sal/textenc/tenchelp.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: tenchelp.c,v $
  10  * $Revision: 1.6 $
  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 "tenchelp.h"
  32 #include "unichars.h"
  33 #include "rtl/textcvt.h"
  34 #include "sal/types.h"
  35
  36 static sal_Bool ImplGetUndefinedAsciiMultiByte(sal_uInt32 nFlags,
  37                                                sal_Char * pBuf,
  38                                                sal_Size nMaxLen);
  39
  40 static sal_Bool ImplGetInvalidAsciiMultiByte(sal_uInt32 nFlags,
  41                                              sal_Char * pBuf,
  42                                              sal_Size nMaxLen);
  43
  44 static int ImplIsUnicodeIgnoreChar(sal_Unicode c, sal_uInt32 nFlags);
  45
  46 sal_Bool ImplGetUndefinedAsciiMultiByte(sal_uInt32 nFlags,
  47                                         sal_Char * pBuf,
  48                                         sal_Size nMaxLen)
  49 {
  50     if (nMaxLen == 0)
  51         return sal_False;
  52     switch (nFlags & RTL_UNICODETOTEXT_FLAGS_UNDEFINED_MASK)
  53     {
  54     case RTL_UNICODETOTEXT_FLAGS_UNDEFINED_0:
  55         *pBuf = 0x00;
  56         break;
  57
  58     case RTL_UNICODETOTEXT_FLAGS_UNDEFINED_QUESTIONMARK:
  59     default: /* RTL_UNICODETOTEXT_FLAGS_UNDEFINED_DEFAULT */
  60         *pBuf = 0x3F;
  61         break;
  62
  63     case RTL_UNICODETOTEXT_FLAGS_UNDEFINED_UNDERLINE:
  64         *pBuf = 0x5F;
  65         break;
  66     }
  67     return sal_True;
  68 }
  69
  70 sal_Bool ImplGetInvalidAsciiMultiByte(sal_uInt32 nFlags,
  71                                       sal_Char * pBuf,
  72                                       sal_Size nMaxLen)
  73 {
  74     if (nMaxLen == 0)
  75         return sal_False;
  76     switch (nFlags & RTL_UNICODETOTEXT_FLAGS_UNDEFINED_MASK)
  77     {
  78     case RTL_UNICODETOTEXT_FLAGS_INVALID_0:
  79         *pBuf = 0x00;
  80         break;
  81
  82     case RTL_UNICODETOTEXT_FLAGS_INVALID_QUESTIONMARK:
  83     default: /* RTL_UNICODETOTEXT_FLAGS_INVALID_DEFAULT */
  84         *pBuf = 0x3F;
  85         break;
  86
  87     case RTL_UNICODETOTEXT_FLAGS_INVALID_UNDERLINE:
  88         *pBuf = 0x5F;
  89         break;
  90     }
  91     return sal_True;
  92 }
  93
  94 int ImplIsUnicodeIgnoreChar( sal_Unicode c, sal_uInt32 nFlags )
  95 {
  96     return
  97         ((nFlags & RTL_UNICODETOTEXT_FLAGS_NONSPACING_IGNORE) != 0
  98          && ImplIsZeroWidth(c))
  99         || ((nFlags & RTL_UNICODETOTEXT_FLAGS_CONTROL_IGNORE) != 0
 100             && ImplIsControlOrFormat(c))
 101         || ((nFlags & RTL_UNICODETOTEXT_FLAGS_PRIVATE_IGNORE) != 0
 102             && ImplIsPrivateUse(c));
 103 }
 104
 105 /* ======================================================================= */
 106
 107 sal_Unicode ImplGetUndefinedUnicodeChar(sal_uChar cChar, sal_uInt32 nFlags)
 108 {
 109     return ((nFlags & RTL_TEXTTOUNICODE_FLAGS_UNDEFINED_MASK)
 110                    == RTL_TEXTTOUNICODE_FLAGS_UNDEFINED_MAPTOPRIVATE) ?
