sal/textenc/unichars.c

   1 /*************************************************************************
   2  *
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * Copyright 2008 by Sun Microsystems, Inc.
   6  *
   7  * OpenOffice.org - a multi-platform office productivity suite
   8  *
   9  * $RCSfile: unichars.c,v $
  10  * $Revision: 1.4 $
  11  *
  12  * This file is part of OpenOffice.org.
  13  *
  14  * OpenOffice.org is free software: you can redistribute it and/or modify
  15  * it under the terms of the GNU Lesser General Public License version 3
  16  * only, as published by the Free Software Foundation.
  17  *
  18  * OpenOffice.org is distributed in the hope that it will be useful,
  19  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  20  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  21  * GNU Lesser General Public License version 3 for more details
  22  * (a copy is included in the LICENSE file that accompanied this code).
  23  *
  24  * You should have received a copy of the GNU Lesser General Public License
  25  * version 3 along with OpenOffice.org.  If not, see
  26  * <http://www.openoffice.org/license.html>
  27  * for a copy of the LGPLv3 License.
  28  *
  29  ************************************************************************/
  30
  31 #include "unichars.h"
  32 #include "osl/diagnose.h"
  33 #include "sal/types.h"
  34
  35 int ImplIsNoncharacter(sal_uInt32 nUtf32)
  36 {
  37     /* All code points that are noncharacters, as of Unicode 3.1.1. */
  38     return (nUtf32 >= 0xFDD0 && nUtf32 <= 0xFDEF)
  39            || (nUtf32 & 0xFFFF) >= 0xFFFE
  40            || nUtf32 > 0x10FFFF;
  41 }
  42
  43 int ImplIsControlOrFormat(sal_uInt32 nUtf32)
  44 {
  45     /* All code points of <http://www.unicode.org/Public/UNIDATA/
  46        UnicodeData.txt>, Version 3.1.1, that have a General Category of Cc
  47        (Other, Control) or Cf (Other, Format).
  48      */
  49     return nUtf32 <= 0x001F
  50            || (nUtf32 >= 0x007F && nUtf32 <= 0x009F)
  51            || nUtf32 == 0x070F /* SYRIAC ABBREVIATION MARK */
  52            || nUtf32 == 0x180B /* MONGOLIAN FREE VARIATION SELECTOR ONE */
  53            || nUtf32 == 0x180C /* MONGOLIAN FREE VARIATION SELECTOR TWO */
  54            || nUtf32 == 0x180D /* MONGOLIAN FREE VARIATION SELECTOR THREE */
  55            || nUtf32 == 0x180E /* MONGOLIAN VOWEL SEPARATOR */
  56            || nUtf32 == 0x200C /* ZERO WIDTH NON-JOINER */
  57            || nUtf32 == 0x200D /* ZERO WIDTH JOINER */
  58            || nUtf32 == 0x200E /* LEFT-TO-RIGHT MARK */
  59            || nUtf32 == 0x200F /* RIGHT-TO-LEFT MARK */
  60            || nUtf32 == 0x202A /* LEFT-TO-RIGHT EMBEDDING */
  61            || nUtf32 == 0x202B /* RIGHT-TO-LEFT EMBEDDING */
  62            || nUtf32 == 0x202C /* POP DIRECTIONAL FORMATTING */
  63            || nUtf32 == 0x202D /* LEFT-TO-RIGHT OVERRIDE */
  64            || nUtf32 == 0x202E /* RIGHT-TO-LEFT OVERRIDE */
  65            || nUtf32 == 0x206A /* INHIBIT SYMMETRIC SWAPPING */
  66            || nUtf32 == 0x206B /* ACTIVATE SYMMETRIC SWAPPING */
  67            || nUtf32 == 0x206C /* INHIBIT ARABIC FORM SHAPING */
