sal/textenc/unichars.c

   1 /* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
   2 /*************************************************************************
   3  *
   4  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   5  *
   6  * Copyright 2000, 2010 Oracle and/or its affiliates.
   7  *
   8  * OpenOffice.org - a multi-platform office productivity suite
   9  *
  10  * This file is part of OpenOffice.org.
  11  *
  12  * OpenOffice.org is free software: you can redistribute it and/or modify
  13  * it under the terms of the GNU Lesser General Public License version 3
  14  * only, as published by the Free Software Foundation.
  15  *
  16  * OpenOffice.org is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19  * GNU Lesser General Public License version 3 for more details
  20  * (a copy is included in the LICENSE file that accompanied this code).
  21  *
  22  * You should have received a copy of the GNU Lesser General Public License
  23  * version 3 along with OpenOffice.org.  If not, see
  24  * <http://www.openoffice.org/license.html>
  25  * for a copy of the LGPLv3 License.
  26  *
  27  ************************************************************************/
  28
  29 #include "unichars.h"
  30 #include "osl/diagnose.h"
  31 #include "sal/types.h"
  32
  33 int ImplIsNoncharacter(sal_uInt32 nUtf32)
  34 {
  35     /* All code points that are noncharacters, as of Unicode 3.1.1. */
  36     return (nUtf32 >= 0xFDD0 && nUtf32 <= 0xFDEF)
  37            || (nUtf32 & 0xFFFF) >= 0xFFFE
  38            || nUtf32 > 0x10FFFF;
  39 }
  40
  41 int ImplIsControlOrFormat(sal_uInt32 nUtf32)
  42 {
  43     /* All code points of <http://www.unicode.org/Public/UNIDATA/
  44        UnicodeData.txt>, Version 3.1.1, that have a General Category of Cc
  45        (Other, Control) or Cf (Other, Format).
  46      */
  47     return nUtf32 <= 0x001F
  48            || (nUtf32 >= 0x007F && nUtf32 <= 0x009F)
  49            || nUtf32 == 0x070F /* SYRIAC ABBREVIATION MARK */
  50            || nUtf32 == 0x180B /* MONGOLIAN FREE VARIATION SELECTOR ONE */
  51            || nUtf32 == 0x180C /* MONGOLIAN FREE VARIATION SELECTOR TWO */
  52            || nUtf32 == 0x180D /* MONGOLIAN FREE VARIATION SELECTOR THREE */
  53            || nUtf32 == 0x180E /* MONGOLIAN VOWEL SEPARATOR */
  54            || nUtf32 == 0x200C /* ZERO WIDTH NON-JOINER */
  55            || nUtf32 == 0x200D /* ZERO WIDTH JOINER */
  56            || nUtf32 == 0x200E /* LEFT-TO-RIGHT MARK */
  57            || nUtf32 == 0x200F /* RIGHT-TO-LEFT MARK */
  58            || nUtf32 == 0x202A /* LEFT-TO-RIGHT EMBEDDING */
  59            || nUtf32 == 0x202B /* RIGHT-TO-LEFT EMBEDDING */
  60            || nUtf32 == 0x202C /* POP DIRECTIONAL FORMATTING */
  61            || nUtf32 == 0x202D /* LEFT-TO-RIGHT OVERRIDE */
  62            || nUtf32 == 0x202E /* RIGHT-TO-LEFT OVERRIDE */
  63            || nUtf32 == 0x206A /* INHIBIT SYMMETRIC SWAPPING */
  64            || nUtf32 == 0x206B /* ACTIVATE SYMMETRIC SWAPPING */
  65            || nUtf32 == 0x206C /* INHIBIT ARABIC FORM SHAPING */
  66            || nUtf32 == 0x206D /* ACTIVATE ARABIC FORM SHAPING */
