 114            || nUtf32 == 0x200C /* ZERO WIDTH NON-JOINER */
 115            || nUtf32 == 0x200D /* ZERO WIDTH JOINER */
 116            || nUtf32 == 0xFEFF; /* ZEOR WIDTH NO-BREAK SPACE */
 117 }
 118
 119 sal_uInt32 ImplGetHighSurrogate(sal_uInt32 nUtf32)
 120 {
 121     OSL_ENSURE(nUtf32 >= 0x10000, "specification violation");
 122     return ((nUtf32 - 0x10000) >> 10) | 0xD800;
 123 }
 124
 125 sal_uInt32 ImplGetLowSurrogate(sal_uInt32 nUtf32)
 126 {
 127     OSL_ENSURE(nUtf32 >= 0x10000, "specification violation");
 128     return ((nUtf32 - 0x10000) & 0x3FF) | 0xDC00;
 129 }
 130
 131 sal_uInt32 ImplCombineSurrogates(sal_uInt32 nHigh, sal_uInt32 nLow)
 132 {
 133     OSL_ENSURE(ImplIsHighSurrogate(nHigh) && ImplIsLowSurrogate(nLow),
 134                "specification violation");
 135     return (((nHigh & 0x3FF) << 10) | (nLow & 0x3FF)) + 0x10000;
 136 }