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