sal/textenc/unichars.cxx

   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 "sal/config.h"
  30
  31 #include "sal/types.h"
  32
  33 #include "unichars.hxx"
  34
  35 bool ImplIsControlOrFormat(sal_uInt32 nUtf32)
  36 {
  37     // All code points of
  38     // <http://www.unicode.org/Public/UNIDATA/UnicodeData.txt>, Version 3.1.1,
  39     // that have a General Category of Cc (Other, Control) or Cf (Other,
  40     // Format):
  41     return nUtf32 <= 0x001F
  42            || (nUtf32 >= 0x007F && nUtf32 <= 0x009F)
  43            || nUtf32 == 0x070F // SYRIAC ABBREVIATION MARK
  44            || nUtf32 == 0x180B // MONGOLIAN FREE VARIATION SELECTOR ONE
  45            || nUtf32 == 0x180C // MONGOLIAN FREE VARIATION SELECTOR TWO
  46            || nUtf32 == 0x180D // MONGOLIAN FREE VARIATION SELECTOR THREE
  47            || nUtf32 == 0x180E // MONGOLIAN VOWEL SEPARATOR
  48            || nUtf32 == 0x200C // ZERO WIDTH NON-JOINER
  49            || nUtf32 == 0x200D // ZERO WIDTH JOINER
  50            || nUtf32 == 0x200E // LEFT-TO-RIGHT MARK
  51            || nUtf32 == 0x200F // RIGHT-TO-LEFT MARK
  52            || nUtf32 == 0x202A // LEFT-TO-RIGHT EMBEDDING
  53            || nUtf32 == 0x202B // RIGHT-TO-LEFT EMBEDDING
  54            || nUtf32 == 0x202C // POP DIRECTIONAL FORMATTING
  55            || nUtf32 == 0x202D // LEFT-TO-RIGHT OVERRIDE
  56            || nUtf32 == 0x202E // RIGHT-TO-LEFT OVERRIDE
  57            || nUtf32 == 0x206A // INHIBIT SYMMETRIC SWAPPING
  58            || nUtf32 == 0x206B // ACTIVATE SYMMETRIC SWAPPING
  59            || nUtf32 == 0x206C // INHIBIT ARABIC FORM SHAPING
  60            || nUtf32 == 0x206D // ACTIVATE ARABIC FORM SHAPING
  61            || nUtf32 == 0x206E // NATIONAL DIGIT SHAPES
  62            || nUtf32 == 0x206F // NOMINAL DIGIT SHAPES
  63            || nUtf32 == 0xFEFF // ZERO WIDTH NO-BREAK SPACE
  64            || nUtf32 == 0xFFF9 // INTERLINEAR ANNOTATION ANCHOR
  65            || nUtf32 == 0xFFFA // INTERLINEAR ANNOTATION SEPARATOR
  66            || nUtf32 == 0xFFFB // INTERLINEAR ANNOTATION TERMINATOR
  67            || nUtf32 == 0x1D173 // MUSICAL SYMBOL BEGIN BEAM
  68            || nUtf32 == 0x1D174 // MUSICAL SYMBOL END BEAM
  69            || nUtf32 == 0x1D175 // MUSICAL SYMBOL BEGIN TIE
  70            || nUtf32 == 0x1D176 // MUSICAL SYMBOL END TIE
  71            || nUtf32 == 0x1D177 // MUSICAL SYMBOL BEGIN SLUR
  72            || nUtf32 == 0x1D178 // MUSICAL SYMBOL END SLUR
  73            || nUtf32 == 0x1D179 // MUSICAL SYMBOL BEGIN PHRASE
  74            || nUtf32 == 0x1D17A // MUSICAL SYMBOL END PHRASE
  75            || nUtf32 == 0xE0001 // LANGUAGE TAG
  76            || (nUtf32 >= 0xE0020 && nUtf32 <= 0xE007F);
  77 }
  78
  79 bool ImplIsPrivateUse(sal_uInt32 nUtf32)
  80 {
  81     // All code points of
  82     // <http://www.unicode.org/Public/UNIDATA/UnicodeData.txt>, Version 3.1.1,
  83     // that have a General Category of Co (Other, Private Use):
  84     return (nUtf32 >= 0xE000 && nUtf32 <= 0xF8FF)
  85            || (nUtf32 >= 0xF0000 && nUtf32 <= 0xFFFFD)
  86            || (nUtf32 >= 0x100000 && nUtf32 <= 0x10FFFD);
  87 }
  88
  89 bool ImplIsZeroWidth(sal_uInt32 nUtf32)
  90 {
  91     // All code points of
  92     // <http://www.unicode.org/Public/UNIDATA/UnicodeData.txt>, Version 3.1.1,
  93     // that have "ZERO WIDTH" in their Character name:
  94     return nUtf32 == 0x200B // ZERO WIDTH SPACE
  95            || nUtf32 == 0x200C // ZERO WIDTH NON-JOINER
  96            || nUtf32 == 0x200D // ZERO WIDTH JOINER
  97            || nUtf32 == 0xFEFF; // ZEOR WIDTH NO-BREAK SPACE
  98 }
  99
 100 /* vim:set shiftwidth=4 softtabstop=4 expandtab: */