 111                RTL_TEXTCVT_BYTE_PRIVATE_START + cChar :
 112                RTL_TEXTENC_UNICODE_REPLACEMENT_CHARACTER;
 113 }
 114
 115 /* ----------------------------------------------------------------------- */
 116
 117 sal_Bool
 118 ImplHandleUndefinedUnicodeToTextChar(ImplTextConverterData const * pData,
 119                                      sal_Unicode const ** ppSrcBuf,
 120                                      sal_Unicode const * pEndSrcBuf,
 121                                      sal_Char ** ppDestBuf,
 122                                      sal_Char const * pEndDestBuf,
 123                                      sal_uInt32 nFlags,
 124                                      sal_uInt32 * pInfo)
 125 {
 126     sal_Unicode c = **ppSrcBuf;
 127
 128     (void) pData; /* unused */
 129
 130     /* Should the private character map to one byte */
 131     if ( (c >= RTL_TEXTCVT_BYTE_PRIVATE_START) && (c <= RTL_TEXTCVT_BYTE_PRIVATE_END) )
 132     {
 133         if ( nFlags & RTL_UNICODETOTEXT_FLAGS_PRIVATE_MAPTO0 )
 134         {
 135             **ppDestBuf = (sal_Char)(sal_uChar)(c-RTL_TEXTCVT_BYTE_PRIVATE_START);
 136             (*ppDestBuf)++;
 137             (*ppSrcBuf)++;
 138             return sal_True;
 139         }
 140     }
 141
 142     /* Should this character ignored (Private, Non Spacing, Control) */
 143     if ( ImplIsUnicodeIgnoreChar( c, nFlags ) )
 144     {
 145         (*ppSrcBuf)++;
 146         return sal_True;
 147     }
 148
 149     /* Surrogates Characters should result in */
 150     /* one replacement character */
 151     if (ImplIsHighSurrogate(c))
 152     {
 153         if ( *ppSrcBuf == pEndSrcBuf )
 154         {
 155             *pInfo |= RTL_UNICODETOTEXT_INFO_ERROR | RTL_UNICODETOTEXT_INFO_SRCBUFFERTOSMALL;
 156             return sal_False;
 157         }
 158
 159         c = *((*ppSrcBuf)+1);
 160         if (ImplIsLowSurrogate(c))
 161             (*ppSrcBuf)++;
 162         else
 163         {
 164             *pInfo |= RTL_UNICODETOTEXT_INFO_INVALID;
 165             if ( (nFlags & RTL_UNICODETOTEXT_FLAGS_INVALID_MASK) == RTL_UNICODETOTEXT_FLAGS_INVALID_ERROR )
 166             {
 167                 *pInfo |= RTL_UNICODETOTEXT_INFO_ERROR;
 168                 return sal_False;
 169             }
 170             else if ( (nFlags & RTL_UNICODETOTEXT_FLAGS_INVALID_MASK) == RTL_UNICODETOTEXT_FLAGS_INVALID_IGNORE )
 171             {
 172                 (*ppSrcBuf)++;
 173                 return sal_True;
 174             }
 175             else if (ImplGetInvalidAsciiMultiByte(nFlags,
 176                                                   *ppDestBuf,
 177                                                   pEndDestBuf - *ppDestBuf))
 178             {
 179                 ++*ppSrcBuf;
 180                 ++*ppDestBuf;
 181                 return sal_True;
 182             }
 183             else
 184             {
 185                 *pInfo |= RTL_UNICODETOTEXT_INFO_ERROR
 186                               | RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
 187                 return sal_False;
 188             }
 189         }
 190     }
 191
 192     *pInfo |= RTL_UNICODETOTEXT_INFO_UNDEFINED;
 193     if ( (nFlags & RTL_UNICODETOTEXT_FLAGS_UNDEFINED_MASK) == RTL_UNICODETOTEXT_FLAGS_UNDEFINED_ERROR )
 194     {
 195         *pInfo |= RTL_UNICODETOTEXT_INFO_ERROR;
 196         return sal_False;
 197     }
 198     else if ( (nFlags & RTL_UNICODETOTEXT_FLAGS_UNDEFINED_MASK) == RTL_UNICODETOTEXT_FLAGS_UNDEFINED_IGNORE )
 199         (*ppSrcBuf)++;
 200     else if (ImplGetUndefinedAsciiMultiByte(nFlags,
 201                                             *ppDestBuf,
 202                                             pEndDestBuf - *ppDestBuf))
 203     {
 204         ++*ppSrcBuf;
 205         ++*ppDestBuf;
 206     }
 207     else
 208     {
 209         *pInfo |= RTL_UNICODETOTEXT_INFO_ERROR
 210                       | RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
 211         return sal_False;
 212     }
 213
 214     return sal_True;
 215 }
 216