  67            || nUtf32 == 0x206E /* NATIONAL DIGIT SHAPES */
  68            || nUtf32 == 0x206F /* NOMINAL DIGIT SHAPES */
  69            || nUtf32 == 0xFEFF /* ZERO WIDTH NO-BREAK SPACE */
  70            || nUtf32 == 0xFFF9 /* INTERLINEAR ANNOTATION ANCHOR */
  71            || nUtf32 == 0xFFFA /* INTERLINEAR ANNOTATION SEPARATOR */
  72            || nUtf32 == 0xFFFB /* INTERLINEAR ANNOTATION TERMINATOR */
  73            || nUtf32 == 0x1D173 /* MUSICAL SYMBOL BEGIN BEAM */
  74            || nUtf32 == 0x1D174 /* MUSICAL SYMBOL END BEAM */
  75            || nUtf32 == 0x1D175 /* MUSICAL SYMBOL BEGIN TIE */
  76            || nUtf32 == 0x1D176 /* MUSICAL SYMBOL END TIE */
  77            || nUtf32 == 0x1D177 /* MUSICAL SYMBOL BEGIN SLUR */
  78            || nUtf32 == 0x1D178 /* MUSICAL SYMBOL END SLUR */
  79            || nUtf32 == 0x1D179 /* MUSICAL SYMBOL BEGIN PHRASE */
  80            || nUtf32 == 0x1D17A /* MUSICAL SYMBOL END PHRASE */
  81            || nUtf32 == 0xE0001 /* LANGUAGE TAG */
  82            || (nUtf32 >= 0xE0020 && nUtf32 <= 0xE007F);
  83 }
  84
  85 int ImplIsHighSurrogate(sal_uInt32 nUtf32)
  86 {
  87     /* All code points that are high-surrogates, as of Unicode 3.1.1. */
  88     return nUtf32 >= 0xD800 && nUtf32 <= 0xDBFF;
  89 }
  90
  91 int ImplIsLowSurrogate(sal_uInt32 nUtf32)
  92 {
  93     /* All code points that are low-surrogates, as of Unicode 3.1.1. */
  94     return nUtf32 >= 0xDC00 && nUtf32 <= 0xDFFF;
  95 }
  96
  97 int ImplIsPrivateUse(sal_uInt32 nUtf32)
  98 {
  99     /* All code points of <http://www.unicode.org/Public/UNIDATA/
 100        UnicodeData.txt>, Version 3.1.1, that have a General Category of Co
 101        (Other, Private Use).
 102      */
 103     return (nUtf32 >= 0xE000 && nUtf32 <= 0xF8FF)
 104            || (nUtf32 >= 0xF0000 && nUtf32 <= 0xFFFFD)
 105            || (nUtf32 >= 0x100000 && nUtf32 <= 0x10FFFD);
 106 }
 107
 108 int ImplIsZeroWidth(sal_uInt32 nUtf32)
 109 {
 110     /* All code points of <http://www.unicode.org/Public/UNIDATA/
 111        UnicodeData.txt>, Version 3.1.1, that have "ZERO WIDTH" in their
 112        Character name.
 113      */
 114     return nUtf32 == 0x200B /* ZERO WIDTH SPACE */
 115            || nUtf32 == 0x200C /* ZERO WIDTH NON-JOINER */
 116            || nUtf32 == 0x200D /* ZERO WIDTH JOINER */
 117            || nUtf32 == 0xFEFF; /* ZEOR WIDTH NO-BREAK SPACE */
 118 }
 119
 120 sal_uInt32 ImplGetHighSurrogate(sal_uInt32 nUtf32)
 121 {
 122     OSL_ENSURE(nUtf32 >= 0x10000, "specification violation");
 123     return ((nUtf32 - 0x10000) >> 10) | 0xD800;
 124 }
 125
 126 sal_uInt32 ImplGetLowSurrogate(sal_uInt32 nUtf32)
 127 {
 128     OSL_ENSURE(nUtf32 >= 0x10000, "specification violation");
 129     return ((nUtf32 - 0x10000) & 0x3FF) | 0xDC00;
 130 }
 131
 132 sal_uInt32 ImplCombineSurrogates(sal_uInt32 nHigh, sal_uInt32 nLow)
 133 {
 134     OSL_ENSURE(ImplIsHighSurrogate(nHigh) && ImplIsLowSurrogate(nLow),
 135                "specification violation");
 136     return (((nHigh & 0x3FF) << 10) | (nLow & 0x3FF)) + 0x10000;
 137 }
 138
 139 /* vim:set shiftwidth=4 softtabstop=4 expandtab: */