  68            || nUtf32 == 0x206D /* ACTIVATE ARABIC FORM SHAPING */
  69            || nUtf32 == 0x206E /* NATIONAL DIGIT SHAPES */
  70            || nUtf32 == 0x206F /* NOMINAL DIGIT SHAPES */
  71            || nUtf32 == 0xFEFF /* ZERO WIDTH NO-BREAK SPACE */
  72            || nUtf32 == 0xFFF9 /* INTERLINEAR ANNOTATION ANCHOR */
  73            || nUtf32 == 0xFFFA /* INTERLINEAR ANNOTATION SEPARATOR */
  74            || nUtf32 == 0xFFFB /* INTERLINEAR ANNOTATION TERMINATOR */
  75            || nUtf32 == 0x1D173 /* MUSICAL SYMBOL BEGIN BEAM */
  76            || nUtf32 == 0x1D174 /* MUSICAL SYMBOL END BEAM */
  77            || nUtf32 == 0x1D175 /* MUSICAL SYMBOL BEGIN TIE */
  78            || nUtf32 == 0x1D176 /* MUSICAL SYMBOL END TIE */
  79            || nUtf32 == 0x1D177 /* MUSICAL SYMBOL BEGIN SLUR */
  80            || nUtf32 == 0x1D178 /* MUSICAL SYMBOL END SLUR */
  81            || nUtf32 == 0x1D179 /* MUSICAL SYMBOL BEGIN PHRASE */
  82            || nUtf32 == 0x1D17A /* MUSICAL SYMBOL END PHRASE */
  83            || nUtf32 == 0xE0001 /* LANGUAGE TAG */
  84            || (nUtf32 >= 0xE0020 && nUtf32 <= 0xE007F);
  85 }
  86
  87 int ImplIsHighSurrogate(sal_uInt32 nUtf32)
  88 {
  89     /* All code points that are high-surrogates, as of Unicode 3.1.1. */
  90     return nUtf32 >= 0xD800 && nUtf32 <= 0xDBFF;
  91 }
  92
  93 int ImplIsLowSurrogate(sal_uInt32 nUtf32)
  94 {
  95     /* All code points that are low-surrogates, as of Unicode 3.1.1. */
  96     return nUtf32 >= 0xDC00 && nUtf32 <= 0xDFFF;
  97 }
  98
  99 int ImplIsPrivateUse(sal_uInt32 nUtf32)
 100 {
 101     /* All code points of <http://www.unicode.org/Public/UNIDATA/
 102        UnicodeData.txt>, Version 3.1.1, that have a General Category of Co
 103        (Other, Private Use).
 104      */
 105     return (nUtf32 >= 0xE000 && nUtf32 <= 0xF8FF)
 106            || (nUtf32 >= 0xF0000 && nUtf32 <= 0xFFFFD)
 107            || (nUtf32 >= 0x100000 && nUtf32 <= 0x10FFFD);
 108 }
 109
 110 int ImplIsZeroWidth(sal_uInt32 nUtf32)
 111 {
 112     /* All code points of <http://www.unicode.org/Public/UNIDATA/
 113        UnicodeData.txt>, Version 3.1.1, that have "ZERO WIDTH" in their
 114        Character name.
 115      */
 116     return nUtf32 == 0x200B /* ZERO WIDTH SPACE */
 117            || nUtf32 == 0x200C /* ZERO WIDTH NON-JOINER */
 118            || nUtf32 == 0x200D /* ZERO WIDTH JOINER */
 119            || nUtf32 == 0xFEFF; /* ZEOR WIDTH NO-BREAK SPACE */
 120 }
 121
 122 sal_uInt32 ImplGetHighSurrogate(sal_uInt32 nUtf32)
 123 {
 124     OSL_ENSURE(nUtf32 >= 0x10000, "specification violation");
 125     return ((nUtf32 - 0x10000) >> 10) | 0xD800;
 126 }
 127
 128 sal_uInt32 ImplGetLowSurrogate(sal_uInt32 nUtf32)
 129 {
 130     OSL_ENSURE(nUtf32 >= 0x10000, "specification violation");
 131     return ((nUtf32 - 0x10000) & 0x3FF) | 0xDC00;
 132 }
 133
 134 sal_uInt32 ImplCombineSurrogates(sal_uInt32 nHigh, sal_uInt32 nLow)
 135 {
 136     OSL_ENSURE(ImplIsHighSurrogate(nHigh) && ImplIsLowSurrogate(nLow),
 137                "specification violation");
 138     return (((nHigh & 0x3FF) << 10) | (nLow & 0x3FF)) + 0x10000;
 139 }